Symposium Organizers
Ingrid Repins, National Renewable Energy Laboratory
Shubhra Bansal, University of Nevada, Las Vegas
Sascha Sadewasser, International Iberian Nanotechnology Laboratory
Edgardo Saucedo, IREC
Symposium Support
Catalonia Institute for Energy Research (IREC)
Dr. Eberl MBE-Komponenten GmbH
First Solar
International Iberian Nanotechnology Laboratory
National Renewable Energy Laboratory
ES14.1: Passivation
Session Chairs
Sascha Sadewasser
Edgardo Saucedo
Tuesday PM, April 18, 2017
PCC North, 200 Level, Room 229 B
11:30 AM - *ES14.1.01
Improving the Open-Circuit Voltage of Cu2ZnSnSe4 Thin-Film Solar Cells via Interface Passivation
Jekyung Kim 1 , Sanghyun Park 1 , Giuk Jeong 1 , Sangwoo Ryu 1 , Jihun Oh 1 , Byungha Shin 1
1 , KAIST, Daejeon Korea (the Republic of)
Show AbstractThe kesterite compound Cu2ZnSn(S, Se)4, (CZTS), consisting of earth-abundant and non-toxic elements, has emerged as a promising photovoltaic material that can overcome issues of the currently dominating thin film solar cell technologies, namely, CdTe and Cu(In, Ga)Se2 (CIGS). The record efficiency of the CZTS solar cells has reached 12.6% using a device structure identical to that used in typical high efficiency CIGS solar cells, that is, with a molybdenum bottom contact, a CdS buffer layer, and double-layered transparent conducting oxide as a top contact. A typical CIGS solar cell contains a MoSe2 layer at the CIGS/Mo interface and it has been claimed that the MoSe2 helps improve the adhesion and electrical conduction between the CIGS and the Mo bottom contact. In CZTS solar cells, similar beneficial roles of the MoSe2 interfacial layer are often assumed although a thorough systematic study has not been performed yet. Due to these known benefits of the MoSe2 interfacial layer, researchers in the field tend to overlook a possibility of the bottom interface as a potential source of recombination centers. Here we have applied a novel scheme to passivate both the bottom and the top interfaces of a CZTS; a thin wide-bandgap dielectric layer with a regular array of openings is inserted between the CZTS and the Mo, and a thin continuous dielectric layer between the CZTS and the CdS buffer. With our interface passivation strategy, we demonstrated a Voc deficit as low as 0.569 V, comparable to the record value (0.560 V) among the reported CZTS solar cells even though full optimization of the passivation structure has not been completed yet. Structural, chemical, and electrical properties of the passivated CZTSe thin film solar cells will be discussed.
If time permits, I will also discuss our group’s recent progress on vacuum-processed binary selenide—such as SnSe and Sb2Se3—for thin film photovoltaic applications, particularly focusing on how to tune carrier concentration of the selenide films via control of process conditions.
12:00 PM - ES14.1.02
Boosting the Open Circuit Voltage of Cu2ZnSnS4 Solar Cells by a Lattice-Matched CeO2 Layer and Theoretical Understanding of Interface Defects
Andrea Crovetto 1 2 , Chang Yan 2 , Mattias Palsgaard 1 , Beniamino Iandolo 1 , Fangzhou Zhou 2 , Tue Gunst 1 , Troels Markussen 3 , John Stride 2 , Jorgen Schou 1 , Kurt Stokbro 3 , Mads Brandbyge 1 , Xiaojing Hao 2 , Ole Hansen 1
1 , Technical University of Denmark, Kgs. Lyngby Denmark, 2 , University of New South Wales, Sydney, New South Wales, Australia, 3 , QUANTUMWISE A/S, Copenhagen Denmark
Show AbstractThe open circuit voltage of state-of-the-art Cu2ZnSn(S,Se)4 solar cells with a low S content appears to be limited by bulk recombination. Therefore, at the current stage of development, the standard CdS buffer layer of Cu(In,Ga)Se2 solar cells is sufficient for obtaining a high-quality heterointerface with Cu2ZnSn(S,Se)4. Conversely, pure-sulfide Cu2ZnSnS4 solar cells are plagued by interface recombination, possibly due to a combination of an unfavorable band alignment with CdS, a large lattice mismatch with CdS, and unfavorable properties of the Cu2ZnSnS4 surface. As a consequence, the open-circuit voltage deficit and efficiency are consistently inferior in Cu2ZnSnS4 solar cells than in Cu2ZnSnSe4 solar cells.
In this work, we tackle the interface recombination problem of Cu2ZnSnS4 solar cells by a combination of theory and experiment. First, we verify by density functional theory that detrimental states within the band gap are expected on Cu2ZnSnS4 surfaces, but not necessarily on Cu2ZnSnSe4 surfaces. This means that band alignment is not the only important parameter when searching for a new buffer material for Cu2ZnSnS4 solar cells. Instead, an ideal buffer material should also remove the native defect states from the Cu2ZnSnS4 surface (passivation role) and form a high-quality interface with Cu2ZnSnS4 without introducing new defect states (by lattice matching and epitaxial growth).
To solve those problems, we propose and test experimentally CeO2 as a novel buffer layer material in Cu2ZnSnS4 solar cells. The major advantage of CeO2 is its nearly perfect lattice match with Cu2ZnSnS4 (0.4%), in contrast to the poor lattice match of CdS (7%). CeO2 is a non-toxic compound that is already used in the fields of catalysis and solid oxide fuel cells. Ce is more earth-abundant than Sn and about as earth-abundant as Cu and Zn (source: U.S. Geological Survey Fact Sheet 087-02).
Here we demonstrate that CeO2 can be easily grown on Cu2ZnSnS4 by chemical bath deposition. For growth temperatures as low as 50°C, we observe epitaxial growth of CeO2 on Cu2ZnSnS4 by transmission electron microscopy. Furthermore, CeO2 films show good surface coverage, low spurious phase content, and a nearly optimal band alignment with Cu2ZnSnS4 as measured by x-ray photoemission spectroscopy.
We then make a first attempt to include CeO2 in the Cu2ZnSnS4 solar cell architecture by inserting a thin CeO2 layer between Cu2ZnSnS4 and the usual CdS buffer. The result (over four solar cell batches up to 7% efficiency) is a reproducible open circuit voltage boost compared to the baseline case with a pure CdS buffer layer. The efficiency is also improved in three batches out of four. We argue that this is due to formation of a less defective heterointerface with a lower recombination velocity. By examining the calculated band structure of CeO2, we discuss why CeO2 can be an excellent thin passivation layer for Cu2ZnSnS4 but cannot be used a stand-alone buffer layer.
12:15 PM - ES14.1.03
Understanding the Full Effects of Rear Contact Passivation in CIGS Solar Cells
Pedro Salome 1 , Bart Vermang 2 7 , Rodrigo Ribeiro-Andrade 1 5 , Jennifer Teixeira 4 , Manuel Mendes 3 , Siraz Haque 3 , Nicoleta Nicoara 1 , Juan Gonzalez 5 , Joaquim Leitao 4 , Marika Edoff 6 , Sascha Sadewasser 1
1 , INL, Braga Portugal, 2 , imec, Heverlee Belgium, 7 Electrical Engineering, KU Leuven, Leuven Belgium, 5 Physics, UFMG, Minas Gerais Brazil, 4 Physics, University of Aveiro and I3N, Aveiro Portugal, 3 Materials Science, i3N/CENIMAT, Caparica Portugal, 6 Engineering Sciences, Uppsala University, Uppsala Sweden
Show AbstractAfter a decade of almost no improvements, in the last 6 years the electrical performance of Cu(In,Ga)Se2 (CIGS)-based solar cells has increased significantly from 19.9 % to 22.8 %. Part of this increase has been the development of a post-deposition treatment that, among many effects, lowers the recombination at the CIGS/buffer layer interface. These developments show that passivation of interfaces is a topic of the utmost importance for the continued improvement of the electrical performance of CIGS solar cells. The introduction of a point contact structure based on nano-openings of Al2O3 at the rear contact of the solar cell has already demonstrated an effective passivation effect in CIGS solar cells. In this paper, we fabricated ultrathin CIGS devices with and without the rear point contact structure. We studied the effects of the passivation layer using four techniques: transmission electron microscopy (TEM) and electron holography, photoluminescence (PL), optical simulations, and solar cell characterization. The rear-passivated 240 nm thick CIGS solar cells show a light-to-power conversion efficiency of 9 %, which is a clear improvement from the value of 7.5 % obtained for the reference cell without the passivation layer. In electron holography measurements, we observed changes to the phase of an electron wave function, which are related to the electric potential created by the Al2O3 layer that is creating an electric field at the CIGS rear interface that has the potential to repel carriers. From PL measurements with different excitation wavelengths we conclude that in spite of the devices being prepared with the same bulk CIGS, its electronic properties are different. Numerical electromagnetic simulations (Lumerical FDTD software) have been employed to analyse the CIGS light absorption and photogeneration, and explain the 1.1 mA/cm2 increase in Jsc for the passivated samples.
12:30 PM - ES14.1.04
Improved CdTe Solar-Cell Performance with An Evaporated Te Layer before the Back Contact
Andrew Moore 1 , Tao Song 1 , Christina Moffett 1 , James Sites 1
1 , Colorado State University, Fort Collins, Colorado, United States
Show AbstractForming an ohmic back contact with CdTe solar cells has long been a challenge due to the high electron affinity of CdTe and apparent lack of a metal with a matching work function. One common strategy to mitigate the CdTe/metal Schottky barrier is to highly dope the back surface of CdTe with a Cu treatment. However, such treatment can introduce stability issues due to the migration of Cu towards the front junction and the compensation of the p-type doping (CuCd) with an interstitial site (Cui). To lessen the negative impact from Cu, an evaporated Te buffer layer was introduced between CdTe absorber and its metallic contact. With such a Te buffer layer, the fabricated cells show better cell performance even with less Cu, primarily due to the gain in fill factor (FF). Simulation suggests that the Te layer forms a smaller hole barrier, which facilitates the hole collection, and thus induces better current collection and FF.
To better understand the physical mechanisms of this contacting scheme, measured material properties were incorporated into simulations for the comparison with the electrical measurements. Hall measurements showed that the carrier density of Te was on the order of 1018 cm-3. Current-Voltage-Temperature measurements were done to extract the back-barrier height. The samples with the Te layer had an approximately 0.1 eV lower barrier than our traditional contacts, consistent with published measurements of the Te band energies. Additionally, XPS was used to confirm the valence band offset (VBO) between Te and CdTe. The evaporated Te layer was shown to be a highly p-type material with a favorable VBO to CdTe, which mitigates the barrier to hole current.
A series of solar cells were fabricated with the Te layer and a range of Cu treatments. SIMS depth profiles were used to confirm the amount of Cu in the devices. These solar cells were electrically characterized by current-voltage, capacitance-voltage and quantum efficiency measurements. The devices with the Te layer and less Cu showed superior and more uniform performance over the substrate. While the incorporation of any amount of Cu treatment produced VOC values of approximately 850 mV, lower-Cu treatments produced average fill factors of 75.4 ± 0.5% verses 72.6 ± 1.3% for traditional Cu treatment on nominally identical devices. The Te contact layer allows for the use of less Cu in the device which results in better performance and stability.
12:45 PM - ES14.1.05
On the Effects of Chalcogen Excess for High Efficiency Kesterite Solar Cells
Douglas Bishop 1 , Priscilla Antunez 1 , Suarabh Singh 1 , Talia Gershon 1 , Teodor Todorov 1 , Oki Gunawan 1 , Richard Haight 1
1 , IBM T.J. Watson Research Center, Yorktown Heights, New York, United States
Show AbstractOne of the key assumptions guiding kesterite fabrication has been to anneal under significant excess chalcogen overpressure. The chalcogen excess (in addition to SnS/Se) has been shown to be critical to avoid decomposition and is thought to reduce chalcogen vacancies. To address these concerns, annealing is often performed under near infinite chalcogen supply, with concern only for excessive MoS/MoSe formation which has been reported to be detrimental to device performance. We demonstrate that chalcogen excess can also negatively impact the defect structure of kesterite absorbers, and in fact, careful optimization and limits to chalcogen excess are a requirement for high efficiency devices.
We fabricated CZTSe and CZTSSe absorbers with different levels of excess for sulfur and selenium and characterized the defect structure and device results. While it is known that too little chalcogen is detrimental, photoluminescence measurements show that too much chalcogen can also result in both deeper radiative defects and also increased non-radiative defects which limit carrier lifetime. We examine and explain the changes in defect structure with selenium quantity and correlate them to changes in device results and grain structure.
Annealing temperature and the back contact interface are the two other primary variables that dramatically affect the delicate chalcogen balance. We demonstrate the role of a back contact reaction barrier and show how annealing temperature dramatically effects the chalcogen balance. We conclude by highlighting the key chalcogen supply and annealing principles which have enabled IBM’s high efficiency devices.
ES14.2: Absorber Characterization
Session Chairs
Matthias Maiberg
Byungha Shin
Tuesday PM, April 18, 2017
PCC North, 200 Level, Room 229 B
2:30 PM - *ES14.2.01
Correlative Optical Microscopy of Charge Carrier Lifetimes, Mobilities and Space Charge Fields in Thin-Film Chalcogenide Solar Cells
Darius Kuciauskas 1 , Ana Kanevce 1
1 , National Renewable Energy Laboratory, Golden, Colorado, United States
Show AbstractRecombination limits the efficiency of solar cells, and understanding/reducing recombination is the focus of many research efforts. One effective approach for identifying defect locations is based on carrier generation with a focused laser beam in well-defined regions in the solar cell. Two-photon excitation (2PE) enables analysis beyond the surface region, and time-resolved photoluminescence (TRPL) acquisition enables studies of interfaces, space charge region, and the absorber bulk. Optical microscopy with excitation beam focused to a radius less than the carrier diffusion length allows analysis of carrier transport characteristics. In addition, correlative measurements of Second Harmonics Generation (SHG) are a novel optical probe for space charge fields in chalcogenide solar cells. Our results show that optical signals due to Electric Field Induced Second Harmonics (EFISH) can be used to analyze band bending at the surface, near Grain Boundaries (GBs), and at stacking fault extended defects in II-VI semiconductors. We use technology computer-aided design (TCAD) modeling to analyze spectroscopic and microscopic experimental data, and to evaluate the impact of defect and transport characteristics on the solar cells. We will present experimental and modeling results for single crystal, epitaxial, and polycrystalline CdTe. The goal is to identify and quantify efficiency-liming defects in solar cells, such as defects due to interfaces or GBs. We will also compare electrooptical properties of various types of chalcogenide solar cells, including CdTe, CIGS, kesterite, and perovskite.
3:00 PM - ES14.2.02
Expanding the Theory of the Diode Factor from Devices towards Semiconductor Layers Using the Example of Cu(In,Ga)Se2
Finn Babbe 1 , Conrad Spindler 1 , Susanne Siebentritt 1
1 , University of Luxembourg, Belvaux Luxembourg
Show AbstractThe diode factor gives indication of the dominant recombination process in a solar cell. The concept of a diode factors is only defined for devices, here we extend it to semiconductor layers by defining an “optical” diode factor. In photoluminescence (PL) measurements the PL intensity follows a power law in dependence of the excitation intensity. The exponent depends on competing recombination channels. The logarithm of the PL intensity is a measure for the quasi Fermi level splitting, which in turn indicates the maximum open circuit voltage that can be achieved with the absorber under investigation. Thus the PL power law is very similar to a SunsVoc plot and the slope gives the optical diode factor. We compare the diode factors from SunsVoc and from IV-curves to the exponent of PL measurements. The diode factors agree well for Cu-rich Cu(In,Ga)Se2 solar cells. Thus the diode factor is on the one hand an electrical quantity of the device, on the other hand it is already determined by the recombination channels of the absorber.
3:15 PM - ES14.2.03
Structural Defects and Lateral Composition Inhomogeneities in Cu(In,Ga)Se2 Layers Grown by Multistage Co-Evaporation
Enrico Avancini 2 , Shiro Nishiwaki 2 , Debora Keller 1 , Thomas Weiss 2 , Romain Carron 2 , Thomas Feurer 2 , Alejandro Filippin 2 , Benjamin Bissig 2 , Stephan Buecheler 2 , Ayodhya Tiwari 2
2 Laboratory for Thin Films and Photovoltaics, Empa, Dübendorf Switzerland, 1 Electron Microscopy Center, Empa, Dübendorf Switzerland
Show AbstractCo-evaporated Cu(In,Ga)Se2 (CIGS) layers grown by 3-stage or multi-stage processes result in polycrystalline material with varying [Ga]/([Ga]+[In]) (GGI) ratio across the thickness. Lateral inhomogeneities of the GGI ratio are expected to influence solar cell performance as they lead to lateral fluctuations of the bandgap grading profile and of the band alignment at the CdS/CIGS junction. Structural defects in the surface region of CIGS, such as voids, could further affect the CdS/CIGS junction quality in a detrimental manner. In this contribution, we report the presence of voids and lateral inhomogeneities in graded CIGS absorbers, as observed by Scanning Electron Microscopy (SEM) and Transmission Electron Microscopy (TEM) coupled with cross-section Energy-dispersive X-ray Spectroscopy (EDX) mapping. CIGS layers were grown by low-temperature (substrate temperature below 450°C) multi-stage co-evaporation with Sodium- and Rubidium- fluoride post-deposition treatments, and further processed into complete solar cells yielding photovoltaic conversion efficiencies above 19%. The evolution of the CIGS morphology is investigated throughout the late stages of the co-evaporation process. It is found, that the voids start to form simultaneously with the segregation of an inhomogeneous layer of copper selenide on the surface after the first stoichiometry point. Due to the use of lower substrate temperatures, the copper selenide layer at this stage is more likely to form islands as compared to the high-temperature (e.g. ~550 °C) processes. The influence of Cu excess after the first stoichiometry point, the rate of Cu evaporation and the Se pressure are investigated in relation to the elemental distribution and void formation. The growth of the chemical-bath deposited Cadmium Sulfide buffer layer on the void surfaces is found to depend on the void size and morphology. Using two-dimensional device simulation, we analyzed the impact on the final device performance of laterally inhomogeneous bandgap profiles, interface band alignments, and voids with passivated and unpassivated surfaces.
3:30 PM - ES14.2.04
Interpreting Time-Resolved Photoluminescence for Chalcopyrite and Kesterite Absorbers and Devices
Charles Hages 1 , Serguj Levcenco 1 , Alex Redinger 1 , Thomas Unold 1
1 , Helmholtz Zentrum Berlin, Berlin Germany
Show AbstractTime-resolved photoluminescence (TRPL) is a useful technique to investigate charge carrier dynamics in photovoltaic absorbers and solar cells. Useful material parameters, such as minority carrier lifetime and mobility, are often reported from TRPL measurements. Additionally, photoluminescence (PL) decay times are often used as a metric for absorber quality during process optimization. However, accurately interpreting the PL decay time can be quite complicated as the charge carrier dynamics are influenced by various recombination, transport, and charge trapping effects. Such metrics are particularly relevant for chalcopyrite materials, which often employ complicated band structures for high performance, as well as for kesterite materials, where short PL decay times are commonly reported for this highly defective material. Therefore, careful consideration is needed when analyzing TRPL data of such non-ideal absorbers and devices to correctly interpret the limiting PL decay times and absorber properties.
In this work, voltage-, temperature-, and intensity-dependent TRPL data for chalcopyrite and kesterite devices and absorbers are analyzed to determine the origin of the PL decay time. Additionally, a supercontinuum excitation source is used to investigate the excitation energy-dependence of the PL decay time. Such measurements allow us to distinguish the various PL decay rate limiting mechanisms for these materials. We find chalcopyrite absorbers – including CuInSe2 (CISe) and Cu(In,Ga)Se2 (CIGSe) – from a variety of processing conditions and compositional profiles follow the expected behavior for Shockley–Read–Hall (SRH) recombination limited decay times. The impact of carrier diffusion and surface recombination are demonstrated. In contrast, TRPL analysis of kesterite devices – including Cu2ZnSnSe4 (CZTSe) and Cu2ZnSn(S,Se)4 (CZTSSe) from various processing conditions – suggests that PL decay times are not directly related to the recombination lifetime for CZTSSe. A model for minority carrier trapping is proposed to explain the measured TRPL behavior for kesterites. These results illustrate the importance of measurement conditions and careful analysis when interpreting TRPL data, and further elucidate the defect related device limitations in kesterite solar cells.
3:45 PM - ES14.2.05
Origin of Band-Tails in Kesterite
Germain Rey 1 , Gerardo Larramona 2 , Stephane Bourdais 2 , Gilles Dennler 2 , Susanne Siebentritt 1
1 Laboratory for Photovoltaics, University of Luxembourg, Belvaux Luxembourg, 2 , IMRA Europe S.A.S., Sophia Antipolis France
Show AbstractKesterite Cu2ZnSn(Sx,Se1-x)4 (CZTSSe), an interesting absorber material for thin film solar cells, is subject to significant band tailing to which the large open-circuit voltage (VOC) deficit can be attributed. The VOC deficit, observed unanimously for kesterite-based solar cells, is the current bottleneck to further improvement of the photovoltaic properties of this material. The tails are caused by disorder in the crystal: electrostatic potential fluctuations caused by compensating defects or band gap fluctuations.
To study the nature of the band tails we performed temperature and intensity dependent photoluminescence (PL) studies in CZTSSe absorbers. These measurements allow conclusions on the nature of the tailing: whether it is caused by electrostatic potential fluctuations or by band gap variations on the nano-scale. Particularly the change of the low energy broadening of the emission with increasing intensity indicates that the tails are mostly due to band gap fluctuations and only to a smaller degree to electrostatic fluctuations.
A noticeable feature of the kesterite structure is its disorder by Cu-Zn exchanges in the Cu and Zn containing planes located at z=1/4 and z=3/4 of the unit cell. Since the degree of disorder has a significant influence on the band gap, it could be assumed that Cu-Zn disorder is responsible for the tails. To study this relationship we measured the optical absorption in Se-rich CZTSSe and pure selenide CZTSe samples with different Cu-Zn ordering degrees. Spectrophotometry and PL were used to access a wide range of the absorption coefficient values, and especially PL was able to provide accurate measurements of the tail-induced low absorption. We find Urbach tail behaviours that reaches deep into the band gap. For both CZTSSe and CZTSe, the comparison of the absorption spectra did not reveal any differences in terms of band tailing for fully disordered kesterite and partially ordered kesterite despite the effort in reaching high degree of order. The Zn-Cu disorder does not appear as the main cause of the tailing in kesterite.
ES14.3: Interfaces and Contacts
Session Chairs
Sergio Giraldo
Clemens Heske
Tuesday PM, April 18, 2017
PCC North, 200 Level, Room 229 B
4:30 PM - *ES14.3.01
(Zn,Mg)O Transparent Electrode and Buffer Layer for Junction Control in Cu(In,Ga)(Se,S)2 Solar Cells
Takashi Minemoto 1 , Jakapan Chantana 1 , Takuya Kato 2 , Hiroki Sugimoto 2
1 , Ritsumeikan University, Shiga Japan, 2 , Solar Frontier K. K., Atsugi Japan
Show AbstractImprovement in junction quality is crucial to achieve high conversion efficiency in thin-film solar cells. In Cu(In,Ga)(S,Se)2 (CIGSSe) solar cells, the quality of junction between buffer and absorber layers is well known to be important and the adjustment of the conduction band offset (CBO) of the layers is needed. In addition to the layers, we report the importance of the CBO control of transparent electrode and absorber layers by theoretical analysis in this study. The conduction band controls of both transparent electrode and buffer layer realize total band alignment control throughout the devices. Also, based on the analysis, we experimentally fabricated CIGSSe solar cells with the total band alignment control. Ultimately, the device structure of (Zn,Mg)O:Al/(Zn,Mg)O/(Cd,Zn)S/CIGSSe/Mo/Glass demonstrated higher values in all solar cell parameters compared to the similar structured cell but replaced (Zn,Mg)O:Al with ZnO:Al. The effect was verified in 20%-efficient device level. This study is the first experimental demonstration of the importance of the band control of transparent electrode in CIGSSe solar cells.
5:00 PM - ES14.3.02
Benefit of Textured CIGS Cells for Low Reflecting Nanogrid Application
Joop van Deelen 1 , Marco Barink 1
1 , TNO, Eindhoven Netherlands
Show AbstractWe have previously shown results on the benefit of enhancing the front contact with a metallic micro grid. However, increased conductivity of the front contact comes with an optical penalty as the metallic grid reflects the light proportionally to the surface coverage. A nanowire grid reduces the reflection, but also does not conduct as well as a micro grid, because the height of the nanogrid is lower and thereby the total amount of metal. For this reason, a new design is proposed by introducing a texture by which we can make the metal invisible.
The optical behaviour of nanogrid on top of CIGS cells for both flat and for textured layers was modeled using COMSOL FEM solver software. It was found that for a textured CIGS cell, the doped ZnO front contact can be supplemented by a metallic nanogrid virtually without optical losses. This is possible when the nanowires are carefully positioned and aligned with the texture. Therefore, this highly conductive and transparent front contact requires tuning of texture design and nanogrid dimensions. We show the optimized textures as well as general design rules derived from data spanning a wide variety of texture and wire dimensions. The texture CIGS stack has virtually no reflection (< 2%), in spite of the nanowire grid. In other words, black CIGS was obtained even with a metal front surface coverage of 20% or more.
5:15 PM - *ES14.3.04
Alternative Buffer and Front Contact Layers for Thin-Film Chalcogenide Cells
Yaroslav Romanyuk 1 , Johannes Loeckinger 1 , Timo Jaeger 1 , Lukas Greuter 1 , Stephan Buecheler 1 , Ayodhya Tiwari 1
1 , EMPA, Duebendorf Switzerland
Show AbstractThe highest efficiency CIGS and CZTS cells typically feature a CdS buffer and i-ZnO/Al:ZnO bilayer transparent front contact. This configuration, however, may present two hurdles for the widespread deployment of the chalcogenide technologies, and specifically the presence of Cadmium and a low corrosion resistance of the ZnO-based contact. Numerous efforts have been taken to develop ZnS-based buffers and to improve the front contact stability by incorporating new dopants into ZnO or replacing it with more corrosion resistant TCOs [1].
The talk will cover several recent developments on alternative buffer and TCO layers for chalcogenide thin film solar cells. Zn(S,O) buffers grown by fast chemical bath deposition using thioamide based sulphide precursors result in conversion efficiencies comparable to the reference growth with thiourea process but allow a 40-fold reduction in precursor concentration [2]. A combination with sputtered ZnMgO interlayers is very effective to reduce the metastable behavior of devices with the alternative buffer. On the TCO side, H- and Zn-doped In2O3 (IOH and IZO) layers can be deposited at room temperature with a reduced sputter damage [3]. Thanks for the high carrier mobility in as-deposited amorphous IOH, the optical absorption in the near IR range can be reduced, and the IOH electrode helps to improve the open circuit voltage by 20 mV. Importantly, non-encapsulated IOH devices exhibit a slower degradation than AZO contacts during heat-light soaking tests, whereas even better corrosion resistance is expected for IZO contacts.
[1] T. Feurer et al, Prog. Photovolt: Res. Appl. (2016) DOI: 10.1002/pip.2811.
[2] J. Löckinger et al., J. Optics 18, 2040 (2016).
[3] T. Jäger et al., J. Appl. Phys. 117, 205301 (2015).
5:45 PM - ES14.3.05
Alkali Doping in Solution Processed Kesterite Solar Cells
Stefan Haass 1 , Raquel Caballero 2 , Christian Andres 1 , Yaroslav Romanyuk 1 , Ayodhya Tiwari 1
1 , Empa, Swiss Federal Laboratories for Materials Science and Technology, Laboratory for Thin Films and Photovoltaics, Duebendorf Switzerland, 2 , Universidad Autónoma de Madrid, Madrid Spain
Show AbstractAlkali doping in kesterite solar cells has shown positive effects on morphological and electronic properties. Especially sodium doping has routinely been applied by most groups through Na diffusion from soda-lime glass or by intentionally adding a Na compound[1]. Other alkali metals like lithium and potassium have also shown beneficial effects yet not identical to sodium[2]. Until now the alkali dopant that results in the highest improvement in photovoltaic properties has not been clearly identified and actually different groups report for diverse alkali dopants to work best for them[2-3].
Here we present a systematic study of the bulk doping of DMSO solution-processed kesterite absorbers with alkali metals (Li, Na, K, Rb, Cs). All alkali metals have a significant impact on the morphology of the kesterite absorbers by increasing the grain size and improving the crystallinity as reported earlier. Alkali metals affect the tin content during the annealing step of the absorber synthesis in a different extent. Importantly, a different Cu/Sn metal ratio is required to achieve the highest device performance for each alkali dopant. Therefore, we investigate a comprehensive matrix of samples with a wide range of Cu/Sn metal ratios and different doping concentrations with a set of advanced characterization methods such as SIMS, JV-T, AS, EQE and TRPL. For Li-, Na- and K-containing devices efficiencies above 10-11% are obtained, whereas Rb and Cs doping yields less efficient devices. The open circuit voltage and minority carrier lifetime are increasing for Li, Na and K, but are decreasing for Rb and Cs. Correlations of device parameters with ionic radii and concentration of the alkali dopants will be discussed.
[1] T. Gershon, B. Shin, N. Bojarczuk, M. Hopstaken, D. B. Mitzi, S. Guha, Adv. Energy Mater. 2015, 5.
[2] H. Xin, S. M. Vorpahl, A. D. Collord, I. L. Braly, A. R. Uhl, B. W. Krueger, D. S. Ginger, H. W. Hillhouse, Phys. Chem. Chem. Phys. 2015.
[3] Y.-T. Hsieh, Q. Han, C. Jiang, T.-B. Song, H. Chen, L. Meng, H. Zhou, Y. Yang, Adv. Energy Mater. 2016.
ES14.4: Poster Session I: Passivation, Absorber and Advanced Characterization, Interfaces and Contacts
Session Chairs
Ingrid Repins
Edgardo Saucedo
Wednesday AM, April 19, 2017
Sheraton, Third Level, Phoenix Ballroom
9:00 PM - ES14.4.01
Mapping Disordered Nanodomains in Cu2(Zn,Sn)S4 and Cu2(Zn,Sn)Se4-xS4
Dennis Pruzan 2 , Jeffery Aguiar 1 2 , Brian Devener 3 , Mehmet Erkan 4 , Akira Nagaoka 2 4 , Kenji Yoshino 5 , Helio Moutinho 6 , Mowafak Al-Jassim 6 , Mike Scarpulla 2 4 7
2 Department of Materials Science and Engineering, University of Utah, Salt Lake City, Utah, United States, 1 , Idaho National Laboratory, Idaho Falls, Idaho, United States, 3 Analysis Laboratory, University of Utah, Salt Lake City, Utah, United States, 4 Department of Electrical Engineering, University of Utah, Salt Lake City, Utah, United States, 5 Department of Applied Physics and Electronic Engineering, University of Miyazaki, Miyazaki Japan, 6 , National Renewable Energy Laboratory, Golden, Colorado, United States, 7 Department of Physics and Astronomy, University of Utah, Salt Lake City, Utah, United States
Show AbstractSecondary phases in photovoltaics are known to degrade photovoltaic device performance. The role of growth techniques remains a particular area of interest in addressing issues of recombination and passivation in devices associated with the presence of secondary phase formation. Despite reaching device efficiencies greater than 12%, Cu2ZnSnS4 (CZTS) is a sustainable alternative energy material used in thin-film photovoltaics prone to the presence of secondary phases. This disorder is believed to account for the larger-amplitude potential and bandgap fluctuations and a sizable part of the open-circuit voltage (VOC) deficit in CZTS and related Cu2ZnSn(S,Se)4 (CZTSSe) devices, yet the origins and length scales of these fluctuations have not been fully elucidated. Amongst correlating growth and secondary phase formation, mapping the presence of phases is not only a crux in photovoltaic research, but ideally can provide all the available bonding arrangements, defects, and chemistries for various materials spanning a multitude of functions beyond photovoltaics.
In this presentation, we plan to report our detailed microscopy studies focusing on composition variations within bulk multicrystals of CZTS grown by the travelling heater method (THM) and co-evaporated CZTSSe. In our slow-cooled, solution grown CZTS crystals we find direct evidence for spatial composition fluctuations of amplitude <1 at% (~5x1020 cm-3). However, rather than being homogeneously-distributed we find a characteristic 20 nm length scale for these fluctuations our slow cooled materials, which sets a definite length scale for band gap and potential fluctuations. Similarly, in evaporated CZTSSe using transmission X-ray microscopy tomography, we find and show evidence of micron-scale copper to zinc anti-correlations over a previously inaccessible combination of resolution and sample size that is consistent with the length scale of grains. These findings yield further insight the large open-circuit voltage deficits regularly seen in CZTS devices and propose root causes for achieving compositional homogeneity in these materials.
9:00 PM - ES14.4.02
Nondestructive High-Power-High-Temperature Raman Spectroscopy for Probing Microscopic Structural Variations in CZTSe Alloys
Qiong Chen 1 , Sergio Bernardi 2 , Yong Zhang 1
1 Department of Electrical and Computer Engineering, and Energy Production and Infrastructure Center (EPIC), University of North Carolina at Charlotte, Charlotte, North Carolina, United States, 2 , CEA, LITEN, Grenoble France
Show AbstractEven though absorber layers fabricated by different methods may yield comparable efficiencies and appear to be similar under conventional probes, such as low power Raman, PL, XRD, they could in fact be quite different in their microscopic structures. We have developed a novel nondestructive spectroscopy approach, high-power-high-temperature (HPHT) Raman spectroscopy, which is capable of revealing the microscopic structural variations of complex alloys like CZTSe over a large area. CZTSe films prepared by sputtering and co-evaporation methods were examined and compared in both lateral and depth directions. HPHT Raman spectroscopy uses a tightly focused CW laser to illuminate the material with a power density that is just high enough to induce local structural change but not to the extent major material ablation, then identify the changes in Raman features compared with the spectrum before the illumination, under low power condition. In addition to red shift and intensity drop of the major CZTSe Raman modes, different new secondary peaks appeared in different samples after HP illumination, for example, 242 cm-1 (MoSe2) in the sputtered film and but 238 cm-1 (t-Se) in the co-evaporated films. Dominant MoSe2 was found near the interface of substrate and absorber layer in the sputtered film, and HP illumination also led to CZTSe and Mo reaction. However, MoSe2 was not observed in the co-evaporated sample. In general, HP illumination brought qualitatively different changes to the CZTSe samples, not only in the CZTSe Raman peaks but also in the secondary phases, which suggests that there are some subtle microscopic difference between the two types of samples. In addition, 2D Raman mapping revealed larger spatial extension of the local heating effect caused by HP illumination in sputtered film, which also indicates that two nominally similar films might have different thermal conductivities. High temperature measurement, which offers uniform heating as opposed to local heating with high power, further enhances the capability of the approach.
9:00 PM - ES14.4.03
Effects of CdCl2 Treatment and Br2/MeOH Etching on the Absorption of CdTe Thin Films as Measured by Photothermal Deflection Spectroscopy
Jordan Andrews 1 2 , Mario Beaudoin 2 , Prakash Koirala 3 , Xinxuan Tan 3 , Robert Collins 3 , Stephen O'Leary 1
1 School of Engineering, University of British Columbia Okanagan, Kelowna, British Columbia, Canada, 2 AMPEL, University of British Columbia, Vancouver, British Columbia, Canada, 3 Department of Physics and Astronomy and Center for Photovoltaics Innovation and Commercialization, University of Toledo, Toledo, Ohio, United States
Show AbstractThe optical bandgap and Urbach edge parameter of two sets of solar cell quality, polycrystalline thin film CdTe samples have been measured using photothermal deflection spectroscopy (PDS) to compare the effects of the deposition process and post-deposition treatments on the absorption onset. One set of samples was grown by RF magnetron sputtering while another set was grown by closed space sublimation. Absorption measurements were performed on as-grown samples, samples annealed in a CdCl2 environment, and samples further treated by etching with Br2/MeOH which leaves a Te-rich surface layer.
X-ray diffraction measurements were performed on the samples in the powder (powder XRD) and in the grazing incidence (GIXRD) geometries to determine crystallite orientations. For sputtered samples, the preferred surface orientation of polycrystalline growth is the <111> plane. After the sample has been CdCl2 treated, the intensity of the <111> peak is greatly reduced, and other peaks appear indicating that the sample has become randomly-oriented polycrystalline. Further treatment with Br2/MeOH leaves the sample unchanged from this polycrystalline form. The samples grown by closed space sublimation do not have a preferred <111> orientation, and are randomly-oriented polycrystalline both before and after the treatments.
Photothermal deflection spectroscopy is a powerful technique that allows the measurement of very low absorption coefficients in thin films that are inaccessible by conventional light transmission techniques. The results from PDS are analyzed in two separate ways. To determine the bandgap, Eg, and width of the absorption edge, E0, an Urbach form of the absorption coefficient, α, as a function of the energy of incident light, hν, α(hν)∝ αgexp[(hν-Eg)/E0] is assumed where h is Planck’s constant and αg is the absorption coefficient at the bandgap. This expression is then used in a multiple reflection solution of the PDS heat equation based on the Rosencwaig-Gersho-Fernelius formalism. E0 is clearly reduced when the films transition from <111> oriented to randomly-oriented polycrystalline upon CdCl2 treatment, and shows an additional small decrease when the sample is etched in the Br2/MeOH solution. For these sputtered films, E0 decreases from 40 meV for <111>-oriented films before CdCl2 treatment, to 23 meV after CdCl2 treament, and to 18 meV after etching. To confirm these PDS results, we also applied the Jackson approximation to extract α(hν) directly from the PDS spectra. We used this α(hν) as a starting solution in our multiple reflection model to simulate the PDS results; we find that the Jackson approximation yields Eg and E0 results comparable to our full model but overestimates the absorption from the states below the bandgap. Direct comparison of this α shows that below the bandgap the absorption is greatly increased after the Br2/MeOH etching due to a well-defined layer of amorphous Te on the surface.
9:00 PM - ES14.4.04
Simulation Study of Photoluminescence Lifetime and Device Efficiency of Cu(In,Ga)Se2 Thin Film Solar Cells
Jose Fabio Lopez Salas 1 , Michael Richter 1 , Stephan J. Heise 1
1 Energy and Semiconductor Research Laboratory, University of Oldenburg, Oldenburg Germany
Show AbstractTime resolved photoluminescence (TRPL) is a contactless and nondestructive method of investigating charge carrier lifetimes in Cu(In,Ga)Se2 thin film solar cells. Therefore it is a promising tool to assess material quality during intermediate production steps. A correlation between photoluminescence lifetime and device efficiency has been reported before by several groups. Here we present a simulation analysis of the specific mechanisms behind this correlation. Our simulation model for Cu(In,Ga)Se2 thin film solar cells is capable of reproducing the results of measurements of TRPL, current-voltage characteristics and quantum efficiency spectra. We provide evidence for the validity of this model by simulating experimental results of TRPL under varying excitation intensities as well as modelling the changes in current voltage characteristics and quantum efficiency spectra after conditioning via light-soaking. With this model we then proceed to analyze the impact of specific simulation parameters on device efficiency and photoluminescence lifetime. For this step we choose parameters that are highly relevant for the dynamics of minority charge carriers in the absorber layer. Among these parameters we include the density of defect states at the interface between CdS and Cu(In,Ga)Se2, as well as the density and energy of trap states near the conduction band in the Cu(In,Ga)Se2 absorber material. The goal is to understand the impact of different mechanisms such as interface recombination, trapping and SRH recombination on both the TRPL signal and the device performance, thereby also providing a method of identification of dominating mechanisms limiting the device efficiency.
9:00 PM - ES14.4.05
High Fidelity Polycrystalline CdTe/CdS Heterostructures via Molecular Dynamics
Rodolfo Aguirre 1 , Jose Chavez 2 , Xiao Zhou 2 , Sergio Almeida 1 , David Zubia 1
1 , The University of Texas at El Paso, El Paso, Texas, United States, 2 , Sandia National Laboratories, Livermore, California, United States
Show AbstractThe structural characteristics of polycrystalline CdTe solar cells play an important role in determining their electronic performance. Understanding of the structure at the meso and atomic scale is difficult to achieve through experiments alone including advanced Atomic Probe Tomography and Transmission Electron Microscopy. Alternatively, molecular dynamics simulations of polycrystalline growth of CdTe/CdS heterostructures have been performed. First, CdS was deposited on an amorphous CdS substrate, forming a polycrystalline film. Subsequently, CdTe was deposited on top of the polycrystalline CdS film. Cross-sectional images show grain formation at early stages of the CdS growth. During CdTe deposition, the CdS structure remains almost unchanged. Concurrently, CdTe grain boundary motion was detected after the first 24.4 nanoseconds of CdTe deposition. With the elapse of time, this grain boundary pins along the CdS/CdTe interface, leaving only a small region of epitaxial growth. CdTe grains are larger than CdS grains in agreement with experimental observations in the literature. Crystal phase analysis shows that zinc blende structure dominates over the wurtzite structure inside both CdS and CdTe grains. Defect analysis shows the formation of various dislocations at the grain boundaries and inside grains. Composition analysis shows Te and S diffusion to the CdS and CdTe films, respectively. These simulated results may stimulate new ideas for studying and improving CdTe solar cell efficiency.
9:00 PM - ES14.4.06
Charged Grain Boundaries Reduce the Open Circuit Voltage of Polycrystalline Solar Cells—An Analytic Description
Benoit Gaury 1 2 , Paul Haney 1
1 , National Institute of Standards and Technology, Gaithersburg, Maryland, United States, 2 , University of Maryland, College Park, Maryland, United States
Show AbstractAnalytic expressions are presented for the dark current-voltage relation J(V) of a pn+ junction with positively charged columnar grain boundaries with high defect density. These expressions apply to non-depleted grains with sufficiently high bulk hole mobilities. The accuracy of the formulas is verified by direct comparison to numerical simulations. Numerical simulations further show that the dark J(V) can be used to determine the open circuit potential Voc of an illuminated junction for a given short circuit current density Jsc . A precise relation between the grain boundary properties and Voc is provided, advancing the understanding of the influence of grain boundaries on the efficiency of thin film polycrystalline photovoltaics like CdTe and Cu(In, Ga)Se2 .
9:00 PM - ES14.4.07
Admittance-Spectroscopy Model of CIGS-Based Solar Cells
Kazimierz Plucinski 1
1 , Military University of Technology, Warsaw Poland
Show AbstractA model for analyzing the current-voltage, conductance-frequency and capacitance-frequency characteristics of a CIGS-based solar cell is developed.
The effects of various parameters such as intensity of light, electron and hole lifetime and trapping have been taken into account. The parameters of the trapping state, like cross-section based on the conductance-frequency and capacitance-frequency characteristics measured in different temperature are determined. Diffusion length for various configurations, taking into account free carriers during their lifetime and the influence of the pump-fluence, is modeled. The results are compared with reported experimental data.
9:00 PM - ES14.4.08
XESCA—X-Ray Emission Spectroscopy for Chemical Analysis
Sang Jun Lee 1 2
1 , SLAC, Menlo Park, California, United States, 2 , Stanford University, Menlo Park, California, United States
Show AbstractX-ray Photoelectron Spectroscopy is an ubiquitous tool in natural sciences, providing elemental and chemical sensitivity to a large community of scientists in both academia and industry. However, due to the limited mean free path of electrons, this technique remains highly surface sensitive (few nm), and only through sputtering (an intrusive technique) can deeper layers be uncovered. Moreover, the data acquisition imposes quite strict sample environments that requires vacuum and associated instrumentation and personnel overhead.
We have recently commissioned a transition edge sensor (TES) based spectrometer at Stanford Synchrotron Radiation Laboratory (SSRL) that has the ability to explore new paradigms in soft x-ray spectroscopy, achieving sensitivity of sub-mMol concentrations in aqueous/organic solvents, sub-percent sensitivity for monolayer films immersed in a solvent, solid matrix, or high-pressure gas, and sensitivity to concentrations <1019/cm3 for defects and dopants in condensed phase samples.
Here we present a new detection scheme that makes use of the ability to separately analyze each pixel in the TES sensor array, in which the position of the pixels are mapped onto the angular exit angle from the sample, providing a non-intrusive depth profiling using X-ray Emission, with chemical sensitivity. We demonstrate the applicability of this approach for CIGS-based absorbers, in which the elemental composition and and chemical state near the interface (with the emitter layer CdS) plays an important role in determining the overall efficiency.
9:00 PM - ES14.4.09
Surface Phases in (Ag,Cu)(In,Ga)Se2 Semiconductors Determined by Ultraviolet and X-Ray Photoemission Spectroscopy
Kevin Jones 1 , William Shafarman 2 , Robert Opila 1
1 Material Science, University of Delaware, Newark, Delaware, United States, 2 , Institute of Energy Conversion at University of Delaware, Newark, Delaware, United States
Show AbstractThe compositional variations and electronic properties of the surfaces of thin films in solar devices are important for device performance. In this work, we investigated (Agx-1Cux)(Iny-1Gay)Se2 absorber (ACIGS), in which we find a considerable reduction in Cu content with respect to the bulk near the surfaces in Ga-poor compositions for ACIGS absorbers, and evidence of a phase transition from Cu-poor to Cu-rich phases in Ga rich conditions as a function of Ga/(Ga+In) ratio. The valence band states of the base material CuInSe2 and alloys with Ag and Ga were investigated in the region of the Cu3d-Se4p, Ag4d-Se4p and Ga4p-Se4p bands using ultraviolet photoelectron spectroscopy (UPS). The UPS data showed that the valence band maxima (VBM) shifted further below the Fermi level when substituting Cu with Ag atoms, with little evidence of Ag4d-Se4p hybridization (with no Ga present). When Ga was present however, strong evidence of Cu3d-Se4p hybridization appears, as shown by the increase in Cu3d nonbonding and Cu3d-Se4p antibonding peaks. While keeping the Ag/(Ag+Cu) ratio constant at 0.27 and increasing Ga content, the overall surface transitioned from a Cu-poor phase to a Cu-rich phase, which was evident from the subtracted-photoemission valence band spectra, closely resembling the calculated density of states. This increase in Cu content was validated by X-ray photoemission spectroscopy (XPS). Thus, the surface phases seen on ACIGS are a strong function of the bulk composition.
9:00 PM - ES14.4.10
Evaluation of Electrical Characteristics of Cu(In,Ga)(S,Se)2 Thin Films Prepared by a Two-Step Sputtering/Annealing Large-Scale Fabrication Process
Juran Kim 1 , Trang Nguyen 1 , Seokhyun Yoon 1 , Chan-Wook Jeon 2 , William Jo 1
1 Department of Physics, Ewha Womans University, Seoul Korea (the Republic of), 2 School of Chemical Engineering, Yeungnam University, Gyeongsan Korea (the Republic of)
Show AbstractWe characterized the electrical properties of Cu(In,Ga)(Se,S)2 (CIGSSe) thin films grown by a two-step sputter and selenization/sulfurization process to make an ideal double-graded band gap and improve the power conversion efficiency. In this research, two CIGSSe light absorber layers – one is from normal panel (efficiency 14.5%), and the other from abnormal one (efficiency 13%) – were investigated to obtain local electrical and structural characteristics utilizing micro-Raman scattering spectroscopy and Kelvin probe force microscopy (KPFM). In particular, we can obtain grain boundary potential and surface work function through the KPFM measurement. The results indicate that the thin film from the abnormal sample shows non-uniformity on the surface. While the normal sample has unified CIGSSe work function peak distribution, the abnormal sample contains not only CIGSSe, but secondary phases such as Cu2Se on the surface. Furthermore, positively charged surface potential is formed near the grain boundaries (GBs), and the surface potential of the normal sample is higher than that of the abnormal sample. This means that the negative band bending near the GBs is stronger in the normal sample. Therefore, we can say the carrier behavior on the absorber surface has been improved, expecting better conversion efficiency of the solar cells.
9:00 PM - ES14.4.11
Surface and Local Electronic Properties of (Ag,Cu)2ZnSn(S,Se)4 Photovoltaic Absorbers Grown by 2-Step Processes
Juran Kim 1 , Gee Yeong Kim 1 , Trang Nguyen 1 , Seokhyun Yoon 1 , Youngill Kim 2 , Kee-Jeong Yang 2 , Dae-Hwan Kim 2 , William Jo 1
1 Department of Physics, Ewha Womans University, Seoul Korea (the Republic of), 2 Convergence Research Center for Solar Energy, Daegu Gyeongbuk Institute of Science &Technology, Daegu Korea (the Republic of)
Show AbstractCu2ZnSn(S,Se)4 (CZTSSe) have been paid attention as rare-metal-free light absorber layer for thin-films solar cells. As the open-circuit voltage (VOC) is usually lower than that of Cu(In,Ga)Se2 (CIGS), it needs to be improved. Also, for the ideal band gap of 1.4 eV, it is known to be able to tune the band gap from 1.0 eV for CZTSe to 1.35 eV for Cu2ZnGeSe4 (CZTSe). The band gap tuned absorber is expected to show better performance. In this research, we applied 10nm-thickness of Ag on Mo back-contact layer, and then deposited precursor layer in the stack of Cu/Sn/Zn. After that, the precursors were gone through sulfo-selenization. The samples doped with Ag shows power conversion efficiency of ~10%. For characterization of local electrical and structural properties, we utilized conductive atomic force microscopy (c-AFM), Kelvin probe force microscopy (KPFM) and micro-Raman scattering spectroscopy. The phase alteration of the (Ag,Cu)2ZnSnSe4 thin films are observed when Ag is substituted with Cu. The grain size has been increased, expecting the decrease in the defect states. Also, after KCN etching, the work function distribution shows unified peaks, removed the residue of S, Se, or Ag from the post annealing process. Consequently, we are able to comprehend the band alignment between the absorber layer and a buffer layer. Moreover, we can suggest the optimized band gap tuning and band gap grading conditions for the higher efficiency photovoltaic cells.
9:00 PM - ES14.4.12
Revisiting Thin-Film Absorption Coefficient Measurement from Photoluminescence
Germain Rey 1 , Conrad Spindler 1 , Maurice Nuys 2 , Reinhard Carius 2 , Shuyi Li 3 , Charlotte Platzer-Bjorkman 3 , Susanne Siebentritt 1
1 Laboratory for Photovoltaics, University of Luxembourg, Belvaux Luxembourg, 2 Institut für Energie und Klimaforschung, Forschungszentrum Jülich GmbH, Jülich Germany, 3 , Angström Solar Center, Uppsala Sweden
Show AbstractPhotoluminescence (PL) is a powerful method to investigate band-tail and defect induced absorption. Because the emissivity and the absorptance of a material are related to each other, the PL emission can be used to gain information about the optical absorption of a semiconductor and measure its absorption coefficient (α). The rate of photon emission rate is given by generalized Planck’s law, while the photon absorption rate is given by Beer’s law, leading to a continuity equation for the photon current density with two source terms. In order to obtain the PL emission outside of the material, the continuity equation has to be integrated over the thickness of the material with appropriate boundary conditions. For self-supported samples like wafer, the solution has been given in Ref [1] and because the PL emission explicitly depends on α, PL measurements were used to determine α on Si and GaAs wafers. Here we extend the aforementioned method to thin-films and thin-film stacks, where substrate and multilayer structure lead to non-trivial boundary conditions that can be treated in the framework of the transfer matrix method [2]. We applied these techniques to determine α on various chalcogenide thin-films (CuGaSe2, CuInSe2 and Cu2ZnSnSe4) and compared the results to other methods like photo-thermal deflection spectroscopy and spectrophotometry with good agreement. We also stress the importance of the boundary condition correction when the PL signal arises from band tails. The determination of α from PL has the advantage to be accurate for a high dynamic range down to very low absorption (10-1 cm-1), thus it is a very convenient tool to evaluate the quality of a semiconductor thin-film by investigating its sub-band gap absorption.
[1] Daub, E. & Würfel, P. Ultralow Values of the Absorption Coefficient of Si Obtained from Luminescence Phys. Rev. Lett. 1995, 74, 1020-1023.
[2] Carius, R., Becker, F., Wagner, H. and Zettler, J.-T. Electroluminescence and Transport in a-Si:H p-i-n Diodes at Room Temperature, MRS Proceedings 1993, 297, 357.
9:00 PM - ES14.4.13
Role of Defects in Cu(In,Ga)Se2 Solar Cells with CdS and Zn(1-x)Sn(x)O(y) Buffer Layers—Microscopy and Photoluminescence Study
Jennifer Teixeira 1 , Pedro Salome 2 , R. Ribeiro Andrade 3 , Jan Keller 4 , Juan Gonzalez 3 , Tobias Torndahl 4 , Sascha Sadewasser 2 , Joaquim Leitao 1
1 , Univ de Aveiro, Aveiro Portugal, 2 , International Iberian Nanotechnology Laboratory, Braga Portugal, 3 Departmento de Física, Universidade Federal de Minas Gerais, Belo Horizonte Brazil, 4 Ångström Solar Center, Solid State Electronics, Uppsala University, Uppsala Sweden
Show AbstractThe investigation of Cd-free buffer layers in Cu(In,Ga)Se2 (CIGS) solar cells has been intensively studied, motivated by environmental concerns and electrical performance. On one hand Cd is toxic and on the other hand the low bandgap energy (2.4-2.5 eV) of the traditional CdS buffer layer causes optical losses which lower the electrical performance. In comparison with the commonly used CdS, alternative compounds should have better electrical properties, larger bandgap, non-toxic elements, suitable conduction band line-up to the absorber and to the front contact, and a low density of recombination defects at the absorber/buffer layer interface. In this work, thin film solar cells based on CIGS, where just the buffer layer material is changed, are fabricated and studied. The two studied buffer layers are CdS and Zn1-xSnOy (ZTO). Probe aberration corrected scanning transmission electron microscopy (Cs-STEM), energy dispersive X-ray spectroscopy (EDS) and photoluminescence (PL) analysis were performed. For the CdS sample, TEM analysis shows areas of the CIGS that were deprived of Cu and enriched with Cd, whereas, in the same location, the CdS is deprived of Cd and enriched with Cu. In the case of the ZnSnO sample, we have not detected any diffusion of Zn or Sn into the CIGS or diffusion of CIGS elements into the ZTO. PL measurements show a clear influence of electrostatic fluctuating potentials on the electronic structure on both samples. However, this influence is higher for the CdS sample. After a chemical removal of the CdS layer, the PL is less affected by fluctuating potentials and resembles the ZTO emission. Such occurrence suggests a lower density of ionised defects in accordance with a smaller concentration of elements diffused from the CdS into the CIGS. The TEM results may explain why the PL measurements indicate that the CdS sample has a higher number of interface defects, compared with ZTO, as seen by the increased electrostatics fluctuating potentials. The beneficial properties of ZTO are also discussed.
9:00 PM - ES14.4.14
Effect of Varying the Cu Content on Cu(In,Ga)Se2 Solar Cells
Bruno Alves 1 , Pedro Salome 2 , Jennifer Teixeira 1 , Piotr Szaniawski 3 , Viktor Fjallstrom 3 , Marika Edoff 3 , Sascha Sadewasser 2 , Joaquim Leitao 1
1 , Univ de Aveiro, Aveiro Portugal, 2 , International Iberian Nanotechnology Laboratory, Braga Portugal, 3 Ångström Solar Center, Solid State Electronics, Uppsala University, Uppsala Sweden
Show AbstractThe amount of Cu on Cu(In,Ga)Se2 (CIGS) has a profound influence on the opto-electronic and on the structural properties of these layers. Such effect is also reflected in the electrical performance of resulting thin film solar cells. Most of the fundamental studies found in the literature are performed either in specimens based in crystals or in conventional co-evaporated thin films. In the former case, the specimens lack the polycrystalline nature of the thin films that are used in solar cells. In the case of thin films produced by co-evaporation processes, like the three-stage growth process or other in-line processes, they usually present thin films that have a [Ga]/([Ga]+[In]) depth profile and/or Cu-rich to Cu-poor transitions during the different growth stages. To understand in detail the effect of variations in the Cu content, it is necessary to isolate this effect from the variation of other elements and to avoid Cu-transitions during its growth. Hence, in this work, we prepared a set of CIGS samples using flat evaporation rates which produced films with a constant value of [Ga]/([Ga]+[In]) (GGI) in depth and without any Cu-transitions. The samples were studied using several techniques: X-ray fluorescence (XRF), J-V, external quantum efficiency (EQE), x-ray diffraction (XRD), scanning electron microscopy/energy dispersive spectroscopy (SEM/EDS), photoluminescence (PL), and Raman spectroscopy. The [Cu]/([Ga]+[In]) (CGI) contents studied were 0.53 (10.5 % efficiency), 0.71 (13.8% efficiency), 0.85 (15.1% efficiency) and 0.98 (14.5% effciency) with a fixed GGI of 0.3. For lower Cu contents, the electrical performance lowers together with a lowering of the crystal quality and an increase of the effect of fluctuating potentials, as observed by PL. EQE measurements show that between CGI values of 0.53 and 0.85, the bandgap energy values increase with increasing Cu content. A relation between the electrical performance and the morphological and opto-electronic properties for multi-crystalline samples with depth-flat Ga ratios and no Cu-transitions is demonstrated.
9:00 PM - ES14.4.15
Characterization of CIGS Solar Cells through Glow Discharge Optical Emission Spectrometry and Differential Interferometry Profiling
Matthieu Chausseau 1 , Anais Loubat 2 , Simon Richard 3 , Muriel Bouttemy 2 , Kayvon Savadkouei 1 , Philippe Hunault 1 , Sofia Gaiaschi 3 , Patrick Chapon 3 , Arnaud Etcheberry 2
1 , HORIBA Scientific, Edison, New Jersey, United States, 2 , Lavoisier Institute of Versailles, Versailles France, 3 , HORIBA Scientific, Palaiseau France
Show AbstractGlow Discharge Optical Emission Spectrometry (GD-OES) provides direct measurement of the chemical composition of materials as a function of depth, with nanometer resolution and the capability to measure both thin and thick layers.
It consists in a pulsed radiofrequency glow discharge plasma source that is sputtering a large area of the material of interest and real time detection by a high resolution optical spectrometer of the sputtered species excited by the same plasma. All elements from H to U can be measured using this technique.
The use of an advanced pulsed RF source allows the measurements of both conductive and non-conductive samples, addressing a wide range of applications for materials science.
It is known that the variation of composition, the diffusion of impurities and the nature of interfaces are critical for the performance of devices making GD-OES an essential tool to optimize engineering processes.
Data obtained on CIGS cells before and after chemical etching will be presented, including data obtained with the Differential Interferometry Profiling (DiP), GD-OES most recent innovation, allowing for real time measurement of layers thicknesses. Thicknesses obtained using DiP will be compared with data from standard layer thickness measurement technology.
9:00 PM - ES14.4.16
Luminescence Detection of the 0.8eV Defect
Conrad Spindler 1 , Susanne Siebentritt 1
1 , University of Luxembourg, Belvaux Luxembourg
Show AbstractPhotocapacitance measurements from literature show a defect at about 0.8eV from the valence band which appears in all Cu(In,Ga)Se2 films independent of the Ga-content. The level is not observed in admittance spectroscopy. It has been argued that this defect can only be observed under illumination, since it represents the -1/-2 transition of the CuIII antisite defect. Therefore the transition should be observable by photoluminescence. In fact, an extension of our photoluminescence set-up allows to detect low energy transitions. We observe a defect level around 0.7-0.8 eV in numerous studied samples. The transition energy appears to be independent of the Ga/(Ga+In) ratio. Since this defect level is close to mid-gap in CuGaSe2 (Eg=1.68 eV) it can also play a major role for the open circuit voltage deficit in wide-gap compounds of Cu(In,Ga)Se2.
9:00 PM - ES14.4.17
Role of Pre-Layer Mo Films in Microstructural and Morphological Properties of Over-Layer CIGS Films
Hamda Al-Thani 1 , Falah S. Hasoon 1
1 , National Energy & Water Research Center (NEWRC), Abu Dhabi United Arab Emirates
Show Abstract
This study focuses on establishing a microstructural and morphological correlation between Cu(In,Ga)Se2 (CIGS) films and its precursor layer of Molybdenum (Mo). Therefore, in this work variations in the morphology and microstructural properties of Mo thin films, grown on SLG substrates using DC planar magnetron sputtering, were induced systematically by varying the growth sputtering pressure from 0.6 to 16 mT with a sputtering power density of 1.2 W/cm2. Subsequently, under fixed deposition conditions (deposition rate and substrate temperature), a growth of CIGS films was carried out on the previous Mo coated SLG substrates, using the physical vapor deposition (PVD) technique, and the 3-stage growth process. High-resolution scanning electron microscopy (HRSEM) and atomic force microscopy (AFM) were used to examine the CIGS films morphology and surface roughness. X-ray diffraction (XRD) was applied to study in detail the microstructure of CIGS films. Where, the film crystal structure including the preferred orientation and lattice parameter were determined by the θ/2θ XRD technique and by applying Cohen’s least-square method. Also, the grazing incidence X-ray diffraction technique (GIXRD) was employed to depth profile the out-of-plane crystallographic texture in the CIGS films. Results showed that the Mo film microstructural properties have a strong influence on the CIGS film microstructural properties and in providing a suitable environment to supply the CIGS film with the Na from the SLG substrate.
Preliminary, the results showed that: for low Mo film growth sputtering pressure (P ≤1 mT); the CIGS film had a lamellar type microstructure accompanied by tightly packed, faceted grains. XRD patterns of CIGS film indicated secondary phase segregation with low diffraction lines intensities. The CIGS film unit cell sizes were larger than those of the standard powder sample. θ/2θ XRD revealed that the films exhibited a near-random texture, while GIXRD showed (112) preferred orientation near the surface. Whereas, for intermediate Mo film growth sputtering pressure (8 ≥ P ≥ 2 mT); the CIGS film unit cell size continued decreasing, reaching a minimum size as the Mo film sputtering pressure reached the upper limit of this range of sputtering pressure. The CIGS film had {(220),(204)} preferred orientation indicated by θ/2θ XRD data, whereas GIXRD data indicated (112) preferred orientation near the film surface, and {(220)/(204)} preferred orientation within the film. Last but not least, for high Mo film growth sputtering pressure (P > 8 mT). The unit cell exhibits further relaxation in both ao and co lattice parameters. θ/2θ XRD showed that the CIGS film have {(220),(204)} preferred orientation, while GIXRD showed no evidence of a change in this texture near the surface. However, further data and interpretation of the results of this research work will be presented in the detailed paper, where a relation between Mo and CIGS films shall be emphasized.
9:00 PM - ES14.4.18
Compositional Influence of the Cationic Solution in CZTS Thin Films Deposited by SILAR for Solar Cells Applications
Shadai Lugo Loredo 1 , Raquel Garza Hernandez 1 , Francisco Servando Aguirre Tostado 1
1 , Centro de Investigación en Materiales Avanzados S. C. Unidad Monterrey, Apodaca Mexico
Show AbstractCu2ZnSnS4 (CZTS) is a promising quaternary absorber material for thin film solar cells due to the non-toxicity and elemental abundance of its constituents. CZTS thin films were deposited by Successive Ionic Layer Adsorption and Reaction (SILAR) method on top Mo-coated glass substrates. The effect of metals concentration on the structural, chemical and morphological properties of the CZTS were investigated. Thin films were prepared using metal chlorides as the cationic precursor and sodium sulfide as the anionic precursor, applying 120 cycles and a thermal annealing at 450 C under N2 under a mixture of N2/S2 flow. Several techniques, including X-ray diffraction (XRD), Scanning Electron Microscopy (SEM), X-ray Photoelectron Spectroscopy (XPS) and Raman Spectroscopy have been used to interrogate the chemical composition and crystalline structure of the CZTS films. The ratio [Cu/Zn+Sn] and [Zn/Sn] were determined by XPS Ar+ sputter depth profiling analysis, revealing that for the annealed films these ratios were around of 1-1.5 and 0.5-1, respectively. Also, the high stability of complexes formed in the solution improves the homogeneity and uniformity observed by SEM. XRD analysis revealed the formation of tetragonal Cu2ZSnS4 crystal structure. The films obtained showed appropriate values for their applications as an absorbing layer in photovoltaics structure of the type glass/Mo/CZTS/CdS/i-ZnO/ITO.
9:00 PM - ES14.4.19
Photoconductive Properties of Nano-Flakes Assembled Porous Microspheres CuInS2: Cd2+, V3+ Thin Films via Hydrothermal Method on Spray Seed Coated Substrates
Logu T 2 3 , Ramesh Raliya 3 , Shalinee Kavadiya 3 , Soundarrajan Palanivel 2 , Pratim Biswas 3 , Arunava Gupta 1 , K. Sethuraman 1 2
2 Physics, Madurai Kamaraj University, Madurai, Tamilnadu, India, 3 Department of Energy, Environmental and Chemical Engineering, Washington University in St. Louis, St. Louis, Missouri, United States, 1 , University of Alabama, Tuscaloosa, Alabama, United States
Show AbstractPristine, cadmium (Cd) and vanadium (V) doped CuInS2 (CIS) Nano-Flakes assembled Porous Microspheres (NFPMs) thin films were grown on spray seed coated substrates using hydrothermal method. The structural, morphological, optical and electrical properties of pristine, Cd and V doped NFPMs CIS thin films have been systematically examined. The scanning electron microscopic (SEM) images show the well-interconnected NFPMs changed to Elongated Nano-Flakes assembled Porous Microspheres (ENFPMs) morphology with respect to V and Hard Microspheres (HMs) due to Cd doping level and also the size of the nanoflakes in the microspheres increases with V doping level. Pristine, Cd and V doped NFPMs CIS thin films exhibit the body centered tetragonal crystal structure along with polycrystalline nature which is characterized by means of X-ray diffraction (XRD). The presence of dopant and host elements is confirmed by X-ray photoelectron spectroscopic (XPS) analysis. The decreases of optical energy gap values are observed using UV- Vis- NIR absorption spectra according to V and Cd doping level into the CIS lattice sites. The photoelectric response of the sample has also been studied which shows significant photo current for the V doped ENFPMs CIS thin films. It anticipates that this kind of self-assembled V doped ENFPMs CIS thin films will give huge interest for use as an absorber layer in solar cell devices.
9:00 PM - ES14.4.20
Evaporated CdIn2S4 Buffer Layer for Kesterite Solar Cells
Leo Choubrac 1 , Ludovic Arzel 1 , Guy Brammertz 2 3 , Marc Meuris 2 3 , Bart Vermang 4 5 , Erik Ahlswede 6 , Thomas Schnabel 6 , Nicolas Barreau 1
1 , Institut des Matériaux Jean Rouxel (IMN), Université de Nantes, Nantes France, 2 , imec Division IMOMEC – Partner of Solliance, Dipenbeek Belgium, 3 , Institute for Material Research (IMO) Hasselt University, Diepenbeek Belgium, 4 , imec – Partner in Solliance, Heverlee Belgium, 5 , Department of Electrical Engineering (ESAT), KU Leuven, Heverlee Belgium, 6 , Zentrum für Sonnenenergie- und Wasserstoff-Forschung, Stuttgart Germany
Show AbstractCdIn2S4 material was recently suggested as new buffer material for thin film solar cells, since it was observed to be formed at the interface of KF-treated CIGS/CdS [1].
Moreover, improved VOC and 12.7% efficient cells have been reported for CZTSSe/CdS/In2S3 devices [2]. The authors of that work report that RTA annealing is mandatory to get the VOC improvement. The role of this RTA annealing is to induce some indium migration into the CdS and the absorber layers. It might happen that some CdIn2S4 is formed during this process, and this might be the reason of VOC deficit reduction.
In addition, CdIn2S4 material has higher band gap than CdS, suggesting potential improvement of current through reduced absorption in the buffer layer. Moreover, it can be deposited by PVD at moderate substrate temperatures (200°C).
CZTSe/CdIn2S4 and CZTSSe/CdIn2S4, CZTSe/CdS and CZTSSe/CdS solar cells were prepared (no ARC). Then, the impacta of aging and annealing on VOC, collection length and efficiencies have been studied. Main results are:
Aging and annealing can improve collection length, VOC and efficiency of CdS-buffer references, leading to improved-CdS devices
Any annealing/aging doesn’t lead to such improvements on CdIn2S4 devices
CZTSSe/CdIn2S4 best cell has 7.5 % efficiency & 430 mV VOC, ie slightly improved open circuit voltage compare to the CZTSSe/improved-CdS (419 mV, 8.9 %)
CZTSe/CdIn2S4 best cell has 6.7 % efficiency & 320 mV Voc. 370 mV cells have been obtained but with low FF. The reference CZTSe/improved-CdS is 8.5% efficient with 400 mV VOC
Main conclusions are:
CdIn2S4 is a promising buffer material for kesterite solar cells, and there is still room for approaching CdS efficiencies
By comparing the effect of annealing/aging on CdIn2S4 vs CdS, we can provide a better understanding of those critical phenomena
Acknowledgement:
This work was performed within the SWInG project that has received funding from the European Union’s Horizon 2020 research and innovation programme under grant agreement No 640868.
[1] Lepetit et al, EMRS IEEE Journal of photovoltaics, DOI: 10.1109/JPHOTOV.2016.2589365
[2] Kim et al, Advances Materials, DOI: 10.1002/adma.201402373
9:00 PM - ES14.4.21
Back Contact Functionalization by Transition Metal Oxides for Kesterite Solar Cells
Moises Espindola-Rodriguez 1 , Dioulde Sylla 1 , Haibing Xie 1 , Yudania Sanchez 1 , Markus Neuschitzer 1 , Sergio Giraldo 1 , Victor Izquierdo 1 , Alejandro Perez-Rodriguez 1 2 , Edgardo Saucedo 1 , Marcel Placidi 1
1 , Catalonia Institute for Energy Research (IREC), Sant Adrià de Besòs Spain, 2 , IN2UB, Barcelona Spain
Show AbstractSeveral studies highlight that Mo is probably not the optimal metallic back contact for Kesterite, due to the non-optimum Kesterite/Mo band alignment that is partially bridged over by the interfacial formation of MoS(e)2; evidence of thermodynamic unstability during the thermal processing. In this work we present the proof of concept of introducing functionalized transparent conductive oxides (FOX) as electrical back contact for Cu2ZnSn(S,Se)4 (CZTSSe) solar cells in search of chemically stable interfaces. Additionally, FOX substrates, extends the applicability of Kesterite solar cells by substituting the standard glass/Mo back contact towards FOX back contacts for bifacial solar cell concepts. In this work several FOX back contacts are analyzed within the Glass/FOX/CZTSSe/CdS/i-ZnO/ITO solar cell layering.
In the first part of the work the functionalization of the back contact was achieved by the sputtering of thin layers (5-20 nm) of Mo or Mo:Na, on the top of SnO2:F (FTO) substrates. A record device on a FTO/Mo:Na functionalized substrate with efficiency (Jsc, Voc) as high as 6.3%-front (28 mA/cm2, 408 mV), 1.1%-rear (6.1 mA/cm2, 341 mV) and 7.7%-bi (35 mA/cm2, 412 mV) was obtained through the optimization of the annealing temperature as well as the Ge and Na doping parameters. In the second part of the work, various transition metal oxides like MoO2, MoO3, NiO, TiO2, V2O5, CuO, and Co3O4 are used to functionalize the back contact of the kesterite solar cells by the insertion of a thin (5-20 nm) semi-transparent oxide layer (by thermal evaporation) onto clean FTO substrates. The results reveal that the device with the MoO3 hole-extractor layer performs better than the other transition metal oxides; with an efficiency (Jsc, Voc) of 3.8%-front (27 mA/cm2, 343 mV), and 4.2%-bi (30 mA/cm2, 344 mV).
Comparing the EQE of the standard Mo and new FOX back contacts, an improved carrier collection at long wavelengths is systematically observed for the last case, indicating a less recombination at the back interface and/or a better band alignment between the FOX contacts and kesterite. Then, by the implementation of the functionalized substrates presented in this work, we demonstrate a completely new route towards better carrier collection, opening new perspectives for solving one of the possible limiting factors of kesterite devices efficiency: the non-optimized back contact. The currently limiting factor of this novel approach for better carrier collection is the high series resistance coming either from the lower conductivity of the TCO with respect to Mo, and/or from the partial selenization of the FTO substrate. Possible technological solutions will be discussed.
9:00 PM - ES14.4.22
Electronic Structure of Cu(In,Ga)(S,Se)2 Surface and CdS/Cu(In,Ga)(S,Se)2 Interface
Suehiro Kawamura 1 , Kenta Kawasaki 1 , Keisuke Isowaki 1 , Shin'ichi Takaki 1 , Takuya Shimamura 1 , Takuya Kato 2 , Hiroki Sugimoto 2 , Norio Terada 1
1 , Kagoshima University, Kagoshima, Kagoshima, Japan, 2 , Solar Frontier K. K., Atsugi, Kanagawa, Japan
Show AbstractHigh conversion efficiencies above 22% have been recently achieved by modification of the surface nature of the Cu(In, Ga)(S, Se)2 (CIGSSe) absorber and/or of the buffer-structure in the CIGSSe-based solar cells. Experimental determination of electronic structure of the CIGSSe surface and the buffer/CIGSSe interface in such cells are essential for further optimizations. In the present study, we have studied the surface nature of the partially sulfurized CIGSSe absorber and the band alignment at the interface between CdS buffer and the CIGSSe by in-situ X-ray, ultraviolet photoemission and inverse photoemission spectroscopy.
The CIGSSe layers were fabricated on Mo-coated glass substrates by selenization and subsequent sulfurization of metal precursors. The cells using the identically prepared CIGSSe and CdS buffer by chemical bath deposition (CBD) showed the efficiency above 20%. Surfaces of the CIGSSe were cleaned by using an Ar+ beam with a low kinetic energy of 50–100 eV. This treatment made residual levels of contaminations equivalent to those on the CIGSSe without air-exposure, whereas any change in chemical bond natures of the constituents was undetectable. Then, the CdS layers were deposited by in-situ step-evaporation. The thick CdS showed identical electronic structure with the CBD-CdS. Band offsets at the interface were determined from the energy differences in the band edges of the CIGSSe surface and the thick CdS, and the interface induced band bending (iibb) correction. The compositional ratio of S/(S+Se) and (S+Se)/(Cu+In+Ga) of the CIGSSe surface are 0.6 and 1.0, respectively. Valence band maximum (VBM) and conduction band minimum (CBM) relative to Fermi level of this surface are -0.70 and +0.8 eV, respectively. The high sulfur substitution mainly causes the expansion of band gap energy, the deepening of VBM and weakened p-type feature. The deep VBM should act as the hole-reflector that is effective to reduce the interface recombination. The CBM of the 30–50 nm thick CdS on the CIGSSe is +0.4 eV. So, a large and downward iibb is necessary for forming the “spike” type conduction band alignment. In the early stage of the CdS-deposition, the electronic band slightly rises by +0.11 eV. Contrarily, further deposition of the CdS up to 30–50 nm makes the Cd core signals deeper by 0.55 eV. Consequently, the iibb at this interface is downward and large of 0.44 eV, which is consistent with the weak p-type feature of the CIGSSe surface. Since the CBM of the thick CdS is lower than that of the CIGSSe surface by 0.4 eV, the conduction band offset (CBO) at this interface is slightly positive of +0.04 eV. These results reveal that the conduction band alignment at the CdS/CIGSSe interface is almost “flat” or “small spike”. This connection is beneficial to both suppressing interface recombination and current loss. Although the CBO is slightly smaller than the optimum, the observed electronic structure is consistent with the high performance.
9:00 PM - ES14.4.23
Band Alignment of CdS/Cu2ZnSnSe4 Heterointerface and Solar Cell Performances
Takehiko Nagai 1 , Shinho Kim 1 , Hitoshi Tampo 1 , Kang Min Kim 1 , Hajime Shibata 1 , Shin'ichi Takaki 2 , Kenta Kawasaki 2 , Suehiro Kawamura 2 , Takuya Shimamura 2 , Koji Matsubara 1 , Shigeru Niki 1 , Norio Terada 2
1 Research Center for Photovoltaics, National Institute of Advanced Industrial Science and Technology (AIST), Tsukuba, Ibaraki, Japan, 2 Graduate School of Science and Engineering, Kagoshima University, Korimoto, Kagoshima, Japan
Show AbstractIn this study, we studied the relationship between electronic structure of the CdS and Cu2ZnSnSe4 (CTZSe) heterointerface and cell performances. Especially, we experimentally evaluated the electronic structure of CdS/CZTSe heterointerface by using the x-ray photo-emission spectroscopy (XPS), inversed photo-emission spectroscopy (IPES) and ultra-violet photo-emission spectroscopy (UPS) techniques, whose measurement systems connected with CdS deposition chamber via transport chamber in high vacuum. Becasue the band alignment at the heterojunction interface plays a crucial role for the development of high performance solar cells.
We deposited the CZTSe absorber layers on Mo coated soda-lime glass by using the co-evaporation technique, and they were thermally annealed by another furnace. Prior to introduction to the load lock chamber (LL), CZTSe films were treated by the NH3 aqueous solution in an Ar gas circulated glove box combined with the LL. In each step of CdS deposition, we measured the XPS, UPS and IPES without air exposure. The band offsets have an error tolerance of ±0.15 eV in our analysis.
We measured the XPS signals originating from the Cu, Zn, Sn and Cd at the CdS film thickness with a few nm to estimate the magnitude of the interface induced band bending (iibb) of the CZTSe absorber layer (iibbCZTSe) and the CdS buffer layer (iibbCdS). These peaks shifted toward the deeper binding energy with increasing the CdS thickness. This result suggests a downwards bands bending of CZTSe and CdS layers. Here we notice that the iibbCdS at CdS thickness of 3 nm, which corresponds to the energy difference between the binding energy of core level of the Cd(3d5/2) at the CdS thickness of 3 nm and that of 30 nm in XPS signal, is really large and becomes +0.45 eV. The total iibb (iibbCZTSe + iibbCdS) of CdS/CZTSe results in +0.62~0.66 eV. Moreover, we determined the valence band maximum (VBM) and the conduction band minimum (CBM) of CZTSe film and deposited CdS film on CZTSe by using the UPS and IPES. The VBM drastically shifts by CdS deposition, whereas the CBM slightly shifts. Finally, we evaluated the magnitude of the conduction band offset (CBO) and the valence band offset (VBO) at CdS/CZTSe interface were +0.50 eV and +0.85 eV, respectively. We concluded that the conduction band alignment at CdS/CZTSe interface has a large spike structure because this CBO is positive value. Surprisingly, despite of building such a CBO, the CdS/CZTSe heterojunction solar cells with relatively high conversion efficiencies (~8 %) can be obtained. Moreover, we reveal that the conversion efficiency can be improved by rising up the CBM and/or reduction of CBO, whereas the dominant progress process is improvement of film quality in bulk CZTSe. The relationship between solar cell performance and band alignment of CdS/CZTSe and /CZTGeSe will be discussed.
9:00 PM - ES14.4.24
Cu2ZnSn(S,Se)4 Surface Modification by Epitaxial Growth of Al(OH)3 Nanolayers—Impact on the Devices Efficiency
Haibing Xie 1 , Yudania Sanchez 1 , Pengyi Tang 1 2 , Moises Espindola-Rodriguez 1 , Maxim Guc 1 , Lorenzo Calvo-Barrio 3 4 , Simon Lopez-Marino 1 , Yu Liu 1 , Juan Morante 1 4 , Andreu Cabot 1 5 , Victor Izquierdo 1 , Jordi Arbiol 2 5 , Alejandro Perez-Rodriguez 1 4 , Edgardo Saucedo 1
1 , IREC, Barcelona Spain, 2 , Catalan Institute of Nanoscience and Nanotechnology (ICN2), CSIC and The Barcelona Institute of Science and Technology (BIST), Bellaterra, Barcelona, Spain, 3 , Centres Científics i Tecnològics de la Universitat de Barcelona (CCiTUB), Barcelona Spain, 4 , IN2UB, Departament d’Electrònica, Universitat de Barcelona, Barcelona Spain, 5 , ICREA, Barcelona, Barcelona, Spain
Show AbstractVoc deficit is considered as the main hurdle for kesterite (Cu2ZnSn(S,Se)4 or CZTSSe) solar cells, with band tailing and bulk defects as the prevalent causes. However, for S containing kesterite solar cells such as Cu2ZnSn(SxSe1-x)4 (0 < x < 1) and Cu2ZnSnS4 types, the CZTSSe/CdS interface can be another plausible reason, being systematically reported as the source of dominant recombination even in highly efficient devices, including the current world record kesterite solar cell. A “cliff” like conduction band offset (CBO) is not likely responsible for the interface recombination in the record kesterite solar cell because it has an optimal “spike” CBO. Thus, the interface defects caused by dangling bonds or structural defects, e.g., lattice mismatch between buffer layers and absorbers could be a possible reason. In addition, FF deficit is another hurdle for kesterite solar cells, which is related to Voc deficit, series and shunt resistance loss. In particular, low shunt resistance is systematically reported in solar cells, due to shunt paths caused by structural damages or defects like pinholes, trenches, stacking faults and non-coverage points of buffer layers in the interface region. Therefore, engineering the CZTSSe/CdS interface could be a reasonable and effective way to reduce the Voc and FF deficit and further improve the efficiency of CZTSSe solar cells.
In this work, CZTSSe/CdS interface is engineered by epitaxial growth of Al(OH)3 nanolayers using a facile wet chemical route based on aqueous solution of AlCl3 and thioacetamide (AT). After the AT chemical treatment, in average the Voc and FF of the devices are improved relatively by about 7% and 15%, respectively, while Jsc is almost unaffected, leading to approximately 25% increased efficiency. A Se-rich CZTSSe device with 9.1% maximum efficiency (neither metallic grid, nor ARC were used) and FF over 69% (comparable to the world record CZTSSe device) was obtained after systematic optimization of the concentration and soaking time of the wet chemical treatment. The substantial Voc and FF enhancement is mainly associated to less interface recombination and reduced shunt paths, as well as higher surface band gap and shallower donor levels, supported by J-V curves and temperature dependence of J-V and PL measurements. Atomic resolution HAADF-STEM combined with electron energy loss spectroscopy (EELS) mapping reveal the Al distribution and further the epitaxial relationship of Al(OH)3 with CZTSSe and CdS, indicating the benign and effective interface passivation by Al(OH)3 nanolayers. Finally, insights into efficiency limiting and beneficial factors related to the bulk and the front/back interfaces, including structural defects like grain boundaries and stacking faults as well as secondary phases/inclusions/voids will also be presented. This wet chemical route might open new perspectives for interface engineering to decrease Voc and FF deficit for high efficiency CZTSSe solar cells.
9:00 PM - ES14.4.25
Zn1-xSnxOy/Cu2ZnSnS4 Interface and Its Chemical Structure Studied by Soft X-Ray Spectroscopies
Bridget Elizan 1 , Dirk Hauschild 2 3 , Tove Ericson 4 , Tobias Torndahl 4 , James Carter 1 , Monika Blum 1 , Wanli Yang 5 , Charlotte Platzer-Bjorkman 4 , Lothar Weinhardt 1 2 3 , Clemens Heske 1 2 3
1 Department of Chemistry, University of Nevada, Las Vegas, Las Vegas, Nevada, United States, 2 Institute for Photon Science and Synchrotron Radiation, Karlsruhe Institute of Technology, Eggenstein-Leopoldshafen Germany, 3 Institute for Chemical Technology and Polymer Chemistry, Karlsruhe Institute of Technology, Karlsruhe Germany, 4 Ångström Solar Center, Solid State Electronics, Engineering Sciences, Uppsala University, Uppsala Sweden, 5 Advanced Light Source, Lawrence Berkeley National Laboratory, Berkeley, California, United States
Show AbstractThin-film solar cells based on Cu(In,Ga)Se2 (CIGSe) absorbers have already reached efficiencies above 22% [1]. In contrast, Cu2ZnSnS4 (CZTS) kesterite-based thin-film solar cells, using a promising alternative absorber material expected to exhibit similar optical properties, have not yet reached their full potential. The “standard” CdS buffer layer, originally developed for CIGSe devices, is difficult to process in a dry inline production and has been shown to exhibit a cliff-like conduction band offset with CZTS. Thus, alternatives to CdS buffers are being developed. Zn1-xSnxOy (ZTO) is one potential replacement for the CdS buffer layer [2]. However, for an optimization of the electronic properties, a detailed knowledge of the buffer/absorber interface is required. As a first step in the development of a comprehensive picture of this interface, we present a depth-resolved picture of the chemical structure of the ZTO/CZTS heterojunction using soft x-ray emission spectroscopy (XES, bulk-sensitive) and x-ray photoelectron spectroscopy (XPS, surface sensitive). For this purpose, ZTO layers were deposited by atomic layer deposition (ALD) onto CZTS films [2]. We find evidence for a Cu-poor absorber surface, Zn and Sn in an oxidized environment upon buffer layer deposition (as expected), and sulfur and copper-related signals for the thickest buffer layer. The results will be discussed in view of their impact on the electronic interface properties and hence the performance of an overall ZTO/CZTS-based thin film solar cell device.
[1] Jackson et al., Phys. Status Solidi RRL, 1–4 (2016)
[2] Platzer-Björkman et al., APL 107, 243904 (2015)
9:00 PM - ES14.4.26
Bi-Facial CdTe Thin Film Solar Cells Using Nanowire Back Contact for Flexible Applications
Yongbeom Kwon 1 , Eun Woo Cho 1 , Donghwan Kim 1 , Jihyun Kim 1
1 , Korea University, Seoul Korea (the Republic of)
Show AbstractRecently solar cell industry is in the spotlight as a part of future. Especially CdTe thin film solar cells have attractive properties including near optimal band gap (~1.5 eV) for solar absorption, low-cost and simple fabrication process and flexibility, therefore, good for various applications. Reported record of CdTe solar cell efficiency is 22.1% and there is still room to enhance cell efficiency further. Most superstrate structured solar cells focus on the light absorbed from the front contact side, and it doesn’t use the reflected sunlight directed towards the back contact side. Research on bi-facial solar cells with transparent electrodes on the backside could be useful to enhance solar cell efficiency by utilizing both front and back contact sides.
In this work, we used copper nanowire (CuNW), silver nanowire (AgNW) and ITO to fabricate CdTe solar cells with transparent back contact. By using nanowires, contact problem caused by rigid ITO in flexible solar cells could be solved. Sputtering was used to deposit ITO and ZnO layers in order as front contact on glass. CdS, as a window layer, and CdTe layer were deposited by chemical bath deposition (CBD) and close spaced sublimation (CSS), respectively, on glass/ITO/ZnO/CdS substrate. Either CuNW, AgNW or mixture solution was spray coated on CdTe layer and ITO was deposited onto nanowires to form transparent back contact. Density of nanowires and ratio of mixture solution were controlled to find optimum conditions for high efficiency of bi-facial solar cells. The cell performance was maintained even after bending tests because nanowire acts as a bridge that connects damaged ITO layer. The detailed experiments and results will be discussed.
9:00 PM - ES14.4.27
From Bandstructure to Bandalignment—A Study on Chalcopyrite Surfaces
Christian Pettenkofer 1 , Andreas Popp 1
1 , Helmholtz-Zentrum Berlin, Berlin Germany
Show Abstract
Most studies on chalcopyrite interfaces are carried out on technologically prepared interfaces. In this study we start from idealized single crystalline interfaces prepared by MBE, MOMBE etc. under very well defined UHV conditions and investigated in situ by UPS, XPS, LEED, STM.
Starting from the electronic and morphological structure of CuInSe2 (100) and (112) surfaces as measured by LEED and ARUPS we try to understand the interface properties of chalcopyrite junctions. In particular we report on our attempts to model the junction in chalcopyrite thin films by well defined interfaces to clarify the influence of grain boundaries, lateral inhomogenities and chemical variations across and aside the contact plane. Chalcopyrites of the Type CuYX2 (Y=Ga,In; X=S,Se) were grown by MBE as single crystalline samples in various orientations and were studied by surface analytical tools like XPS, UPS, LEED, STM and XPEEM in situ. Especially the application of synchrotron radiation in photo-emission experiments is an extremely powerful tool to gain insight into the morphology and structure of hetero contacts. In a single deposition experiment it is possible to determine the band alignment, band bending, chemical reacted interfaces and their crystalline structure with high accuracy. By following the development of the contact phase to ZnO, ZnSe, ZnS step by step in an UHV environment, all properties of the interface are determined on an atomic scale with high resolution. Beside the formation of an ordered vacancy compound of the absorber the existence of various interfacial layers are detected and their influence on the parameters of a device is discussed.
For CuInSe2 the conversion to Cu poor interface layers is observed by SRXPS induced by the formation of the interface to ZnSe buffer layers. The development of Cu-poor surface phases was discussed by Zunger et al. and is here detected unambiguously by its LEED signature and its band structure E(k) in comparisson to the chalcopyrite.
To determine the band alignment valence band spectra have to be recorded to obtain the valence band onset. Here we will show that the right value can only be obtained by using synchrotron radiation as the correct position of the valence band in k-space has to be determined at the Γ-point.
We will demonstrate the behavior and the specific properties of CuInSe2 CuGaSe2 and CuInS2 and their reactions and junction formation to ZnSe and ZnO.
The complete band alignment from the absorber to the defect-layer to a ZnSe buffer-layer and ZnO will be given for various surface orientations. CuGaSe2 films could be grown from separate Cu, Ga and Se sources or by using the GaAs substrate at elevated temperatures as a Ga source. This result explains also the perfect epitaxial growth of CuInSe2 on GaAs despite the lattice mismatch as at growth temperature of 600°C a graded GaAs- CuGaInSe2 junction is grown.
9:00 PM - ES14.4.28
Cadmium Tin Oxide and Zinc Magnesium Oxide Prepared by Hollow Cathode Sputtering for CdTe Photovoltaics
Alan Delahoy 1 , Shou Peng 2 , Payal Patra 3 , Surya Manda 1 , Akash Saraf 1 , Yunfei Chen 1 , Xuehai Tan 1 , Ken Chin 1
1 , CNBM New Energy Materials Research Center, Department of Physics, New Jersey Institute of Technology, Newark, New Jersey, United States, 2 , China Triumph International Engineering Co. Ltd., Shanghai China, 3 , New Jersey Innovation Institute, Newark, New Jersey, United States
Show AbstractThe efficiency of solar cells based on thin film CdTe depends strongly on the nature and properties of the front stack layers (TCO and buffer). This work targets a high mobility, low absorption TCO and a wide gap buffer layer offering minimal blue loss and an optimized ΔEcabs-buff band offset. Hollow cathode sputtering (HCS) is a versatile, scalable process for metal oxide deposition featuring low lattice damage. Cadmium tin oxide (CTO) has been prepared by HCS with an as-deposited average transmittance of over 86% and an electron mobility μe of 45 cm2/Vs. Both properties are slightly improved after high temperature processing during CTO/CdS/CdTe preparation. CdTe solar cells with CTO front TCO exhibit a Jsc of over 23 mA/cm2. Zinc magnesium oxide (ZMO) can be employed to form high efficiency CdS-free CdTe solar cells having the structure TCO/ZMO/CdTe. ZMO has likewise been prepared by HCS. With a Mg content of 18% (at.), Al-doped ZMO has an as-deposited bandgap of about 3.7 eV and a mobility of 5.2 cm2/Vs (μe 50 cm2/Vs for ZnO:Al). The lower mobility of ZMO relative to AZO is attributed to an increased effective mass and to disorder scattering. This work represents the first use of HCS to prepare the mixed metal oxides CTO and ZMO. Structural characterization of the layers is also planned. Work is currently under way to prepare undoped ZMO for comparison purposes and to further optimize CdTe solar cells incorporating these and other HCS layers.
9:00 PM - ES14.4.29
ALD of ZnO1-xSx Buffer Layer Films
Samual Wilson 1 , Zhi Li 1 , Ho Ming Tong 1 , Ryan Kaczynski 2 , Timothy Anderson 3
1 , University of Florida, Gainesville, Florida, United States, 2 , Global Solar, Tucson, Arizona, United States, 3 , University of Massachusetts Amherst, Amhurst, Massachusetts, United States
Show AbstractCdS is predominantly used as the n-type heterojunction partner (buffer layer) for thin film p-type absorber layers (e.g., CIGS, CdTe, and CZTS). Alternative buffer layer materials have been explored to achieve a more optimal band alignment, reduce Cd content, and avoid wet batch processing via chemical bath deposition (CBD). Zinc oxysulfide (ZnO1-xSx) is a potential buffer layer material that allows adjustment of the band alignment via composition, gives flexibility in the deposition method and obviously avoids Cd. The wider bandgap of ZnO1-xSx provides higher transmission through the buffer layer. Furthermore, it should be possible to grown thinner films with ALD than with CBD and still achieve complete conformal coverage to further increase transmission.
In this work, ZnO1-xSx thin films were deposited, characterized, and device tested as an alternative buffer layer material. The film was deposited using a custom ALD reactor using reactants diethylzinc, H2S, and H2O. The temperature and pulse time ALD windows were first established. ZnO1-xSx films were then deposited across the full composition by varying the ratio oxide to sulfide steps in a super cycle (i.e. (diethylznic-purge-H2O-purge)x(diethylznic-purge-H2S-purge)y, where x and y are number of times each Zn-O and Zn-S cycle are respectively executed), and the thickness was varied by number of super cycles. XPS studies of the deposited films indicates that the concentration of sulfur in the buffer layer film deviated negatively from the precursor pulse ratio x/(x+y). The influence of the sulfur-oxygen ratio on the optical properties of ZnO1-xSx thin films was investigated. The minimum bandgap can be observed on the film with 1:1 H2O/H2S pulse ratio, corresponding sulfur composition x=0.03 in the film. Further increasing the sulfur composition leads to the higher bandgap energy. Additionally, the optimum thickness of ZnO1-xSx for CIGS was explored. Test cells were completed by sputter depositing aluminum-doped zinc oxide (AZO) on the ZnO1-xSx ALD buffer layer, followed by deposition of an Al/Ni contact grid using e-beam evaporation. The thickness of the buffer layer shows monotonically increasing correlation with the number of super cycles. Cross sectional SEM images reveal good conformity, uniformity and coverage was achieved with 39 cycles of deposition (<30 nm). A 17% improvement of VOC was observed when the number of cycles reduced from 91 to 39.
9:00 PM - ES14.4.30
DFT Calculations and XPS Analysis of the Adsorption of Transition Metal-Citrate Complexes on a CdS Surface for the Deposition of Cu2ZnSnS4 Thin Films
Raquel Garza Hernandez 1 , Shadai Lugo Loredo 1 , Mario Sanchez Vazquez 1 , Francisco Servando Aguirre Tostado 1
1 , Centro de Investigación en Materiales Avanzados S. C. Unidad Monterrey, Nuevo Leon Mexico
Show AbstractThe adsorption of M-citrate (M= Cu, Sn, Zn) complexes and sulfide anions on CdS surface has been investigated using theoretical calculations based on density functional theory (DFT) and experimentally using X-ray photoelectron spectroscopy (XPS) measurements. A theoretical study based on the geometric optimization and the energy interaction of metal-citrate and sulfide anions with a CdS surface is the key to understand the nucleation and reaction mechanisms of Cu2ZnSnS4 (CZTS) thin films when they are grown by the Successive Ionic Layer Absorption and Reaction (SILAR) method. Several orientations of the metal-citrate complexes on the surface of the adsorbent have been considered. XPS experiments were performed on an ESCALAB250Xi system using a monochromatic Al-Kα radiation (1486.7 eV) with an energy resolution of 0.25 eV as the excitation source. The initial stages of the growth of CZTS thin films was studied by controlling the SILAR deposition cycles (from 1 to 10). The XPS analysis revealed the presence of C-O bonds related to citrate ions used as complexing agent what proves that there is a mechanism of ion exchange of citrate by sulfide ions. The copper citrate complex remains preferably in the surface, even when the amount of copper is 5 times smaller than Sn and Zn. After 10 SILAR cycles the Cu:Zn:Sn:S stoichiometry ratio reached is approximately 2:1:1:4.
9:00 PM - ES14.4.31
A Comparative Study of CdTe/CdS Junction Activation Using MgCl2 and CdCl2
Gonzalo Angeles-Ordonez 1 , Eulisis Regalado-Perez 1 , Clara Mendoza Gonzalez 1 , Marisol Gutierrez 1 , Jose Jeronimo-Rendon 1 , Martin Reyes-Banda 1 , Nini Mathews 1 , Xavier Mathew 1
1 , Instituto de Energías Renovables, Universidad Nacional Autónoma de México, Temixco, Morelos, Mexico
Show AbstractCdTe/CdS junction activation was performed using MgCl
2 and CdCl
2 with the intention to replace the traditional and toxic CdCl
2 in the solar cell processing. In order to obtain better control on the chloride salt loading over the films, we opted for the dipping and annealing procedure, where same molar solutions of both salts were used. Emphasis was given to explore the elemental profiles of S, O, Cl, Mg, and Cu across the CdTe as well as at the junction. Sulfur diffusion into the bulk of CdTe and evidence of alloying at the junction were more prominent in the case of devices activated with CdCl
2 indicating that CdCl
2 is more efficient in catalyzing the recrystallization. However, the device parameters were comparable in both cases or slightly better for devices activated with MgCl
2. The overall performance of devices activated with MgCl
2 is explained as due to less recombination losses at the interface, lower series resistance and higher carrier concentration at the back contact region. In addition, it was observed that the dispersion in efficiency values were smaller for devices activated with MgCl
2 indicating better spatial uniformity.
Acknowledgement: This work is part of the project CeMIE-Sol 207450/P25
* Corresponding author (X. Mathew); e-mail:
[email protected]9:00 PM - ES14.4.32
Structural and Optical Properties of CdSxSe1-x Thin-Film Chalcogenide Glass
Pawan Kumar 1 3 , Rajesh Katiyar 1 , Aravind Kumar 2 , Balram Tripathi 1 , Ram Katiyar 1
1 Department of Physics, University of Puerto Rico San Juan, San Juan, Puerto Rico, United States, 3 Department of Physics, Gurukula Kangri University Haridwar, Haridwar India, 2 Department of Physics, Kalindi College, University of Delhi, Delhi India
Show AbstractCdSxSe1-x Chalcogenide glass was prepared by melt quenching technique. A thin film of the as-prepared material on the glass substrates is deposited by thermal evaporation technique under vacuum better than 10-5 torr. The samples were structurally characterized using X-ray diffraction (XRD), Energy Dispersive X-Ray (EDX), & Differential Scanning Calorimetric (DSC). Optical parameters, such as refractive index (n), extinction coefficient (k) were calculated by using Transmittance Spectra in the range of 800nm - 1800nm.
Keywords: Chalcogenide glasses, Optical band gap, dielectric constant, EDX & DSC.
9:00 PM - ES14.4.33
Deposition Kinetics of Zinc Oxide Thin Films by Magnetron Sputtering
Yifei Sun 1 , Nadia Foo Kune 1 , James Doyle 1
1 , Macalester College, Saint Paul, Minnesota, United States
Show AbstractAluminum-doped zinc oxide ZnO:Al is of great interest as a transparent electrode for chalcogenide solar cells. Here we present a study of the deposition kinetics of radio frequency magnetron ZnO:Al as a function of working gas pressure and substrate temperature. Characterization methods include growth rate, x-ray diffraction, optical transmission, resistivity, chemical composition using energy dispersive spectroscopy, and morphological characterization by atomic force microscopy and scanning electron microscopy. At low substrate temperatures the oxygen to zinc ratio varies from stoichiometric at low pressures through a minimum at around 5 milliTorr (target-substrate distance 15 cm) followed by a slow recovery to stoichiometric composition at higher pressure. We interpret this trend as a transition from ballistic arrival of precursors at low pressure to diffusive arrive of precursors at higher pressure. With increasing substrate temperature these trends are not as pronounced, implying temperature activated surface reactions play an important role in film composition. A simple kinetic model is proposed to explain these trends.
9:00 PM - ES14.4.34
Effects of Gas Flow Rate on the Properties of SnO2 Thin Films Deposited by RF Sputtering
Muntaser Al-Mansoori 1 , Sahar Al-Shaibani 2 , Ahlam Al-Jaeedi 2 , Jisung Lee 1 , Daniel Choi 1 , Falah S. Hasoon 2
1 Department of Mechanical and Materials Engineering, Masdar Institute of Science and Technology, Abu Dhabi United Arab Emirates, 2 , National Energy and Water Research Center (NEWRC), Abu Dhabi United Arab Emirates
Show AbstractSince last decade the solar cell industry is growing rapidly, with an annual growth rates excess of 30% per year, and Photovoltaic (PV) technology is one of the key answers for a better sustainable future. Important components in the structure of common PV cells are the window and buffer layers. These layers are made out of transparent conductive oxide (TCO) materials, and understanding the nature of TCO materials helps to lead to the scientific advancements of this field. A widely applied TCO is tin oxide (SnO2); used specially in energy applications such as efficient window coating, batteries, and PV cells. Advantage of using SnO2 comes from its high stability, and it is inexpensive in terms of raw materials and processing techniques. In this study, we present our experiment of depositing thin-film SnO2 layer (2000-3000 Å) on soda lime glass (SLG) substrate by utilizing magnetron rf-sputtering technique from SnO2 target while varying the Ar inert gas flow rate and oxygen content. Sputtering is one of the preferred thin film deposition techniques compared to others because it enables good adhesion, scalability, larger range of material choices, and better step coverage. The deposited films’ structural, electrical, and optical properties are characterized using Atomic Force Microscopy (AFM), X-Ray Diffraction (XRD), X-Ray Photoelectron Spectroscopy (XPS), Hall Effect, and UV-Vis-NIR Spectrometer, respectively. Also, Dektak Surface Profilometer was used to determine the films thickness. Preliminary results showed that sheet resistance values obtained for the SnO2 films spanned across four orders of magnitude in their as-deposited amorphous conditions, and transmittance values extending from 47 to 90% within the visible light spectrum. However, in the detailed paper more results shall be explained in relation to the films structural and electrons interactions behavior.
9:00 PM - ES14.4.35
Micropatterned Oxide Layers for Front Contact Passivation in CdTe-Based Thin-Film Photovoltaics
Jason Kephart 1 , Anna Kindvall 1 , Darius Kuciauskas 2 , Desiree Williams 1 , Seth Thompson 1 , W. S. Sampath 1
1 , Colorado State University, Fort Collins, Colorado, United States, 2 , NREL, Golden, Colorado, United States
Show AbstractWhile silicon photovoltaics predominate the current market, CdTe is the second most common absorber material. CdTe has lower cost-per-Watt and superior low-light and high-temperature field performance, and this direct-gap material requires approximately 1% as much material as silicon to absorb the same amount of light. Recently, record efficiency of over 22% has been reported for small-area devices by alloying the absorber with CdSe. In polycrystalline thin-films, the effective carrier lifetime is dependent on the quality of the front interface, rear interface, bulk, and grain boundaries. Addressing all of these aspects of the cell will be necessary for polycrystalline devices to continue to improve device voltage and efficiency.
One strategy to passivate semiconductor interfaces is with an insulating oxide. In CdTe, an improvement in VOC and efficiency was previously demonstrated by eliminating the CdS window layer and growing CdTe directly on (Mg,Zn)O. In this case, tuning the conduction band alignment is necessary for maximum efficiency. Adding a large, positive conduction band offset or “spike” at the front interface could have an additional passivating effect. In silicon, oxides such as ALD-deposited Al2O3 passivate the surface effectively by creating a surface charge which induces band-bending. In general, the material which provides optimal band alignment for electron collection may not produce the most passive interface, and a versatile photolithography-based process has been developed to test the passivation effect of several high-bandgap oxide materials. This allows micron-scale point electron contacts to (Mg,Zn)O with varying coverage of passivating oxide. The ratio of point contact area to passivation layer area can be varied over a wide range to observe the effects of passivation and lateral diffusion of carriers.
Films and devices were prepared with continuous coverage oxide layers in order to probe interface properties by photoluminescence and time-resolved photoluminescence. Patterned passivation layers were prepared, and in addition to current-density voltage measurements, laser-beam induced current (LBIC) and electroluminescence will be applied to probe the spatial collection of carriers for point contacts of varying size and spacing.
9:00 PM - ES14.4.36
CIGS Performance Enhancement by Texturisation
Joop van Deelen 1 , Marco Barink 1 , Marcel Simor 1 , Karine van der Werf 2 , Wim Soppe 2
1 , TNO, Eindhoven Netherlands, 2 Thin Film Photovoltiacs, ECN, Eindhoven Netherlands
Show AbstractMost research on texturisation of solar cells has been devoted to Si based cells. Our results show that a significant gain in current density can be achieved for CIGS cell stacks, although its benefit is smaller than for Si cased cells, because of the lower index of refraction of the materials in the CIGS layer stack. Moreover, the opto-electronical properties of the absorber material do not force towards device thicknesses of less than 2 micron as is the case with thin film Si. However, ultra-thin CIGS has emerged recently and light management technology is necessary to compensate for optical losses.
We have made CIGS cells on textured substrates (i.e. back side texturing) and even though the features of the texture are quite small, the cell stack with 1 µm CIGS follows the texture and a 60% drop in reflection was obtained, in spite of the a non-optimized texture used. These results are compared with reduced reflection obtained in other experiments whereby nano-texturing was applied only to a transparent sol gel on top of the (not-encapsulated) front side of the solar cell.
We have also performed optical modeling of textured CIGS stacks over a wide variety of height and periods of the texture. In addition to mappings of the whole parameter window we show generic trends in performance with texture, as well as detailed optical behavior by analysis of absorption (of each layer) and reflection spectra. By detailing the losses, we can explain the increase in CIGS absorption. In the overview we include generic trends in terms of mA/cm2 as derived from the spectra. This way an easy insight is given in the critical texture parameters and how it affects the solar cell performance. It was found that the height/period ratio of the texture is a key factor. Moreover, a complex interplay between the different optical losses, which varies from texture to texture, determine the final effective CIGS absorption. Also the losses are subject to generic trends, which have been revealed during modeling with a systematic variation of the texture dimensions. By choosing the right texture, the reflection can be virtually diminished and parasitic absorption can be reduced.
9:00 PM - ES14.4.37
Investigation of Crystal Growth Mechanisms of the CuGaSe2 - CuAlSe2 Single Crystals Grown by Chemical Vapor Transport
B.V. Korzun 1 , S. Rywkin 1 , A. Yanavitski 1 , V.R. Sobol 2 , G. Gurieva 3 , S. Schorr 3
1 , The City University of New York, Borough of Manhattan Community College, New York, New York, United States, 2 , Belarusian State Pedagogical University, Minsk Belarus, 3 , Helmholtz-Zentrum Berlin fur Materialen und Energie, Berlin Germany
Show AbstractCopper gallium diselenide (CuGaSe2) and copper aluminum diselenide (CuAlSe2) belong to the I-III-VI2 group of compounds and are extensively studied as active elements in optical filters and as absorbing materials in solar cells. To optimize their properties, it is necessary to develop and characterize the technology of the growth of alloys based on these materials. The goal of the present paper is to investigate crystal growth mechanisms of the CuGaSe2 – CuAlSe2 single crystals grown by chemical vapor transport using iodine as a transporting agent.
Synthesis of the initial ternary compounds CuGaSe2 and CuAlSe2 was performed in quartz ampules by melting the elements at a temperature that exceeds the melting point of the compound by 20 K. After preparation of the initial compounds, single crystals of seven alloys of the (CuGaSe2)1-x (CuAlSe2)x system with molar part of CuAlSe2 (x) equaling to 0.20, 0.40, 0.50, 0.60, 0.70, 0.80, and 0.90 were grown by chemical vapor transport in an evacuated quartz tube using iodine as a transporting agent. Typical dimensions of the grown plate-like single crystals were about 10 mm x 5 mm x 0.5 mm with a well-developed surface (112).
The CuGaSe2 - CuAlSe2 single crystals were investigated by means of X-ray diffraction, scanning electron microscopy, and Raman spectroscopy. Raman measurements in backscattering configuration were performed at room temperature using a WITec alpha 300R confocal Raman imaging system with an excitation Nd:YAG laser at a wavelength of 532 nm and a 20x objective lens.
Two groups of the CuGaSe2 - CuAlSe2 single crystals were investigated. The first group consisted from the samples prepared by the traditional version of chemical vapor transport but without applying the last stage of growth when the intermediate products are transferring into a zone with a source material. The second group consisted of single crystals of the first group exposed to the chemical etching. This allowed for characterization of the intermediate products and establishing crystal growth mechanisms.
It was found that transport of chemical elements from the zone with a source material into a crystallization zone also transfer iodides (copper, aluminum) and vapor of elements (selenium). Usually iodides of copper and aluminum are doped and contain up to several percent of doping elements. The rate of transfer of gallium is lower and it leads to the shift of the chemical composition of single crystals towards the CuAlSe2-based alloys. Also shifts in frequencies and changes in relative intensities of some modes in different cluster regions of the CuGaSe2 - CuAlSe2 single crystals were found. These can be explained by the variation of chemical composition inside of the homogenous region of the corresponding single crystals.
Acknowledgment. V.R. Sobol would like to thank the Belarusian Republican Foundation for Fundamental Research for partial financial support of the studies under project F15MLD-025.
9:00 PM - ES14.4.38
Mössbauer Studies of the CuFeS2-δ - CuInS2 System
B.V. Korzun 1 , V.R. Sobol 2 , M. Myndyk 3 , V. Sepelak 4 , K.D. Becker 3
1 , The City University of New York, Borough of Manhattan Community College, New York, New York, United States, 2 , Belarusian State Pedagogical University, Minsk Belarus, 3 , Braunschweig University of Technology, Institute of Physical and Theoretical Chemistry, Braunschweig Germany, 4 , Karlsruhe Institute of Technology, Institute of Nanotechnology, Eggenstein-Leopoldshafen Germany
Show AbstractMultinary semiconducting compounds with crystal structures of chalcopyrite CuFeS2 are at the focus of current research as absorbing materials in solar cells. To obtain these materials with optimal physical characteristics it is necessary to know the processes of phase formation during preparation of such materials from chemical elements or binary/ ternary compounds. The goal of this paper is to study the interaction in the CuFeS2-δ - CuInS2 system by means of Mössbauer studies and X-ray powder diffraction (XRPD).
The X-ray studies were carried out using monochromatic Cu Ka-radiation (1.5406 Å, step size 0.01° or 0.04°, counting time 10 s). Room-temperature 57Fe Mössbauer spectra were taken in transmission geometry using a 57Co/Rh γ–ray source. The velocity scale was calibrated relative to 57Fe in Rh.
The initial elements for the preparation of CuInS2 and CuFeS2-δ ternary compounds were 99.9998% copper, 99.9997% indium, 99.999% iron, and 99.9999% sulfur. The synthesis of the initial ternary compounds CuInS2 and CuFeS2-δ (with δ = 0 and 0.10) was performed in quartz ampoules by melting from the elements. Thirteen alloys of the (CuFeS2-δ)1-x - (CuInS2)x system with molar part of CuInS2 (x) equaling to 0.01, 0.03, 0.05, 0.125, 0.25, 0.375, 0.50, 0.625, 0.75, 0.875, 0.95, 0.97, and 0.99 were prepared. The required amounts of powder of the corresponding ternary compounds CuInS2 and CuFeS2-δ were weighted, mixtured and homogenized. Samples of the alloys were prepared by melting, where the mixtures of ternary compounds were heated up to the 1383 K temperature, which exceeds by 20 K the melting point of the compound with the highest melting point (CuInS2).
The investigated samples revealed two different types of the Mössbauer spectra for the CuInS2-based alloys (singlet) and the CuFeS2-δ-based alloys (sextet). The analysis of the intermediate alloys confirmed the limited solubility in the CuFeS2-δ - CuInS2 system and the limits of solubility were found. The influence of the variation of chemical composition of chalcopyrite CuFeS2-δ with changing δ from 0 to 0.10 on the limits of solubility in the CuFeS2-δ - CuInS2 system was determined. A full description of the alloys by XRPD and Mössbauer studies allowed to make recommendations for the technology of preparation of alloys in the CuFeS2-δ - CuInS2 system.
Acknowledgment. V.R. Sobol would like to thank the Belarusian Republican Foundation for Fundamental Research for financial support of the studies within project F15MLD-025.
Symposium Organizers
Ingrid Repins, National Renewable Energy Laboratory
Shubhra Bansal, University of Nevada, Las Vegas
Sascha Sadewasser, International Iberian Nanotechnology Laboratory
Edgardo Saucedo, IREC
Symposium Support
Catalonia Institute for Energy Research (IREC)
Dr. Eberl MBE-Komponenten GmbH
First Solar
International Iberian Nanotechnology Laboratory
National Renewable Energy Laboratory
ES14.5/ES11.5: Joint Session: Tandem Devices
Session Chairs
Richard King
Yaroslav Romanyuk
Wednesday AM, April 19, 2017
PCC North, 200 Level, Room 221 ABC
9:00 AM - *ES14.5.01/ES11.5.01
Efficiency Potential and Recent Activities of High Efficiency and Si Tandem Solar Cells
Masafumi Yamaguchi 1 , Hiroyuki Yamada 3 , Yasuhiro Katsumata 2
1 , Toyota Technological Institute, Nagoya Japan, 3 , NEDO, Kawasaki Japan, 2 , JST, Kawasaki Japan
Show AbstractThe present status of R&D for various types of solar cells is presented by over viewing research and development projects for solar cells in Japan as the Project Leader of the previous PV R&D Project “High Performance Photovoltaic System Technology Development for the Future” under the NEDO (New Energy and Industrial Technology Development Organization of Japan) and as the Research Supervisor of the Research Filed “Creative Clean Energy Generation by using Solar Energy” under the JST (Japan Science and Technology Agency). Developments of high efficiency solar cells such as 44.4% (under concentration) and 37.9% (under 1-sun) InGaP/GaAs/InGaAs 3-junction solar cells by Sharp, 25.1% crystalline Si hetero-junction back-contact (HBC) solar cells by Sharp, 20.9% CIGS solar cells by Solar Frontier, and 11.9% dye-sensitized solar cells by Sharp have been demonstrated under the previous NEDO PV R&D Project. 15.0% efficiency has also been attained with 1cm2 Perovskite solar cell by NIMS under the JST Project. Most recently, 26.3% crystalline Si HBC solar cells by Kaneka, 22.3% CIGS solar cells by Solar Frontier, and 18.2% Perovskite solar cell by NIMS have been achieved under the present NEDO PV R&D Project.
In order to create future clean energy infra structures based on photovoltaics, further development of PV science and technology is necessary. One of most important R&D issues is to develop high efficiency and low cost solar cells. This paper also presents analytical results for efficiency potential of high-efficiency solar cells such as crystalline Si, GaAs, GaAs/Si, CIGSe and CdTe solar cells based on external radiative efficiency (ERE), open-circuit voltage loss and fill factor loss. Crystalline Si solar cells have efficiency potential of 28.5% by improvement in ERE from around 1% to 20%. GaAs and GaAs/Si cells have efficiency potential of 29.7% and 27.4% by improvements in ERE from 22.5% to 30% and from 0.1% to 1%, respectively. CIGSe and CdTe solar cells have potential efficiencies of 26.5% by improvements in ERE from 0.5% to 10% and from around 0.1% to 5%, respectively. Efficiency potential of future generation solar cells such as CZTS and CZTSSe, MQW and QD, Perovskite and Ferroelectric solar cells is also discussed. Recent activities of Si tandem solar cells such as III-V/Si tandem solar cells are also overviewed. Because III-V/Si tandem solar cells have great potential of high- efficiency and low-cost and are great candidate for Solar EV applications.
9:30 AM - *ES14.5.02/ES11.5.02
Development of High Gap Ge- and Si-Based Kesterite-Like Solar Cells for Tandem Applications
Guy Brammertz 1 , Sylvester Sahayaraj 1 , Zijian Huang 1 , Samaneh Ranjbar 1 , Bart Vermang 2 , Marc Meuris 1 , Jef Poortmans 2
1 , imec - Division of IMOMEC, Leuven Belgium, 2 , imec, Leuven Belgium
Show AbstractSimulations show that for tandem solar cell applications with Si bottom layer cells, the top absorber band gap should ideally be around 1.8 eV1. Ge- and Si-based Kesterite-like materials such as Cu2ZnGe(S,Se)4 and Cu2Zn(Sn,Si)Se4 have band gap energies in that range. In the present contribution we have investigated the potential of these type of materials as solar cell absorbers. We have fabricated high gap Ge- and Si-based thin film absorber layers on Mo/glass substrates by a two step selenization or sulfurization process. First, we sequentially evaporate a metal layer stack on a Mo/glass substrate, followed by selenization or sulfurization in either a 10% H2Se in N2, 100% H2S or an elemental Se environment at temperatures varying from 460 to 560°C. Investigated materials in this work are Cu2ZnSiSe4, Cu2Zn(Sn,Si)Se4, Cu2SiS3, Cu8SiS6 and Cu2ZnGe(S,Se)4.
Our studies show that, despite a very remarkable crystallinity and photoluminescence response at 1.8 eV of the Cu8SiS6 absorber, none of the Si-based materials shows any photocurrents in a standard solar cell configuration with a CdS buffer layer.
On the other hand, Cu2ZnGe(S,Se)4 with a CdS buffer layer does show relatively good solar cell behavior with efficiencies in excess of 5 % and very good crystallinity of the absorber layer. Besides a still relatively large Voc deficit, the main limitation of the efficiency seems to be a large series resistance of the order of 5 Ohm cm2. A possible reason for the large series resistance could be a backside contact problem with the Mo and we have therefore studied the effect of different backside contacts such as TiN, TiW, Ti, Cr or Al on the solar cell efficiency. We have furthermore analyzed the devices using spectral response measurements, capacitance-voltage measurements as well as temperature dependent current-voltage and photoluminescence measurements, trying to gain some insight in the main recombination pathways in the solar cells.
1 T. P. White, N. N. Lal, and K. R. Catchpole, IEEE Journal of Photovoltaics 4, 208-214 (2014).
10:00 AM - ES14.5.03/ES11.5.03
NIR-Transparent Perovskite Solar Cell for Flexible All-Thin-Film Tandem Devices
Stefano Pisoni 1 , Fan Fu 1 , Thomas Feurer 1 , Stephan Buecheler 1 , Ayodhya Tiwari 1
1 , EMPA, Dubendorf Switzerland
Show AbstractThe outstanding photovoltaic properties and large bandgap of organometal halide perovskite solar cells make them attractive as top cells in tandem structures. We have already demonstrated 20.5% efficiency perovskite/Cu(In,Ga)Se2 (CIGS) tandem solar cells on rigid glass substrates in a four-terminal configuration. However, the use of flexible and lightweight foil as a substrate opens up the possibility for roll-to-roll manufacturing of high efficiency tandem devices in future. Here we report the development of NIR-transparent flexible perovskite solar cells for their application as top cells on highly efficient CIGS bottom cells, realizing flexible and lightweight tandem devices. Perovskite solar cells in planar heterojunction configuration were developed, avoiding the use of high-temperature sintered mesoporous structures. By employing an appropriate electron transport multilayer stack and suitable transparent conductive oxide materials, NIR-transparent flexible perovskite photovoltaic devices, with stabilized efficiency of 12.2% and NIR transmittance greater than 78%, were achieved in the preliminary development work. Flexible perovskite/CIGS tandem device with 18.2% efficiency was measured in 4-terminal configuration. In order to further improve tandem performances, current efforts are directed for achieving enhanced NIR-transparency and higher efficiency with perovskite layers of bandgap better matched for double-junction tandem devices with CIGS and CuInSe2 (CIS) bottom cells. Special emphasis is placed on the development of low temperature processing methods, including optimization of the two-step perovskite deposition, suitable for temperature-sensitive flexible substrates. The paper would present a comprehensive work on the development of high efficiency flexible perovskite/CIGS thin film tandem devices and discuss processing and performance related challenges.
10:15 AM - ES14.5.04/ES11.5.04
Infrared-Tuned Silicon Bottom Cell for 23.6%-Efficient Perovskite/Silicon Tandem
Zhengshan Yu 1 , Mathieu Boccard 1 , Peter Firth 1 , Kevin Bush 2 , Axel Palmstrom 2 , Stacey Bent 2 , Michael McGehee 2 , Zachary Holman 1
1 , Arizona State University, Tempe, Arizona, United States, 2 , Stanford University, Stanford, California, United States
Show AbstractAmorphous silicon / crystalline silicon heterojunction (SHJ) solar cell is an excellent bottom cell because of its high open-circuit voltage, which results from the separation of the highly recombination-active (ohmic) contacts from the silicon absorber bulk, and because of its dominant performance-loss mechanism under the standard solar spectrum—parasitic absorption of blue light in the front amorphous silicon (a-Si:H) layers—is irrelevant in tandems.
We report our design and fabrication of “IR-tuned SHJ cells” which are designed to serve as bottom cells in perovskite/silicon tandem applications. In this cell structure, the front surface was polished to facilitate perovskite deposition, its rear surface was textured to scatter infrared light. On the rear side, which accounts for the main IR light loss due to parasitic absorption in the rear TCO/metal stack, we insert porous, nanoparticulate films as low-refractive-index layers between silicon bulk and silver, and fabricate nanoparticle/silver rear reflectors. We vary the porosity and thus the refractive index (n =1.1–1.5) of the nanoparticle films, which are deposited by a controllable aerosol spray process, and investigate their effectiveness in reducing infrared parasitic absorption in the solar cells. Optical test structures incorporating films with the highest n exhibit an internal reflectance of over 99%, which means almost eliminating parasitic loss.
On top of this silicon cell, a cesium formamidinium lead halide perovskite cell was deposited with a new, nickel oxide hole contact. This cell also featured a buffer layer deposited by atomic layer deposition at its front side; this layer prevents sputter damage during deposition of the front transparent conductive oxide electrode.
Combing both sub-cells, we report a two-terminal monolithic perovskite/silicon tandem solar cell with an efficiency that exceeds that of either sub-cell and that of the record single-junction perovskite cell. This cell reached an efficiency of 23.6%, certified at NREL, and has been stable at that value for 500 hours of continuous illuminated operation.
10:30 AM - ES14.5.05/ES11.5.05
Large-Area Scalable Perovskite/Silicon Multi-Junction Solar Modules
Manoj Jaysankar 1 , Ulrich Paetzold 2 1 , Weiming Qiu 1 , Tamara Merckx 1 , Tom Aernouts 1 , Robert Gehlhaar 1 , Maarten Debucquoy 1 , Jef Poortmans 1
1 , imec, Leuven Belgium, 2 , Karlsruhe Institute of Technology, Karslruhe Germany
Show AbstractCrystalline silicon (c-Si) photovoltaics is the dominant technology today for solar power generation. However, the power conversion efficiency of c-Si solar cells is approaching the theoretical limit. Recently, hybrid organic-inorganic perovskite based solar cells, owing to their remarkable lab-scale efficiency, bandgap tuneability, and low cost of fabrication, have attracted a great deal of attention. Moreover, multi-junction solar cells employing perovskite and c-Si solar cells bear the exciting potential to surpass the efficiency limit of market-leading single-junction c-Si solar cells besides being cost-efficient compared to other multi-junction solar cell technologies. However, scaling up this technology and maintaining high efficiency over large areas is challenging, as evidenced by the small-area (< 1.5 cm2) perovskite/Si multi-junction solar cells reported so far. For the economic viability of the perovskite/Si multi-junction technology, an efficient transition from lab-scale cells to industrial-scale modules is crucial.
In this work, we present four-terminal perovskite/c-Si multi-junction solar modules that are fully scalable to commercial solar module dimensions. We first demonstrate a module-on-cell architecture in which a semi-transparent methylammonium lead triiodide perovskite solar module is stacked onto an interdigitated back contact c-Si bottom solar cell of identical aperture area. By a detailed opto-electronic analysis we investigate the impact of the transparent electrodes and the design of the semi-transparent perovskite solar module on the overall performance of the four-terminal multi-junction solar module. With a combination of optimised transparent electrodes and efficient module design, our perovskite/c-Si multi-junction solar modules yield power conversion efficiencies of 23% on 4 cm2 aperture area. In a second demonstrator, we scale up the aperture area to 16 cm2 by going to a module-on-module architecture and achieve power conversion efficiencies of 20%. Both efficiencies represent record results for such sizes. Furthermore, by quantifying and addressing the various losses in our four-terminal solar modules, we demonstrate the feasibility of achieving perovskite/c-Si multi-junction solar modules that can outperform stand-alone c-Si solar cells. Our multi-junction solar module concept is compatible with industrially scalable fabrication techniques thus, enabling the production of high-performance large-area perovskite/silicon multi-junction solar modules.
10:45 AM - ES14.5.06/ES11.5.06
Study of Polycrystalline MgxCd1-xTe/MgyCd1-yTe Double Heterostructures for Tandem Solar Cell Applications
Calli Campbell 1 2 , Cheng-Ying Tsai 1 3 , Yong-Hang Zhang 1 3
1 Center for Photonics Innovation, Arizona State University, Tempe, Arizona, United States, 2 School for Engineering of Matter, Transport and Energy, Arizona State University, Tempe, Arizona, United States, 3 School of Electrical, Computer and Energy Engineering, Arizona State University, Tempe, Arizona, United States
Show AbstractPolycrystalline cadmium telluride (CdTe) and silicon solar cells are the two most mature photovoltaic technologies, each with impressive single junction efficiency records (22.1% and 25.6% respectively) and relatively high manufacturability and economy. In order to achieve higher efficiencies than either of these technologies can achieve alone, a tandem configuration is necessary. Connecting cells in series is the most practical method from a fabrication standpoint, however limits the bandgaps which can be used together in order to achieve maximum efficiency. Detailed balance calculations determine that the ideal bandgap for a top subcell in series configuration with Si (Eg = 1.1 eV) is 1.7 eV, slightly higher that of CdTe (Eg = 1.5 eV). It is found that alloying CdTe with Mg shifts the bandgap up with relatively little Mg incorporation to achieve Eg = 1.7 eV. Previous work in the group growing bulk monocrystalline 1.7 eV Mg0.13Cd0.87Te/Mg0.5Cd0.5Te double heterostructures (DHs) by molecular beam epitaxy (MBE) on lattice matched InSb(001) substrates reveals thin films with high structural and optical quality. Solar cells featuring an n-type Mg0.13Cd0.87Te/Mg0.5Cd0.5Te absorber layer and a p-type a-Si hole contact reveal promising preliminary results including a current record efficiency of 11.2% and implied open circuit voltage of 1.3 V.
In this study, polycrystalline MgxCd1-xTe bulk layers and MgxCd1-xTe/MgyCd1-yTe double heterostructures will be grown in a molecular beam epitaxy (MBE) chamber for precise environmental control. Real-time in-situ analysis by Reflection High Energy Electron Diffraction (RHEED) and ex-situ structural and optical characterization, including X-ray diffraction and photoluminescence (PL) spectroscopy, will monitor the material samples as substrate temperature and material flux are varied. Polycrystalline double heterostructure samples will be grown to reveal if wider bandgap barriers reduce recombination and increase PL intensity in the polycrystalline bulk. Preliminary polycrystalline MgCdTe solar cells will be fabricated and electrically characterized.
11:00 AM - ES14.5/ES11.5
BREAK
ES14.6: Buffer Layer and Alkali Treatments
Session Chairs
Michelle Mezher
Pedro Salome
Wednesday PM, April 19, 2017
PCC North, 200 Level, Room 229 B
11:30 AM - *ES14.6.01
Design of Optimal Buffer Layers for CIGS Thin-Film Solar Cells
Vincenzo Lordi 1
1 , Lawrence Livermore National Laboratory, Livermore, California, United States
Show AbstractA major hurdle to achieving maximum efficiency in CuInGaSe2 (CIGS) thin-film solar cells is associated with the choice of buffer layer deposited between the absorber and transparent contact. The prototypical buffer layer, CdS, exhibits very favorable electrical properties, but parasitically absorbs light from the blue end of the solar spectrum. Optimized alternative buffer layer materials with wider band gap and good transport properties are thus desired, but have been difficult to develop.
Here, we describe a combined synthesis, materials characterization, and theory effort to design optimal buffer layers based on the (Cd,Zn)(O,S) quaternary alloy system. Optimization of buffer composition, as well as properties of the absorber/buffer interface, was performed in light of several competing requirements, including band gap, conduction band offset, dopability, interface quality, and film crystallinity. Process variations to control the film and interface quality were explored to achieve the optimum buffer.
Theoretical calculations were performed using hybrid density functional theory at the atomistic scale to predict the properties of materials across the quaternary composition range and to search for optima. Devices were fabricated in MiaSolé Hi-Tech’s production line, which uses a continuous all-sputter process without vacuum breaks to deposit layers onto a flexible stainless steel substrate. Compositional and morphological analyses of the film stack were performed in cross-section using aberration-corrected scanning transmission electron microscopy (STEM) with energy-dispersive X-ray spectroscopy (EDS) and electron energy loss spectroscopy (EELS). Secondary ion mass spectrometry (SIMS) was used for comparative depth profiles averaged over larger areas. Comparison to devices fabricated using chemical bath deposition for the buffer layer was also performed.
Process variations were developed to produce layers with controlled crystallinity, varying from amorphous to fully epitaxial, depending primarily on oxygen content. The degree of elemental intermixing between the buffer and absorber, particularly involving Cd and Cu, also was found to depend on process conditions and significantly affects device performance. Secondary phase formation at the interface was observed for some conditions. Characterization of the experimental degrees of freedom allows targeted fabrication of theoretically predicted optimal material stacks, while using theory to guide experiments.
Prepared by LLNL under Contract DE-AC52-07NA27344.
12:00 PM - *ES14.6.02
Highly Efficient Solar Cells Based on Full PVD Processed CIGSe/CdIn2S4 Heterojunction
Nicolas Barreau 1
1 , Université de Nantes, Nantes France
Show AbstractSolar cells based on polycrystalline Cu(In,Ga)Se2 (CIGSe) thin films have reached the outstanding level of performance of 22.6 %. Such record efficiency has been achieved by treating the completed absorber with heavy alkali fluorine post deposition treatment (PDT) before forming the electrical junction using chemical bath deposited (CBD)CdS. Such a solution based technique is nevertheless likely to slow industrial production down although it appears to be the most reliable and absorber-tolerant for efficient junction formation.
The recent investigation of CIGSe (KF-PDT) / (CBD) CdS interface of the best solar cells fabricated in our group (~ 20 %) led us to the conclusion that the release of indium selenide compounds at the absorber surface, thanks to KF-PDT, enhances the formation of materials such as CdIn2S4; the latter thiospinel was already extensively investigated by the end of 1970ies as single crystals for electro-optical applications.
The present contribution aims at showing that the use of similar compounds grown by physical vapour deposition (PVD) onto as co-evaporated CIGSe layers allows the achievement of cells with similar output voltage and higher currents than the standard (CBD)CdS. It will also be demonstrated that a fine control of CIGSe/CdIn2S4 interface characteristics yields similar junction quality devices, overpassing the performance of CIGSe/(CBD)CdS cells.
These results will be presented as additional puzzle pieces to the understanding of high quality junction formation in polycrystalline hetero-structures, showing highly efficient full-PVD CIGSe-cells compatibility with in-line production processes.
12:30 PM - ES14.6.03
Impact of the Heavy Alkali Fluoride Post-Deposition Treatment on the Electronic Structure of the CdS/Cu(In,Ga)Se2 Interface in High-Efficiency Thin-Film Solar Cells
Dirk Hauschild 1 2 3 , Dagmar Kreikemeyer-Lorenzo 1 , Philip Jackson 4 , Theresa Friedlmeier 4 , Dimitrios Hariskos 4 , Friedrich Reinert 3 , Michael Powalla 4 , Clemens Heske 1 2 5 , Lothar Weinhardt 1 2 5
1 Institute for Photon Science and Synchrotron Radiation (IPS), Karlsruhe Institute of Technology (KIT), Eggenstein-Leopoldshafen Germany, 2 Institute for Chemical Technology and Polymer Chemistry (ITCP), Karlsruhe Institute of Technology (KIT), Eggenstein-Leopoldshafen Germany, 3 Experimental Physics VII, University of Würzburg, Würzburg Germany, 4 , Zentrum für Sonnenenergie- und Wasserstoff-Forschung Baden-Württemberg (ZSW), Stuttgart Germany, 5 Department of Chemistry and Biochemistry, University of Nevada, Las Vegas (UNLV), Las Vegas, Nevada, United States
Show AbstractWith the introduction of an alkali fluoride post-deposition treatment (AF-PDT) for Cu(In,Ga)Se2 (CIGSe) absorber layers, world-record efficiencies exceeding 22 % have recently been achieved [1,2]. In the present contribution, we report on the electronic structure at the CdS/CIGSe heterojunction for thin-film solar cells with and without a heavy AF-PDT. For this purpose, we use surface-sensitive x-ray and ultraviolet photoelectron spectroscopy (XPS and UPS), as well as inverse photoemission spectroscopy (IPES) to derive a detailed understanding of the impact of the AF-PDT on the chemical and electronic properties of the CIGSe surface and its interface to the CdS buffer layer. We find that the AF-PDT modifies the growth of the CdS buffer layer [3], and leads to an additional band bending in the CIGSe absorber upon interface formation. In addition, a detailed analysis of the UPS and IPES spectra to determine valence band maxima and conduction band minima derives a flat CdS/CIGSe conduction band alignment when taking the interface-induced band bending (as observed in XPS) into account. In this contribution, these findings will also be discussed in view of the observed increase in energy conversion efficiency of corresponding thin-film solar cell devices.
[1] Jackson et al., Phys. Status Solidi RRL 10, 583–586 (2016)
[2] Solar Frontier, Proc. 43rd IEEE Photovoltaic Specialists Conference (PVSC), in print
[3] Friedlmeier et al., Proc. 43rd IEEE Photovoltaic Specialists Conference (PVSC), in print
12:45 PM - ES14.6.04
Comparison of the Interface Formation of Cu(In,Ga)Se2 with RbF Post-Deposition Treatment and CdS and ZnS Buffer Layers
Nicoleta Nicoara 1 , Philip Jackson 2 , Dimitrios Hariskos 2 , Wolfram Witte 2 , Sascha Sadewasser 1
1 , International Iberian Nanotechnology Laboratory, Braga Portugal, 2 , Zentrum für Sonnenenergie-und Wasserstoff-Forschung Baden-Württemberg (ZSW), Stuttgart Germany
Show AbstractRecent improvements in the efficiency of Cu(In,Ga)Se2 (CIGSe) thin film solar cells are mainly attributed to alkali post-deposition treatments (PDT) [1]. This process modifies the optoelectronic properties of CIGSe providing an improved surface for subsequent buffer deposition [2,3] and favoring a reduction of the buffer thickness, as compared with non-PDT interfaces [4]. Knowledge related to the chemical and electronic properties of CIGSe/buffer interfaces has been gained previously, but issues related to the buffer growth mechanism and its effect on the interface quality are still not well understood. In this contribution, we report on the initial stages of CdS and ZnS buffer formation on CIGSe absorbers subjected to the recently introduced RbF PDT [5]. Kelvin probe force microscopy, providing spatially resolved surface potential images, is used to analyze the evolution of the interface electronic properties as a function of the buffer thickness. Our results suggest a poor (i.e. non-uniform) interface quality up to ~3nm CdS, and the early stage formation of a pn-junction starting from this coverage, although layer inhomogeneity is still observed even for a larger thickness (~20nm) as shown by significant variations in the surface photovoltage (SPV) of the junction. More homogeneous deposition is observed for ZnS, as suggested by the thickness-independent and relatively low SPV values up to 3nm thickness. A clear signature of a complete, n-type layer is observed for 20nm thick ZnS. These results provide relevant insight into the interface formation in view of the quest to further reduce the buffer layer thickness (e.g. in an attempt to increase the collection in the blue region). Our findings on the absorber/buffer interface formation reveal the underlying mechanism for limitations in the reduction of the buffer thickness, even when an alkali PDT is applied.
[1] P. Jackson et al., Phys. Status Solidi RRL 9, 28 (2015).
[2] A. Laemmle et al. Phys. Status Solidi RRL 7, 631 (2013).
[3] F. Pianezzi et al., Phys. Chem. Chem. Phys. 16, 8843 (2014).
[4] A. Chirila et al., Nature Materials 12, 1107 (2013).
[5] P. Jackson et al., Phys. Status Solidi RRL 10, 583 (2016)
ES14.7: Defects and Disorder
Session Chairs
Shubhra Bansal
Stephan Lany
Wednesday PM, April 19, 2017
PCC North, 200 Level, Room 229 B
2:30 PM - ES14.7.01
Optical Properties and Band Structures of Cu-Deficient Phases, Cu(In,Ga)3Se5 and Cu(In,Ga)5Se8 in Cu-Poor Cu2Se-(In,Ga)2Se3 Pseudo-Binary System
Takahiro Wada 1 , Tsuyoshi Maeda 1 , Mai Watanabe 1
1 , Ryukoku Univ, Otsu Japan
Show AbstractCu-deficient phases such as Cu(In,Ga)3Se5 and Cu(In,Ga)5Se8 are useful for controlling the valence band offset (ΔEv) at CdS/Cu(In,Ga)Se2 (CIGS) interface and at grain boundary in the CIGS solar cells. Therefore, Cu-poor CIGS compounds, Cu(In,Ga)3Se5 and Cu(In,Ga)5Se8 have been attracted attention as the key compounds for high-efficiency CIGS solar cells. We have studied crystallographic and optical properties and band structures of CuInSe2, CuIn3Se5, and CuIn5Se8 phases in Cu-poor Cu2Se-In2Se3 pseudo-binary system [1] and studied those of CuIn3(S,Se)5 and CuGa3(S,Se)5 systems [2]. In this symposium, we report crystallographic and optical properties and band structures of Cu(In,Ga)3Se5 and Cu(In,Ga)5Se8.
We prepared Cu(In1-yGay)Se2 and Cu-poor Cu-(In,Ga)-Se (CIGS) phases such as Cu(In1-yGay)3Se5 and Cu(In1-yGay)5Se8 in the composition of (1-x)Cu2Se-(x)(In1-yGay)2Se3 with 0.5 ≤ x ≤ 1.0, y=0.0, 0.25, 0.50, 0.75 and 1.0. Their crystal structures were analyzed by Rietveld refinement using X-ray diffraction data. The crystal structure of the Cu-poor Cu-(In,Ga)-Se samples change from tetragonal chalcopyrite-type to tetragonal stannite-type through the mixed phase of chalcopyrite-type and stannite-type structures. The band-gap energies were determined by diffuse reflectance spectra. The band-gap energies of the Cu-poor Cu-(In,Ga)-Se samples increase in a stepwise manner with decreasing the Cu/(In,Ga) ratio. The energy levels of the valence band maximums (VBMs) were estimated from the ionization energy by photoemission yield spectroscopy (PYS) measurements. The electron affinity, energy level of conduction band minimum (CBM) could be also determined by adding the value of the optical band gap to the energy level of the VBM. The energy levels of the VBM of the Cu-poor Cu-(In,Ga)-Se samples decrease significantly with decreasing the Cu/(In,Ga) ratio. To understand the electronic structure of Cu-poor Cu-(In,Ga)-Se compounds, we performed first-principles band structure calculations of stannite-type CuIn5Se8, CuGa5Se8 and reference compounds, tetragonal chalcopyrite-type CuInSe2 and CuGaSe2, by using HSE06 nonlocal screened hybrid density functional. We will discuss relation between chemical composition of chalcopyrite and stannite-type phases and their optical properties and electronic structures.
[1] T. Maeda, W. Gong, and T. Wada, “Crystallographic and optical properties and band structures of CuInSe2, CuIn3Se5, and CuIn5Se8 phases in Cu-poor Cu2Se-In2Se3 pseudo-binary system”, Jpn. J. Appl. Phys. 55, 04ES (2016).
[2] K. Ueda, T. Maeda and T. Wada, “Crystallographic and optical properties of CuIn3(S,Se)5 and CuGa3(S,Se)5 systems”, submitted to Thin Solif Films.
2:45 PM - ES14.7.02
Investigation of Carrier Transport in CuInGaSe2 by Highly Spatially, Spectrally and Time Resolved Cathodoluminescence Microscopy
Mathias Mueller 1 , Martin Mueller 1 , Torsten Hoelscher 2 , Setareh Zahedi-Azad 2 , Matthias Maiberg 2 , Frank Bertram 1 , Roland Scheer 2 , Juergen Christen 1
1 Institute of Experimental Physics, Otto-von-Guericke-University, Magdeburg Germany, 2 FG Photovoltaik, Martin-Luther-University, Halle-Wittenberg Germany
Show AbstractCuInGaSe2 (CIGSe) is a heavily disordered material with a huge amount of point defects and extended defects like grain boundaries and dislocations. Despite impressive efficiencies exceeding 22 % there are still open questions regarding transport of carriers and the influence of inhomegeneities.
To gain a deeper understanding highly spatially, spectrally, and time resolved cathodoluminescence (CL) measurements have been performed on polycrystalline CIGSe. The samples were grown on Mo-coated soda lime glass in a single-stage-process and neither a buffer layer nor a window layer were applied. Absorbers with varying Cu/III-ratios (CGI: 0.73 and 0.86) and therefore varying grades of disorder were investigated. CL measurements were carried out in a home-built setup based on a modified scanning electron microscope.
The characteristic temperature dependence of the diffusion length of excess carriers together with their lifetime gives access to the temperature dependence of the carrier mobility yielding information about the underling scattering mechanisms in the material.
The CIGSe backside was covered with 220 nm thick rectangular titanium masks, allowing electrons to penetrate the sample through this masks, while completely absorbing the luminescence of the subjacent CIGSe. The structure facilitates the independent realization of undisturbed time resolved transient measurements on the unmasked CIGSe layer (far away from any masking) together with cw CL-linescans perpendicular to the mask’s edge for the identical sample. In CL-linescans only a fraction of generated excess carriers are able to reach the uncovered CIGSe surface area by lateral transport within their diffusion length and lifetime and contribute to the CL intensity. Luminescence of carriers recombining below the metal masks is blocked, i.e. absorbed by the mask. Simultaneously, the decay in time resolved measurements yields the initial life time. The temperature dependence (4.5 K - 300 K) of both, lifetime as well as diffusion length, is determined and gives access to the temperature dependent mobility.
With decreasing temperature from 300 K to 4.5 K the diffusion length increases from 6 µm to 27 µm (CGI 0.86) and 23 µm (CGI 0.73). Simultaneously, carrier lifetime increases from 20 ns (@125 K) to 48 ns (@4.5 K) and from 3 ns (@125 K) to 37 ns (@4.5 K), respectively. The resulting mobilities follow a power law with µ ∝ T-0.86 (CGI 0.86) and µ ∝ T-0.43 (CGI 0.73), which reveals scattering at neutral defects as the dominant mechanism, reaching up to 340,000 cm2/Vs at 4.5 K.
3:00 PM - ES14.7.03
Optoelectronic Properties of Bulk Single-Crystal (Ag,Cu)2ZnSnSe4 Alloys
Michael Lloyd 1 2 , Douglas Bishop 3 , Brian McCandless 2 , Richard Haight 3 , Robert Birkmire 2 1
1 Materials Science and Engineering, University of Delaware, Newark, Delaware, United States, 2 , Institute of Energy Conversion, Newark, Delaware, United States, 3 , IBM TJ Watson Research Center, Yorktown Hts., New York, United States
Show AbstractSpatially inhomogeneous band-tail formation has been shown to cause VOC limiting effects on solar cell performance [1]. Evidence is mounting that associates the existence of Cu-Zn antisite pairs within the Cu2ZnSn(S,Se)4 kesterite system (CZTS) to such deleterious defects [2]. The critical temperature of disorder in CZTS is relatively low for kinetic processes, fundamentally limiting the effects of thermal treatments on antisite suppression [3]. Recently, there has been increased interest in alloying CZTS with partial and full substitutions for either copper or zinc. One promising substitution is AgCu which is predicted to raise the formation energy of I - II antisite defects and subsequently result in a more easily ordered system [4].
Investigations into the fundamental optoelectronic properties of single crystalline (AgxCu1-x)2ZnSnSe4 materials will be presented. Crystals of millimeter size scales with compositions of x = 0.1 and x = 0.3 are successfully grown via a solid state process similar to that demonstrated in reference [5]. Specialized surface treatments are developed to mitigate the effects of surface damage resulting from crystal polishing. Hall characterization demonstrates p-type conductivity for all compositions explored with carrier densities ranging from 1010 to 1017 cm-3 for crystal compositions in the range of x = 0.3 to x = 0.1. Photoluminescence (PL) spectra are analyzed to explore the effects of silver incorporation on the presence of band-tailing effects. The dependence of PL spectra on temperature and illumination intensity show evidence of remaining disorder in as-grown Ag alloyed crystals. Spectra for Ag containing specimens will be compared to those of pure Cu kesterite single-crystals which have been subjected to thermal ordering treatments.
[1] U. Rau and J. H. Werner, “Radiative efficiency limits of solar cells with lateral band-gap fluctuations,” Appl. Phys. Lett., vol. 84, no. 19, p. 3735, 2004.
[2] T. Gokmen, O. Gunawan, T. K. Todorov, and D. B. Mitzi, “Band tailing and efficiency limitation in kesterite solar cells,” Appl. Phys. Lett., vol. 103, no. 10, pp. 101–106, 2013.
[3] K. Rudisch, Y. Ren, C. Platzer-Björkman, and J. Scragg, “Order-disorder transition in B-type Cu2ZnSnS4 and limitations of ordering through thermal treatments,” Appl. Phys. Lett., vol. 108, no. 23, 2016.
[4] E. Chagarov, K. Sardashti, A. C. Kummel, Y. S. Lee, R. Haight, and T. S. Gershon, “Ag2ZnSn(S,Se)4: A highly promising absorber for thin film photovoltaics.,” J. Chem. Phys., vol. 144, no. 10, p. 104704, 2016.
[5] M. A. Lloyd, D. Bishop, O. Gunawan, and B. McCandless, “Fabrication and performance limitations in single crystal Cu2ZnSnSe4 solar cells.,” in Photovoltaic Specialists Conference (PVSC), 2015 42th IEEE, 2015.
3:15 PM - ES14.7.04
Role of Nanoscale Disordering in Photovoltaics
Jeffery Aguiar 1 2 , Dennis Pruzan 2 , Brian Devener 3 , Sudhajit Misra 4 , Mehmet Erkan 4 , Akira Nagaoka 2 5 , Kenji Yoshino 5 , Helio Moutinho 6 , Mowafak Al-Jassim 6 , Mike Scarpulla 2 4 7
1 , Idaho National Laboratory, Idaho Falls, Idaho, United States, 2 Materials Science and Engineering, University of Utah, Salt Lake City, Utah, United States, 3 Analysis Laboratory, University of Utah, Salt Lake City, Utah, United States, 4 Department of Electrical Engineering, University of Utah, Salt Lake City, Utah, United States, 5 Department of Applied Physics and Electronic Engineering, University of Miyazaki, Miyazaki, Miyazaki, Japan, 6 , National Renewable Energy Laboratory, Golden, Colorado, United States, 7 Department of Physics and Astronomy, University of Utah, Salt Lake City, Utah, United States
Show AbstractRecently, a number of thin-film photovoltaics absorber materials such as Cu2ZnSn(S,Se)4 (CZTSSe), Cu(In, Ga)Se2 (CIGS), CdTe, and inorganic organic perovskite semiconductors have attracted broad interest because of their common highly variable and complicated relationships between atomic ordering, chemistry, and device properties. Due to the inherent complexities associated with standardized growth conditions common to each of these photovoltaics, especially concerning deposition techniques, studies on the bulk nature of slowly, rapidly cooled, and tempered semiconductors can serve as a baseline comparison for elucidating the role of growth—especially concerning the presence of extra phases and defects. Comparing different growth techniques and subsequent annealing therefore remains an area of interest in addressing issues of recombination and passivation in devices.
Studying material interfaces naturally provides the breadth of knowledge for comparing materials grown; as a function of annealing, partial vapor pressures, and gas environments. The differences in material structure and chemistry are presumably caused by the low vapor pressures that leads to the decomposition of materials. For example, in Cu2ZnSnS4 (CZTS) secondary phases such as Cu2SnS3, ZnS, and others are often difficult to detect, and ultimately hinder the performance of these solar cells. In light of these phases, interest remains as to whether this behavior is associated with differences in sputter deposition, thermal co-evaporation, or solution processing of materials (including CdTe, CIGS, and inorganic organic perovskites), or an inherent property of these material-types.
In this work, we will present on the latest results spanning several photovoltaics, including CZTS to resolve and correlate secondary phases in photovoltaics, with different growth techniques and subsequent post growth annealing. We report the presence of secondary phases based on the latest advances made in the materials characterization techniques to differentiate nanoscale differences in material chemistry based on developing our analysis.
ES14.8: Alkali Incorporation
Session Chairs
Nicolas Barreau
Stefan Haass
Wednesday PM, April 19, 2017
PCC North, 200 Level, Room 229 B
4:30 PM - *ES14.8.01
Thermodynamic Limitations for Alkali Metals in Cu(In,Ga)Se2
Dimitrios Hariskos 1 , Michael Powalla 1 2
1 , Zentrum für Sonnenergie- und Wasserstoff-Forschung Baden-Württemberg, Stuttgart Germany, 2 , Karlsruher Institut für Technologie (KIT), Karlsruhe Germany
Show AbstractThe efficiency of Cu(In,Ga)Se2 (CIGS) -based solar cells could be continuously increased up to 22.6% employing alkali metal dopants as Na, K, Rb, Cs*). The alkali metals are supplied to the CIGS layer from the glass substrate during deposition, from precursor layers, or by a post deposition treatment. It is repeatedly reported that the alkali metal distribution in CIGS is not homogenous. Independent of the alkali metals used, their concentration at grain boundaries is much higher than in the grains.
In this contribution we give an overview about thermodynamic limitations for alkali metals in CIGS based on literature data. We apply the concept of non-miscibility of phases for alkali metals in CIGS and explain their segregation at grain boundaries, the formation of clusters in CIGS grains, the sporadic formation of microstructures in the CIGS layer (hotspots, nodules), and the formation of secondary phases with ordered structures. We demonstrate on examples from the literature that the necessity to drive the three-stage process in the way it was empirically developed is partly a result of these thermodynamic limitations.
*) P. Jackson, R. Wuerz, D. Hariskos, E. Lotter, W. Witte, and M. Powalla, Effects of heavy alkali elements in Cu(In,Ga)Se2 solar cells with efficiencies up to 22.6%, Phys. Status Solidi RRL 10, No. 8, 583–586 (2016) / DOI 10.1002/pssr.201600199
5:00 PM - ES14.8.02
KF Post-Deposition Treatment of Industrial Cu(In,Ga)(S,Se)2 Thin-Film Surfaces—Modifying the Chemical and Electronic Structure
Michelle Mezher 1 , Lorelle Mansfield 2 , Kim Horsley 1 , Monika Blum 1 , Wanli Yang 3 , Robert Wieting 4 , Lothar Weinhardt 1 5 6 , Kannan Ramanathan 4 , Clemens Heske 1 5 6
1 , University of Nevada, Las Vegas, Las Vegas, Nevada, United States, 2 , National Renewable Energy Lab, Golden, Colorado, United States, 3 , Lawrence Berkeley National Lab - Advanced Light Source, Berkeley, California, United States, 4 , STION, Menlo Park, California, United States, 5 , Institute for Photon Science and Synchrotron Radiation, Karlsruhe Institute of Technology, Karlsruhe Germany, 6 Institute for Chemical Technology and Polymer Chemistry, Karlsruhe Institute of Technology, Karlsruhe Germany
Show AbstractEver since 2014, when EMPA set a new world record efficiency of 20.4% for Cu(In,Ga)Se2 (CIGSe) thin-film photovoltaic devices1, KF-post deposition treatments (KF-PDTs) have become a “hot topic” in the chalcopyrite community, especially once ZSW increased the record to 21.7% shortly thereafter2. Most recently, ZSW achieved a new record efficiency (22.6%) due to the inclusion of a Rb-PDT3. Despite the advances in efficiencies, the impact of alkali-PDT on the CIGSe device is not fully understood, with various groups publishing numerous, sometimes conflicting, results. It is the goal of this study to reveal the chemical and electronic structure of alkali (i.e., Na and K) treated CIG(S)Se surfaces for devices from both, industry and research-lab environments. This adds two aspects to the discussion, namely the impact on sulfur-containing absorbers, as well as on absorbers taken from an industrial process.
Sample sets prepared at STION (CIGSSe) and NREL (CIGSe) were investigated by electron spectroscopy, including x-ray (XPS) and ultraviolet (UPS) photoelectron spectroscopy, as well as inverse photoemission (IPES) at UNLV. Furthermore, soft x-ray emission spectroscopy (XES) was performed at Beamline 8.0.1 of the Advanced Light Source. XPS and XES provide information on the chemical structure at the surface and in the near-surface bulk region. The valence and the conduction band extrema with respect to the Fermi Energy (EF) are derived from UPS and IPES spectra.
We find that nominally identical treatments can result in considerable differences between samples, suggesting the presence of “hidden” parameters. As will be presented, such differences can be found in the degree of surface Cu-depletion, the S/Se ratio, and the valence band maximum (VBM) and conduction band minimum (CBM) energies. For KF-PDT, we find a surface bandgap widening. The findings will be discussed in view of their relevance for industrially-prepared chalcopyrite photovoltaic devices, especially when using sulfur to tailor the properties of the absorber/buffer interface.
(1) Chirilă, A.; Reinhard, P.; Pianezzi, F.; Bloesch, P.; Uhl, A. R.; Fella, C.; Kranz, L.; Keller, D.; Gretener, C.; Hagendorfer, H.; Jaeger, D.; Erni, R.; Nishiwaki, S.; Buecheler, S.; Tiwari, A. N. Nat. Mater. 2013, 12 (12), 1107–1111.
(2) Jackson, P.; Hariskos, D.; Wuerz, R.; Kiowski, O.; Bauer, A.; Friedlmeier, T. M.; Powalla, M. Phys. Status Solidi RRL – Rapid Res. Lett. 2015, 9 (1), 28–31.
(3) Jackson, P.; Wuerz, R.; Hariskos, D.; Lotter, E.; Witte, W.; Powalla, M. Phys. Status Solidi RRL – Rapid Res. Lett. 2016. DOI: 10.1002/pssr.201600199.
5:15 PM - ES14.8.03
Na-Diffusion Enhanced P-Type Conductivity in Cu(In,Ga)Se2—A New Mechanism for Efficient Doping in Semiconductors
Shiyou Chen 1 , Zhen-Kun Yuan 2 , Hongjun Xiang 2 , X.G. Gong 2 , Ji-Sang Park 3 , Su-Huai Wei 4
1 , East China Normal University, Shanghai, CA, China, 2 , Fudan University, Shanghai China, 3 , National Renewable Energy Laboratory, Golden, Colorado, United States, 4 , Beijing Computational Science Research Center, Beijing China
Show AbstractIncorporation of Na (K) is now a standard process in the fabrication of the chalcopyrite Cu(In,Ga)Se2 (CIGS) and kesterite Cu2ZnSn(S,Se)4 (CZTSSe) thin film solar cells and was found very important for achieving high efficiency. It is believed that its beneficial effect results partially from the increase of the hole concentration after Na (K) doping. However, how the Na (K) doping increases the hole concentration is not clear, although this problem has been intensively studied in the past 15 years. To further improve the efficiency of the CIGS solar cells (to over 25%), there is an urgent need to unveil the mechanism of the hole concentration increase in the Na (K) doped CIGS.
Using first-principles calculations, we find that the Na dopant is most likely to take place of Cu (forming NaCu) and is electrically inactive in CIGS, so the observed increase of hole concentration can not be understood according to the traditional doping theory. Here we propose a new mechanism. A high concentration of NaCu are formed in CIGS (Na diffuses into the CIGS grains) at a Na-rich and high-temperature growth environment. However, during cooling, the equilibrium concentration (solubility) of Na in the bulk of CIGS becomes lower, and NaCu is not thermodynamically stable any more in the CIGS grains at low temperature, so Na tends to diffuse out of the CIGS grains. Since Na is a fast diffusor in bulk CIGS with a small migration energy, Na can diffuse out of the CIGS grains and leaves the Cu sites unoccupied, then a high concentration of Cu vacancies (VCu) are formed within the grains. The subsequent rinsing in water can further facilitate the formation of VCu. Because the water rinsing can significantly dissolve Na at the CIGS surface and decrease the chemical potential of Na (the environment becomes more and more Na-poor, so the formation energy of Na dopants in CIGS lattice increases), the out-diffusion of Na from the CIGS grain interiors is enhanced. Since the surfaces and grain boundaries of CIGS films are usually Cu-depleted (more Cu poor than in the grains), so less Cu can diffuse back from the surfaces or grain boundaries into the CIGS grains, then a higher concentration of VCu and hole carriers are formed within the grains. Compared to Na, K has a lower solubility in the bulk of CIGS. Besides, after K doping a K-enriched layer is formed on the CIGS surface, which hinders K diffusion out of the CIGS grains, so less VCu can be formed and the increase of the hole concentration is smaller, in good agreement with the experiments.
Based on this dopant-diffusion mechanism, we can also understand the p-type conductivity enhancement in Na-doped CZTSSe, and may design new strategies for achieving efficient p-type or n-type bipolar doping in wide-gap semiconductors.
Reference:
Yuan, Z.-K., Chen, S., Xie, Y., Park, J.-S., Xiang, H., Gong, X.-G. and Wei, S.-H., Adv. Energy Mater., doi:10.1002/aenm.201601191
5:30 PM - ES14.8.04
Impact of Different Alkali Post Deposition Treatments on the Formation of the Zn(O,S)/Cu(In,Ga)Se2 Interface
Thomas Kunze 1 , Philip Jackson 2 , Dimitrios Hariskos 2 , Andreas Siebert 1 , Claudia Hartmnann 1 , Roberto Felix Duarte 1 , Xeniya Kozina 1 , Dominic Gerlach 3 , Yoshiyuki Yamashita 3 , Toyohiro Chikyow 3 , Shigenori Ueda 4 , Regan Wilks 1 5 , Wolfram Witte 2 , Marcus Baer 1 5 6
1 Renewable Energy, Helmholtz-Zentrum Berlin für Materialien und Energie GmbH, Berlin Germany, 2 , Zentrum für Sonnenenergie- und Wasserstoff-Forschung Baden-Württemberg (ZSW), Stuttgart Germany, 3 Nano-Electronic Materials Unit/MANA, National Institute for Materials Science (NIMS), Tsukuba Japan, 4 Synchrotron X-Ray Group, National Institute for Materials Science (NIMS), Kouto Japan, 5 Energy Materials In-Situ Laboratory Berlin (EMIL), Helmholtz-Zentrum Berlin für Materialien und Energie GmbH, Berlin Germany, 6 Institut für Physik und Chemie, Brandenburgische Technische Universität Cottbus-Senftenberg, Cottbus Germany
Show AbstractSolar cells based on Cu(In,Ga)Se2 (CIGSe) chalcopyrite thin films are considered to be an alternative to silicon-wafer based devices. Recent progress has led to a significant performance leap of chalcopyrite-based thin film solar cells. The record conversion efficiencies reach 22.6%. The latest boost in efficiency was achieved by replacing the KF- by an RbF-post deposition treatment (PDT). [1] Usage of the even heavier alkali Cs in a PDT had no additional effect on the efficiency. Nevertheless, CIGSe cells based on absorbers that underwent a KF- or CsF-PDT also exhibit efficiencies above 20%. As a next step to further improve device efficiency Zn(O,S) – a non-toxic and more transparent substitute to the standard CdS buffer – has been employed in cell production.
In order to understand the beneficial effects of alkali treatments on the chemical and electronic absorber structures, their influence on the Zn(O,S) buffer layer growth, and the properties of the new Zn(O,S)/CIGSe interface, we studied three sample series in which different Zn(O,S) layer thicknesses were deposited on KF-, RbF-, or CsF-PDT CIGSe absorbers grown by co-evaporation. The thickness of the buffer layer was stepwise increased from 0 to 20 nm by interrupting the chemical bath deposition after different times. The chemical and electronic structure of the resulting (partially deeply) buried Zn(O,S)/CIGSe interfaces was then non-destructively and depth-dependently probed by photoemission using soft and hard x-rays.
For the bare absorbers, we always find a Cu-deficient surface composition. However, the Na surface content decreases with increasing alkali-PDT mass. Furthermore, we find that the position of the valence band maximum (VBM) moves away from the Fermi level with increasing surface sensitivity for all absorbers, with a larger VBM for the RbF and CsF-PDT CIGSe compared to the KF treated absorber. In addition, the Zn(O,S) buffer layer grows significantly faster on the RbF and CsF treated CIGSe than on the KF-PDT absorber. In our contribution, we will attempt to relate these results to the performance of corresponding solar cell devices.
[1] Effects of heavy alkali elements in Cu(In,Ga)Se2 solar cells with efficiencies up to 22.6%, P. Jackson, R. Wuerz, D. Hariskos, E. Lotter, W. Witte, and M. Powalla, Phys. Status Solidi RRL 10, No. 8 (2016) 583
5:45 PM - ES14.8.05
Improvement of Open Circuit Voltage in Cu2ZnSnSe4 Solar Cells by Surface Treatment
Hitoshi Tampo 1 , Kang Min Kim 2 , Shinho Kim 1 , Hajime Shibata 1 , Shigeru Niki 1
1 , National Institute of Advanced Industrial Science and Technology, Tsukuba Japan, 2 , Korea Institute of Industrial Technology, Gangreung Korea (the Republic of)
Show AbstractCurrently, the certified conversion efficiency of Cu2ZnSn(S,Se)4 (CZTSSe) (12.6%) is less than that of Cu(In,Ga)Se2 (22.6%), and many researchers are therefore striving to address this gap. One of the most significant issues for kesterite solar cells including CZTSSe in achieving this objective is a large open circuit voltage (Voc) deficit (Eg/q – Voc), which is attributed to the poor qualities of the CZTSSe bulk and heterointerface.
To solve the above problem, a simple and promising method is to improve the bulk quality of absorption layers, e.g., by replacing host materials, such as by a Ge-incorporated Cu2Zn(Sn,Ge)Se4 (CZTGSe) alloy, and we have demonstrated a high conversion efficiency CZTGSe solar cell (12.3%) with low Voc deficit [1]. Another method, as recently applied by several groups, is conducting alkali metal incorporation in CZTSSe-based material. We have demonstrated improvement of minority lifetime from 2 to 15 ns with Na incorporation in Cu2ZnSnSe4, and the efficacy of 9.57% was obtained due to the Na effects [2]. Furthermore, conversion efficacy of 10.7% was obtained using a same deposition system of CZTSe layers for NaF deposition. However, the CZTSe surface (CdS/CZTSe heterointerface) recombination was found to limit the minority carrier lifetime, which imply the limitation of the efficacy improvement.
In this study, we demonstrate Voc improvement with various CZTSe surface treatments. The treatments are conducted after a thermal treatment of CZTSe absorbers, and sulfur treatment resulted in the improvement of Voc and conversion efficiency of 11.2% was obtained (Voc: 0.430 V, Jsc: 37.6 mA/cm2, FF: 0.692).
References
[1] S. Kim, K.M. Kim, H. Tampo, H. Shibata, and S. Niki, Appl. Phys. Express 9, 102301 (2016).
[2] H. Tampo, K.M. Kim, S. Kim, H. Shibata, and S. Niki (submitted)
ES14.9: Poster Session II: Defects, Degradation and Stability, Chalcopyrite Growth, Tandem Devices and Alkali Incorporation
Session Chairs
Shubhra Bansal
Ingrid Repins
Thursday AM, April 20, 2017
Sheraton, Third Level, Phoenix Ballroom
9:00 PM - ES14.9.01
Using Diffusion-Reaction Simulation to Study Light Soaking Effect in CdTe Solar Cells
Da Guo 1 , Abdul Shaik 1 , Dragica Vasileska 1
1 School of Electrical, Computer and Energy Engineering, Arizona State University, Tempe, Arizona, United States
Show AbstractNearly all PV technologies exhibit changes in device performance under extended illumination, or “light soaking”, although the magnitude and the trend of these changes are not always the same among different technologies. Experiments on both commercial modules and research cells based on CdTe technology have shown improvement of cell performance under light soaking conditions for up to 20 hours. Many accredited such phenomena to the passivation of traps and migration of Cu ions.
In this work, we employed a self-consistent one-dimensional (1D) diffusion-reaction simulator to study the migration and passivation of Cu related dopants in CdTe solar cell as a function of soaking conditions. The employed 1D simulator was originally developed to study the incorporation of Cu and compensating mechanisms among Cu dopants in CdTe material. It solves diffusion-reaction equations both for the free carriers and the point defects in time-space domain self-consistently with global Poisson equation.
A standard ZnTe/CdTe/CdS structure solar cell was simulated with simplified Cu defect distribution in this work. Our simulation shows that both deionization of Cu dopants and migration of charged Cu donors under light soak could cause device performance enhancement. The simulation result also suggests that 10-14cm2/s diffusivity of Cu interstitials (or any other mobile donors) could explain the 20-hour long light soaking effect at 65oC. To further investigate such phenomena, full simulation that includes defect-defect reactions is needed.
9:00 PM - ES14.9.03
Analysis of Waiting Times between CIGS and CdS and In-Diffusion of Na on the Properties of Cu(In,Ga)Se2 Materials and Solar Cells
Pedro Salome 1 , Jennifer Teixeira 2 , Miguel Cardoso 2 , Viktor Fjallstrom 3 , Marika Edoff 3 , Nicoleta Nicoara 1 , Joaquim Leitao 2 , Sascha Sadewasser 1
1 , INL, Braga Portugal, 2 Departamento de Física and I3N, Universidade de Aveiro, Aveiro Portugal, 3 Engineering Sciences, Uppsala University, Uppsala Sweden
Show AbstractIn this work we investigate two important industrially relevant issues and their impact on the electro-optical parameters of materials and devices. After CIGS is deposited, there is a small period of time that allows for the deposition of the CdS layer without a big influence of air exposure on the CIGS layer. Long waiting periods, with the CIGS exposed to air, will cause unwanted oxidation of the CIGS. Another issue is the amount of Na diffusion which, if soda-lime-glass is used, is extremely dependent on the substrate growth temperature, the Mo permeability for Na and/or the amount of Na-containing precursors used in pre-deposition or post-deposition treatments. In this work we studied three samples: a reference sample (15% of power conversion efficiency), prepared according to standard industrial procedures; a sample exposed to air during a 48-hours after the growth of the CIGS (12% efficiency), in order to simulate a problem in e.g. the buffer layer deposition tool; and a sample grown on an Al2O3 substrate (7% efficiency), that simulates an extreme problem of insufficient Na-diffusion from the substrate. While the reference sample and the oxidized sample show the same crystalline structure, as shown by XRD and Raman, the Na-free sample showed a different crystalline orientation with lower crystalline quality. SEM characterization confirmed that the Na-free sample has a smaller grain size than the other samples. PL measurements on the reference sample and on the oxidized sample showed a single, broad, and asymmetric band, whereas on the Na-free sample, a second component at lower energies is identified. With the increase of the excitation power, a huge blueshift is observed for the PL on the three samples, which clearly shows the influence of fluctuating potentials on the electronic levels’ structure of the CIGS. This behaviour is in accordance with a high doping and strong compensation levels, typically observed in Cu-poor CIGS samples. From the dependence on the temperature of the PL, non-radiative de-excitation mechanisms are discussed. This study allows a better understanding of the consequences of common industrial problems on the properties of the CIGS-based solar cells and can be used for early detection of production problems related to surface oxidation and/or Na deficiency.
9:00 PM - ES14.9.04
Na-Assisted Grain Growth in Cu2ZnSnS4 Nanoparticle Thin Films for Solar Cell Applications
Sara Engberg 1 , Andrea Crovetto 2 , Ole Hansen 2 3 , Yeng Ming Lam 4 , Jorgen Schou 1
1 DTU Fotonik, Technical University of Denmark, Roskilde Denmark, 2 DTU Nanotech, Technical University of Denmark, Lyngby Denmark, 3 CINF, Technical University of Denmark, Lyngby Denmark, 4 School of Materials Science and Engineering, Nanyang Technological University, Singapore Singapore
Show AbstractWe have studied the effect of Na in Cu2ZnSnS4 nanoparticle thin films [1]. The as-synthesized CZTS nanoparticles were inherently ligand-free [2], which allows us to use polar solvents, such as water and ethanol. Another advantage of these particles is that the user- and environmentally-friendly NaCl salt can be directly dissolved in controllable amounts. This further circumvents the need for later incorporation of dopants, or a ligand-exchange step to functionalize the surface of the nanoparticles. In addition, the homogeneous distribution of Na in the ink allows uniform grain growth within the deposited absorber layer.
By including Na in the nanoparticle ink, micron-sized grains throughout the whole absorber are achieved after annealing in a sulfur atmosphere at 600°C. The absorber layer appeared to be of full density, and no closed porosity could be detected. In addition, the photoluminescence signal increased by a factor of 200 after Na-inclusion. Without Na, the grains were very difficult to sinter, the film was porous, and the photoluminescence was low. A concentration of Na/(Cu+Zn+Sn)=30% was necessary for the densification of the absorber, which is significantly higher than that used in other Na-doped CZTS systems. The annealed films were found to be of the desired Cu-poor and Zn-rich composition.
We also found that a sulfidation temperature above 550°C was required. At 550°C, NaCl-crystals appeared on the surface of the thin films, suggesting an incomplete transformation of Na into the liquid phase Na2Sx-additive during sintering. At this temperature, grain growth was only detected in close proximity to the NaCl regions. It was also observed that the NaCl crystals could be easily removed by a quick water rinse, but that this treatment reduced the photoluminescence signal. This is relevant as it is customary to leave the absorber layer in a water-based solution after annealing before buffer layer deposition.
[1] Sara Engberg, Andrea Crovetto, Stela Canulescu, Ole Hansen, Yeng Ming Lam, and Jørgen Schou, Na-assisted grain growth in CZTS nanoparticle thin films for solar cell applications, Submitted 2016.
[2] Naghmeh Mirbagheri, Sara Engberg, Andrea Crovetto, Søren Simonsen, Ole Hansen, Yeng Ming Lam, and Jørgen Schou, Synthesis of ligand-free CZTS nanoparticles via a facile hot injection route, Nanotechnology, 27 (2016).
9:00 PM - ES14.9.05
In-Ga Interdiffusion in CIGS—The Roles of Potassium and Selenium
Diego Colombara 1 , Conrad Spindler 1 , Florian Werner 1 , Anne-Marie Goncalves 2 , Susanne Siebentritt 1 , Phillip Dale 1
1 Physics and Materials Science Research Unit, Universitè du Luxembourg, Belvaux Luxembourg, 2 Institut Lavoisier, Universitè de Versailles, Versailles France
Show Abstract
D. Colombara1, C. Spindler1, F. Werner1, A.-M. Gonçalves2, S. Siebentritt1 and P. J. Dale1
Ga/(In+Ga) depth gradient engineering is crucial to achieve the highest power conversion efficiencies in CIGS photovoltaic devices. Unfortunately, appropriate compositional profiles can only be achieved by means of a suitably high processing temperature or by a complex sequence of depositions at lower temperature [1]. One reason for this is that sodium, a much needed extrinsic dopant, has the drawback of hindering In-Ga interdiffusion in polycrystalline CIGS.
Studying monocrystalline CIS/GaAs diffusion couples helps discerning the diffusion processes occurring within the grains from those occurring at the grain boundaries [2]. Recent findings obtained by post-deposition annealings in a sodium-rich atmosphere have shown that sodium appears to promote In-Ga interdiffusion in CIGS films free from grain boundaries, suggesting that sodium occupation of the grain boundaries may act as a barrier for intergrain diffusion of In and Ga [3].
Here, the study is extended to the effects of potassium and selenium. As per the case of sodium, both potassium and selenium induce a blue-shift of the photoluminescence peak compared to untreated samples. This may be due to enhanced intragrain In-Ga interdiffusion. The effect of selenium has been reported similarly for polycrystalline CIGS [4] and is subtle compared to that of the alkali metals. Contrarily, potassium (like sodium) is known to hinder In-Ga interdiffusion in polycrystalline films [5]. This confirms that grain boundaries play a dominant role in diffusion processes [6].
References:
[1] F. Pianezzi et al., “Defect formation in Cu(In,Ga)Se2 thin films due to the presence of potassium during growth by low temperature co-evaporation process,” Journal of Applied Physics, vol. 114, no. 19, p. 194508, Nov. 2013.
[2] D. J. Schroeder, G. D. Berry, and A. A. Rockett, “Gallium diffusion and diffusivity in CuInSe2 epitaxial layers,” Applied Physics Letters, vol. 69, no. 26, pp. 4068–4070, Dec. 1996.
[3] D. Colombara, “Does Na hinder or enhance In-Ga interdiffusion in CIGS?,” presented at the SPIE Optics + Photonics for Sustainable Energy, San Diego, 2016.
[4] B. J. Mueller, B. Opasanont, V. Haug, F. Hergert, S. Zweigart, and U. Herr, “Influence of selenium amount on the structural and electronic properties of Cu(In,Ga)Se2 thin films and solar cells formed by the stacked elemental layer process,” Thin Solid Films, vol. 608, pp. 62–70, Jun. 2016.
[5] R. Wuerz, A. Eicke, F. Kessler, S. Paetel, S. Efimenko, and C. Schlegel, “CIGS thin-film solar cells and modules on enamelled steel substrates,” Solar Energy Materials and Solar Cells, vol. 100, pp. 132–137, May 2012.
[6] A. Laemmle, R. Wuerz, T. Schwarz, O. Cojocaru-Mirédin, P.-P. Choi, and M. Powalla, “Investigation of the diffusion behavior of sodium in Cu(In,Ga)Se2 layers,” Journal of Applied Physics, vol. 115, no. 15, p. 154501, Apr. 2014.
9:00 PM - ES14.9.06
Enhanced Performance of Cu2ZnSn(S,Se)4 Solar Cells with Introducing Interfacial Alkali Fluoride Layers
Cheng-Ying Chen 1 2 , Wei-Chao Chen 1 2 , Bandiyah Sri Aprillia 1 2 3 , Naili Saidatin 1 2 3 , Ruei-San Chen 3 , Kuei-Hsien Chen 2 1 , Li-Chyong Chen 1
1 , National Taiwan University, Taipei Taiwan, 2 , Institute of Atomic and Molecular Science, Academia Sinica, Taipei 106, Taiwan, Taiwan, 3 , Graduate Institute of Applied Science and Technology, National Taiwan University of Science and Technology, Taipei, Taiwan, Taiwan
Show AbstractCu2ZnSn(S,Se)4 (CZTSSe) could be the earth-abundant (i.e low cost) and low toxicity alternative compounds for the commercialized Cu(In,Ga)(S,Se)2 thin-film solar cells. To achieve this possible, many efforts applied to the absorbers, band alignments, front and back interfaces/contacts must be performed for improving CZTSSe performance. [1,2,3]
In this work, we demonstrated the high efficiency CZTSSe solar cells by introducing an interfacial alkali fluoride layers (< 5nm NaF, LiF) between the absorber (i.e., CZTSSe) and the buffer layer (i.e., CdS) after sulfo-selenization processes without post-annealing. In the statistically studies (10 cells), the alkali fluoride layer increase power conversion effcicincy from 6.27% to 7.0%, short circuit current density (Jsc) from 24.6 mA/cm2 to 26.8 mA/cm2 and open circuit voltage (Voc) from 480 mV to 490 mV, possibly resulting from the defect passivation on grain boundaries [2]. The defect energy levels of the absorber measured by admittance spectroscopy decrease from 224 to 117 meV with increasing Na or Li contents. Finally, a 7.7 % efficient CZTSSe solar cell with Voc of 480 mV, Jsc of 27.3 mA/cm2 and fill factor (FF) of 58 % was obtained.
The morphology, elemental composition, and distribution of the absorber layers are being examined by scanning electron microscopy (SEM), time-of-flight secondary ion mass spectroscopy (TOF-SIMS), Raman spectroscopy, and by combining transmission electron microscopy (TEM) with electron energy loss spectroscopy (EELS).
References
[1] V. Tunuguntla, W.C. Chen, P.H. Shih, I. Shown, Y.R. Lin, C.H. Lee, J.S. Hwang, L.C. Chen and K.H. Chen, J. Mater. Chem. A, 2015,3, 15324-15330
[2] Y.R. Lin, V. Tunuguntla, S.Y. Wei, W.C. Chen, D. Wong, C.H. Lai, L.K. Liu, L.C. Chen and K.H. Chen, Nano Energy, 2015, 16, 438
[3] W.C. Chen, C.Y. Chen, V. Tunuguntla, S.H. Lu, C. Su, C.H. Lee, K.H. Chen and L.C. Chen, Nano Energy (DOI: http://dx.doi.org/10.1016/j.nanoen.2016.09.022) (2016)
9:00 PM - ES14.9.07
Composition Dependent Cation Ordering Characteristics of Thin-Film CZTS
Katharina Rudisch 1 , Alexandra Davydova 1 , Charlotte Platzer-Bjorkman 1 , Jonathan Scragg 1
1 Solid State Electronics, Uppsala University, Uppsala Sweden
Show AbstractThe low open-circuit voltage is the main limiting factor for state-of-the-art devices based on Kesterite CZTS absorbers, and Cu-Zn disorder in CZTS may be one of the possible causes of this. Cu-Zn disorder is abundant in Zn-rich (B-type) CZTS, which has been studied extensively over the last few years. Thermal treatments showed positive effects on the Cu-Zn ordering. But up to now, a reduction of the open circuit voltage deficit in devices with increased Cu-Zn order could not be shown. One reason could be the severe kinetic limitations in the production of highly ordered B-type CZTS through slow-cooling treatments [1]. On the other hand, CZTS samples prepared by solid state synthetic routes with near-stoichiometric [2] and A- or (A+B) type compositions [3] have been shown to feature increased cation order. However, no comprehensive study evaluating the ordering kinetics in dependence of the CZTS stoichiometry type has been performed for thin films so far.
Here we report for the first time about the ordering characteristics of CZTS thin films for a broad composition range. We encompass primarily E- (Sn-rich), A- (Cu-poor and Zn-rich), B and F-type (Sn-poor) material by characterization of compositionally graded samples. The influence of the stoichiometry type on the critical temperature and the ordering kinetics is studied with resonant Raman spectroscopy, via the secondary order parameter Q. After slow cooling, large variations in the behavior of Q are seen in different compositional ranges within the samples, which are in clear dependence on the stoichiometry type. Particularly, Q values were significantly higher in (F+B)- and (E+A)-type regions ( ≈ 3), and reached its minimum in the (A+B)-type region (≈ 1). Our “baseline” anneal in S-containing atmosphere was compared with a new anneal process featuring a significantly higher Sn activity. A sample from the latter process showed strongly increased Q values in the (E+A)-type regions (≈ 4-5), which could be attributed to an increased solubility of E- and A-type defect complexes under the modified anneal conditions.
The fundamental reasons for the variations in Q with stoichiometry type are discussed in the context of the order-disorder transition dependence on the different defect complexes, and the influence of defect complexes on the Raman spectrum itself. Our study reveals important basic materials characteristics that are of relevance for production of highly ordered CZTS thin films for future device applications.
[1] K. Rudisch, Y. Ren, C. Platzer-Björkman, J. Scragg, Appl. Phys. Lett. 108, 231902 (2016)
[2] A. Ritscher, M. Hoelzel, M. Lerch, J. Solid State Chem. 238, 68-73 (2016)
[3] M. Paris, L. Choubrac, A. Lafond, C. Guillot-Deudon, and S. Jobic, Inorg. Chem. 53, 8646 (2014)
9:00 PM - ES14.9.08
Antisite Defects in Cu2ZnSn(S,Se)4—Local and Long Range Order
Laura Schelhas 1 , Kevin Stone 1 , Glenn Teeter 2 , Steven Harvey 2 , Ingrid Repins 2 , Michael Toney 1
1 , SLAC National Laboratory, Menlo Park, California, United States, 2 , National Renewable Energy Laboratory, Golden, Colorado, United States
Show AbstractThe interest in Cu2ZnSn(S,Se)4 (CZTS) for photovoltaic applications is motivated by the similarities to Cu(In,Ga)Se2 (CIGS) while being comprised of non-toxic and earth abundant elements. However, CZTS suffers from a Voc deficit, where the open-circuit voltage (Voc) is much lower than expected based on the band gap [1, 2]. The cause of this deficit is debated, but may be the result of a high concentration of point-defects in the CZTS lattice. Theory predicts the presence of Cu on Zn and Zn on Cu antisites (CuZn and ZnCu, respectively) as well as point defect clustering. Recently, Scragg et al [3] and Rey et al [4] have reported the observation of a low temperature order/disorder transition by Raman spectroscopy in both Sulphur and Selenium CZTS films. This transition is reported to describe the ordering of Cu and Zn atoms in the CZTS kesterite crystal structure. In addition, NMR has shown significant Cu and Zn site disorder in CZTS powders [5].
To directly determine the level of CuZn, and ZnCu defects (Cu and Zn sublattice ordering), we have used resonant X-ray diffraction (REXD), a site and element specific probe of long range order [6]. We used CZTSe films annealed just below and quenching from just above the transition temperature; based on previous work, the Cu and Zn sublattice should be ordered and highly disordered, respectively. Our data show that there is some Cu and Zn ordering near the low temperature transition identified by Scragg and Rey[3,4], but significantly less than complete chemical order expected from these local probes. To understand both our REXD results and the NMR and Raman results, we present a structural model that involves antiphase domain boundaries and accommodates the excess Zn within the kesterite lattice [7].
[1] A. Polizzotti, I. L. Repins, R. Noufi, S.-H. Wei, D. B. Mitzi, Energy Environ. Sci. 6, 3171 (2013).
[2] S Bourdais et al, Adv Ener Mater. 1502276 (2016).
[3] J.J.S. Scragg, L. Choubrac, A. Lafond, T. Ericson, C. Platzer-Björkman, Appl. Phys. Lett. 104, 041911 (2014).
[4] G. Rey, A. Redinger, J. Sendler, T. P. Weiss, M. Thevenin, M. Guennou, B. El Adib, S. Siebentritt, Appl. Phys. Lett. 105, 112106 (2014).
[5] L Choubrac et al., Phys.Chem.Chem.Phys. 15,10722 (2013).
[6] K.H. Stone, S.T. Christensen, S. Harvey, G. Teeter, I.L. Repins, M.F. Toney, Appl. Phys Lett., in press (2016).
[6] LT Schelhas et al., in preparation.
9:00 PM - ES14.9.09
Can Deep Defects Limit the Open Circuit Voltage of Cu(In,Ga)Se2 Solar Cells?
Conrad Spindler 1 , Susanne Siebentritt 1
1 , University of Luxembourg, Luxembourg Luxembourg
Show AbstractOur recent photoluminescence measurements [1] have indicated two deep electron traps in Indium-free CuGaSe2 thin films. Comparisons with recent theoretical predictions suggest the GaCu antisite as the most likely intrinsic defect involved in both transitions. We show here that one dominant deep defect transition remains almost constant in energy when Indium is added to grow Cu(In,Ga)Se2 compounds with high Ga/(Ga+In) ratios. An increasing amount of Indium decreases the band gap and leads to a lower distance of the deep electron trap to the conduction band. The experimental results are in accordance with theory, where GaCu antisites are expected to be much shallower in In-rich compounds. This behavior contributes to the open circuit voltage deficit in Ga-rich Cu(In,Ga)Se2. The deep defect acts as a recombination center for electrons, whereas in In-rich compositions thermal activation to the conduction band dominates and thus reduces non-radiative recombination through the defect level.
[1] Appl. Phys. Lett. 109, 032105 (2016)
9:00 PM - ES14.9.10
Atomic-Scale Study of Grain Boundaries in CdTe
Fatih Sen 1 , Eric Schwenker 1 4 , Tadas Paulauskas 2 , Ce Sun 3 , Christopher Buurma 2 , Moon Kim 3 , Robert Klie 2 , Maria Chan 1
1 , Argonne National Laboratory, Argonne, Illinois, United States, 4 , Northwestern University, Evanston, Illinois, United States, 2 , University of Illinois Chicago, Chicago, Illinois, United States, 3 , University of Texas at Dallas, Dallas, Texas, United States
Show AbstractThe high efficiency and low manufacturing cost enables CdTe to be one of the most promising thin film photovoltaic (PV) material. The practical efficiencies of polycrystalline CdTe photovoltaic cells are still well below the theoretical limit, indicating possible room for improvement. A fundamental understanding of the role of imperfections and grain boundaries on the electronic structure of CdTe at the atomistic-level may lead to efficiency improvements. First principles modeling and electron microscopy investigations are two most beneficial techniques in developing such a fundamental understanding. In the present work, we built atomistic grain boundary and dislocation core models directly from the STEM images using image analysis methods and crystallographic information at the interface. First principles density functional theory (DFT) calculations are used to compute the electronic structures of grain boundary models. We incorporate various anion and cation substitutional dopants at the grain boundaries. Thermodynamics of dopants and point defects that can exist on or near grain boundaries are investigated, and pertaining changes in electronic structure are reported. The implications of these electronic structure changes at grain boundaries on photovoltaic performance, and corresponding strategies to improve performance by doping are discussed.
ACKNOWLEDGEMENT: We acknowledge funding from the DoE Sunshot program under contract # DOE DEEE007545. Use of the Center for Nanoscale Materials was supported by the U. S. Department of Energy, Office of Science, Office of Basic Energy Sciences, under Contract No. DE-AC02-06CH11357.
9:00 PM - ES14.9.11
Molecular Dynamics Study of Grain Boundaries within CdTe Thin Films Grown on CdS Substrates
Jose Chavez 1 , Xiao Zhou 1 , Rodolfo Aguirre 2 , Sergio Almeida 2 , David Zubia 2
1 , Sandia National Laboratories, Livermore, California, United States, 2 Electrical and Computer Engineering, University of Texas at El Paso, El Paso, Texas, United States
Show AbstractMolecular dynamics simulations have been applied to study grain boundaries in CdTe films grown on CdS substrates containing pre-existing planar defects. Several pre-existing planar defects were considered including high symmetry ∑111 coherent twins, ∑112 double position twins, and stacking faults, as well as low symmetry bi-crystal configurations with fifteen different grain to grain orientation combinations. CdTe thin film deposition was simulated on top of these structures and the film’s microstructure was characterized. In particular, crystal lattice and defect detection algorithms were employed to study grain boundary structure, grain boundary and bulk elemental composition, dislocation line morphology, and grain size. The analyses revealed that substrate grain boundaries propagated into the epilayer for most of the cases studied. The film grain boundary polarity was preserved when compared to the underlying substrate defect. In addition, we found that depending on substrate orientation, dislocations can exhibit in either threading or misfit directions. The predicted atomic structures of the defects are in good agreement with experimental data reported in literature. Our simulations provide new understanding on the atomic scale structures of lattice mismatched epilayers especially when the underlying substrates contain defects. This understanding can guide first principles modeling and experimental efforts towards the design and fabrication of high performance devices.
Acknowledgement - Sandia National Laboratories is a multimission laboratory managed and operated by Sandia Corporation, a wholly owned subsidiary of Lockheed Martin Corporation, for the US Department of Energy's National Nuclear Security Administration under contract DE-AC04-94AL85000. This work was performed under a DOE Project No. EE0005958. SAND2016-10367 A.
9:00 PM - ES14.9.12
Structural Trends in Chalcopyrite Based Intermediate Band Absorber Materials
Julien Marquardt 1 2 , Alexandra Franz 1 , Christiane Stephan 2 3 , Susan Schorr 1 2
1 EM-ASD, Helmholtz-Zentrum Berlin für Materialien und Energie, Berlin Germany, 2 , Freie Universität Berlin, Berlin, Berlin, Germany, 3 , Bundesanstalt für Materialforschung und -prüfung , Berlin Germany
Show AbstractBy now, the progress in manufacturing Cu(Ga,In)(S,Se)2 absorbers used for thin film solar cells led to conversion efficiencies of more than 22% [Solar Frontier, 2015]. In general, compound semiconductors own the advantage of adjusting the band gap by changing the composition of the solid solution, as in CuIn1-xGaxSe2. By establishing an intermediate band is one possibility to optimize the utilization of solar energy [Luque and Martí, 2001]. The Shockley-Queisser limit for chalcopyrite type semiconductors is at 32% [Scheer and Schock, 2011], with an intermediate band gap it is proposed, that the efficiency can be raised up to 63.3%, in ideal conditions [Luque et al. 2006]. In this study presented here CuGaS2 which has the widest band gap in the solid solution series of Cu(Ga,In)(S,Se)2 is used as basic compound to establish the intermediate band. Martí et al. (2008) proposed different transition elements, such as Ti4+/3+ and Fe3+/2+ which may cause an intermediate band within the energy band gap of CuGaS2.
This study aims in establishing the solubility limits of Mn2+, Cr3+as well as Fe3+ in CuGaS2. For the latter a higher solubility in comparison to Ti4+ was proposed from thermodynamic calculations [Palacios et al. 2008]. The initial composition of the studied solid solutions follows the pseudo-binary section of CuGaS2 and Cr2S3 or MnS, respectively and also as a direct exchange of Ga3+ and Mn2+ as well as Ga3+ and Fe3+. Thus different series of powder samples have been prepared by solid state reaction of the elements (900 - 950°C) in evacuated silica tubes. Before an additional annealing (900 - 950°C for 300h) the samples were grounded and pressed to pellets. Different substitutions were utilized, (CuGaS2)x-1 - (Cr2S3)x as well as (CuGaS2)x-1 - (MnS)x and as direct exchange (CuGaS2)x-1 - (CuMnS2)x and (CuGaS2)x-1 - (CuFeS2)x. The powder samples were analyzed in terms of structural properties by X-ray and neutron diffraction as well as chemical composition and phase content by wavelength dispersive spectroscopy (WDX) using an electron microprobe. In addition we have performed photoluminescence measurements on single phase samples.
The presentation will summarize structural trends of lattice parameters, tetragonal distortion and tetragonal deformation in dependence on the Mn, Cr and Fe content. Furthermore the photoluminescence data of the single phase samples will be shown. Moreover the solubility limits of 0.0055 apfu chromium in CuGaS2 from (CuGaS2)x-1 - (Cr2S3)x and the solubility limit of 0.17 apfu in CuGaS2 is (CuGaS2)x-1 - (CuMnS2)x will be discussed.
9:00 PM - ES14.9.13
Defects in Copper Indium Gallium Aluminum Diselenide (CIAGS) Films and Their Impact on Photovoltaic Device Performance
Mandip Sibakoti 1 , Sreejith Karthikeyan 1 , Sehyun Hwang 1 , Timothy Bontrager 1 , Stephen Campbell 1
1 Electrical Engineering, University of Minnesota, Minneapolis, Minnesota, United States
Show AbstractCIGS based photovoltaic devices have garnered attention due to their high device efficiency and potential for low cost manufacturing. Research to date has shown that the highest efficiency devices (>20%) are based on band gaps in the 1.1 - 1.3eV range achieved in films where the [Ga]/[In+Ga] ratio is between 0.1 and 0.3. The CIGS band gap can be increased by increasing the [Ga]/[In+Ga] ratio. However, as the band gap is increased, the short circuit current falls off and the VOC saturates. The saturation of VOC at larger band gaps in Cu(In, Ga)Se2 devices presents a major challenge in developing high efficiency solar devices for tandem solar applications. Recently we have incorporated Al during a single stage deposition process to form CIAGS absorbers. DLTS measurements have shown that these absorbers have demonstrated a lower bulk trap density than Ga-rich CIGS at comparable bandgaps. Although the density of bulk traps is lower, we have observed the same VOC limitation for wide gap devices. In this work, the electronic properties of polycrystalline CIAGS films with varying band gaps are studied to better understand the limitations on device performance. Temperature dependent IV, CV and DLCP measurements are applied to investigate the dominant recombination mechanism limiting VOC in these devices. The nature of the bulk levels are investigated using Admittance Spectroscopy. Based on our analysis and device simulation results, we attribute the VOC saturation in wide gap CIAGS devices to increased recombination rate at the buffer/absorber heterointerface.
9:00 PM - ES14.9.14
Electronic Transitions in Highly Doped and Compensated Chalcopyrites and Kesterites
Jennifer Teixeira 1 , Pedro Salome 2 , Bruno Alves 1 , Sascha Sadewasser 2 , Joaquim Leitao 1
1 , Univ de Aveiro, Aveiro Portugal, 2 , International Iberian Nanotechnology Laboratory, Braga Portugal
Show AbstractDue to recent efficiency improvements in thin-film photovoltaic
technologies, Cu(In,Ga)Se2 (CIGS) and Cu2ZnSnS4 (CZTS) have received an
increased attention in the study of their opto-electronic limitations in
terms of device performance. Further improvements to the electrical
performance of the solar cells depend on the understanding of
fundamental physical properties, namely, the electronic levels’
structure. Different models have been considered for the assignment of
the radiative transitions in CIGS and CZTS, namely: donor acceptor pair
recombination (DAP), quasi-DAP recombination (QDAP), and fluctuating
potentials. In this work, we focus on the self-doped CIGS and CZTS
semiconductors . Both of these compounds have a high doping level and a
strong compensation. The radiative and non-radiative recombination
channels are strongly influenced by electrostatic fluctuating potentials
along the film. As a consequence of these properties, bound states for
electrons do not occur for single donors but only for sufficiently large
clusters of them. The behavior of donors is quite different from that of
acceptors: holes can bind to individual acceptor defects. Such
discussion is sometimes discarded in the literature but it is of utmost
importance for a full understanding of the electronic structure of the
compounds. Photoluminescence (PL) results for Cu-poor CIGS and Cu-poor
CZTS thin films are presented. The observed radiative transitions are
discussed in the scope of the previous theoretical analysis. The DAP
model is unable to explain the excitation power dependence of the PL.
The influence of fluctuating potentials is shown, discussed and proven
to fully explain all of the experimental observations.
9:00 PM - ES14.9.15
The Defects Generated in Magnetron Sputtered Thin-Film CdTe Solar Cells that Limit Performance and Cause Delamination
Piotr Kaminski 1 , Ali Abbas 1 , Sibel Yilmaz 1 , John Walls 1
1 , Loughborough University, Loughborough United Kingdom
Show AbstractThe use of high rate magnetron sputtering for the fabrication of thin film CdS/CdTe solar cells is a potentially attractive alternative to the lower energy techniques such as Close Space Sublimation or Vapour Transport Deposition. Magnetron sputtering is an industrially capable technique which deposits thin films with exceptional uniformity. Uniformity reduces the amount of material necessary for complete light absorbtion thereby reducing materials cost. Uniformity is also an essential requirement for some applications such as those on semi-transparent window products. Work on the development of a viable sputtering process has been conducted for many years in a number of laboratories word-wide. However, the performance of sputtered devices has not matched that generated by the lower energy deposition methods. Voids are often observed at the CdS/CdTe junction and within the bulk of the CdTe absorber. Catastrophic delamination is often observed usually occurring at the junction. These effects are only observed after the CdCl2 activation treatment. Using High Resolution Transmission Electron Microscopy (HRTEM), we have discovered that unusual nanoscale defects are responsible for the formation of voids and subsequent device delamination. We have also observed the dynamics of defect growth using an in situ heating stage in the TEM. Using these atomic scale observations has provided a clear pathway to process improvement.
9:00 PM - ES14.9.16
A Review—Metastability, Potential Induced and Damp Heat Degradation and Recovery in CIGS Solar Cells
Shubhra Bansal 1 , Krishna Solasa 1 , Eric Ng 1
1 , University of Nevada, Las Vegas, Las Vegas, Nevada, United States
Show AbstractAbility to determine degradation rates of PV modules is important towards achieving the SunShot goals of $0.06/kWh and beyond as both degradation rates and product lifetime are key components of module price or LCOE calculations. Cu(In,Ga)Se2 photovoltaic devices have shown record efficiency of 21.7% and are globally manufactured at gigawatt levels. However, CIGS devices exhibit light or thermally induced metastability effects, which are not well understood. The paper will discuss a review of light, heat and bias induced transient effects in CIGS solar cells and correlate to the knowns and unkowns of stabilization of power output, transient effects and recovery during potential induced and damp heat degradation. A pareto of possible mechanisms contributing to transient effects in CIGS devices will also be discussed. The key focus will be on following effects:
(1) Understanding of light, heat and bias induced transient behavior: One of the most puzzling properties of high-efficiency CIGS devices is the appearance of metastabilities, which have accompanied the development of CIGS devices since its early days. These metastabilities are in most cases reversible parameter drifts of primarily Voc and FF with extremely long time constants. It is widely accepted that CIGS-based solar cells exhibit drifts in device parameters depending on illumination and temperature. However, the mechanisms are not well understood as a function of device processing parameters.
(2) Understanding of potential induced degradation: Several PV technologies have been found to be prone to in-field degradation related to the high system voltage in PV installations with serial connected modules. This effect is referred to as potential induced degradation (PID) due to potential-difference-driven degradation mechanisms. For CIGS, some studies suggest increase in series resistance of ZnO, while others stipulate Na concentration and ioinic mobility in glass as reasons for PID. It has also been observed that degradation can be reversed with light soak or applied bias. The known mechanisms of PID and subsequent recovery with light soak will be reported and gaps will be identified.
(3) Understanding of damp heat degradation: Similar to PID, it has been observed that degradation during damp heat testing of CIGS is reversible with current injection or light soaking. Again the known mechanisms of damp heat degradation and subsequent recovery as a function of processing variables will be reviewed.
The goal of the paper is to provide a comprehensive literature review and community feedback on knowns and unknowns of the fundamental mechanisms of metastability, potential induced and damp heat degradation and recovery in CIGS solar cells. The review should provide a basis for exchange of ideas and assist in identifying gaps for future research.
9:00 PM - ES14.9.17
Anomalous Reverse Breakdown of CIGS Devices—Theory and Simulation
Marco Nardone 1 , Saroj Dahal 1
1 , Bowling Green State University, Bowling Green, Ohio, United States
Show AbstractCopper indium gallium diselenide (CIGS) photovoltaic devices exhibit some interesting and anomalous characteristics under reverse bias. Experimentally, reverse breakdown voltages (Vbr) of around -4 V in the dark and -2 V in light have been observed in pristine CIGS cells. There is variation among cells and ‘defective’ spots can have much lower Vbr. Key experimental findings include: (1) the breakdown voltage decreases with increasing temperature when T > 200 K and increases with temperature when T < 200 K; (2) Vbr decreases with increasing light intensity and photon energy; and (3) Vbr decreases when sodium-rich glass substrate is used. In this work we describe theories that may explain the observations and we develop a 2D CIGS device simulation platform to calculate the reverse current-voltage (IV) characteristics. Several candidate physical models are possible and require further investigation. At the temperatures of interest, avalanche multiplication would occur at breakdown voltages greater than those observed for CIGS and the significant scattering processes in this polycrystalline material make it quite unlikely. Direct band-to-band tunneling is often considered a possibility when Vbr is close to 4 or 6 times the band gap energy and the temperature coefficient of Vbr is negative, which is the case here when T > 200 K. However, it is unlikely that under normal circumstances high enough doping concentrations are achieved in CIGS to create a sufficiently large field and thin energy barrier. It was shown that a Poole-Frenkel mechanism could reproduce the observed temperature and light dependence of the reverse IV characteristics of CIGS cells but further investigation is required. Trap-assisted tunneling and hopping conduction are also candidate mechanisms that require quantitative analysis. Another possibility is the presence of semi-shunts, which are conductive filaments that extend part-way through the absorber material. Field enhancement and tunneling at the tip of the semi-shunts can lead to reverse breakdown current that increases exponentially with voltage. Understanding the reverse bias behavior is important because it may be related to shading-induced failure of CIGS modules and it can also elucidate some aspects of the fundamental behavior of these devices.
9:00 PM - ES14.9.18
Clarification of Proton- and Electron-Irradiated Degradation Mechanism of Cu2ZnSnS4 Solar Cells
Mutsumi Sugiyama 1 , Satoru Aihara 1 , Yosuke Shimamune 2 , Hironori Katagiri 2
1 , Tokyo University of Science, Chiba Japan, 2 , National Institute of Technology, Nagaoka College, Nagaoka Japan
Show AbstractA kesterite Cu2ZnSnS4 (CZTS) solar cell shows relatively high conversion efficiency. The use of nontoxic and abundant elements is the main reason in this solar cell, as an alternative to Cu(In,Ga)Se2 (CIGS) or CdTe solar cells. Revealing the improving/degradation properties of solar cells after solar cell fabrication are very important for industrialization. These days, several improving/degradation mechanism of CIGS solar cell such as light soaking, heat soaking, proton- or electron-irradiation effect have been reported. On the other hand, there is almost no report dealing such effects on CZTS solar cell. In fact, the mechanism of solar cell improving/degradation has gradually become known as a complicated phenomenon.
Among several properties, proton- and electron- irradiation effects are essential and fundamental for several solar cells such as Si, GaAs, and CIGS. From a scientific viewpoint, an investigation of the proton- and electron- irradiation effects on CZTS thin films is needed to reveal the degradation mechanism of a CZTS solar cell. On the other hand, from an industrial viewpoint, this investigation is needed to clarify the long-term reliability for an accelerated test of practical applications. Therefore, in this study, the effects of proton- and electron- irradiation on the properties of a CZTS solar cell and several thin films composed of a CZTS solar cell will be investigated. We will reveal "where" and "how" the degradation of CZTS solar cell occurs.
Conventional CZTS solar cells were used for irradiation experiments. Proton and electron irradiation were carried out using an electron accelerator and ion implantation equipment, respectively. The proton-irradiation energy was fixed at 380keV and the fluence was varied between 1E12 and 3E16 cm-2. On the other hand, the electron-irradiation energy was fixed at 2MeV and the fluence was varied between 1E14 and 2E17 cm-2.
The normalized efficiency of the CZTS solar cell is investigated as a function of irradiation fluence. The parameters tended to decrease with increasing proton and electron irradiation fluence in the case of a fluence greater than approximately the order of 10E13 cm-2 and 10E15 cm-2, respectively. These trends are approximately consistent with previous reports of CIGS solar cells. As a result, CZTS solar cell shows excellent radiation tolerance for proton and electron like a CIGS solar cell. These characteristics make these solar cells extremely useful not only for conventional practical use but also for space applications.
9:00 PM - ES14.9.19
Compare Corrosion Resistance of Aluminum-Doped Zinc Oxide (AZO) and Gallium-Doped Zinc Oxide (GZO) Films Depending on the Hydrogen Content
Soo Ho Cho 1
1 Material Science Engineering, Koreatech, Cheonan Korea (the Republic of)
Show AbstractAluminum-doped zinc oxide (AZO) and Gallium-doped zinc oxide (GZO) films are promising transparent conducting material which can substitute for the commercially using indium tin oxide with its cheapness, safety to human body, easy supplying and chemical inertness. Especially, for the electrode of photovoltaic device such as copper indium gallium selenide(CIGS) solar cell that can be endured in the atmospheric humidify environment, its corrosion resistance is very important. AZO and GZO are the n-type semiconducting material to which trivalence Al and Ga are doped in the divalence Zn. And the electron conductivity of the hexagonal wurtzite material is very dependent on the oxygen vacancies. The hydrogens in the AZO and GZO influence a lot to the concentration of oxygen vacancies. So in this study, the influence of the hydrogen content on the structural, optical and electrical and especially compare AZO to GZO on the corrosion resistance by RF magnetron sputtered TCO films was investigated. As the hydrogen increased, the sheet resistance improved and over some point then decreased. Corrosion resistance measured by using potential stat polarization test showed a more dependent characteristic on the microstructure and grain boundaries of the film. corrosion resistance is effected of grain boundaries. Corrosion will occur form grain boundaries. Grain boundaries rate increase, grain size decrease. Important of AZO and GZO micro structural. AZO and GZO are hexagonal wurtzite materials, but chemical composition different. In crystalline of AZO and GZO, the grain boundaries of small grain with much hydrogen led a low polarization resistance mean a bad corrosion resistance, while in amorphous with much higher hydrogen it didn’t make a valid difference in corrosion resistance.
9:00 PM - ES14.9.20
Low Band Gap Cu(In,Ga)Se2 Absorber Layers for Current Matched Perovskite/CIGS Tandem Solar Cells
Thomas Feurer 1 , Fan Fu 1 , Enrico Avancini 1 , Stefano Pisoni 1 , Johannes Loeckinger 1 , Benjamin Bissig 1 , Shiro Nishiwaki 1 , Romain Carron 1 , Thomas Weiss 1 , Stephan Buecheler 1 , Ayodhya Tiwari 1
1 , Empa, Dübendorf Switzerland
Show AbstractMulti-junction solar cells are promising to reduce the thermalization losses that limit single-junction solar cell efficiency. In case of monolithic design current matched solar cells of different bandgap are grown on top of each other and interconnected in series, simplifying the structure and minimizing parasitic losses of additional interlayers. In order to achieve current matching with state of the art perovskite top cells with energy band gap of ~1.6 eV the energy band gap of the bottom cell should be ~1 eV with high spectral response in the near infrared region (NIR). In the chalcogenide absorber family the CuInSe2 compound fulfills the first requirement of the energy band gap; however, the NIR response is generally low. The weak NIR response originates from a low minority carrier lifetime besides the absence of drift forces (space charge region and/or back surface field). In this work we report our efforts to improve the charge carrier collection by implementing a back surface field introducing Ga grading towards the back with minimal increase in the optical band gap of the absorber. We have investigated the use of back contact passivation layers in order to reduce recombination on the Mo/CIGS interface and increased doping in CIS by different post deposition treatments resulting in improvements in PV parameters, namely VOC and fill factor. Such CIGS solar cells are then combined with highly efficient perovskite top-cells into tandem devices. We show that in such a combination current matching between top- and bottom-cell is possible and that a significant improvement over the single cell efficiencies can be achieved.
9:00 PM - ES14.9.21
Comparison of Low Bandgap CuInSe2 Alloys for Tandem Solar Cells
Nicholas Valdes 2 1 , William Shafarman 2 1
2 Department of Materials Science and Engineering, University of Delaware, Newark, Delaware, United States, 1 , Institute of Energy Conversion, Newark, Delaware, United States
Show AbstractLow bandgap Cu(In,Ga)Se2 (CIGS) is promising for the bottom cell in a tandem solar cell configuration, especially since its bandgap can be tuned to the ideal value for the bottom cell (≤ 1.1 eV) [1]. However, high efficiency CIGS devices using co-evaporated absorbers typically have bandgaps around 1.2 eV. Because bandgap tuning of CIGS occurs by alloying CuInSe2 with Ga, Ag, or S, it is essential to determine what alloy concentrations are needed for superior performance of CIGS solar cells with bandgaps ≤ 1.1 eV.
In this study, the impact of Ga and Ag alloy concentrations on the device performance and material properties of CuInSe2 based solar cells with Eg < 1.1 eV is investigated. CuInSe2 and its alloys with Ga only, Ag only, and both Ga and Ag (([Ga]/([Ga]+[In]) ≈ 0.05, [Ag]/([Ag]+[Cu]) ≈ 0.2) were grown using a three-stage elemental co-evaporation process. Films with Ag have larger grain sizes as in previous work with high Ga content [2] but alloy concentration did not have an effect on film orientation determined by x-ray diffraction. On complete devices, J-V measurements show an increased Voc with both Ga and Ag concentration, while Jsc increases with Ag alloying. The best device in this study to date had 13.6% efficiency with Voc = 498 mV, Jsc = 37.0 mA cm-2, and FF = 74.0%. External quantum efficiency show improved current collection in the long wavelength region due to the addition of Ag. This is advantageous for bottom cell performance in tandems, since the cell is only illuminated with long wavelength light. This improvement in collection is attributed to a larger space charge width due to a reduction in defect density from Ag incorporation [3]. All combinations of alloys in this study show low diode factors (1.2 to 1.4) and low Urbach energies determined by the QE tails. Therefore, optimal Voc, Jsc, FF, and efficiency can be obtained with proper Ga and Ag alloying while maintaining devices with minimal disorder. Further improvement is expected from the use of a potassium fluoride post-deposition treatment, and the effects of this treatment on low bandgap CuInSe2 based devices will be reported.
References
[1] T.J. Coutts, J.S. Ward, D.L. Young, K.A. Emery, T.A. Gessert, and R. Noufi, Prog. Photovolt: Res. Appl 11, 359 (2003).
[2] L. Chen, J. Lee, and W. Shafarman, IEEE J. Photovolt 4 (1), 447 (2014).
[3] P.T. Erslev, J. Lee, G.M. Hanket, W.N. Shafarman, and J.D. Cohen, Thin Solid Films 519, 7296 (2011).
9:00 PM - ES14.9.22
Cu(In,Ga)Se2-Based Monolithic Tandem Solar Cell with Open-Circuit Voltage over 1 V
Jae-Hyung Wi 1 , Won Seok Han 1 , Dae-Hyung Cho 1 , Woo-Jung Lee 1 , Chae-Woong Kim 2 , Chaehwan Jeong 2 , Jae Ho Yun 3 , Yong-Duck Chung 1 4
1 , Electronics and Telecommunications Research Institute (ETRI), Daejeon Korea (the Republic of), 2 , Korea Institute of Industrial Technology (KITECH), Gwangju Korea (the Republic of), 3 , Korea Institute of Energy Research (KIER), Daejeon Korea (the Republic of), 4 , Korea University of Science and Technology (UST), Daejeon Korea (the Republic of)
Show AbstractThin film multi-junction solar cell has been considered as the most promising structure for the next-generation photovoltaic (PV) devices. Double-junction solar cells based on polycrystalline thin films such as Cu(In,Ga)Se2 (CIGS) are desirable due to the controllability of the band-gap. In double-junction CIGS tandem cells, two solar cells including different In and Ga compositions, absorbs different wavelength regions of the solar spectrum. The CuGaSe2 (CGS, Eg ≈ 1.7 eV) absorber is used for top cell because CGS absorbs short wavelengths (< 800 nm), and the In-rich CIGS (Eg ≈ 1.0 ~ 1.2 eV) absorber is used for bottom cell because the it absorbs long wavelengths (< 1200 nm).
According to some reports, there are challenging technical issues in CIGS-based monolithic tandem solar cells such as durable CIGS bottom cell and low temperature of CGS top cell. The best cell efficiency of monolithic tandem cell was recorded of 2.7% with 0.759 V by W. N. Shafaman et al, until now. In this case, a bias light was not utilized in the external quantum efficiency (EQE) measurement, allowing distinguishable response of each sub-cell.
In this study, we fabricated the CIGS bottom cell in the double-junction solar cells, where the CGS is used as the top cell. For the monolithic tandem structure, the CGS (1 μm) layer was grown by a co-evaporation technique at substrate temperature of 400 oC on AZO (varied-thickness)/sputtered-Zn(O,S)/CIGS/Mo/SLG. The sputtered-Zn(O,S) buffer layer of bottom cell was used to endure the high-temperature in the top cell process. The AZO layer works as both back contact to the CGS top layer and front contact to the CIGS bottom layer. The best photovoltaic performances of Voc = 1.02 V, Jsc = 6.21 mA/cm2, FF = 51.3%, and efficiency = 3.24% were obtained from the tandem device using a 50 nm thick AZO interconnection layer. The EQE measurements distinguished each light-response from the top cell and the bottom cell using different light bias of a long wavelength (> 950 nm) and a short wavelength (< 550 nm), respectively. The EQE showed that the tandem device successfully collected the currents from the each cells. Furthermore, the Voc over 1 V was achieved, which indicates that the individual sub-cells were well electrically interconnected. The current matching could be improved by optimizing the optical and electrical properties of the both top and bottom cells.
9:00 PM - ES14.9.23
Textured CIGS/Perovskite Tandem Cells
Joop van Deelen 1 , Marco Barink 1
1 , TNO, Eindhoven Netherlands
Show AbstractWith the emergence of perovskite cell technology, there is an increased interest in thin-film heterojunction tandem cell. Because of the potentially low temperature and facile fabrication of perovskite cells, these would be good candidate to put on top of a CIGS cell. Such tandem would combine absorber materials with the highest efficiencies in thin film cells, with favourable and tunable band gaps, requiring only large scale coating processes. Just as already proven for Si based cells, also this type of cell architecture could benefit from texturisation.
For this purpose, the optical characteristics and the performance of perovskite/CIGS thin film tandem solar cells were studied by optical modeling. We took the ‘standard’ bandgaps of 1.55 Ev for the perovskite and 1.12 Ev for the CIGS, as these are available materials that have shown record efficiencies and their current densities are almost matching in the tandem cell. The texture dimensions were systematically varied over a wide range with periods from 400 nm to 1500 nm and heights from 0 to 3500 nm.
This way mappings of the performance of the individual cells, tandem cells and optical losses with texture dimensions were obtained including several maxima within the parameter window. Moreover, general trends could be derived and analysis of the data shows not only effective reduction of reflection, but also how light capturing in specific layers can be enhanced locally by using a texture. Small textures can focus the light in the perovskite, while larger textures lead to increased incoupling in the CIGS layer, even for similar height/period ratios of the texture. Current matching can, therefore be tuned, without changing the layer thickness.
Texturisation can improve the light incoupling in the perovskite/CIGS tandem cell up to 8 rel.%. Texture optimization was performed for 4-terminal tandem and 2-terminal tandem, which each have their specific demands and thereby their own optimum texture and efficiency gain. The 4-terminal tandem cell was calculated to be 26.8% for flat layers (including a fill factor of 0.8 and 4% reflection from the cover glass), which increased to 29.0% for the optimal texture. The 2-terminal cell increased from 26.5% to 28%.
9:00 PM - ES14.9.24
Understanding the Effect of Stainless Steel Substrates on Flexible CIGS Solar Cell Performance
Tara Nietzold 1 , Michael Stuckelberger 1 , Bradley West 1 , Barry Lai 2 , Joerg Maser 2 , Dmytro Poplavskyy 3 , Rouin Farshchi 3 , Jeff Bailey 3 , Mariana Bertoni 1
1 , Arizona State University, Tempe, Arizona, United States, 2 Advanced Photon Source, Argonne National Laboratory, Argonne, Illinois, United States, 3 , MiaSole Hi-Tech Corp., Santa Clara, California, United States
Show AbstractThe deposition of Cu(In,Ga)Se2 (CIGS) solar cells on stainless steel (SS) substrates has been of interest for quite some time due to the compatibility of thin-film deposition methods with flexible substrates and module efficiencies capable of 17% [1]. When compared to solar cells on glass substrates, flexible CIGS solar cells can be used in a wide array of custom applications that take advantage of their reduced weight, portability and low production costs through large-scale roll-to-roll processing [2, 3].
Studies on the influence of stainless steel on the performance of CIGS/Mo/SS devices have predominantly focused on metal impurity diffusion into the devices and have mostly covered the diffusion of elements from the stainless steel, particularly iron, into the CIGS layer [3]. Little is known regarding the effect of the substrate’s composition, structure and topology on the final device performance.
In this work, we perform a correlative pixel-to-pixel analysis at the nanoscale comparing X-ray fluorescence (XRF) versus X-ray beam induced current and voltage (XBIC and XBIV) on industry relevant flexible CIGS devices. Through 2-dimensional mapping of XRF spectra and XBIC/XBIV, we correlate elemental distributions of the absorber layer to the effect they have on charge collection within the complete solar cell. We extend this analysis to relate the distributions of majority elements in stainless steel, such as iron and chromium, and the influence that compositional variations have on electrical behavior.
We have observed regions of positive and negative correlations between different cation concentration, stainless steel element distribution and XBIC/XBIV results. Particularly, preliminary results show micrometer-size dark, low performing regions in the electrical data that correspond to variations within the stainless steel. This contribution looks to provide insight into the different effects that the stainless steel substrate composition, structure and topology have on nanoscale correlations between the elemental distributions in the CIGS absorber layer and solar cell performance.
[1] MiaSolé Launches Flex Series Modules At SPI. Retrieved October 14, 2016, from http://www.pv-magazine.com
[2] Kessler, F., et al. (2005). Approaches to flexible CIGS thin-film solar cells. Thin Solid Films.
[3] Shi, C., et al. (2009). Cu(In,Ga)Se2 solar cells on stainless-steel substrates covered with ZnO diffusion barriers. Solar Energy Materials and Solar Cells
9:00 PM - ES14.9.25
Predicting Ga and Cu Profiles in Co-Evaporated Cu(In,Ga)Se2 Using Modified Diffusion Equations and a Spreadsheet
Ingrid Repins 1 , Steven Harvey 1 , Karen Bowers 1 , Stephen Glynn 1 , Lorelle Mansfield 1
1 , National Renewable Energy Laboratory, Lakewood, Colorado, United States
Show AbstractCu(In,Ga)Se2 (CIGS) photovoltaic absorbers frequently develop Ga gradients during growth. These gradients vary as a function of growth recipe, and are important to device performance. The occurrence of such gradients and their effect on devices has been the subject of numerous experimental studies. The formation of the gradients has been explained qualitatively by the faster of diffusion of In (compared to Ga) toward regions of the film that are lower in concentrations of group III atoms. To date, implementation of Ga profiles is largely based on guided empiricism: Small changes are made to existing growth recipes based on qualitative understanding of the diffusion, then the film is grown according to the modified recipe, the new Ga profile is measured using profiling spectroscopy, and the process is repeated. This cycle can be a time-consuming process for the experimentalist.
Two issues have prevented prediction of Ga profiles using the simple diffusion equations that have been very useful for describing the behavior of atoms in other photovoltaic structures, e.g. the motion of boron in silicon. First, In and Ga atoms occupy the same lattice sites and thus diffuse interdependently. A treatment of such diffusion using the classic Fickian approach requires a knowledge of the chemical potential of the Cu-In-Ga-Se system as a function of composition, which is at present only partially documented. Second, some reported values for diffusion coefficients in the Cu-In-Ga-Se system vary by orders of magnitude, and even these values only partially map the composition space.
In this work, we show how the diffusion equation can be modified to account for site sharing between In and Ga atoms. This derivation constitutes a new approach that can be used to predict consequences of growth conditions, yet without knowing the chemical potential of the Cu-In-Ga-Se system. The analysis has been implemented in an Excel spreadsheet, and outputs predicted Cu, In, and Ga profiles for entered deposition recipes. A single set of diffusion coefficients and activation energies are chosen, such that simulated elemental profiles track with published data and those from this study. Extent and limits of agreement between elemental profiles predicted from the growth recipes and the spreadsheet tool are demonstrated. It is noted where the chosen diffusion coefficients and activation energies fall within the bounds of literature values, and where they differ. Based on insight from the growth simulation, qualitative arguments about the relationship between growth recipe and elemental profiles are refined.
Limits of this approach are discussed. Among other things, it applies only to Se-rich environments, does not account for changes in diffusion coefficients due to Na or morphology. On the other hand, the method does not require assumptions about rates, phases formed, or system chemical potential. The spreadsheet will be available for distribution.
9:00 PM - ES14.9.28
Pulsed Laser Deposition (PLD) of the CZTS Absorber Material for Solar Cells with up to 5.2% Efficiency
Andrea Cazzaniga 1 , Andrea Crovetto 1 , Stela Canulescu 1 , Rebecca Ettlinger 1 , Eugen Stamate 1 , Joan Ramis Estelrich 1 , Nini Pryds 1 , Yan Chang 2 , Kaiwen Sun 2 , Hao Xiaojing 2 , Ole Hansen 1 , Jorgen Schou 1
1 , TU Denmark, Roskilde Denmark, 2 , University of New South Wales, Sydney, New South Wales, Australia
Show AbstractCZTS (Cu2ZnSnS4 ) is a promising material for a solar cell absorber and consists of abundant and environmentally friendly elements. The efficiency of a cell based on this material has increased from 2% in 2001 to 9.4 % in 2015 [1].
Pulsed laser deposition (PLD) is usually considered as a technique, by which a material with a complicated stoichiometry can be transferred from a target to a growing film in a vacuum chamber. It is widely used in scientific labs, in particular for production of complex oxide films [2]. PLD is a non-equilibrium film deposition technique, and since the energy source is outside the deposition chamber, the parameter space is huge for all physical parameters such as the background gas pressure, the substrate temperature, the target-substrate distance and the laser fluence.
For complicated compounds which are produced from sintered, multi-element targets, the composition of the films can deviate significantly from that of the target, depending on the laser fluence. By tuning the laser fluence with a 248-nm excimer laser beam on a sintered CZTS target (2CuS:ZnS:SnS), the metal ratios in the film can be controlled and the film made slightly copper-poor. An ultra-thin CZTS absorber(400 nm) was deposited in high vacuum (p < 5 10-6 mbar) on a Mo-coated soda lime glass substrate at room temperature. At low fluence below 1 J/cm2 the copper content in the film is low, while at high fluence the films become copper-rich. At a fluence 0.7 J/cm2 we have achieved a ratio Cu/(Sn+Zn) ~ 0.85 which is in the right regime for an operating solar cell [3,4].
A drawback of PLD is the generation of micron-sized droplets on the surface which typically originates from violent heating of the surface and the adjacent layers by the laser beam. At the low fluence, 0.6-0.8 J/cm2 used here, the number of droplets turned out to be relatively low, (and any sign of droplets turned out to disappear after sulfurization).
The sulfurization and the final deposition of the buffer and window layers were carried out at University of New South Wales, Sydney, Australia [3]. The best cell with an effective area of 21 mm2 and only a 400-nm thick CZTS layer had a remarkably high efficiency of 5.2 %, a Voc = 616 mV, a Jsc = 17.6 mA/cm2 and a fill factor of 48 %. This efficiency is the highest one from a cell obtained with an absorber layer of CZTS produced by PLD until now [3]. Despite the ultra-thin absorber, there are no signs that material and junction quality are significantly lower than that of thicker absorbers: grain size, carrier lifetimes, collection efficiency, shunt resistance, and dark saturation current are all similar to (thicker) benchmark CZTS solar cells.
[1] S. Tajima et al. , Prog. Photov. Res. Appl. (2016), in press.
[2] Pulsed laser deposition of thin films, ed. R. Eason, Wiley (2007)
[3] A. Cazzaniga, A. Crovetto et al. submitted to Prog. Photov. Res. Appl.
[4] K. Ito Copper Zinc Tin Sulfide-based thin film solar cells, Wiley (2015)
9:00 PM - ES14.9.29
Pulsed Laser Deposition of Thin Films of Chalcogenides
Stela Canulescu 1 , Jorgen Schou 1
1 , Technical University of Denmark, Roskilde Denmark
Show AbstractCu2ZnSnS4 (CZTS) is an emerging p-type quaternary chalcogenide material consisting of non-toxic and abundant elements. The complex composition of CZTS makes it difficult to obtain a single phase material with conventional thermal growth techniques. In this paper we have investigated the synthesis of kesterite CZTS thin films by laser ablation of a sintered target in vacuum. The deposition of CZTS was assisted by a sulfur evaporation beam that aims to compensate for possible sulfur losses in the growing films. The phase formation of the CZTS deposited with one-step pulsed laser deposition approach was examined using both X-ray diffraction and Raman spectroscopy, both of which are shown to be complimentary tools for phase identification. A significant increase in the grain size as well as structural ordering in CZTS was observed with increasing synthesis temperature. We demonstrate that by controlling the synthesis temperature, high purity CZTS phase can be obtained at a substrate temperature of 4500C.
9:00 PM - ES14.9.30
The Effect of Annealing Temperature on CIGS Solar Cell
Yasir Alrikabi 1 , Tar-pin Chen 1
1 , University of Arkansas at Little Rock, Little Rock, Arkansas, United States
Show AbstractOur work is focusing mainly on the effect of annealing temperature of two different annealing methods like sulfurization under vacuum with sulfur powder and selenization in Ar filled tube with selenium powder on the structural morphological, optical and electrical characteristics of CIGS thin films solar cells prepared by sputtering method from quaternary CIGS target. XRD measurements show that the films are in crystalline structure. Annealing temperature in both Sulfurization under vacuum and Selenization in Se ambient atmosphere improving the crystalline structure, and the grain size of the prepared films. From the absorption and transmission spectrum we have been able to calculate the band gap energy which increases in the range between 0.97-1.31 eV with the annealing temperature increase. I-V curve of CIGSe device show a better performance than the CIGS device that could be because of the grain size is grown better.
9:00 PM - ES14.9.31
Photoluminescence Study of Pentanary Ag-Containing Chalcopyrite Solar Cells
Abhinav Chikhalkar 1 , Aymeric Maros 2 , William Shafarman 3 , Richard King 2
1 School for Engineering of Matter, Transport & Energy (SEMTE), Arizona State University, Tempe, Arizona, United States, 2 School of Electrical, Computer and Energy Engineering (ECEE), Arizona State University, Tempe, Arizona, United States, 3 Institute of Energy Conversion (IEC), University of Delaware, Newark, Delaware, United States
Show AbstractCu(In,Ga)Se2 (CIGS) is one of the most promising materials for high efficiency thin-film photovoltaics. It has many desirable characteristics such as a high absorption coefficient, a tunable bandgap between 1.05eV and 1.68eV and low recombination grain boundaries for low Ga compositions. It has been found that adding silver to the alloy allows a higher Ga mole fraction to be incorporated into the films while maintaining a low recombination rate and a good surface morphology for solar cells [1]. These properties make (Ag,Cu)(In,Ga)Se2 (ACIGS) an interesting candidate for the top cell material in low-cost, flat-plate, high-efficiency tandem solar cells.
This study is among the first to probe the defect physics of this extremely promising pentanary compound for tandems using temperature and power dependent Photoluminescence (PL) spectroscopy. To focus on effects that stem from the addition of Ag, a comparative study of ACIGS cells with the corresponding Ag-free CIGS composition cells is presented. For this work, high efficiency (~16.5%) ACIGS cells with varying Ag:Cu ratio were fabricated at IEC using a three-stage process [2]. The baseline structure of the cells was soda-lime glass/Mo/ACIGS/CdS/ZnO/ITO. The (Cu+Ag):(In+Ga) ratio was below 0.90 and the Ga:(Ga+In) ratio was varied to obtain the desired bandgap.
From the PL spectra, four major peaks corresponding to band-to-band (BB), band-to-tail (BT) and band-to-impurity (BI) transitions are observed at 1.18 eV, 1.11 eV, 1.06 eV and 1.00 eV respectively; the latter two corresponding to BI transitions. A linear increase in energy position of the BT peak with excitation powers and eventual saturation of that increase at higher intensities suggest a donor-acceptor pair recombination mechanism [3]. Variation in the relative PL intensity and the blue shift of the BT peak over a range of excitation powers, points to the mechanism of filling up of the shallow hole states in the valence band tail and could be used to determine the defect density. The temperature dependence of the integrated PL intensity was used to determine the activation energy of donor defects. It was observed that at low temperatures the BT peak prevails, while at higher temperatures the BB peak starts to dominate. This behavior confirms the localization of holes in the shallow band tails.
Further characterization with admittance spectroscopy and quantum efficiency measurements to establish a correlation between these defect characteristics and the device performance will also be presented.
[1] W.N. Shafarman, C. Thompson, J. Boyle, G. Hanket, P. Erslev and J. D. Cohen, Proc. 35th IEEE Photovoltaic Specialists Conference, 325 (2010).
[2] L. Chen, J. Lee, and W.N. Shafarman, IEEE J. Photovoltaics 4, 447 (2014).
[3] M. Wagner, I. Dirnstorfer, D.M. Hofmann, M.D. Lampert, F. Karg, and B.K. Meyer, Phys. Status Solidi A 167, 131 (1998).
9:00 PM - ES14.9.32
Monolithic Tandem Devices Demonstrating over 1.2 V Voc Using a Wide-Bandgap Chalcopyrite Absorber
Kim Horsley 1 , Alex DeAngelis 1 , Thomas Hellstern 2 , Thomas F. Jaramillo 2 , Nicolas Gaillard 1
1 , Hawaii Natural Energy Institute, Honolulu, Hawaii, United States, 2 , Stanford University, Stanford, California, United States
Show AbstractWhile power conversion efficiencies (PCE) over 20% have been demonstrated with single-junction photovoltaics (PV) using chalcopyrites, these efficiencies are ultimately limited by the theoretical Shockley-Queisser limit. Multi-junction (MJ) PV devices offer one way to surpass current efficiencies and theoretical limitations, and have already demonstrated PCE over 40% utilizing costly III-V materials.
Furthermore, future power grids will need to include not only efficient solar energy converters, but also advanced systems for energy storage, such as hydrogen-based storage and conversion systems (fuel cells). Photo-induced water splitting is a promising technology identified by the Department of Energy for renewable hydrogen generation. Solar-to-hydrogen (STH) efficiencies greater than 15% have already been demonstrated with III-V MJ devices, integrated either as PV-electrolysis systems (STH = 24 %) or monolithically integrated as photoelectrochemical (PEC) devices (STH = 16 %). In both approaches, the use of MJ devices was integral to generate the voltage required to split water into molecular hydrogen and oxygen (> 1.23 V).
Here, we report our efforts to create a cost-effective chalcopyrite-based MJ device for high-efficiency photovoltaics and PEC water splitting. There are a number of obstacles present to achieving these high efficiency devices, which we aim to address. We will first report on our strategy to develop a temperature-resistant TCO intermediate window layer which can withstand the deposition condition for the top junction. We demonstrate the effect of high temperature annealing (500 °C) on the optoelectronic properties of several TCO candidates; indium tin oxide (ITO) and indium molybdenum oxide (IMO). We will also report on the solid-state properties of silicon p/n junctions coated with our TCOs as a function of post-deposition annealing temperature, to simulate the effect of the wide-bandgap material synthesis. Finally, our results on monolithic Si/chalcopyrite tandem devices using ITO and IMO as intermediate TCO window will be presented, including an open circuit voltage > 1.2 V, achieved with CuGaSe2/Si structures. Results thus far were achieved using a CdS buffer, with expected non-ideal band energetics for wide-bandgap chalcopyrites. Potential avenues for future improvement, especially focusing on the wide-bandgap chalcopyrite junction, will be discussed.
Symposium Organizers
Ingrid Repins, National Renewable Energy Laboratory
Shubhra Bansal, University of Nevada, Las Vegas
Sascha Sadewasser, International Iberian Nanotechnology Laboratory
Edgardo Saucedo, IREC
Symposium Support
Catalonia Institute for Energy Research (IREC)
Dr. Eberl MBE-Komponenten GmbH
First Solar
International Iberian Nanotechnology Laboratory
National Renewable Energy Laboratory
ES14.10: Advanced Characterization
Session Chairs
Daniel Abou-Ras
Thomas Kunze
Thursday AM, April 20, 2017
PCC North, 200 Level, Room 229 B
9:30 AM - *ES14.10.01
Alternative Buffer Materials and Their Electronic and Chemical Properties in Cu(In,Ga)(S,Se)2 and CdTe Thin-Film Solar Cells
Clemens Heske 1 2
1 , University of Nevada Las Vegas (UNLV), Las Vegas, Nevada, United States, 2 , Karlsruhe Institute of Technology (KIT), Eggenstein-Leopoldshafen Germany
Show AbstractAmong the various technologies for thin-film solar cells, Cu(In,Ga)(S,Se)2- (CIGSSe-) and CdTe-based systems rank prominently, having reached conversion efficiencies well above 20%. At first glance, then, one would think that “we are done”. However, the opposite is true: much remains to be understood and optimized, in particular when alternative buffer materials (i.e., materials that deviate from the traditional materials in CdTe- or CIGSSe-based devices) are to be employed.
Using a tool chest of electron and soft x-ray spectroscopic methods, it is possible to unravel (some of) the secrets of the alternative buffers and their interfaces with the various solar cell absorber materials. This tool chest includes lab-based photoelectron spectroscopy (PES) with x-ray and UV excitation (XPS and UPS, respectively), inverse photoemission spectroscopy (IPES), and x-ray-excited Auger electron spectroscopy (XAES). These techniques are complemented by soft x-ray emission (XES) and absorption (XAS) spectroscopy using high-brilliance synchrotron radiation (in our case at Beamline 8.0.1 of the Advanced Light Source, Berkeley Lab). While the electron-based techniques are very surface-sensitive, XES and XAS are photon-in-photon-out techniques that probe the bulk region near the surface. In the talk, experiments to gain insights into alternative buffer materials and their absorber interfaces will be presented, in particular in view of band alignment and intermixing behavior, and the impact of these properties on the performance of corresponding solar cell devices will be discussed.
10:00 AM - ES14.10.02
Synchrotron-Based In Situ Characterization of CuInxGa1-xSe2 Solar Cells—Nanoscale Performance under Operating Conditions
Michael Stuckelberger 1 , Bradley West 1 , Tara Nietzold 1 , Barry Lai 2 , Joerg Maser 2 , Mariana Bertoni 1
1 , Arizona State University, Tempe, Arizona, United States, 2 , Argonne National Laboratory, Lemont, Illinois, United States
Show AbstractPolycrystalline absorber layers have a huge potential for low-cost electricity generation. However, the performance of solar cells with polycrystalline thin-film absorber layers such as CuInxGa1-xSe2 (CIGS), is very often limited by inhomogeneously distributed nanoscale defects[1] and stoichiometric inhomogeneities at grain boundaries, which are the limiting factor in thin-film solar cells where grain sizes are often on the order of the diffusion length[2].
For decades, polycrystalline absorbers like CdTe and CIGS with adjacent junction and contact layers have been optimized following an “Edisonian” approach, which has led to remarkable efficiencies. However, the next step in device optimization to increase the module-level efficiencies requires fine tuning and greater fundamental understanding of the limiting defects. Promising approaches include in-situ and operando methods using correlative synchrotron-based microscopy techniques such as X-ray fluorescence (XRF) and X-ray beam induced current (XBIC)[3]. Using these, functional materials and their defects can be understood and designed rather than optimized by empiricism.
We have modified a heating stage used for in-situ tracking of the elemental distribution during CIGS growth at nanoscale by XRF[4] such that we can now simultaneously measure the XBIC signal of complete solar cells under different operating conditions, which gives access to the charge collection efficiency with nanoscale lateral resolution. Furthermore, we have developed the X-ray beam induced voltage (XBIV) measurement technique. By measuring the voltage instead of the current, the sensitivity to topological inhomogeneities decreases, whereas the sensitivity to bandgap and recombination variation increases. Applying a bias voltage during XBIC measurements, we have access to the charge collection efficiency at any power point, at any measurement spot.
In this presentation, we demonstrate for the first time correlative XBIC, XBIV and XRF results of CIGS solar cells under real operating conditions for which we varied temperature, bias light, and bias voltage. In particular, we highlight how the role of different nanoscale defects and macroscopic underperforming regions change with temperature. We will share experimental details and discuss the potential of the combination of XRF with XBIC and XBIV, which enables full J-V curve measurements at nanoscale resolution under operating conditions coupled with high elemental sensitivity.
[1] Bertoni, M. I. et al. Energy Environ. Sci. 4, 4252 (2011)
[2] Siebentritt, S. Sol. Energy Mater. Sol. Cells 95, 1471–1476 (2011)
[3] West, B. et al. J. Synchrotron Radiation – In Press (2016)
[4] Chakraborty, R. et al. Rev. Sci. Instrum. 86, 113705 (2015)
10:15 AM - ES14.10.03
In Situ Nano Elemental Mapping to Visualize the Growth of Cu(In,Ga)Se2 Thin Films
Bradley West 1 , Michael Stuckelberger 1 , Robert Lovelett 2 , Sina Soltanmohammad 2 , Barry Lai 3 , Joerg Maser 3 , William Shafarman 2 , Mariana Bertoni 1
1 School of Electrical, Computer and Energy Engineering, Arizona State University, Tempe, Arizona, United States, 2 Department of Chemcial and Biomolecular Engineering, University Of Delaware, Newark, Delaware, United States, 3 Advanced Photon Source, Argonne National Laboratory, Argonne, Illinois, United States
Show AbstractThe ability to measure and visualize film properties during growth allows for direct insight to how grains are formed, how temperature impacts elemental segregation/phase formation, and how quickly diffusion processes and ripening take place. It is necessary to understand the behavior of a material after growth, but for its engineering, it is even more powerful to understand how the material achieved these properties. This is particularly true for complex quaternary systems such as Cu(In,Ga)Se2. Many current in-situ characterization techniques end up averaging over a large area, limiting the ability to visualize where the processes take place. In order to fill this gap, we have designed a temperature and atmosphere controlled in-situ stage that can be installed at the advanced photon source beamline 2-ID-D and is compatible with nano-x-ray fluorescence (XRF) measurements at temperatures up to 600°C with ~200 nm spatial resolution [1].
In this work, we have collected high-resolution 7 x 7 μm2 x-ray fluorescence maps with a step size of 200 nm and 50 ms dwell time. Maps were collected throughout the duration of industrially relevant CIGS precursor reaction growth processes. Sputtered precursor layers of Cu, In, Ga, with elemental ratios of [Ga]/[Ga+In] = 0.3 and [Cu]/[In+Ga] = 0.8, are capped with 10 μm of evaporated Se. Following the process reported by Berg et al., these films are then rapidly heated to 600 °C for 25 min and quickly cooled to room temperature [2]. From these data we are able to track effects like co-segregation and ripening over time. We have found that during the first ~12 minutes Ga and In tend to have a correlated segregation behavior, and after that time they correlate negatively though the remainder of the growth. The negative correlation is expected based on stoichiometry. We are also able to track individual particles to observe phenomena such as ripening. For example, we are able to track the concentration of multiple copper particles through time and evaluate the rate at which they grow or dissolve into the film – Fluxes of Cu ~2 ng●μm2/s have been observed. Measuring these same properties with varied atmosphere and time-temperature profiles gives unique insight into how much control and tuning is possible during growth. Subsequent evaluation of the electrical performance allows us to build a kinetic-based model for growth optimization.
[1] R. Chakroborty, et al., Review of Scientific Instruments (2015)
[2] D. M. Berg, et al., IEEE PVSC proc., Denver (2014)
10:30 AM - ES14.10.04
Nanoscale Photovoltaic Performance throughout 3-Dimensions in CdTe Solar Cells via Tomographic AFM
Yasemin Kutes 1 , Justin Luria 1 , Katherine Atamanuk 1 , Andrew Moore 2 , Lihua Zhang 3 , Eric Stach 3 , Bryan Huey 1
1 , University of Connecticut, Storrs, Connecticut, United States, 2 Physics, Colorado State University, Fort Collins, Colorado, United States, 3 Center for Functional Nanomaterials, Brookhaven National Lab, Upton, New York, United States
Show AbstractAlthough many solar cell technologies are industrially mature, questions still often remain about the role of microstructure and interfaces on ultimate device properties. This is largely due to experimental challenges in measuring photovoltaic performance with the requisite spatial resolution, especially throughout the thickness of a device stack. Computed Tomography (CT) based on x-rays, TEM, or SEM are increasingly employed in such circumstances, providing 3-d maps of chemical composition, grain sizes, and grain orientation. For 3-d mapping of properties, however, CT-AFM is uniquely introduced leveraging hundreds of serial sections through the 2 um thickness of a CdTe absorbing layer. With contrast arising from measuring photocurrents as a function of bias with and without in-situ illumination, correlations between microstructure and properties are clearly revealed at the nanoscale. This includes confirmation for CdTe of extensive electron conductivity localized at some (but not all) grain boundaries. But more importantly, a network of hole-conduction pathways along planar defects is also observed, supported by complementary TEM studies. Significant heterogeneities for inter- and intra- granular interfaces are reported as well, altogether leading to 10x variations in local solar cell performance. Such novel insight into these interconnected mechanisms may enable future microstructural engineering to further optimize the performance and reliability of CdTe and other polycrystalline solar cells.
10:45 AM - ES14.10.05
Studies of Atom Disorder in Cu2ZnSnSSe4 and Ag2ZnSnSSe4 Alloys
David Cherns 1 , Ian Griffiths 1 , Michael Lloyd 2 , Brian McCandless 2 , Talia Gershon 3 , Richard Haight 3 , Douglas Bishop 3
1 Physics, University of Bristol, Bristol United Kingdom, 2 Institute of Energy Conversion, University of Delaware, Newark, Delaware, United States, 3 , TJ Watson Research Center, Yorktown Heights, New York, United States
Show AbstractCu2ZnSnSSe4 (CZTSSe) has been considered as an absorber material to replace CuInGaSe2 (CIGS) in thin film solar cells owing to its similar crystal and electronic properties, and the relative abundance of its constituent elements. However, the highest efficiencies for CZTSSe cells are around 10%, much less than for CIGS equivalents. It has been suggested that cation antisite defects, such as CuZn, ZnCu, SnCu etc may be responsible. We have used scanning transmission electron microscopy (STEM) at sub-atomic resolution to identify high densities of extended boundaries in Cu2ZnSnS4 films which contain planes of atoms in antisite positions. These so-called antisite domain boundaries may be charged, providing potential barriers and scattering centres for charge carriers.
In this paper we use high resolution STEM to compare Cu2ZnSnSSe4 and Ag2ZnSnSSe4 films. The substitution of Ag for Cu is of interest as there should be a larger energy barrier for exchange of Ag and Zn ions compared to that for Cu and Zn, thus inhibiting the formation of antisite defects in the Ag-substituted alloys. To investigate antisite disorder, we have oriented individual crystals in the STEM along the kesterite [010] direction such that, for perfectly ordered crystals, the individual end-on atom columns contain a single atom species. By taking high angle annular dark field (HAADF) images, which give atomic number contrast, we aim to distinguish atom columns containing Sn, Ag and Zn in Ag2ZnSnSSe4. In contrast, Cu and Zn columns in Cu2ZnSnSSe4 are virtually indistinguishable. The HAADF images of Ag2ZnSnSSe4 films allowed us to identify the Ag columns, and to confirm a kesterite structure. Samples of Ag2ZnSnSSe4 powders grown at 500°C and then either quenched into an ice bath to potentially freeze in cation disorder or cooled at 3°C/hr to allow ordering to take place showed similar results, with both showing some evidence of cation disorder. In contrast, although the Ag2ZnSnSSe4 films contained antisite domain boundaries, the densities of these defects were much lower than in Cu2ZnSnSSe4 films.
The paper will compare the STEM evidence for cation disorder in Cu2ZnSnSSe4 and Ag2ZnSnSSe4 films. The extent to which Ag reduces the formation of antisite defects will be discussed.
ES14.11: Defects
Session Chairs
Igor Sankin
Dragica Vasileska
Thursday PM, April 20, 2017
PCC North, 200 Level, Room 229 B
11:30 AM - ES14.11.01
Three-Dimensional Lifetime Tomography Reveals How CdCl2 Improves Recombination throughout CdTe Solar Cells
Edward Barnard 1 , Benedikt Ursprung 1 , Eric Colegrove 2 , Helio Moutinho 2 , Nicholas Borys 1 , Brian Hardin 3 , Craig Peters 3 , Wyatt Metzger 2 , P James Schuck 1
1 , Lawrence Berkeley National Lab, Berkeley, California, United States, 2 , National Renewable Energy Laboratory, Golden, Colorado, United States, 3 , PLANT PV, Alameda, California, United States
Show AbstractAccurately measuring and untangling bulk, interface, and grain-boundary carrier recombination is one of the greatest challenges in evaluating thin-film materials in electro-optical devices. Here, we report previously-unknown insight into how CdCl2 passivates CdTe photovoltaic films by characterizing how local, sub-surface defects such as buried interfaces and grain boundaries affect carrier recombination – dynamics that are inaccessible with traditional optical techniques. Using two-photon lifetime tomography, we map with diffraction-limited spatial resolution the optically-excited carrier lifetimes in polycrystalline CdTe photovoltaic devices. These novel and unique three-dimensional (3D) maps reveal that CdCl2 treatment of CdTe solar cells – a ubiquitous processing step for these materials – suppresses non-radiative recombination and enhances carrier lifetimes throughout the film with substantial improvements particularly near grain boundaries. Moreover, we demonstrate that the CdCl2 treatment effectively passivates the critical buried interface at the p-n junction, leaving the excited state dynamics of the treated films significantly more homogeneous. These detailed microscopic lifetime tomography maps provide a clear, subsurface understanding of the photophysical changes that occur throughout polycrystalline CdTe devices. The results immediately impact routes for improving solar cell efficiencies, and generally demonstrate how 3D lifetime tomography can characterize previously hidden dynamics in a wide range of optoelectronic materials and devices.
11:45 AM - ES14.11.02
Correlations of Grain-Boundary Character with Electrical and Optoelectronic Properties of CuInSe2 Thin Films
Daniel Abou-Ras 1 , Norbert Schafer 1 , Thorsten Rissom 1 , Madeleine Kelly 2 , Jakob Haarstrich 3 , Carsten Ronning 3 , Gregory Rohrer 2 , Anthony Rollett 2
1 , Helmholtz-Zentrum Berlin, Berlin Germany, 2 Department of Materials Science and Engineering, Carnegie Mellon University, Pittsburgh, Pennsylvania, United States, 3 Institut für Festkörperphysik, Friedrich Schiller Universität Jena, Jena Germany
Show AbstractThin-film solar cells based on polycrystalline Cu(In,Ga)Se2 absorbers exhibit record conversion efficiencies of up to 22.6%. There is still a lack of a quantitative connection between the grain-boundary character distribution (GBCD) and the corresponding electrical and optoelectronic properties. The present work uses microstructural data from a CuInSe2 thin film acquired by electron backscatter diffraction (EBSD) to evaluate the GBCD. The most prominent features of the GBCD of CuInSe2 are Σ3 twin boundaries and the Σ9 and Σ27a symmetric tilt grain boundaries. Moreover, combining EBSD with electron-beam-induced current and cathodoluminescence (measurements on the same identical area) on a CuInSe2/Mo/glass stack provide the means to relate the grain-boundary character with the corresponding electrical and optoelectronic signals across the grain boundary. In part, determining this relationship is accomplished by means of correlation analysis using measurement data from more than 100 grain boundaries. However, the crystallographic, electrical and optoelectronic data showed no strong correlations, which is attributed to atomic reconstruction found in atomic planes adjacent to planar defects in polycrystalline CuInSe2 thin films and corresponding reductions of excess charge densities at these defects. The present work has been published in: D. Abou-Ras et al., Acta Mater. 118 (2016) 244.
12:00 PM - ES14.11.03
Phosphorus, Arsenic and Antimony Diffusion and Doping in Polycrystalline CdTe
Eric Colegrove 1 , Ji-Hui Yang 1 , Steven Harvey 1 , John Moseley 1 , Soren Jensen 1 , Helio Moutinho 1 , Joel Duenow 1 , David Albin 1 , Su-Huai Wei 2 , Mowafak Al-Jassim 1 , Wyatt Metzger 1
1 , National Renewable Energy Laboratory (NREL), Lakewood, Colorado, United States, 2 , Beijing Computational Science Research Center, Beijing China
Show AbstractRecent high voltage single-crystal (sX) CdTe devices have been achieved in large part because of vastly improved absorber material quality. Group V elements phosphorus, arsenic, and antimony are promising as p-type, anion site, dopants that may enable absorber material with high hole concentration and long minority carrier lifetime.
In this work, P, As, and Sb are investigated as p-type dopants in polycrystalline (pX) CdTe using ex-situ processing. Group V elements are diffused into sX and pX samples in Cd-rich conditions. Dynamic and time-of-flight secondary ion mass spectroscopy are used to measure 1D, 2D, and 3D dopant incorporation profiles. Analytical models are used to fit diffusion profiles and extract temperature dependent diffusion parameters for bulk and grain boundary (GB) diffusion mechanisms. Arrhenius behavior of diffusion parameters is compared to density functional theory calculations to evaluate possible atomistic mechanisms for diffusion. All group V elements exhibit slow bulk diffusion through Te sites and significantly faster GB diffusion, but P also exhibits fast, interstitial, bulk diffusion. While significantly higher concentrations of group V elements are found at GBs, dopants can diffuse into the grain interiors of pX films.
The temperature dependent diffusion parameters are used in analytical models to simulate incorporation profiles for multiple diffusion conditions. Capacitance-Voltage is used to measure hole concentrations in typical superstrate device stacks following these ex-situ group V diffusion processes and compared to modeled incorporation profiles to evaluate activation percentages. Greater than 5x1015 cm-3 hole concentrations can regularly be achieved with As and P; lower concentrations can be achieved with Sb. Activation percentages for all group V dopants incorporated by diffusion are similar. Between 0.1% and 1% of the incorporated dopant is contributing to the net hole density suggesting that the low hole concentrations achieved with Sb is a result of lower incorporation. While the doping concentrations for P and As are encouraging, the other 99% of the group V elements are not activated and can be detrimental to device performance. Cathodoluminescence and electron-beam induced current measurements of device cross sections indicate that aggressive conditions can create significant recombination at GBs and gain interiors. Time resolved photoluminescence measurements fit with a bi-exponential function at times exhibit a short τ1 component of < 1ns but a longer τ2 component between 1 and 20ns. One explanation could be that the grain boundaries are recombination centers due to the high concentrations of group V elements while the grain interiors are effectively doped with relatively few deep defects. These results suggest group V elements can effectively dope pX CdTe with few compensating defects, however a tradeoff exists between increased doping by ex-situ processes and lifetime.
12:15 PM - ES14.11.04
Influence of Defects Interactions on the Properties of CdTe:Cl,Cu Solar Cell Absorber
Dmitry Krasikov 1 , Igor Sankin 1 , Andenet Alemu 1
1 , First Solar Inc, Perrysburg, Ohio, United States
Show AbstractThe maturing of thin-film CdTe PV technology has encouraged significant research effort on studying the formation and properties of point defects in CdTe solar absorbers. This work investigates possible association of point defects in CdTe absorbers and the impact of the resulting complexes on performance and stability of thin-film CdTe PV devices.
After systematic analysis of defect reactions in 216-atoms supercell using range-separated hybrid HSE06 functional, we found a number of complexes with negative association enthalpies including some donor-donor complexes, although it is usually supposed that only donor-acceptor association is possible. We establish that all major complexes formed in CdTe after Cl-treatment and Cu-doping stages are either neutral or demonstrate donor character in p-type CdTe. Based on our analysis we suggest that association of point defects plays an important role in doping compensation, passivation of recombination centers and performance instabilities and thus, should not be overlooked in the analysis of solar cell performance. Using monomolecular and bimolecular defect chemistry reactions formalism, we build a kinetic mechanism for defects interactions and feed this mechanism with parameters obtained from first principles calculations. We use this kinetic mechanism to simulate the kinetics of defect chemistry reactions occurring in CdTe absorber during fabrication and stress-measurements to explain the experimentally observed phenomena.
12:30 PM - ES14.11.05
Voltage Dependent Admittance Spectroscopy for the Detection of near Interface Defect States and the Extraction of the Doping Density
Thomas Weiss 1 , Shiro Nishiwaki 1 , Benjamin Bissig 1 , Enrico Avancini 1 , Stephan Buecheler 1 , Ayodhya Tiwari 1
1 Laboratory for Thin Films & Photovoltaics, Empa - Swiss Federal Laboratories for Materials Science and Technology, Dübendorf, Zürich, Switzerland
Show AbstractAdmittance spectroscopy is capable of detecting defect states within the bandgap of semiconductors and is widely applied for absorber layers employed in photovoltaic devices. However, generally these measurements are performed without intentional bias voltage across the device and therefore are limited to the detection of defect states which cross the Fermi level. Applying a dc bias during the admittance measurements allows shifting the Fermi level with respect to the band edges and therefore enables the detection of defect states further inside the bulk as well as closer to the interface to the emitter.
We conducted admittance measurements under varying dc bias voltages on Cu(In,Ga)Se2 based solar cells. Absorber layers were treated with different alkaline post deposition treatments (PDT) and yielded efficiencies above 18 % (w/o anti-reflective-coating). A capacitance step is observed in forward biased admittance measurements, which can be attributed to a near interface defect state. This capacitance step is not observed under zero bias conditions. The interpretation of such a defect state confined to the near interface region is corroborated by SCAPS simulations.
CIGS based solar cells are based on a heterointerface between the absorber and the buffer layer and are therefore prone to exhibit interface defect states which might limit the solar cell performance. In fact, PDT treatments of various alkalis are applied to modify the near interface region of the absorber layer and opened up the possibility to grow solar cell devices beyond 22% efficiency. Therefore, being able to detect defect states in the absorber region close to the hetero interface by forward biased admittance spectroscopy is an important tool to further optimize the absorber front surface region.
Additionally, the deduction of the doping density from voltage dependent admittance spectroscopy will be discussed. Especially the capacitance contribution in forward bias due to the near interface defect state can be accounted for. Therefore, by identifying the SCR capacitance from the admittance spectrum (for each bias voltage) it is possible to fit its voltage dependence and therefore obtain the built-in voltage and the doping density. Such discrimination between various capacitance contributions is generally not possible using CV measurements.
12:45 PM - ES14.11.06
Structural and Compositional Dependence of the CdTexSe1-x Alloy Layer Photoactivity in CdTe-Based Solar Cells
Jonathan Poplawsky 1 , Wei Guo 1 , Naba Paudel 2 , Amy Ng 3 , Karren More 1 , Donovan Leonard 1 , Yanfa Yan 2
1 , Oak Ridge National Laboratory, Oak Ridge, Tennessee, United States, 2 , The University of Toledo, Toledo, Ohio, United States, 3 , Naval Research Laboratory, Washington, District of Columbia, United States
Show AbstractBand gap grading is a common method to increase the solar cell efficiency of Cu(In,Ga)Se2 (CIGS) solar cells, but significant band gap grading is not reported often in the literature for CdTe based solar cells. Sulfur diffusion from the CdS layer creates a photoactive CdTexS1-x alloy layer at the CdS/CdTe interface region, which results in a band gap gradient and enhances the photocurrent in the long wavelength regime due to bowing effects. However, S diffusion into CdTe is limited due to the large lattice mismatch between CdS and CdTe, and there is not enough S diffusion to be effective in increasing the photocurrent. Using a CdSe window layer is more effective at creating a band gap gradient because Se can easily diffuse into the CdTe layer due to the relatively close lattice constants of CdTe and CdSe, and create a CdTexSe1-x alloy with high Se contents. In fact, the published external quantum efficiency data of the world-record CdTe solar cell suggests that this device uses a Se graded CdTexSe1-x alloy layer for band gap grading to increase the short-circuit current and overall device efficiency. Therefore, an engineered band gap gradient within CdTe solar cells is required to create high-efficiency CdTe-based solar cells. A better understanding of the photoactivity, structure, and composition dependence of the CdTexSe1-x alloy layer is important to improve these devices further. To this end, CdTe solar cells have been fabricated with varying thickness CdSe window layers (50 nm, 100 nm, 200 nm, and 400 nm) to study Se diffusion during closed spaced sublimation (CSS) growth of CdTe and subsequent heat treatments. The nano- to meso-scale structural, compositional, and photoactivity relationships of the CdTexSe1-x alloys were studied using transmission electron microscopy (TEM), atom probe tomography (APT), and scanning electron microscopy (SEM) electron beam induced current (EBIC). The results show that the CdTexSe1-x layer photoactivity is highly dependent on the crystalline structure of the alloy (zincblende versus wurtzite), which is also dependent on the Se and Te concentrations. In addition, nanoscale regions with increased Se content were found in the CdTexSe1-x alloy with high concentrations of Se. These results will help direct future fabrication techniques to produce more efficient CdTe-based solar cells using a CdTexSe1-x alloy layer gradient.
This research was supported by the US Department of Energy (DOE) Office of Energy Efficiency and Renewable Energy, Foundational Program to Advance Cell Efficiency (F-PACE), and ORNL’s Center for Nanophase Materials Sciences (CNMS), which is a U.S. DOE Office of Science User Facility.
ES14.12: Degradation, Stability and Modules
Session Chairs
Thursday PM, April 20, 2017
PCC North, 200 Level, Room 229 B
2:30 PM - *ES14.12.01
Scaling Up Solar
Becca Jones-Albertus 1
1 SunShot Initiative, U.S. Department of Energy, Washington, District of Columbia, United States
Show AbstractIn 2011, when solar power composed less than 0.1% of the U.S. electricity supply, the U.S. Department of Energy (DOE) launched the SunShot Initiative with the goal of making solar electricity cost-competitive with conventional electricity sources, to enable widespread solar deployment. Today, solar power supplies 1% of U.S. electricity demand, with the potential to be more than 10% of electricity by 2030. This talk will discuss the potential for solar-generated electricity, and the technology advances needed to enable that potential.
To further drive down the levelized cost of solar energy, photovoltaic performance and reliability must continue to improve while manufacturing and installation costs decrease. A module and system-level approach is critical to identifying and addressing the most significant levers in performance and reliability, as well as cost reduction; thus SunShot is increasing its investments in PV module and system research and development. Research priorities for thin film chalcogenide photovoltaic modules and systems include reliability and degradation science (e.g., reverse bias degradation), increasing energy yield and improving performance prediction.
3:00 PM - *ES14.12.02
Metastability and Reliability of CdTe Solar Cells
Dragica Vasileska 1 , Igor Sankin 2 , Da Guo 1 , Daniel Brinkman 3 , Christian Ringhofer 1 , Andrew Moore 4 , James Sites 4
1 , Arizona State University, Tempe, Arizona, United States, 2 , First Solar, Perrysburg, Ohio, United States, 3 , SJSU, San Jose, California, United States, 4 , Colorado State University, Fort Collins, Colorado, United States
Show AbstractThe record efficiencies of thin-film CdTe technology are still ten absolute percent lower than the Shockley-Queisser limit. As short-circuit current density (JSC) is approaching the theoretical limit, both open-circuit voltage (VOC) and fill factor (FF) are far below the theoretical limits for most devices. Although VOC larger than 0.9V have been reported for single crystal (sx-) CdTe solar cells, low VOC still limits the performance of polycrystalline CdTe devices.
Since VOC is a strong function of the doping concentration in the absorber layer, better understanding of doping mechanism and defects formation is a must. Like most common dopants in px-CdTe, Copper (Cu) forms multiple species of defects including interstitial donors (Cui), substitutional acceptors on Cd site (CuCd) and tightly-bounded complexes such as Cui-CuCd and Cdi-CuCd. Resulting amount of uncompensated acceptor impurities is usually three or four orders of magnitude smaller than the total atomic Cu concentrations, which limits the VOC of Cu-doped CdTe solar cells significantly. Low VOC of px-CdTe solar cells is also believed to be due to large defect density, and short minority carrier lifetime in the absorber layer. In addition, the self-compensated active Cu dopants provide active defect (recombination centers) in CdTe material as well, which results in poor minority carrier lifetime that once again connects Cu with the low VOC presented in px-CdTe PV cells.
Although total Cu concentration profiles can be measured by the secondary ion mass spectrometry (SIMS) technique, the concentration of different species of Cu (mainly Cui and CuCd) generally cannot be identified. Theoretical concentrations of the related defects were estimated from charge neutrality equation with formation energies of defects obtained from First Principles calculations. Given this, gaining a better understanding of Cu migration in CdTe is of crucial importance in order to further enhance the performance of CdTe solar cells.
Moreover, PV modules (multiple solar cells electrically connected) are expected to function properly for more than 25years, in order to provide electricity at proper cost. However, due to the fast diffusion rates of Cu atoms, the gentle balance between mutually compensating Cu impurities could be subject to temporal changes causing metastabilities observed in CdTe solar cells, which also makes the predictive simulation of device performance more important.
Thus, gaining a better understanding of mechanisms that govern formation and interactions between Cu-related defects is of crucial importance for further advancement of the CdTe photovoltaics.
3:30 PM - ES14.12.03
Reduced Degradation of Cu(In,Ga)Se2 Devices through Modification of ZnO:Al Surfaces
Lorelle Mansfield 1 , Samuel Sprawls 3 , Rachael Matthews 4 , Emily Pentzer 4 , Ina Martin 3 2 , Timothy Peskek 2 , Roger French 2
1 , National Renewable Energy Lab, Golden, Colorado, United States, 3 Department of Physics, Case Western Reserve University, Cleveland, Ohio, United States, 4 Department of Chemistry, Case Western Reserve University, Cleveland, Ohio, United States, 2 Department of Materials Science and Engineering, Case Western Reserve University, Cleveland, Ohio, United States
Show AbstractA known failure mode in Cu(In,Ga)Se2 (CIGS) photovoltaics (PV) is degradation of the transparent conductive oxide top contact, aluminum-doped zinc oxide (AZO), through delamination and increasing resistivity. In this project, we deposited ultrathin organofunctional silane modifiers that covalently bond to the oxide surface, with a goal of doubling the lifetime of CIGS PV devices. We have already shown that damp-heat-induced degradation of AZO films on glass is mitigated by deposition of an ultrathin layer of 3-aminoproplytriethoxysilane (APTES) [1]. Only a minimal change in performance was detected in the CIGS devices treated with APTES compared to untreated devices. Additional silane modifiers were tested, including a monofunctional amine-terminated silane (APDMES, 3- aminopropyldimethlyethoxysilane), octadecyltrichlorosilane (ODTS), and octadecyltriethoxysilane (OTES). The chemistry of the leaving group is shown to be critical to successful surface modification; the acidic byproduct of the chlorosilane reaction etches the AZO film.
Select modifiers were chosen for extended study. CIGS devices with modified and unmodified AZO were encapsulated with ethylene-vinyl acetate (EVA) and front glass. The samples were exposed to accelerated-aging conditions including damp heat (85 °C/85% relative humidity), thermal cycling (85 °C to -40 °C, 6 cycles per day), and outdoor testing in Cleveland, OH. In order to provide statistically significant results, we plan to test at least 192 small-area CIGS solar cells over the course of the project. For this presentation we will focus on study design, exposure conditions, and initial results for treated and untreated AZO films and CIGS devices.
[1] I. T. Martin, T. M. Oyster, L. M. Mansfield, R. Matthews, E. B. Pentzer, R. H. French, and T. J. Peshek, “Interfacial Modifiers for Enhanced Stability and Reduced Degradation of Cu(In,Ga)Se2 Devices,” in 43rd IEEE Photovoltaic Specialists Conference, Portland, OR, 2016.
3:45 PM - ES14.12.04
Study of Time-Resolved Photoluminescence on Cu(In,Ga)Se2 Absorbers with Varying Cu-Content
Torsten Hoelscher 1 , Setareh Zahedi-Azad 1 , Matthias Maiberg 1 , Roland Scheer 1
1 , Martin Luther University Halle-Wittenberg, Halle Germany
Show AbstractIn recent publications, it has been shown that time-resolved photoluminescence (TRPL) on Cu(In,Ga)Se2 (CIGSe) may strongly be affected by minority carrier trapping. However, information on the origin of the traps regarding material properties are still missing. Based on the determined energies, the traps observed by TRPL may be due to Cu-vacancies. In this work, we follow this idea and measure time-resolved photoluminescence on CIGSe absorbers with different Cu-content
0.61 ≤ [Cu]/([Ga]+[In]) ≤ 0.9. The absorbers are grown by co-evaporation in a one-stage process. Following XRD and SEM data, the crystal growth is similar to that of a standard three-stage process. The open-circuit voltage of all samples is comparable around 600 mV, for which reason a similar photoluminescence decay is expected. However, the TRPL differs and the decay curves are strongly bi-exponential for low CGI and become more mono-exponential with increasing copper-content. The measurement of TRPL under increased excitations and temperatures leads to a reduction of the second decay, which is an indication for TRPL governed by minority carrier trapping. Additionally, the photoluminescence decay curves exhibit a strong degradation after illumination of the absorbers with high photon fluxes. A change of the position of excitation on the sample reestablishes the undegraded TRPL, which argues for a degradation induced by light. This is also supported by the photoluminescence after exposure of the samples to white light. Already after a few minutes of illumination, the time-constant of the photoluminescence decay is reduced and approaches the time-resolution of the setup.
ES14.13: Industry
Session Chairs
Dmitry Krasikov
Lorelle Mansfield
Thursday PM, April 20, 2017
PCC North, 200 Level, Room 229 B
4:30 PM - *ES14.13.01
Comprehensive Solution for Defect Chemistry in II-VI Photovoltaics
Igor Sankin 1 , Dmitry Krasikov 1 , Andenet Alemu 1 , Christian Ringhofer 2 , Da Guo 3 , Daniel Brinkman 4 , Dragica Vasileska 3 , Markus Gloeckler 1
1 , First Solar, Inc, Perrysburg, Ohio, United States, 2 School of Math and Stat Sciences, Arizona State University, Tempe, Arizona, United States, 3 School of Electrical, Computer and Energy Engineering, Arizona State University, Tempe, Arizona, United States, 4 Department of Mathematics and Statistics, San Jose State University, San Jose, California, United States
Show AbstractAs thin-film CdTe PV industry strives to reduce the cost of renewable energy, the decades-old questions related to underlying mechanisms of device performance and stability became increasingly expensive to ignore. The most notorious among such questions are those related to formation of high and stable p-type absorber doping that does not impede carrier lifetimes. Multiple incarnations of Cu introduction/diffusion/activation process to form p-type in CdTe absorbers have been practiced for decades because of the relative simplicity and broad process window. Yet, the process cannot produce p-type concentrations above 1e15 cm-3 without penalties in device stability, which raises multiple questions. For example, what are the reasons behind poor (~1%) activation of Cu acceptor in CdTe? Why the presence of Cl reduces the maximum achievable free-carrier concentration in Cu-doped devices? Why adding more Cu may temporarily improve open-circuit voltage, but also triggers stability problems? What is the impact of Cu doping on recombination losses? Decades of research have proven that such questions could not be answered without comprehensive analysis of numerous mechanisms involved in convoluted interactions. A typical Cu diffusion/activation in Cl-treated CdTe absorbers involves tens of different species engaged in even larger number of defect reactions. More recently, the composition of II-VI absorbers has become a subject of intensive bandgap engineering targeting stronger light absorption. As a result, the properties of defect species in graded ternary alloys become functions of position-dependent stoichiometry, adding to the complexity of underlying defect chemistry.
To answer the growing challenge, we developed a kinetic defect simulator that predicts formation and evolution of electrically active centers in II-VI absorbers during fabrication and field operation. The numerical core of the developed simulator solves a system of partial differential equations defined for each defect species and coupled with a global Poisson equation. The correctness of the solution depends on initial and boundary conditions as well as the correctness of low-level parameters used to calculate kinetic rate constants. Since direct high-fidelity measurement of low-level parameters used in kinetic equations is generally unavailable, first-principle computations are employed to provide the most intelligent initial guess for the required values. Indirect verification of simulation results is carried out by simulating device performance and stability related to specific process/stress conditions that, in turn, is verified against experimental data. The implemented approach links the atomistic-scale computations of material properties to the actual electrical behaviors of II-VI semiconductors through comprehensive simulation of defect chemistry responsible for performance and stability of thin-film PV devices.
5:00 PM - *ES14.13.02
CIGS Solar Cell Research in Solar Frontier—Progress and Current Status
Takuya Kato 1
1 , Solar Frontier, Kanagawa Japan
Show AbstractAs the largest Cu(In,Ga)(Se,S)2 (CIGS) manufacturing company with more than 1 GW/year capacity, Solar Frontier has continuously improved the performance of CIGS module products as well as the efficiencies of submodule and small-area cell in the laboratory. Recently we have achieved certified efficiencies of 22.3% and 22.0% on CdS-buffered and Cd-free cells, respectively, and that of 18.6% on a Cd-free mini-module. The new Cd-free record efficiency was accomplished by implementing the following techniques: i) alkaline metal treatment of the absorber surface and ii) (Zn,Mg)O layer deposited by atomic layer deposition as an alternative highly resistive n-type layer for the conventional ZnO deposited by metal-organic chemical-vapor deposition. Obviously, the first technique was developed to imitate the treatment known as KF post-deposition treatment (PDT) demonstrated on a co-evaporated CIGS films by Empa. Because our CIGS film is formed by the two-step process including sulfurization, we developed our own K-based treatment process feasible for our absorber. In consequence, the optimized treatment constantly boosts Voc by 20 – 30 meV. Though the mechanism of this treatment is still under investigation, so far our results indicate that an increase of the carrier concentration and a reduction of recombination rates at the interface and in the space charge region are responsible for the Voc improvement. The second technique was mainly intended to enhance the light absorption at the short wavelength region by widening the bandgap of the window layer. However, in practice, both Jsc and Voc were enhanced by introducing (Zn,Mg)O layer. We confirmed that Jsc was actually increased by the EQE enhancement at the short wavelength region and that Voc was improved by a reduction of the interface recombination. The details of the characterization of these two techniques will be discussed.
5:30 PM - ES14.13.03
In-Line Alkali Post-Deposition Treatment for Flexible CIGS Solar Cell Manufacturing
JinWoo Lee 1 , Ryan Kaczynski 1 , Jane Van Alsburg 1 , Nguyet Nguyen 1 , Yejiao Wang 1 , Baosheng Sang 1 , Urs Schoop 1 , Jeffrey Britt 1 , Christopher Thompson 2 , William Shafarman 2
1 , Global Solar Energy, Tucson, Arizona, United States, 2 , Institute of Energy Conversion, Newark, Delaware, United States
Show AbstractCIGS solar cells have the attractive features of high champion cell efficiency and low capital expenditure per watt, but there are still great challenges and opportunities to reduce performance gap between small-area champion lab cells and large-area mass production modules. Major factors in this disparity are insufficient capability to measure, optimize, and control key properties of CIGS absorber layer such as recombination loss mechanism and secondly, lack of quickly adapting recent laboratory advances into the manufacturing line.
Recently, world record CIGS solar cell efficiency breakthroughs have been demonstrated with alkali post-deposition treatments, which involve NaF, KF, and RbF on top of the CIGS films. The alkali post-deposition treatments are hypothesized to assist in p-type doping, and to reduce interface recombination loss. Global Solar Energy, with a proven technology in high-throughput, large-area evaporation source, is in a unique position to bring this laboratory advance to the manufacturing line. We have achieved 18.3% cell efficiency by varying the KF evaporation rate for CIGS solar cells grown on flexible SS substrates in a CIGS production line. Na was incorporated in the CIGS material by an in-situ NaF precursor layer and KF was in-line evaporated at the end of CIGS deposition under Se ambient condition. Nearly 1% efficiency has been improved by KF post-deposition treatment, which resulted from Voc being boosted by 16 mV and FF by 1.2 % compared to the baseline control from the same CIGS run.
One of the unique aspects of KF post-deposition treatment to the CIGS solar cells is the redistribution of alkali elements in their compositional depth profiles. While the liquid phase of Cu-Se at the end of the second stage may promote Na accumulation in the middle of the CIGS layer (so called "hump"), K seems to replace the Na at both the CdS/CIGS interface and at the hump position. On the other hand, K is observed to accumulate at the CIGS/Mo interface without disrupting Na profiles. We observed Cu/(In+Ga) and Ga/(In+Ga) ratio to increase at CIGS surface, which was measured by glow discharge optical emission spectroscopy depth profiling. Material and device analyses will be presented for characterizing recombination at different depth profiles in order to identify efficiency loss mechanisms and reveal fundamental device properties.
5:45 PM - ES14.13.04
Alkali Element Interdependence in CIGS Relating to Depth Distribution of Na/K and Ga/(Ga+In)
Lars Stolt 1 2 , Erik Wallin 1 , Sven Sodergren 1 , Tobias Jarmar 1 , Marika Edoff 2
1 , Solibro Research AB, Uppsala Sweden, 2 Solid-State Electronics, Uppsala University, Uppsala Sweden
Show AbstractIn the recent couple of years post deposition treatment (PDT) of CIGS thin films for photovoltaic devices with potassium has proven to lead to a significant solar cell performance gain. It has also been shown that a similar effect is achieved with rubidium or cesium. In fact, all recent record laboratory cell efficiencies have been achieved using such treatments. In this contribution we report experiments with co-evaporated CIGS films with post deposition treatments using evaporated KF. By studying films grown on normal soda-lime glass and glass with high K (low Na) concentration, with and without a Na precursor layer deposited on the Mo back contact prior to CIGS deposition, we provide evidence that the incorporation and distribution of K after the PDT treatment is strongly dependent on the Na content and distribution prior to PDT. The K from the KF PDT diffuses into the CIGS material at a much higher rate at higher Na concentrations and is also predominantly found in regions with higher Na presence. Conversely, the surface concentration of K is higher for films with low Na content, where most of the potassium from the KF PDT stays on the surface. Comparing CIGS films with the same two types of substrates we also observe that the Ga/In interdiffusion is clearly affected. It is known from previous work that the presence of Na reduces Ga/In interdiffusion so that steeper bandgap (Ga/(Ga+In)) gradients can be achieved. In this work we find that a similar, but stronger, suppression of Ga/In interdiffusion occurs in K-enriched CIGS material. Analyzing complete solar cell devices, it is demonstrated that it is possible to obtain a positive effect on solar cell performance using K-rich substrates without K PDT, clearly seen in enhanced Voc, without the use of a K PDT process. In such samples, an accumulation of K in the surface region similar to CIGS with K PDT is observed. The main analytic methods in this work are glow discharge optical emission spectroscopy (GDOES) and characterization of solar cells by light I-V and EQE.
ES14.14: Poster Session III: Novel Chalcogenide Absorber Materials and Kesterite Absorber Growth and Devices
Session Chairs
Sascha Sadewasser
Edgardo Saucedo
Friday AM, April 21, 2017
Sheraton, Third Level, Phoenix Ballroom
9:00 PM - ES14.14.01
Effect of Mg Incorporation on Solution-Processed Kesterite Solar Cells
Raquel Caballero 1 , Stefan Haass 2 , Christian Andres 2 , Yaroslav Romanyuk 2 , Ayodhya Tiwari 2
1 , Univ Autonoma-Madrid, Madrid Spain, 2 , EMPA, Duebendorf Switzerland
Show AbstractThe substitution of Zn with Mg in Cu2ZnSn(S,Se)4 (CZTSSe) has recently been reported in view of potential photovoltaic applications but the observations are quite contradictory [1-4]. Cu2MgSnS4 thin films grown by ultrasonic co-spray pyrolysis showed p-type conductivity and a band-gap energy of 1.76 eV [2]. In contrary, n-type conductivity was observed for (Cu2-xMgx)ZnSnSe4 bulk materials with x = 0.1-0.4, which was attributed to the formation of the donor-type MgCu antisite defects [3]. The formation of stable Cu2MgSn(S,Se)4 was calculated based on the density functional theory [1], whereas a complete phase separation was predicted in [4].
The objective of this work is to investigate the influence of Mg on kesterite solar cells. CZTSSe thin films were grown using the solution approach with DMSO solvent followed by annealing in selenium atmosphere [5]. Two series of experiments were carried out to introduce Mg in the absorber bulk. In the first one, Mg bulk doping of CZTSe thin films was achieved by adding Mg(CH3COO)2.2H2O to the precursor solution, which corresponds to nominal 5 at% with respect to all metals Mg doping. The second experiment involved alloying of Zn with Mg to evaluate possible effects on the absorber band-gap for Cu2Zn1-yMgySnSe4 thin films with y = 0.5 and 1. All absorbers were systematically characterized with SIMS, XRD, SEM/EDX, PL, TRPL and the processed solar cells were analyzed with temperature-dependent JV, EQE and admittance spectroscopy.
Mg-doped kesterite solar cells with efficiency up to 6.6 % were obtained by adding 5% magnesium. These devices were compared with intentionally Na-doped kesterite cells using identical selenization, buffer and window layers. Smaller lattice parameters, a higher band-gap energy of 1.06 eV, shorter minority carrier lifetimes and a carrier concentration in the range of 1016 cm-3 were estimated for Mg-doped CZTSe devices. A higher Mg concentration was detected on the absorber surface by SIMS. In the second experiment, the decomposition of the Cu2Zn1-yMgySnSe4 compounds into secondary Cu2SnSe3, MgSe(2) and SnSe2 phases was detected by XRD and EDX as predicted theoretically in [4], which did not result in any working solar cells.
We believe Mg doping introduced by post-deposition treatments should be investigated as a promising outlook for creating buried p-n homojunction which might lead to enhancements in the device efficiency.
[1] Zhong et al., Thin Solid Films 603 (2016) 224.
[2] Guo et al., Mat. Letters 172 (2016) 68.
[3] Kuo et al., J. Solid State Chem. 215 (2014) 122.
[4] Wang et al., Chem. Mat. 26 (2014) 3411.
[5] Haass et al., Adv. Energy Mat. 5 (2015) 1500712.
9:00 PM - ES14.14.02
Kesterite Solar Cells Grown by Sulfurization of Co-Evaporated Cu2ZnSnSe4 Thin Films
Raquel Caballero 1 , Noelia Martín 1 , Guillermo Garcia 1 , Yudania Sanchez 2 , Marcel Placidi 2 , Ivan Fernandez 4 , Jose Manuel Merino 1 , Fernando Briones 3 , Maximo Leon 1
1 , Univ Autonoma-Madrid, Madrid Spain, 2 , IREC, Barcelona Spain, 4 , Nano4Energy S.L.N.E., Madrid, Madrid, Spain, 3 , Instituto de Microelectrónica Madrid-CSIC, Tres Cantos, Madrid Spain
Show AbstractA maximum efficiency of 12.6 % has been achieved for Cu2ZnSn(S,Se)4 (CZTSSe) solar cells grown by hydrazine-based solution deposition process [1]. Cu2ZnSnSe4 (CZTSe) solar cells with 11.6 % performance were produced by co-evaporation of the elements followed by annealing with excess selenium under N2 atmosphere. Despite the lower performance obtained using vacuum-growth processes, generally these lead to a higher quality of the semiconductor material and reproducibility, which is fundamental for an industrial implementation.
In this work, CZTSe thin films were deposited by co-evaporation of Cu, Sn, ZnSe and Se at different substrate temperatures and evaporation rates onto soda-lime glass (SLG) and Mo/SLG substrates. Diverse multi-stage co-evaporation processes were studied. After that, the kesterite layer was annealed with excess sulfur under Ar atmosphere. All absorber films, before and after sulfurization, were systematically characterized with SEM, EDX, GIXRD, transmittance and reflectance and the processed solar cells were analyzed with JV and IQE measurements.
The goal of this work was to investigate the influence of kesterite growth parameters on the absorber layer properties and final devices. The termination of the co-evaporation process with Sn or Zn, always in the presence of Se, was critical to achieve the absorber with the proper composition. The effect of Cu concentration and substrate temperature on structural, morphological and optical properties of the kesterite layer was investigated. A nominal growth temperature of 325° C was identified as the optimum to produce high crystallinity thin films. Low out-diffusion of Na from the SLG substrate was detected at this condition. It was also observed that Na diffusion from SLG depended not only on the substrate temperature, but also on the Cu content of the absorber layer.
CZTSSe solar cells with efficiency up to 4.5 % were obtained with an atomic ratio Se/(Se+S) = 0.61 and a band-gap energy of 1.13 eV. These devices will be compared with intentionally Na-doped kesterite cells using identical processing steps (co-evaporation, sulfurization, buffer and window layers) to promote grain growth of absorber thin films and improve photovoltaic parameters [2-3]. The alkali Na is added by evaporation of a NaF layer before or after CZTSe deposition inside the same vacuum chamber.
[1] W. Yang, M.T. Winkler, O. Gunawan, T. Gokmen, T,K, Todorov, Y, Zhu, D.B. Mitzi, Advanced Energy Materials 4 (2014) 1301465.
[2] Y.S. Lee, T. Gershon, O. Gunawan, T.K. Todorov, T. Gokmen, Y. Virgus, S. Guha, Advanced Energy Materials 5 (2015) 1401372.
[3] I. Repins, C. Beall, N. Vora, C. DeHart, D. Kuciaukas, P. Dippo, B. To, J. Mann, W.C. Hsu, A. Goodrich, R. Noufi, Solar Energy Materials and Solar Cells 101 (2012) 154-159.
9:00 PM - ES14.14.03
Exploring the Role of Ge in the Synthesis of Cu2ZnSnSe4:Ge Kesterite Absorbers
Sergio Giraldo 1 , Markus Neuschitzer 1 , Florian Oliva 1 , Paul Pistor 1 , Xavier Alcobe 2 , Victor Izquierdo 1 , Alejandro Perez-Rodriguez 1 3 , Edgardo Saucedo 1
1 , Catalonia Institute for Energy Research (IREC), Sant Adrià de Besòs Spain, 2 , Centres Científics i Tecnològics (CCiTUB), Barcelona Spain, 3 , IN2UB, Barcelona Spain
Show AbstractRecently, several works have shown the positive effect of Ge on Cu2ZnSn(SxSe1-x)4 kesterites, in both doping and alloying approaches. Among the improved properties, it has been demonstrated a longer minority carriers life-time, improved morphology of the layers with larger grains, annihilation of detrimental defects, controlled p-type doping level, etc., ultimately impacting beneficially on the conversion efficiency of the resulting devices.
In the same line of our previous investigations in this matter, in this work we present the study of the mechanisms behind the effect of Ge during the synthesis of Cu2ZnSnSe4 (CZTSe) absorbers with the aim of understanding the positive role of Ge. For this purpose, CZTSe and CZTSe:Ge layers were synthesized using a sequential process, based on the two-step reactive annealing of Cu/Sn/Cu/Zn metallic stacks with and without Ge. The two-step annealing process consists of a first step at 400 °C (1.5 mbar Ar pressure) where the selenide precursors are formed and react to form the CZTSe kesterite; and a second step at 550 °C (1 bar Ar pressure) to crystallize the CZTSe layers, leading to well-crystallized absorbers. In order to investigate the mechanisms behind the formation of CZTSe in the presence or absence of Ge, an experiment was designed consisting of stopping the synthesis reaction, i.e. the reactive annealing, at different stages. The resulting samples were characterized combining SEM/EDX, Raman spectroscopy, XRF, XRD, and also the solar cells fabricated with the final layers.
Intriguingly, the presence of Ge changes drastically the formation mechanism of kesterites from a tri-molecular mechanism towards a hypothetical bi-molecular one, which could lead to a better control of the elemental losses during the annealing process. Furthermore, analyzing the samples obtained stopping the reaction at different temperatures and annealing times, we can confirm a drastic change in the elemental distribution when Ge is present, mainly at the beginning of the selenization, somehow retaining Sn at bottom of the layer mixed with Cu. This was corroborated by XRF measurements where the samples stopped at early selenization stages without Ge showed a notably Sn poor composition, while the Ge-doped samples kept a constant cationic composition during the entire process. Hence, we will propose and describe in more detail the formation mechanism of CZTSe in the presence of Ge, also showing its impact on the morphology of the layers as well as on the secondary phases distribution, and ultimately the positive effect on the solar cell devices obtaining routinely efficiencies around 10%.
9:00 PM - ES14.14.04
Ge Incorporated CZTSe Thin-Film Solar Cell with a Conversion Efficiency of 12.3%
Shinho Kim 1 , Kang Min Kim 2 , Hitoshi Tampo 1 , Hajime Shibata 1 , Shigeru Niki 1
1 , National Institute of Advanced Industrial Science and Technology, Tsukuba Japan, 2 , Korea Institute of Industrial Technology, Gangreung Korea (the Republic of)
Show AbstractKesterite devices, Cu2ZnSnSe4 (CZTSe) and Cu2ZnSn(SxSe1−x)4 (CZTSSe), have been attracting attention because earth abundant and non-toxic absorber is a potential replacement for the Cu(In1-xGax)Se2 (CIGS) in thin film solar cells. CZTSSe solar cell achieved the highest conversion efficiency of 12.6 % with an optical band gap (Eg) of 1.13 eV, where the Eg of CZTSSe (1.0 – 1.5 eV) was controlled by the S/(S+Se) ratio. However, CZTSSe devices have exhibited relatively large open circuit voltage (VOC) deficit (Eg/q-VOC, q: electron charge > 0.6) that increases with increasing the S/(S+Se) ratio, whereas the VOC deficit of CIGS is ~ 0.4.
In this study, we demonstrate Cu2Zn(Sn1-xGex)Se4 (CZTGSe) thin films solar cells; their Eg (1.0 – 1.5 eV) was controlled through the cationic atoms ratio, Ge/(Sn+Ge) in the range of 0 to 1. We achieved high conversion efficiency of 12.3% with a greatly improved VOC deficit (0.583 V) and fill factor (FF = 0.73), which are thought to be the primary reasons for the high conversion efficiency. Therefore, the improved reasons for these two parameters are discussed.
The VOC deficit of 12.3% efficient device exhibited 0.583 with a Eg of 1.11 eV, and was found to be improved through reduced band tailing via control of the Ge/(Sn + Se) ratio. The band tailing caused by compositional disorder is improved with a reduced amount of Ge. Accordingly, CZTGSe with Eg of 1.11 eV exhibited relatively low VOC deficit less than 0.6, which is similar value with high efficient CZTSe devices. Also, CZTGSe showed improved VOC deficit compared to those for CZTSSe with a similar range of Eg. This means that compositional disorder associated with S-incorporation are larger than those associated with Ge-incorporation, which results in greater band tailing of S-incorporated kesterite devices.
In addition, the high FF was mainly induced by a reduced carrier recombination at the absorber/buffer and/or space charge region, whereas parasitic resistive effects on FF were very small. We examined loss of FF in terms of ideality factor (A) and reverse saturation current (J0) and parasitic resistive effects by series and shunt pass. Using the ideal fill factor (FF0), the effect of each parameter were determined separately, because FF0 is affected by only A and J0, and parasitic resistive effects can be ignored. In the case of 12.3% efficient CZTGSe, FF loss caused by resistive effect was similar with 19.9% efficient CIGS device. Instead, FF was significantly influenced by A and J0, which indicates that FF loss was greatly limited by carrier recombination at the absorber/buffer and/or space charge region.
9:00 PM - ES14.14.05
Reaction Kinetics of Cu2ZnSnS4 and Cu2SnS3 Formation from Cu2S, ZnS, and SnS2 Precursors Studied Using Differential Scanning Calorimetry
Elizabeth Pogue 1 , Melissa Goetter 1 , Angus Rockett 2 1
1 Materials Science and Engineering, University of Illinois, Urbana, Illinois, United States, 2 Department of Metallurgical & Materials Engineering, Colorado School of Mines, Golden, Colorado, United States
Show AbstractSolar cells incorporating Cu2ZnSn(S, Se)4 (CZTSSe) reached 12.6 % efficiency in 2013. Further improvements have proven difficult, suggesting that new approaches are needed. A better understanding of the kinetics of phase formation and the phases that form in the Cu-Zn-Sn-S system would help. Here, differential scanning calorimetry (DSC) experiments were performed on mixtures of Cu2-xS, SnS2, and ZnS powders to determine the temperatures at which reactions take place, the energy change involved, and the variability among samples.
Cu2-xS (Alfa Aesar, Cu2S), SnS2, and ZnS were mixed, packed in Al sample pans, and hermetically sealed to prevent Sn and S loss. No reactions with the pan were observed using x-ray diffraction (xrd) and all samples reacted fully during the first temperature ramp. Within each batch from a given precursor mixture there were variations in the reaction onset temperature and peak width measured by DSC. This was attributed to variations in the local precursor concentration. The resulting Cu2SnS3 samples consisted of monoclinic Cu2SnS3 with a small amount of Cu4Sn7S16 inclusions, suggesting that the precursor was slightly SnS2-rich. The diffraction patterns for the CZTS samples were consistent with phase-pure CZTS, although ZnS or tetragonal Cu2SnS3 coexistence could not be excluded.
The reactions to form Cu2SnS3 and CZTS are both complicated and multi-step where peaks often overlap. Others have reported that the final reaction involved in forming CZTS is Cu2SnS3+ZnS -> CZTS.1 The Cu2SnS3 sample began reacting endothermically at 350-370 °C. This is a higher temperature than for CZTS (endothermic, 320-350 °C) suggesting that ZnS mediates a reaction to form Cu2SnS3. The reactions in the ZnS-containing sample end at a lower temperature (highest observed temperature ~380 °C) than those without ZnS (lowest observed temperature ~420 °C). ZnS mediation would not be surprising due to the structural similarities between ZnS and tetragonal Cu2SnS3.
Reaction energies could be extracted in a reproduceable manner for Cu2SnS3 in many cases. The first transition was endothermic. The second transition (370-390 °C), was exothermic with an energy of -1.5 ±0.1 J/g. A series of at least two overlapping endothermic reactions then took place, with a combined energy of 1.5 ±0.2 J/g. A final endothermic reaction (5.5 ±0.5 J/g) finally occurred between 420 and 440 °C.
1. R. Schurr, A. Hölzing, S. Jost, R. Hock, T. Voβ, J. Schulze, A. Kirbs, A. Ennaoui, M. Lux-Steiner, A. Weber, I. Kötschau, and H.-W. Schock: The crystallisation of Cu2ZnSnS4 thin film solar cell absorbers from co-electroplated Cu–Zn–Sn precursors. Thin Solid Films 517(7), 2465 (2009).
9:00 PM - ES14.14.06
Tuning of Stoichiometry and Band Gap of Solution-Processed Cu2ZnGeSxSe4-x Absorbers for Thin-Film Solar Cells
Thomas Schnabel 1 , Mahmoud Seboui 1 , Erik Ahlswede 1
1 , Zentrum fur Sonnenergie- und Wasserstoff-Forschung, Stuttgart Germany
Show AbstractThin-film solar cells with a kesterite-type Cu2ZnGeSxSe4-x (CZGS) absorber are very similar to the extensively studied Cu2ZnSnSxSe4-x (CZTS) and their high band gap of 1.5 – 2.0 eV makes them an interesting material for the application as top cell in tandem solar cells. However, so far only low efficiencies of around 1 % have been reported in literature [1].
In this work, CZGS absorbers were prepared in a two-step process based on a metal salt solution that is deposited onto a molybdenum-coated substrate and subsequently annealed in selenium-containing nitrogen atmosphere. Different metal salt solutions with dimethyl sulfoxide (DMSO), N,N-dimethylformamide (DMF) and water are compared. With respect to the solubility also different metal salts (namely chlorides, iodides, oxides and acetates) were used as starting materials. The structural and morphological properties of the absorbers before and after selenization are compared with an additional focus on the electrical properties of the resulting thin-film solar cells.
The best performing solar cells could be fabricated by the solution based on DMF with efficiencies exceeding 6 % and therefore demonstrating the highest reported efficiency for this material. As a starting point, a Cu-poor and Zn-rich composition was used which has been proven successful for CZTS absorbers, if Ge is substituted for Sn [2]. However, since different secondary phases and defects can be expected, the stoichiometry for CZGS has to be optimized independently. Therefore, different stoichiometries were investigated and their structural, morphological and optoelectrical properties are discussed in detail.
Based on the optimized stoichiometry, additional variations have been performed, namely different annealing parameters, film thickness, and S/(S+Se)-ratio, thereby tuning the band gap.
Since the commonly used CdS buffer layer is expected to have a non-ideal band alignment and also raises environmental concerns, In2S3 fabricated by atomic layer deposition (ALD) is explored as alternative buffer. Here different film thicknesses and subsequent heat treatments were compared, demonstrating promising efficiencies of >3%.
[1] K. Timmo, et al., Proceedings of 28th European Photovoltaic Solar Energy Conference and Exhibition, 2013.
[2] T. Schnabel et al., Solar Energy Materials & Solar Cells 117, 324-328, 2013.
Acknowledgments: This project has received funding from the European Union’s Horizon 2020 research and innovation program under grant agreement No 640868.
9:00 PM - ES14.14.07
The Influence of Synthesis Conditions upon the Single-Phase Region in Cu2ZnSnS4 Thin Films
Alexandra Davydova 1 , Katharina Rudisch 1 , Jonathan Scragg 1
1 Solid State Electronics, Uppsala University, Uppsala Sweden
Show AbstractDuring the last decades, great interest has been directed towards new materials for photovoltaic (PV) applications. One of the promising examples is Cu2ZnSnS4 (CZTS) – a non-toxic, low-cost, earth-abundant material, with a band gap of about 1.5 eV that makes it suitable for the absorber layer in thin film solar cell structures. CZTS solar cells have reached an efficiency of 12.6% (with inclusion of Se). However, to reach the goal of >20% efficiency, a better understanding of the relationship between synthesis conditions and optoelectronic properties is needed. In particular, manufacture of the quaternary CZTS compound is challenging since the CZTS single phase region is apparently very narrow. Slight instability during the synthesis or the heat treatment could thus result in secondary phases and a change in the concentrations of critical defects, which both strongly influence the properties of the absorber layer and the solar cell performance respectively.
In order to get better understanding of the CZTS single phase region, and its dependence on synthesis conditions, we present a combinatorial study of compositionally graded thin films prepared by reactive magnetron sputtering. Grown films were annealed in two different ways: (1) using our “baseline” annealing procedure with high S partial pressure, and (2) using a procedure which additionally contains a high activity of Sn. Each CZTS sample contains large regions of E- (Sn-rich), A- (Cu-poor, Zn-rich), B- (Zn-rich) and F-Type (Sn-poor) material. The structural and optical properties of the combinatorial samples were investigated with multi-wavelength (including resonant) Raman mapping and photoluminescence, combined with compositional mapping by EDX spectroscopy. The multi-wavelength Raman maps allowed us to detect neighboring Cu3SnS4, ZnS, SnS phases, and so indicate the boundaries of the CZTS single phase region “from the outside”; whilst the resonant Raman helps to locate the phase boundaries “from the inside” by revealing structural changes in the CZTS phase (such as a variation in the concentration of Cu/Zn antisite defects and other defect complexes). The results show very good correlation with the expected distribution of the proposed defect complexes (especially the E, A, B, and F stoichiometry types). The maximum solubilities of these defect complexes define the single phase boundary, and could be estimated based on our data. Rather large changes were seen under the different anneal conditions. In particular, annealing with higher Sn activity enlarged the single phase region in comparison with the baseline method. Reasons for this arising from defect equilibrium theory are discussed. Our results suggest possible strategies for improving growth/annealing processes to target a particular CZTS stoichiometry, so that desired properties of the absorber layer could be achieved for successful application in CZTS-based solar cells.
9:00 PM - ES14.14.08
Insights into the Chemistry of Amine-Thiol Solution Processing and Implications for Solution Processing of Chalcogenide Optoelectronic Devices
Caleb Miskin 1 , Priya Murria 2 , Robert Boyne 1 , Laurance Cain 2 , Evan Wegener 1 , Ruihong Zhang 1 , Xin Zhao 1 , Jeffery Miller 1 , Hilkka Kenttamaa 2 , Rakesh Agrawal 1
1 School of Chemical Engineering, Purdue University, West Lafayette, Indiana, United States, 2 Department of Chemistry, Purdue University, West Lafayette, Indiana, United States
Show AbstractSolution processing of electronic materials is an attractive alternative to vacuum based methods due to its promise of high-throughput, roll-to-roll fabrication at low cost. Recently, significant work has been done to advance the so-called molecular precursor route to inorganic thin films. This route involves the dissolution of appropriate precursors to form a homogenous solution, followed by an appropriate coating method, and subsequent heat treatments to promote the growth of a polycrystalline film of the desired final phase. This method has been applied with great success for thin film PV such as Cu(In,Ga)Se2, Cu2ZnSn(S,Se)4, CdTe, and others.
An especially promising solvent system, a mixture of an amine and a thiol, has been identified. Recent work has demonstrated the incredible flexibility of the amine-thiol solvent system in dissolving a host of salts and precursors that are insoluble in either solvent by itself.1,2 The system has found significant use in processing photovoltaic absorber layers, luminescent quantum dot films, and a variety of other thin-film materials, making it a very general solvent system for processing of inorganic thin films.3–7
Currently, little is known about the solution chemistry enabling this powerful solvent. In this study we present electrospray ionization mass spectrometry, X-ray absorption spectroscopy, and Raman spectroscopy results obtained for copper chloride precursors dissolved in thiol-amines to reveal the species formed in solution. In short, they were found to be copper thiolates, copper chlorides, copper thiolate-chlorides, and alkylammonium chlorides. We also show that the final oxidation state of the complexed Cu is +1, even when the initial state of the Cu is +2. Furthermore, we present SEM-EDS, XRD, and Raman spectroscopy studies of annealed films prepared from these solutions to understand the conditions under which precursor impurities (such as chlorides) are removed from the films making them suitable for use in high efficiency devices. We also present results obtained by applying this film preparation method to CdTe,7 CIGS,6 and other metal chalcogenide photovoltaic devices.4,5
References
1. D. H. Webber and R. L. Brutchey, J. Am. Chem. Soc., 2013, 135, 15722–15725.
2. B. C. Walker and R. Agrawal, Chem. Commun., 2014, 50, 8331–8334.
3. Q. Tian, G. Wang, W. Zhao, Y. Chen, Y. Yang, L. Huang and D. Pan, Chem. Mater., 2014, 26, 3098–3103.
4. R. Zhang, S. Cho, D. G. Lim, X. Hu, E. A. Stach, C. A. Handwerker and R. Agrawal, Chem. Commun., 2016, 52, 5007–5010.
5. R. Zhang, S. M. Szczepaniak, N. J. Carter, C. A. Handwerker and R. Agrawal, Chem. Mater., 2015, 27, 2114–2120.
6. X. Zhao, M. Lu, M. J. Koeper and R. Agrawal, J. Mater. Chem. A, 2016, 4, 7390–7397.
7. C. K. Miskin, A. Dubois-Camacho, M. O. Reese and R. Agrawal, J. Mater. Chem. C, 2016, 4, 9167-9171.
9:00 PM - ES14.14.09
Atmospheric Variations on Close Spaced Vapor Transport Deposited SnS Thin Films—Air vs Argon
Jacob Andrade-Arvizu 1 , Maykel Courel 2 , Mario Fidel Garcia-Sanchez 3 , Roberto Gonzalez-Castillo 4 , Osvaldo Vigil-Galan 1
1 Física del Estado Sólido, Escuela Superior de Física y Matemáticas - Instituto Politécnico Nacional (ESFM-IPN), Mexico City Mexico, 2 , Instituto de Energías Renovables - Universidad Nacional Autónoma de México (IER-UNAM), Temixco Mexico, 3 , Unidad Profesional Interdisciplinaria en Ingeniería y Tecnologías Avanzadas - Instituto Politécnico Nacional (UPIITA-IPN), Mexico City Mexico, 4 , Centro de Investigación en Ciencia Aplicada y Tecnología Avanzada del Instituto Politécnico Nacional (CICATA-IPN), Tamaulipas Mexico
Show AbstractIn this work, single-phase, p-type Tin (II) sulfide (SnS) thin films were deposited by employing the Close Spaced Vapor Transport (CSVT) technique from the variation on the evaporation chamber conditions. The impact of air and argon (Ar) atmospheres on thin film properties is presented and compared. The analysis of film properties was performed by X-ray diffraction, Raman spectroscopy, scanning electron and atomic force microscopy, optical measurements and opto-electronic characterization techniques. The different atmospheric conditions impact on the morphological and directional preferred orientation (DPO) of crystal is also presented.
9:00 PM - ES14.14.10
Engineering the Crystal Phase and Morphology of Chalcogenide Nanomaterials for the Next Generation Thin-Film Solar Cells
Xiaoyan Zhang 1 , You Xu 1 , Ningzhong Bao 1
1 , Nanjing Tech University, Nanjing China
Show AbstractAs one of the most investigated chalcogenide nanomaterials, Cu2ZnSnS4 (CZTS) has obtained great attention due to its high absorption coefficient, optimal direct band gap, earth abundant composition, and tunable physical properties. The power conversion efficiency as high as 12.6 % has been obtained using colloidal chemical method. Recently, CZTS nanomaterials have been proved to be a suitable alternative counter electrode (CE) for dye sensitized solar cells (DSSCs), which shows comparable efficiency to Pt. While, most work on CZTS nanomaterial as CE have concentrated on the kesterite structure. Recently, Wu et al. have demonstrated that wurtzite-structured CZTS films show a higher carrier concentration and lower resistivity, which shows to be a more effective CE material to replace Pt. Meanwhile, with recent great progress in graphene, other types of two-dimensional nanomaterials are being more attractive for both fundamental research and potential applications. Thus, fine tailoring the preferred orientation of wurtzite CZTS nanocrystals and its application for DSSC is of great importance.
Based on our previous work on the crystal phase-, size-, morphology-controlled synthesis of ternary and quaternary chalcogenide nanomaterials (CuInxGa1-xS2, Cu2(Zn, Fe, Co)SnS4, etc.), further design and preparation of anisotropic nanosheets are of great research interests. Herein, we report a simple non-injection colloidal chemistry method for the synthesis of wurtzite-derived Cu2ZnSnS4 (CZTS) nanosheets (NSs) and enhanced photoelectrochemical properties. CZTS NSs with thickness down to 3.0 ± 0.5 nm, corresponding to 4-6 layers of unit cells, and lateral dimensions of 200-300 nm were successfully obtained. The morphology of nanocrystals has further been engineered with both nanosheets and nanospheres being obtained. As a proof of concept, the CZTS nanorcystals with various morphologies have been applied as counter electrode for DSSCs. Densely packed CZTS nanosheet thin films demonstrate promising photoelectrocatalytic activity toward triiodide reduction at a rate comparable to platinum and better than spherical CZTS nanoparticles for DSSCs. This is likely due to the reduced charge transfer resistivity (Rct) and comparable series resistivity for CZTS NSs and Pt. The efficient and scalable method for preparation of 2D nanomaterials also paves a new way to other chalcogenide nanomaterials, which shows potential application for solar cell conversion.
9:00 PM - ES14.14.11
Efficient Planar Antimony Sulfide Thin-Film Photovoltaics
Haisheng Song 1 , Jiang Tang 1
1 , Huazhong University of Science and Technology, Wuhan China
Show AbstractCompact thin film of antimony sulfide (Sb2S3) as a promising absorber layer was obtained by rapid thermal evaporation (RTE) rather than conventional chemical bath deposition or atomic layer deposition based methods. The systematical characterizations of Sb2S3 film demonstrated the pure phase, void free and high crystallization. The large grain and preferential growth of Sb2S3 thin film were implemented by crystallization and cooling techniques, respectively. And the post-annealing could also greatly improve the rear contact. The corresponding devices were gradually optimized with a power conversion efficiency of ~4 %, almost three times of planar devices fabricated by vacuum method. Both the non-oxide buffer layer (CdS layer) and free of hole transport layer enabled the high stability of the non-encapsulated planar devices. The high throughput and reliable RTE fabrication technique, environment-friendly and earth-abundant Sb2S3 materials, stable and superior device performances were expected to drive the research progress of Sb2S3 thin film photovoltaics.
9:00 PM - ES14.14.12
Chemically Deposited Solar Cells of Orthorhombic and Cubic Tin Sulfide—An Overview
Ana Rosa Garcia Angelmo 1 , Victoria Elena Gonzalez-Flores 1 , P.Karunakaran Nair 1 , M.T. Santhamma Nair 1
1 , Universidad Nacional Autonoma de Mexico, Temixco Mexico
Show AbstractTin sulfide thin film solar cells mainly of orthorhombic crystalline structure (SnS-ORT) with a band gap of 1.1-1.3 eV have attracted much attention for research and development in recent years. To this has also been added thin film solar cells of tin sulfide belonging to a large cubic structure in which 32 formula units of SnS are contained in an SnS-CUB unit cell with lattice parameter a = 11.5944 Å. While thin films of SnS-ORT are made by many deposition techniques, thin films of SnS-CUB has been made only via chemical methods. The band gap of the latter is very distinct, of nearly 1.7 eV. In this work we present an overview of thin film solar cell structures of these two materials by chemical deposition in the “substrate structure” on stainless steel substrates and in the “superstrate structure” on transparent conductive oxide (TCO) films of SnO2:F (TEC 15). The superstrate structure of SnS-ORT absorber with chemically deposited CdS, TCO/CdS/SnS-ORT/C-Ag, showed open circuit voltage (Voc) of 0.27 V and short circuit current density (Jsc) of 1.13 mA/cm2. In the case of SnS-CUB films in this configuration, TCO/CdS/SnS-CUB/Au, showed a Voc of 0.53 V and a Jsc of 4.18 mA/cm2 resulting in photovoltaic conversion efficiency (h) of 0.8%. The substrate configuration using thin films of SnS-CUB developed on stainless steel (SS), SS/SnS-CUB/CdS/ZnO/ZnO:Al, showed an efficiency h of 1.28 %, Voc of 0.470 V, Jsc of 6.23 mA/cm2 with fill factor (FF) of 0.44. The spectral response studies of this solar cell indicated an external quantum efficiency of 37%. The other parameters of the solar cells were also evaluated. These cells remained stable even under concentrated solar radiation of 16 suns, even though the increase in Jsc was not linear as expected. Photovoltaic modules were produced using the solar cells of SnS-CUB thin films developed in this work. A module consisting of five of these solar cells connected in series showed Voc of 2.6 V, and Isc of 4.6 mA. Two such modules connected in parallel showed Voc of 2.54 V and Isc of 9.8 mA. The performance of the modules was stable even after one year from their fabrication.
9:00 PM - ES14.14.13
Cubic Phase Tin Sulfide Thin Film Obtained by Atomic Layer Deposition and Its Application to Solar Cells
Xizhu Zhao 1 , Sang Bok Kim 1 , Xiabing Lou 1 , Laura Schelhas 2 , Chuanxi Yang 1 , Roy Gordon 1
1 , Harvard University, Cambridge, Massachusetts, United States, 2 , SLAC, Menlo Park, California, United States
Show AbstractTin monosulfide (SnS) has been gaining heat as a promising solar cell absorber in recent years. Cell fabricated with the thermodynamically stable orthorhombic SnS has reached efficiencies over 4% [1]. Yet the orthorhombic crystal structure leads to an indirect bandgap of 1.1 eV and anisotropic electrical properties, which are detrimental to the open circuit voltage as well as efficient carrier transport and collection. Recently, a new cubic phase of SnS has been discovered, which composes of 64 atoms in a unit cell and has a lattice constant of 11.7 Å [2]. Solar cells made with this cubic SnS thin film by chemical bath deposition have reached an efficiency of 1.28% [3].
In this work we report cubic SnS thin film made by atomic layer deposition (ALD) using a new tin precursor, which enables deposition at lower temperatures compared to previous ALD-SnS. Films with thickness between 70 nm and 750 nm were made at temperatures between 80 °C and 180 °C on different substrates including thermal oxide, molybdenum, and single-crystal sodium chloride. The crystal structures were characterized using X-ray diffraction. The electrical and optical properties were studied with Hall measurements, UV-vis spectrophotometry, and photoluminescence (PL).
We show that the crystal structures of the ALD-SnS thin films vary with deposition temperature, film thickness, and substrate material. On thermal oxide or molybdenum, the cubic SnS gradually transforms to orthorhombic as the deposition temperature or film thickness increases. However, on sodium chloride the cubic phase remains dominant at 120 °C and 750nm, showing that the lattice-matching between substrate and film enables the access of the thermodynamically less stable phase. PL results show a peak at 1.7 eV, agreeing with the previously reported bandgap of cubic SnS [3]. Electrical measurements also show promises of this cubic SnS being a good solar cell absorber.
1. P. Sinsermsuksakul, et al., Advanced Energy Material (2014).
2. A. Rabkin, et al., Nano Letters (2015)
3. A. R. Garcia-Angelmo, et al., Phys. Status Solidi A (2015)
9:00 PM - ES14.14.14
Emitter/Absorber Interface Formation in SnS-Based Thin-Film Solar Cells
Leonard Koehler 1 , Riley Brandt 2 , Chuanxi Yang 3 , Evelyn Handick 1 , Thomas Kunze 6 , Xiaxia Liao 1 , Roberto Felix Duarte 1 , Dominic Gerlach 7 , Yoshiyuki Yamashita 7 , Toyohiro Chikyow 7 , Regan Wilks 1 4 , Roy Gordon 3 , Tonio Buonassisi 2 , Marcus Baer 1 4 5
1 Renewable Energy, Helmholtz-Zentrum Berlin für Materialien und Energie GmbH, Berlin Germany, 2 , Massachusetts Institute of Technology, Cambridge, Massachusetts, United States, 3 Department of Chemistry and Chemical Biology, Harvard University, Cambridge, Massachusetts, United States, 6 , Helmholtz-Zentrum Berlin für Materialien und Energie GmbH, Berlin Germany, 7 MANA/Nano-Electronics Materials Unit, National Institute for Materials Science, Tsukuba Japan, 4 Energy Materials In-Situ Laboratory Berlin (EMIL), Helmholtz-Zentrum Berlin für Materialien und Energie GmbH, Berlin Germany, 5 Institut für Physik und Chemie, Brandenburgische Technische Universität Cottbus-Senftenberg, Cottbus Germany
Show AbstractSnS – a promising thin-film solar cell absorber candidate material – benefits from being non-toxic, inexpensive, and based on abundant materials. Interest in SnS-based photovoltaics (PV) has increased due to its less complicated phase diagram compared to quaternary Cu2ZnSnS4 kesterite-based absorbers. Efficiencies of SnS-based solar cells are, however, still below 5% [1]. Especially the interfaces in the corresponding thin-film PV layer stack are suspected to limit device performance.
We have investigated the interface formation between ZnO-based emitter layers and the SnS absorber non-destructively and depth-dependently by using synchrotron-based hard x-ray photoelectron spectroscopy (HAXPES) employing various excitation energies. The studied emitter layers ZnO and ZnO:N of different thickness have been prepared by atomic layer deposition (ALD) directly onto intentionally oxidized SnS absorbers that also were prepared by ALD. We find a significantly stronger change in the chemical environment of Sn at the emitter/absorber interface – including the formation of a metallic phase – at the deposition of the (undoped) ZnO compared to that of the ZnO:N emitter layer. In parallel, we observe a more pronounced interface-induced band bending in that case. Based on our findings, we will draw the complete picture of the chemical and electronic (incl. energy level alignment) structure of the studied ZnO/SnS and ZnO:N/SnS hetero-interfaces with the ultimate goal to relate these results to device performance.
[1] P. Sinsermsuksakul, L. Sun, S. W. Lee, H. H. Park, S. B. Kim, C. Yang, and R. G. Gordon, “Overcoming Efficiency Limitations of SnS-Based Solar Cells,” Adv. Energy Mater., vol. 4, no. 15, Oct. 2014.
9:00 PM - ES14.14.15
Ultrasonic Spray Pyrolysis Deposition of Sb2S3 for Extremely Thin Absorber Solar Cells
Erki Karber 3 1 , Romain Parize 2 , Atanas Katerski 1 , Inga Gromoko 1 , Laetitia Rapenne 2 , Herve Roussel 2 , Estelle Appert 2 , Malle Krunks 1 , Vincent Consonni 2 , Clemens Heske 3 4
3 Department of Chemistry and Biochemistry, University of Nevada, Las Vegas (UNLV), Las Vegas, Nevada, United States, 1 Laboratory of Thin Film Chemical Technologies, Department of Materials and Environmental Technology, Tallinn University of Technology, Tallinn Estonia, 2 , Université Grenoble Alpes, CNRS, LMGP, Grenoble France, 4 , Karlsruhe Institute of Technology (KIT), Eggenstein-Leopoldshafen Germany
Show AbstractExtremely thin absorber (ETA) solar cells utilize light trapping to compensate for the decrease of bulk absorber thickness and, at the same time, benefit from shorter carrier path lengths in the absorber before separation into their respective electrodes. In the present approach, we use a layer of ZnO nanowires as the window layer, a TiO2 protective shell, Sb2S3 as the light absorbing shell, and P3HT filler as the hole conductor to create an ETA solar cell. The thickness of each shell is ~10 nm. ZnO is grown by chemical bath deposition (CBD), TiO2 by atomic layer deposition (ALD), and Sb2S3 by chemical spray pyrolysis (CSP). CSP is a fast wet-chemical deposition method by which well-adhered single-phase crystalline Sb2S3 absorber particles can be grown from raw chemicals at a growth rate of 0.07 nm/s onto planar substrates at around 250° C, using a single step deposition in air under normal pressure, with no use of vacuum or inert gas [1]. For the use of ETA solar cells, we have modified the routine to grow Sb2S3 as a homogeneous conformal shell. We use Field-Emission Scanning Electron Microscopy (SEM) and Transmission Electron Microscopy (TEM)-based experiments (HRTEM, HAADF-STEM, EDS-STEM) for characterization of the core-shell structure; Raman spectroscopy, in-plane and out-of-plane X-ray diffraction for the determination of the phase composition of the absorbing shell; optical absorption measurements, X-ray and photoelectron spectroscopies for characterization of the chemical composition and of the electronic structure of the Sb2S3 shell (XPS, UPS/IPES); as well as current-voltage and quantum-efficiency measurements for characterization of the completed solar cell. The resulting ETA cell with the ZnO/TiO2/Sb2S3/P3HT core-shell structure showed a photo-conversion efficiency of 2.3% with a promising short-circuit current density of 7.5 mA/cm2 and a high open circuit-voltage of 656 mV (one the largest reported values for ZnO nanowire-based ETA solar cells). The findings reveal the potential of Sb2S3 as an absorbing semiconducting shell when coupled with a ZnO-nanowire/TiO2-shell window structure, and the potential of using a simple wet-chemical method for the growth of cost-efficient solar cell absorber materials.
[1] Erki Kärber, Atanas Katerski, Ilona Oja Acik, Arvo Mere, Valdek Mikli, and Malle Krunks. Sb2S3 Grown by Ultrasonic Spray Pyrolysis, and Implementation in a Hybrid Solar Cell. Beilstein Journal of Nanotechnology, 2016, 7, 1662–73.
9:00 PM - ES14.14.16
Perspectives on Antimony Sulfide-Selenide Thin-Film Solar Cells
Fabiola De Bray Sanchez 1 , Jose Campos Alvarez 1 , Oscar GomezDaza 1 , M.T. Santhamma Nair 1 , P.Karunakaran Nair 1
1 , Universidad Nacional Autonoma de Mexico, Temixco Mexico
Show AbstractAmong prospective thin film solar cell materials, antimony chalcogenides have a desirable place owing to their availability in many minerals spread over many parts of the world, low toxicity and low cost. Solid solutions of Sb2SxSe3-x offer a unique opportunity to design solar cell absorber materials with optical band gap within the extremes of 1.88 eV of Sb2S3 and 1.1 eV of Sb2Se3. There have been many recent reports highlighting the stability of solar cell using extremely thin absorber films of Sb2S3 or Sb2Se3, uncommon in other upcoming absorber thin films. In this work we present the fabrication of thin film solar cells made of Sb2SxSe3-x solid solution absorber deposited by thermal evaporation of chemical precipitates. The chemical precipitate is prepared in-house from chemical solution mixture containing antimony salt, thioacetamide and selenosulfate. Typical solar cell structure is simple to make: TCO/Sb2S1.45Se1.55/C-Ag and it remains stable over long periods. Here the TCO is a commercial fluorine-doped SnO2 coating on 3 mm glass; CdS is a chemically deposited thin film of 80 nm in thickness, and the antimony chalcogenide thin film produced by thermal evaporation is of 315 nm in thickness. The antimony chalcogenide thin film has an optical band gap of 1.45 eV, optical absorption coefficient of > 105 cm-1 in the visible region, and a photoconductivity of 10-6 Ω-1 cm-1. Carbon electrodes defining the cell structure are of colloidal graphite paste of 0.35 cm2 in area, painted on the antimony chalcogenide thin film. After the paint has dried, the cell structure is heated at 300 oC in a nitrogen ambient of 20 Torr pressure for 30 min. This process brings-in stability for the cell structure upon operation under sunlight over many hours and days. The cell structure is finished by applying colloidal silver paint and drying it at 80 oC for 30 min. Under standard conditions, the solar cell exhibit open circuit voltage Voc of 0.552 V, Jsc of 19.9 mA/cm2 and conversion efficiency of 4.92%. We also present here the energy level scheme for the junction, and capacitance-voltage characteristics, which enable assessment of the built-in voltage of the junction and carrier concentration in the antimony chalcogenide film.
9:00 PM - ES14.14.17
Aqueous Spray Pyrolyzed ZnO as Orientation Induced Buffer Layer for Sb2Se3 Photovoltaics of Improved Stability
Liang Wang 1 , Dengbing Li 1 , Kanghua Li 1
1 , Huazhong University of Science and Technology, Wuhan China
Show AbstractOptimum band gap, strong absorption, simple composition, low cost and non-toxic constituents promise antimony selenide (Sb2Se3) an appealing light absorber for photovoltaic applications. Moreover, Sb2Se3 showed intrinsically benign grain boundaries when properly aligned because of their peculiar one dimensional crystal structure. Thus, a facile and straightforward method to control Sb2Se3 orientation is in urgent requirement. Here, we provide an easy to implement strategy for orientation control via quasi-epitaxial mechanism. Environmental-friendly ZnO produced by spray pyrolysis was used as the buffer layer instead of toxic CdS to induce Sb2Se3 growth orientation. The results indicate that the orientation of Sb2Se3 show strong correlation to the orientation of ZnO buffer layer, resulted in a ZnO[100]/ Sb2Se3[221] junction with low recombination loss. The combination of high bulk transmission and low interfacial recombination synergistically boost a highest ZnO/ Sb2Se3 device efficiency 5.78% with impeccable devices stability.
9:00 PM - ES14.14.18
Synthesis and Characterization of Sn(Sx,Se1-x)2 Ternary Alloy Thin Films for Photovoltaic Applications
Joshua Fox 1 2 , Xiaotian Zhang 1 2 , Zakaria Al Balushi 1 2 , Joan Redwing 1 2 , Nathan Martin 1 2
1 Materials Science and Engineering, The Pennsylvania State University, State College, Pennsylvania, United States, 2 , Materials Research Institute, State College, Pennsylvania, United States
Show AbstractCrystalline silicon solar cells comprise over 90% of the commercial solar cell market, but their conversion efficiencies are limited to a theoretical maximum of ~29%. One pathway to increase efficiency is to form tandem devices that are comprised of a top wider bandgap absorber material and a bottom c-Si photovoltaic device. A theoretical efficiency approaching 44% can be achieved by using a top cell with a bandgap energy of ~1.7 eV. Candidate materials for the 1.7 eV top absorber include dilute nitride III-Vs, CZTS and ZnSiP2, but none of these have yet emerged as the material of choice. Our research is focused on investigating Sn(SxSe1-x)2 ternary alloys, a layered material system with bandgap energies that span the range from 1.1 eV for SnSe2 to 2.2 eV for SnS2, as a potential top absorber material. In addition to having a bandgap energy in the correct range for tandem devices, Sn(SxSe1-x)2 crystallizes as 2-D sheets, bound to one another via van der Waals forces. This is favorable for tandem solar cell applications due to the lack of required lattice matching between the stacked top and bottom absorbing layers which typically limits the choice of material for multi-junction devices. In addition, Sn and S are earth-abundant elements which is important for the development of a sustainable photovoltaic technology. Despite the intriguing potential of these materials, there have been limited studies on the synthesis and properties of Sn(SxSe1-x)2 thin films.
Our initial studies have focused on the synthesis of SnSe2 and SnS2 thin films as binary components of the Sn-Se-S system via powder vapor transport (PVT). Source powders are evaporated in a heated horizontal quartz tube and transported downstream resulting in thin film deposition. The film characteristics were evaluated as a function of furnace temperature, growth duration, carrier gas flow rate and substrate type. Most significantly, the orientation of SnSe2 and SnS2 platelets are shown to vary depending on substrate type and position within the reactor. Experiments performed with oxide substrates (sapphire and SiO2) show out of plane platelet orientation likely due to the presence of dangling bonds on the surface which promote attachment of the platelet edges. In contrast, growth on epitaxial graphene which provides a well passivated surface resulted in lateral hexagonal platelets with domain sizes up to ~100 µm which can be coalesced to form continuous films with an RMS roughness of ~0.5 nm over a 5x5 µm scan area. Raman spectroscopy was used to confirm the formation of SnSe2 and SnS2 single phase films. Photoluminescence spectroscopy revealed room temperature emission from the SnS2 at ~1.9 eV which is close to the anticipated bulk bandgap energy. The results demonstrate the importance of surface passivation in obtaining high quality films which will be important for integration on Si. The synthesis and properties of Sn(SxSe1-x)2 films is currently underway and will also be discussed.
9:00 PM - ES14.14.19
Towards the Improvement of the Antimony Sulfide-Selenide Sensitized Solar Cells—Effect of Mesoporous TiO2 Annealing and CdS Interlayer
Araceli Hernandez-Granados 1 , Jose Escorcia-Garcia 4 , Diego Perez-Martinez 3 , Jose Garcia Cerrillo 2 , Carmina Menchaca-Campos 1 , Hailin Zhao Hu 2
1 CIICAp, UAEM, Cuernavaca, Morelos, Mexico, 4 CINVESTAV, IPN, Saltillo, Coahuila, Mexico, 3 CBI, UAM-AZC, Mexico City, DF, Mexico, 2 IER, UNAM, Temixco, Morelos, Mexico
Show AbstractA lot of efforts have been done in the search of new materials and device configurations to develop alternative technologies that could exploit efficiently the solar energy-to-electricity conversion, with a reduced ambient impact and at low cost. Among these novel materials is the antimony sulfide (Sb2S3), which has optoelectronic properties useful for its application as an absorbing material. However, it presents recombination of the photo-carriers in the bulk as well as in the surface that limits its application for high efficient solar cells. An alloying between the Sb2S3 and antimony selenide (Sb2Se3) materials, i.e. Sb2(SxSe1-x)3, could overcome these problems by exploiting their optoelectronic properties. Here we present a cheap and easy methodology of obtaining solid state SSSCs by solution deposition methods such as sol-gel, successive ionic layer absorption and reaction (SILAR), and chemical bath, and their further optimization by studying the effect of annealing in the mp-TiO2 layer and by varying the thickness of the CdS interlayer. The mp-TiO2 electrodes were annealed during 1 h at different temperatures from 300 to 600 °C and later sensitized with CdS by SILAR. Then, the ternary absorber was deposited by chemical bath deposition and heated at 300 °C in nitrogen. The results indicate that the XRD pattern of the mp-TiO2, annealed at temperatures below 500 °C, showed an amorphous feature while those heated above 500 °C showed a crystalline structure resembling the Anatase phase. The increase of the annealing improves the crystallinity of mp-TiO2 films as well as the mean crystallite size, from 16 to 18 nm for 500 and 600 °C, respectively. The SEM analysis shows a reduction of the porosity while increasing the annealing temperature with mean particle sizes of ~50 nm. The optical band gap of the mp-TiO2 is close to 3.2 eV. On the other hand, increasing the number of CdS depositions by SILAR improves the optical absorbance of the mp-TiO2 electrodes. Then, the materials were incorporated in the SSSCs using two different conductive substrates, fluorine-doped tin oxide (FTO) and indium tin oxide (ITO) coated glasses. The photovoltaic characterization of the cells showed a better performance with FTO rather than ITO as well as in those cells having a higher content of CdS. The best solar cell, FTO/bl-TiO2/mp-TiO2/CdS/Sb2(SxSe1-x)3/C solar cells, showed a power conversion efficiency of 1.7% under AM1.5 G solar radiation. The solar cells remained stable during one month of evaluation under solar radiation.
9:00 PM - ES14.14.20
Band Gap Grading in Cu2ZnSnSe4 Thin Films Prepared by the Annealing Process of S and Se from Metal Precursor-Based Deposits
Juran Kim 1 , Gee Yeong Kim 1 , Trang Nguyen 1 , Seokhyun Yoon 1 , Youngill Kim 2 , Kee-Jeong Yang 2 , Dae-Hwan Kim 2 , William Jo 1
1 Department of Physics, Ewha Womans University, Seoul Korea (the Republic of), 2 Convergence Research Center for Solar Energy, Daegu Gyeongbuk Institute of Science &Technology, Daegu Korea (the Republic of)
Show AbstractCu2ZnSn(S,Se)4 (CZTSSe) is an attractive candidate absorber material for light absorber layer of solar cells due to its high optical absorption coefficients and the useful band gap (Eg) of 1.4-1.5 eV. Nonetheless, its power conversion efficiency (about 13%) is still far below the 22.3% record efficiency of CIGS thin-film solar cells. To enhance the solar cell performance, different Se, S compounds (Se+S, Se+SeS2) were used for the post process. Using sputtering method, we deposited the precursor films with stacking order of Cu/Sn/Zn on Mo coated soda line substrates. After that, the samples were sulfor-selenized. They have shown similar power converstion efficiency (PCE) of ~11%. Using Kelvin probe force microscopy (KPFM), we could see the local surface potential differences on the thin film before and after KCN etching. Before the etching, the surface potential near the grain boundaries (GBs) were negative, indicating positive band bending. With etching process, however, the negative band bending is formed near the GBs. This advocates the benign effects of GBs on polycrystalline thin films solar cells, which act as the dominant current flow path for the minority carriers. The work function distribution shows more unified CZTSSe peaks (ΦCZTS = 4.6~4.8 eV). As a results, we can say KCN etching can remove the secondary phases, or the residue of Se and S from the post annealing process. Therefore, optimized post process and KCN etching are needed so as to improve the thin film solar cell performance.
9:00 PM - ES14.14.21
Surface and Depth Profile of Phase Identification of Tin Sulfide Thin Films for Earth-Abundant Solar Cell Applications
Juran Kim 1 , Ja-yeong Kim 1 , Seokhyun Yoon 1 , Jeong-yoon Kang 2 , Chan-Wook Jeon 2 , William Jo 1
1 Department of Physics, Ewha Womans University, Seoul Korea (the Republic of), 2 School of Chemical Engineering, Yeungnam University, Gyeongsan Korea (the Republic of)
Show AbstractTin sulfide (SnS) is obtaining a growing interest as a promising earth abundant semiconductor for photovoltaic materials because of its direct band gap (1.3 ~ 1.5 eV), higher absorption coefficient and good carrier mobility. Moreover, SnS is non-toxic and easily controllable chemical stoichiometry. It is known that phase transition occurs depending on the sulfurization temperature. Under lower temperature between 150 and 300 °C, SnS2 and Sn2S3 phases are formed, and over 300 °C SnS phase is favored. We deposited Sn precursor using sputtering method and then they were sulfurized using S vapor at 400 °C. In order to investigate the depth profile of the thin film, we tried dimple-grinding of the absorber film to expose the film sidewall with a shallow angle. We chose 4 points with an equal interval between Mo back-contact layer and the absorber surface, and characterized their electrical and structural properties by Kelvin probe force microscopy (KPFM) and micro-Raman scattering spectroscopy. As going to the Mo back-contact layer, the thin film shows less mixed phase such as SnS2 and Sn2S3. Their work function distribution near the back-contact shows also a work function of ~ 4.9 eV. From the results, we can expect the phase control of tin sulfides for fabrication of high efficiency solar cells.
9:00 PM - ES14.14.22
Detection of Competing Parasitic Phases in Pyrite Photovoltaic Thin Films and Their Influence to Solar Cell Device Performance
Juran Kim 1 , Gee Yeong Kim 1 , Hankyoul Moon 1 , Seokhyun Yoon 1 , Il Wan Seo 2 , Yunsang Lee 2 , Dong Gwon Moon 4 3 , SeJin Ahn 4 , William Jo 1
1 Department of Physics, Ewha Womans University, Seoul Korea (the Republic of), 2 Department of Physics, Soongsil University, Seoul Korea (the Republic of), 4 Photovoltaic Laboratory, Korea Institute of Energy Research, Daejeon Korea (the Republic of), 3 Department of Materials Science and Engineering, Yonsei University, Seoul Korea (the Republic of)
Show AbstractIron pyrite (cubic FeS2) is one of the promising materials for thin-film solar cells due to the suitability of its band gap of ~0.95 eV for earth-abundant materials and its high light absorption coefficient. We investigated FeS2 thin films grown by the non-vacuum spin-coating method [1]. The thin films were annealed under a sulfur atmosphere at different sulfurization temperatures. The phase transformation from marcasite-containing pyrite to pure pyrite occurs between 350 and 400 °C. The structural phase formation on the films depends on the sulfurization temperature. FeS2 thin films with and without the marcasite phase were investigated in terms of their local electrical properties and carrier transport by conductive atomic force microscopy (c-AFM) and Kelvin probe force microscopy (KPFM). Interestingly, the pure pyrite thin film shows less conducting behaviour that the mixed phase sample since the mixed phase thin film has other residues on the surface. However, the marcasite-containing FeS2 thin film has vague GBs and grain formation was poor. As surface current and a negative band bending is formed near the GBs, the single phased FeS2 thin film shows better carrier transport behaviour. Pure pyrite has two major work functions at 4.64 and 4.70 eV, and the mixed phase sample shows multiple peaks below 4.63 eV. Thus, we expect that FeS2 with the pure pyrite phase will perform better as a light absorber in solar cells [2].
References
[1] D. G. Moon, A. Cho, J. H. Park, S. H. Ahn, H. S. Kwon, Y. S. Cho, and S. J. Ahn, J. Mater. Chem. A., 2014, 2, 17779.
[2] J. Kim, G. Y. Kim, H. Moon, S. Moon, I. W. Seo, Y. Lee, D. G. Moon, S. J. Ahn, and W. Jo, RSC Adv., 2016, 6, 81394.
9:00 PM - ES14.14.23
Structural and Electronic Characterization of Tin Calcium Sulfide Alloy Thin Films Made by Atomic Layer Deposition
Chuanxi Yang 1 , Sang Bok Kim 1 , Xizhu Zhao 1 , Laura Schelhas 2 , Xiabing Lou 1 , Roy Gordon 1
1 , Harvard University, Cambridge, Massachusetts, United States, 2 , SLAC National Accelerator Laboratory, Menlo Park, California, United States
Show AbstractRecent advances on tin monosulfide (SnS) based thin-film solar cells have led to efficiencies over 4% [1] using atomic layer deposition (ALD). Despite its strong absorption coefficient, SnS has two inherent flaws as a photovoltaic absorber. First, the indirect band gap of 1.1 eV results in a sub-optimal open circuit voltage. Second, the orthorhombic crystal structure of SnS results in anisotropic optical and electronic properties. Recently, it has been shown that by alloying SnS with calcium using pulsed laser deposition, it can adopt a cubic crystal structure [2]. This transition would induce a direct band gap and isotropic electrical and optical properties.
Here, we report Sn1-xCaxS thin films made by ALD, using a new calcium precursor [3]. Films with different Ca composition (x) are deposited by changing the dosing ratio of Sn to Ca precursors. The deposition temperature was kept between 135 °C and 190 °C. Post-deposition annealing was applied in order to increase the grain size and to reduce structural defects. The crystal structures of Sn1-xCaxS films were characterized by X-ray diffraction (XRD) and transmission electron microscopy (TEM). Their electrical properties were characterized by Hall measurements.
We show that the crystal structures of the alloy films vary with deposition temperature. The transition from orthorhombic to cubic phase occurs at around x=0.25. As-deposited Sn1-xCaxS films have higher resistivity and lower carrier mobility compared to SnS, while annealing at 400 °C reduces their resistivity significantly. However, phase separation of cubic Sn1-xCaxS into orthorhombic SnS and cubic CaS is observed by XRD, indicating the metastable nature of Sn1-xCaxS alloys.
[1] Sinsermsuksakul, Prasert, et al. "Overcoming Efficiency Limitations of SnS-Based Solar Cells." Advanced Energy Materials 4.15 (2014).
[2] Vidal, Julien, et al. "Structural and electronic modification of photovoltaic SnS by alloying." Journal of Applied Physics 115.11 (2014): 113507.
[3] Kim, Sang Bok, et al. "Synthesis of Calcium (II) Amidinate Precursors for Atomic Layer Deposition through a Redox Reaction between Calcium and Amidines." Angewandte Chemie 128.35 (2016): 10384-10389.
9:00 PM - ES14.14.24
Optoelectrical Properties of The Schottky Junction Devices Based on Mo1-xWxSe2
Sum-Gyun Yi 1 , Sung Hyun Kim 1 , Sungjin Park 1 , Dong Gun Oh 1 , Hwan Young Choi 1 , Nara Lee 1 , Young Jai Choi 1 , Kyung-Hwa Yoo 1
1 Department of Physics, Yonsei University, Seoul Korea (the Republic of)
Show AbstractRecently, various heterosctuctures consisting of MoS2, MoSe2, or WS2, have been reported to investigate electrical and optoelectrical properties which are piezo, p-n diode, photodiode and photovoltaic effects. Here, we report on Schottky junction photovoltaic cells based on multilayer Mo1-xWxSe2 with x = 0, 0.5, and 1. To generate built-in potentials, Pd and Al were used as the source and drain electrodes in a lateral structure, and Pd and graphene were used as the bottom and top electrodes in a vertical structure. These devices exhibited gate-tunable diode-like current rectification and photovoltaic responses. Interestingly, compared to MoSe2 or WSe2 FET devices with asymmetric electrode, Mo1-xWxSe2 FET devices yielded enhanced photovoltaic and photocurrent effects, likely because of the greater adjusted band alignment in the Mo0.5W0.5Se2 devices. Furthermore, we showed that Mo0.5W0.5Se2 - based vertical Schottky diodes yield a power conversion efficiency of approximately 16% under 532 nm light and approximately 13% under a standard air mass 1.5 spectrum, demonstrating their remarkable potential for photovoltaic applications.
9:00 PM - ES14.14.25
Preparation and Characterization of In2S3:V Solar Cells
Leonard Waegele 1 , Tanja Jawinski 1 2 , Holger von Wenckstern 2 , Marius Grundmann 2 , Matthias Maiberg 1 , Roland Scheer 1
1 , Martin Luther University Halle-Wittenberg, Halle Germany, 2 , University of Leipzig, Leipzig Germany
Show AbstractThe efficiency of common single junction solar cells is limited by the Shockley-Queisser limit. One approach to overcome the Shockley-Queisser limit is the formation of an intermediate band within the energy band gap of the solar cell’s absorber. This also allows absorption of photons with energies smaller than the band gap. Based on theoretical calculations, a promising candidate for such absorbers may be In2S3 doped with vanadium. In our contribution, we prepare intrinsic In2S3:V absorbers by co-evaporation on n-ZnO:Al back contacts. Raman and in-situ XRD data unveil a growth of the β-In2S3-phase with high crystallinity. No secondary phases are detected. Optical absorption shows subband gap response for V-doped films. On top of the absorbers, p-ZnCo2O4 window layers and thin gold front contacts are deposited, completing the p-i-n structure. The measurements of the current-voltage characteristics in the dark reveal a rectifying behaviour. However, measurements under illumination show a low collection of charge carriers. For deeper understanding, external quantum efficiency (EQE) measurements are performed for different [V]/([V]+[In])-ratios (VVI) of the absorber. It comes out that the quantum efficiency does not change within 0 ≤ VVI ≤ 0.03. Although this missing subband gap response might be due to ideal intermediate band behaviour, we suppose high recombination to suppress 2-step excitation at room temperature in our samples.
9:00 PM - ES14.14.26
Transparent Laminated Electrode for Highly Efficient Chemically Synthesized n-MoS2 /p-Si Heterojunction Solar Cell
Sungbum Kang 1 , Ki Chang Kwon 2 , Ho Won Jang 2 , Kyoung Jin Choi 1
1 , UNIST, Ulsan Korea (the Republic of), 2 Department of Materials Science and Engineering, Seoul National University, Seoul Korea (the Republic of)
Show AbstractTwo-dimensional materials including molybdenum disulfide (MoS
2), graphene have attracted great attention as future photovoltaic materials due to strong sunlight absorption which is one order higher than Si and GaAs commonly used as a solar cell. Using bulk-like 220nm MoS
2 thin film, Shanmugam et al. attained a power conversion efficiency (PCE) of 1.8% based ITO/MoS
2 Schottky junction solar cell with Au electrode.
1 Some papers are reported about n-MoS
2 /p-Si heterojunction solar cell by Hao et al.
2 They deposited fully covered Pd electrode on the top of MoS
2. However, the device with fully covered Pd electrode suffered from low short-circuit photocurrent density due to a lot of light is reflected caused by Pd electrode.
In order to maximize exotic light absorption property of MoS
2 as a photovoltaic application, the transparent but highly conductive electrode is needed. In this work, we demonstrate high efficient n-MoS2 /p-Si heterojunction solar cell by using chemically synthesized few-layers of MoS
2. After synthesized MoS
2 are moved to p-Si, transparent electrode developed by electrospinning and lift off process is laminated. Based on the Transparent laminated electrode (Pd-based nanowire / tapes) / n-MoS2 /p-Si, the record efficiency over 6% is obtained.
[1] M. Shanmugam, C. A. Durcan and B. Yu, Nanoscale, 2012, 4, 7399.
[2] L. Z. Hao, W. Gao, Y. J. Liu, Z. D. Han, Q. Z. Xue, W. Y. Guo, J. Zhu and Y. R. Li Nanoscale, 2015, 7, 8304
*E-mail:
[email protected]9:00 PM - ES14.14.27
Antimony Sulfide-Selenide Thin-Film Solar Cells of 4% Efficiency from Stibnite Mineral
Geovanni Vazquez Garcia 1 , Eira Anais Zamudio Medina 1 , Laura Guerrero Martinez 1 , Oscar Leyva Castrejón 1 , Jacqueline Moctezuma Ortiz 1 , M.T. Santhamma Nair 1 , P.Karunakaran Nair 1
1 , Universidad Nacional Autonoma de Mexico, Temixco Mexico
Show AbstractAntimony sulfide, which is available as powdered stibnite mineral has been used to prepare thin film antimony sulfide-selenide solar cells by vacuum thermal evaporation. The merit of this approach for large scale production of antimony chalcogenide solar cell is that stibnite mineral is of common occurrence in many parts of the world. It is devoid of impurities except for overgrowth of calcite or quartzite, which might be left as residue in thermal evaporation, without affecting the film composition. Solar cells are developed on commercial transparent conductive oxide (TCO) coating of SnO2:F on 2 mm clear glass (TEC 15) on which a CdS thin film film of 100 nm in thickness was added by chemical deposition. The solar cells are of the configuration: TCO/CdS/Sb2SxSe3-x/C-Ag. Here, x = 3 when the evaporation source is solely powdered stibnite mineral but is reduced down to 1.2 when a chemical precipitate of Sb2Se3 is added to the source. Thickness of the Sb2SxSe3-x film has been 250-300 nm. The area of colloidal graphite/colloidal silver electrode is 1 cm2. For solar cells using Sb2S3 thin film, with optical bandgap (Eg) of 1.88 eV, the open circuit voltage (Voc) is 0.650 V, sort circuit current density (Jsc) is 8 mA/cm2 and solar energy conversion efficiency (η) is 1.2 %. With Sb2S1.2Se1.8 thin film, these values are: 0.420 V, 19 mA/cm2 and 4.2% respectively. In this latter case, the external quantum efficiency of the solar cell reaches 78 % at a wavelength of 600 nm. This thin film has crystalline grain size of 22 nm. The electrical conductivity of this antimony sulfide-selenide thin film is 10-8 Ω-1 cm-1, which increases by two orders of magnitude under the illumination. Characteristics of solar cells of intermediate compositions of Sb2SxSe3-x are presented and perspectives are discussed. We find that use of laboratory reagent grade Sb2S3 powder instead of powdered stibnite mineral with some quartz inclusion as seen in x-ray diffraction does not result in better solar cell characteristics.
9:00 PM - ES14.14.28
Versatile Transition Metal Perovskite Chalcogenides as Strong Solar Absorbers
Shanyuan Niu 1 , Huaixun Huyan 1 , Yang Liu 1 , Matthew Yeung 1 , Kevin Ye 1 , Louis Blankemeier 1 , David Singh 2 , Rehan Kapadia 1 , Jayakanth Ravichandran 1
1 , University of Southern California, LA, California, United States, 2 , University of Missouri, Columbia, Missouri, United States
Show AbstractTransition metal perovskite chalcogenides (TMPC) are a new class of versatile semiconductors with high absorption coefficients and desirable optoelectronic properties. Theoretical studies predict band gap tunability from visible to IR range and large absorption coefficient at the band edge. These features are appealing for photovoltaics, especially for the realization of tandem cells. We report a new iodine catalyzed synthesis method to produce high quality polycrystalline samples of TMPCs. Systematic structural, chemical, and stability studies are performed on BaZrS3 in distorted perovskite phase, and SrZrS3 in two room temperature stabilized phases (distorted perovskite and needle like). Band gap engineering over the solar relevant energies is demonstrated by controlling the structure and chemical composition of TMPCs, without any need for alloying. Clear and intense photoluminescence spectra are shown for the first time and establish tunable optical band gap across the visible range. Photovoltaic potential is evaluated by measurement of luminescence efficiency and extraction of quasi Fermi level splitting. One of materials’ external luminescence efficiency approaches 0.2% in the polycrystalline form. Strategies to improve the optoelectronic properties and also to obtain thin films of these materials for device characterizations are also discussed.
9:00 PM - ES14.14.29
Cu2-II-Sn-VI4 (II = Ba, Sr and VI = S, Se) Quaternary Compounds for Earth-Abundant Photovoltaics
Weiwei Meng 1 2 , Feng Hong 1 3 , Zewen Xiao 1 , Wenjun Lin 3 , Jianbo Wang 2 , Yanfa Yan 1
1 , Department of Physics and Astronomy, The University of Toledo, Toledo, Ohio, United States, 2 , School of Physics and Technology, Center for Electron Microscopy, MOE Key Laboratory of Artificial Micro- and Nano-structures, and Institute for Advanced Studies, Wuhan University, Wuhan, Hubei, China, 3 , Department of Physics, International Centre for Quantum and Molecular Structures, SHU-Solar E R & D Lab, Shanghai University, Shanghai China
Show AbstractWe propose Cu2-II-Sn-VI4 (II = Ba, Sr and VI = S, Se) quaternary compounds for earth-abundant solar cell applications. Through density functional theory calculations, we show that these compounds exhibit similar electronic and optical properties to kesterite Cu2ZnSnS4 (CZTS): high optical absorption with band gaps suitable and tunable for efficient single-junction solar cell applications. Moreover, the Cu2-II-Sn-VI4 compounds exhibit defect properties more suitable for photovoltaic applications than those of CZTS. In CZTS, the dominant defects are the deep acceptors, Cu substitutions on Zn sites, which cause non-radiative recombination and limit the open-circuit voltages of CZTS solar cells. On the contrary, the dominant defects in Cu2-II-Sn-VI4 are the shallow acceptors, Cu vacancies, if the compounds are deposited under Cu-poor conditions and S and Se vapors are controlled carefully. These defect properties are similar to those seen in CuInSe2. Therefore, Cu2-II-Sn-VI4 compounds are to be promising candidates for efficient earth-abundant thin-film solar cell and photoeletrochemical water-splitting applications.
9:00 PM - ES14.14.30
Zinc Molybdenum Oxide—A New Solar Absorber
Pramod Ravindra 1 , Eashwer Athresh 1 , Rajeev Ranjan 1 , Srinivasan Raghavan 1 , Sushobhan Avasthi 1
1 , Indian Institute of Science, Bangalore, KA, India
Show AbstractOxides are an attractive family of materials for solar absorption due to their abundance, chemical stability, non-toxicity, ease of fabrication and low cost of deposition. However, oxide-based photo-absorbers have seen limited device success, mainly due to low minority carrier lifetimes and low mobility. Due to spectacular success of lead-based organic-inorganic perovskite, there has been renewed interest in discovering other perovskite-like oxide based solar absorbers. In this paper, we present optical and electronic properties of zinc molybdenum oxide (ZMO), a promising solar absorber.
ZMO films were deposited by pulsed laser deposition (PLD) using phase-pure target. PLD was selected because it yields high-quality stoichiometric films. UV-visible spectroscopy showed that ZMO has a direct bandgap of 1.8 – 2 eV and an absorption coefficient > 104 cm-1 for visible light. So 1000 nm thick ZMO can efficiently absorb almost all above-bandgap photons of the solar spectrum. Besides efficient absorption, reasonable mobility and carrier concentration are other important criteria for a solar absorbers. Majority carrier concentration and mobility were determined using Hall measurements. As-deposited ZMO was unintentionally n-type doped with carrier density of 1017 cm-3. This is expected, as most oxides are known to be n-type in nature due to presence of oxygen vacancies. Electron mobility was measured 0.6 – 0.7 cm2/V-s, high enough for thin-film solar cells, and comparable to mobility reported for other thin-film solar absorbers. UV-photoelectron and UV-visible spectroscopy showed that the conduction band minimum and valence band maximum of ZMO are 4.4 eV and 6.3 eV below vacuum level, respectively. The knowledge of band-position allows us to choose appropriate hole and electron transport layers (ETL) in devices, e.g. TiO2 is expected to be a good ETL for ZMO. Unoptimized single-sided ZMO diodes with FTO/TiO2/ZMO/Au structure show photoconductivity, i.e. increase in current upon illumination. Work is currently underway to engineer a complete solar device with ETL and HTL.
To summarize, we present optical and electrical properties of a new thin-film oxide semiconductor, zinc molybdenum oxide. To our knowledge, this is the first report about deposition and properties of thin films of ZMO. Based on our measurements and observations, we propose ZMO as a suitable absorber in solar cells. Selection of appropriate carrier selective contact materials and device engineering is expected to lead to fabrication of low cost, stable, safe solar cells with high efficiencies.
9:00 PM - ES14.14.31
Antimony Chalcogenide Solar Cells Exceeding 2% Efficiency Prepared by Thermal Evaporation of Core-Shell Precipitates
Angelica Lizbeth Espinosa Santana 1 , Geovanni Vazquez Garcia 1 , Jose Campos Alvarez 1 , P.Karunakaran Nair 1
1 , Universidad Nacional Autonoma de Mexico, Temixco Mexico
Show AbstractAntimony chalcogenide thin film solar cells have by now crossed 5.5% conversion efficiency (η) for solar energy. Even though antimony is the 64th out of 103 elements with respect to its abundance in the Earth's crust, it is contained in substantial quantity in more than 40 minerals. World production of antimony and its compounds exceeds 150,000 metric tons in any year. Application of antimony sulfide-selenide in solar cell is a relatively novel application for this well known element. We present how Sb2(S-Se)3 powder evaporation source may be prepared as a core-shell chemical precipitate, by covering the surface of powdered stibnite (Sb2S3) mineral in a chemical bath originally developed to deposit Sb2SxSe3-x thin films. The chemical bath for the shell-film deposition was prepared from aqueous solutions of potassium antimony tartrate, sodium selenosulfate and thioacetamide. At 80 oC of the deposition bath with the core-powder under constant stirring, a shell-thin film of Sb2SxSe3-x of 200 nm in thickness is deposited on Sb2S3 core of 50-60 μm in diameter. Repeating the deposition for up to 4 times, the selenium content in the core-shell precipitate may be increased. The filtered precipitate serves as a powder source for thermal evaporation in a vacuum unit with provision for substrate heating. In a typical production process, 600 mg of the precipitate is evaporated from the crucible to form a thin film of Sb2SxSe3-x of 250 nm in thickness on SnO2:F (TEC-15)/CdS(80 nm) substrates at 425 oC. The crucible-substrate distance of 40 cm assures uniformity of the thin film over a substrate area of 15 cm x 15 cm. Colloidal graphite paint typically of 9 mm x 9 mm in area (cell area, 0.8 cm2) is applied on the thin film. These electrodes are thermally stabilized upon heating the cell structure at 300 oC for 45 min in a nitrogen atmosphere of 30 Torr pressure. The variation of the thickness of the shell-thin film on the stibnite-core is decisive in the chemical composition (x) of the Sb2SxSe3-x thin film absorber produced. In 1, 2, 3, and 4 depositions, the value of x drops from 2.2 to 1.2. For a composition of x = 2.2, the open circuit voltage Voc of the solar cell is 0.65 V, with short circuit current density Jsc of 6 mA/cm2 and η of 1.2%. As x drops with the number of depositions to 1.6, these values change to Voc, of 0.62 V, Jsc of 8.8 mA/cm2, and η of 2.2 %. As the shell-film thickness increases further, x becomes 1.5, Voc drops to 0.55 V, Jsc of 12.7 mA/cm2, and achieving η of 2.3%. All these solar cells show remarkable stability under solar radiation. The drop in Voc and rise in Jsc with decrease in x takes place because the optical band gap Eg of Sb2SxSe3-x decreases nearly monotonically with the composition, from 1.88 eV for Sb2S3 (x = 3) and 1.1 eV for Sb2Se3 (x = 0).
9:00 PM - ES14.14.32
Antimony Sulfide Thin Films Deposited by Microwave Heating for Hybrid Solar Cells Application
Claudia Martinez-Alonso 1 , Alejandro Baray 2 , Sandra Mayen-Hernandez 1 , Hailin Zhao Hu 2
1 , FQ-UAQ, Queretaro Qro Mexico, 2 , IER-UNAM, Temixco, Morelos Mexico
Show AbstractAntimony sulfide (Sb2S3) is a semiconductor of V-VI group with an orthorhombic crystalline system and a direct permitted band-gap varied from 0.9 to 1.7 eV. It shows an absorption coefficient of 1.8 x105 cm-1 at 450 nm, and electrical conductivity of 10-8-10-9Ω-1cm-1 at room temperature. Because of its semiconductor properties, Sb2S3 could be used in solar cells as a window layer, absorber or sensitizer material. The principal method for Sb2S3 thin films deposition is chemical bath deposition (CBD). But, by this method, the as-deposited films may contain impurities such as Sb2O3, SbOCl, and Sb(OH)3, and showed the amorphous phase as well. The thermal annealing treatment is necessary for those films to eliminate parcially the impurities and achieve the phase transition, from amorphous to crystalline (Stibnite).
In this work, Sb2S3 thin films were deposited by microwave heating method. Antimony trichloride (SbCl3) was used as antimony source, tartaric acid (C4H6O6) as complex agent, sodium hydroxide (NaOH) as buffer, thioacetamide (TA) as sulfur source and water as solvent. The temperature of synthesis was varied from 150 °C to 180 °C, and the reaction time, for 15 to 30 min. The as-deposited Sb2S3 thin films were crystalline , and 15 min were enough to obtain a good film thickness. The atomic ratio of S:Sb was found to be close to 1.5:1 by EDS analysis. The SEM images of those films show the formation of microrods. After the characterization, the Sb2S3 thin films were applied in hybrid solar cells as an absorber-sensitizer material. The structure of the cell was TCO/compact-TiO2/mesoporous-TiO2/Sb2S3/P3HT/C/Au. The preliminary results suggested that the improved external quantum efficiency of the cells with microwave deposited Sb2S3 thin film, compared to those without them, should come from a good adhesion of the Sb2S3 films on mesoporous TiO2 layer. A major control on Sb2S3 film morphology and thickness is necessary to improve the photovoltage and fill factor of the mencioned solar cells.
9:00 PM - ES14.14.33
Comparative Study of the Properties of SnS Thin-Film Absorbers Deposited by Both E-Beam Evaporation Followed by Sulfurization and Rf Sputtering Processes
Vinaya Kumar Arepalli 1 , Tingjian Huang 1 , Jeha Kim 1
1 , Cheongju University, Cheongju Korea (the Republic of)
Show AbstractTin monosulfide (SnS) is of great interest as a promising absorber material for solar cell applications. Because it has high absorption coefficient of ~10 5 cm-1 with both direct (1.21 eV – 1.5 eV) and indirect (1.0 eV – 1.2 eV) band gaps. Moreover, when compared with other conventional absorbers like CdTe, CIS, and CIGS, the selection of ohmic contact for SnS films is more convenient due to its lower work function and electron affinity. SnS thin film absorber has been prepared by several vacuum based and non-vacuum based methods like e-beam evaporation, thermal evaporation, RF sputtering, and spray pyrolysis etc. In this work, we deposited the SnS thin films on Mo/SLG both by rapid thermal process (RTP) sulfurization of Sn layers prepared by electron beam evaporation and by RF sputtering from SnS alloy target. The effect of sulfurization parameters, such as temperature and pressure, on the properties of tin sulfide layers has been investigated for samples grown by e-beam followed by RTP sulfurization step. Whereas, for samples grown by RF sputtering, we have studied the effect of sputtering conditions like argon pressure, power and post annealing parameters on the properties of SnS thin films. The deposited SnS films were characterized by scanning electron microscopy, electron dispersive spectroscopy, X-ray diffraction, micro Raman spectroscopy, and UV-Visible spectroscopy. X-ray diffraction spectra confirmed that a dominant SnS pure phase of herzenbergite crystal structure existed with (040) and (111) as major orientation peaks for samples grown by e-beam followed by RTP sulfurized and RF sputtered simultaneously. We will further discuss the properties of solar cell with structure SLG/Mo/SnS/CdS/i-ZnO/ITO/Al.
9:00 PM - ES14.14.34
Cus Thin Films Doped with Cus and Zns Nanoparticles for Semiconductor Devices Applications
Adriana Garcia Gallardo 1 , Amanda Carrillo 1 , Maria de la Luz Mota 1 , Santos Castillo 2 , Edna De la Cruz 3
1 , Universidad Autonoma de Ciudad Juarez, Ciudad Juarez Mexico, 2 , Universidad de Sonora, Hermosillo, Sonora, Mexico, 3 , CONACyT-Escuela Superior de Física y Matemáticas-Instituto Politécnico Nacional, Altamira, Tamaulipas, Mexico
Show AbstractIn this work the doping effect of Sol-gel synthetized CuS and ZnS nanoparticles (NPs) on top of chemical bath deposited CuS thin films was studied. The doped semiconductor thin films were optically, chemically, structurally and electrically characterized. XRD characterization showed that the homogeneous films were deposited with amorphous structure. The surface roughness of doped films showed smother surface as compared to CuS non-doped thin films. The absorption measurement (UV-Vis) of the deposited CuS films was used to estimate the optical band gap. The energy band gap values are in the range from 2.4 to 1.9 eV depending on doping concentration and chalcogenide type nanoparticles. The electrical conductivity changed five orders of magnitude from 10 (Ωcm)-1 for CuS films to 5.0x106 (Ωcm)-1 CuS nanoparticle doped film. Nanoparticle ZnS doped thin film varied four orders of magnitude 4.0x106 (Ωcm)-1 compared to non-doped films. The outstanding results obtained in the doping of thin films were obtained by low cost and easy soft chemistry techniques: sol gel and chemical bath deposition. These advances on the doping of CuS films make them suitable candidates for the development of semiconductor devices such as photovoltaic cells, transistors and flexible electronics.
9:00 PM - ES14.14.35
Fabrication of Thin Film Solar Devices Using Emergent Materials
Hector Padron Perez 1 , Harumi Moreno Garcia 1
1 , Universidad Autonoma de San Luis Potosi, San Luis Potosí, FDM, Mexico
Show AbstractRecent studies show that photovoltaic (PV) panels made from thin films of Cu(InGa)Se2 and CdTe yield efficiency (h) of 21.7% and 22.0% respectively 1,2, however the scarcity and price of In and Ga involves a high production cost. The use of emerging materials is proposed, they should present values of band gap of 1 to 1.7 eV, being also abundant materials and that PV can be produce by more attractive energy processes
Sb2(S,Se)3 thin films were deposited to the manufacture of solar cells by spin-coating method that is a non-toxic technique, inexpensive and simple vacuum, where a hybrid ink has been used as a precursor, which was generated from nanoparticles of Sb2(S,Se)3.
The project includes the synthesis of nanoparticles and thin films, their characterization and analysis by Raman, X-ray diffraction (XRD), scanning electron microscopy (SEM), atomic force microscopy (AFM). Preliminary results indicate that the nanoparticles corresponds to Sb2(S,Se)3, also the RAMAN modes taken in the measuring corresponds to the material. In the optical analysis it is determined that thin films have a Eg value of 1.6 eV. In the theoretical analysis is determined that solar cells made with Sb2(S,Se)3 may achieve a conversion efficiency (n%) of 39%. It has been developed solar cells obtaining a photovoltaic effect.
1 : T.M. Friedlmeier, et al., High efficiency Cu(In,Ga)Se2 solar cells, Thin Solid Films (2016), http://dx.doi.org/10.1016/ j.tsf.2016.08.021
2 Green MA, Emery K, Hishikawa Y, Warta W, Dunlop ED. Solar cell efficiency tables (version 47). Progress in Photovoltaics: Research and Applications 2015; 1-9
9:00 PM - ES14.14.36
Plasmon Enhanced Photovoltaic Effect in Graphene/MoxW1-xSe2/Pd Vertical Schottky Diodes
Sung Hyun Kim 1 , Sum-Gyun Yi 1 , Sungjin Park 1 , Kyung-Hwa Yoo 1
1 Department of Physics, Yonsei University, Seoul Korea (the Republic of)
Show AbstractWe have developed graphene/MoxW1-xSe2/Pd vertical devices that exhibit diode-like rectification behaviors in dark. Upon illumination, a photovoltaic effect is observed because of the built-in potential caused by the difference in work function between graphene and Pd. However, the photovoltaic efficiency is not high enough due to poor light absorption nature of ultra-thin materials. To improve the photovoltaic effect, Au nanodisks with a diameter of 70 nm are made on graphene/MoxW1-xSe2/Pd by electron-beam lithography and lift-off techniques because Au nanodisks which have diameter below 100 nm are plasmonic resonant to visible light. The plasmonic structure provides hot electrons and near field enhancement, resulting in an increase in photovoltaic efficiency of graphene/MoxW1-xSe2/Pd vertical Schottky diodes.
Symposium Organizers
Ingrid Repins, National Renewable Energy Laboratory
Shubhra Bansal, University of Nevada, Las Vegas
Sascha Sadewasser, International Iberian Nanotechnology Laboratory
Edgardo Saucedo, IREC
Symposium Support
Catalonia Institute for Energy Research (IREC)
Dr. Eberl MBE-Komponenten GmbH
First Solar
International Iberian Nanotechnology Laboratory
National Renewable Energy Laboratory
ES14.15: Novel Chalcogenide Absorber Materials
Session Chairs
Gilles Dennler
Joop van Deelen
Friday AM, April 21, 2017
PCC North, 200 Level, Room 229 B
9:15 AM - ES14.15.01
Potential Resolution to the “Doping Puzzle” in Pyrite FeS2: Carrier Type Determination by Hall Effect and Thermopower
Xin Zhang 1 , Mengqun Li 1 , Jeffery Walter 1 , Liam O'Brien 1 2 , Mike Manno 1 , Frazier Mork 1 , James Kakalios 1 , Eray Aydil 1 , Chris Leighton 1
1 , University of Minnesota, Minneapolis, Minnesota, United States, 2 , University of Cambridge, Cambridge United Kingdom
Show AbstractPyrite FeS2 has outstanding potential as a photovoltaic based on earth-abundant, low-cost, non-toxic constituents, but underperforms dramatically in solar cells. While the full reasons for this are not clear, one important factor is the historical inability to understand and control doping in FeS2. This is exemplified by the widely accepted notion that unintentionally doped single crystals are predominantly n-type, whereas thin films are p-type. Here we provide a potential resolution to this “doping puzzle”, arrived at via Hall effect, thermopower, and resistivity measurements on a large set of FeS2 single crystals and films, spanning five orders of magnitude in mobility. The results reveal three main findings: (i) The highest mobility crystals and films studied in this work are definitively n-type, from Hall and thermopower; (ii) As mobility decreases an apparent crossover to p-type occurs, first in thermopower and then in Hall. This can be understood, however, within a simple picture for the crossover from diffusive to hopping transport clearly evidenced in variable temperature measurements. Apparent p-type behavior can thus be an artifact of hopping, and the prevailing notion of predominantly p-type films should be revised; (iii) Notably universal behavior is found for both crystals and films, suggesting a common dopant, potentially S vacancies. These findings have deep implications for the interpretation of prior work on FeS2 solar cells, and for the design of future studies.
Work supported by the NSF under DMR-1309642, the University of Minnesota NSF MRSEC under DMR-1420013, and the Xcel Energy Renewables Development Fund.
9:30 AM - ES14.15.02
Binary and Ternary Sb-Based Chalcogenide Thin-Film Solar Cells
Liang Wang 1 , Chao Chen 1 , Xinsheng Liu 1 , Bo Yang 1 , Haisheng Song 1 , Jiang Tang 1
1 , Huazhong University of Science and Technology, Wuhan China
Show AbstractExploring low toxicity and low cost absorber materials for highly efficient solar cells receives continuous attention in photovoltaic research. Copper zinc tin sulfoselenide (CZTSSe) is a successful outcome of this long-lasting pursuit. Here we argue that antimony-based binary and ternary chalcogenide including Sb2S3, Sb2Se3, CuSbS2 and CuSbSe2 are promising materials for thin film solar cells. The constituents of these materials are low-cost (Sb has similar price to Cu), relatively earth-abundant (Sb is 5 times abundant as In in the earth crust) and low-toxicity (none of these materials are listed as highly-toxic or carcinogenic by China, Europe or U. S. regulation offices). Furthermore, they demonstrate appropriate band gaps (Sb2S3: 1.68 eV, Sb2Se3: 1.17 eV, CuSbS2: 1.40 eV, CuSbSe2: 1.04 eV) for single junction solar cells, and also show strong absorption coefficient above 105 cm-1 due to the involvement of p-orbital originated from Sb3+ in the optical transition. In this talk we would like to review the recent progress in using these compounds as absorber materials for thin film photovoltaics. Special emphasis will be focused on findings in our group, which include: 1) hydrazine-solution and aqueous solution processed Sb2Se3, CuSbS2 and CuSbSe2 solar cells1-3. The solution chemistry, as well as device performance will be discussed. 2) thermally evaporated Sb2Se3 solar cells with best device efficiency of 5.8%. The influence of oxygen, sulfur and selenium on the film defects, carrier dynamics and device performance will be discussed4,5. 3) superstrate Sb2S3 and Sb2Se3 device produced by rapid thermal evaporation (RTE). The advantages of this fast fabrication strategy, and the key parameters that dictated device performance, and the stability study will be disclosed in this part6,7. Other efforts including Cd-free buffer layers and fundamental parameters of Sb2Se3 and CuSbSe2 will also be included. With its attractive material properties, easy manufacturing, demonstrated efficiency (6.5% certified for Sb2Se3 and 4.6% reported for CuSbSe28) and encouraging device stability, we believe these materials could be potential alternative to CdTe and CuSbSe2 and therefore worth further research efforts.
Reference:
1. J. Tang et. al. Adv. Energy Mater., 2014, 4, 1301846
2. J. Tang et. al. Adv. Energy Mater. 2015, 5, 1501203.
3. J. Tang et. al. Chem. Mater., 2015, 27, 8048.
4. J. Tang et. al. Prog. Photovolt., 2015, 23, 1828-1836.
5. J. Tang et. al. Prog. Photovolt., 2016, under review.
6. J. Tang et. al. Nature Photon., 2015, 9, 409.
7. J. Tang et. al. Sol. Energ. Mater. Sol. C., 2016, 157, 887.
8. A. Zakutayev et. al. Adv. Energy Mater. submitted.
9:45 AM - ES14.15.03
Should KCN Etching be Performed when Cu2Sn1-xGexS3-Based Thin Films are Processed into Solar Cells?
Erika Robert 1 , Diego Colombara 1 , Jessica de Wild 1 , Alex Redinger 2 , Phillip Dale 1
1 , University of Luxembourg, Belvaux, Luxembourg, Luxembourg, 2 Structure and Dynamics of Energy Materials, Helmholtz-Zentrum Berlin für Materialien und Energie, Berlin Germany
Show AbstractWith 6.7% efficiency Cu2Sn1-xGexS3 (CTGS) [1] is an emerging absorber layer of high interest for photovoltaic applications. Ge/Sn alloying allows to tune the bandgap between 0.93 eV (x=0) [2] and 1.57 eV (x=1) [3], which is beneficial for power conversion efficiency.
In this study, CTGS films were produced by annealing sputtered Cu-Sn precursor stacks in the presence of GeS vapor. Grazing incidence XRD shows that the monoclinic polymorph is formed, but also a shift of the CTGS peaks to lower angles with increasing incidence angle. This indicates that the Ge amount is decreasing towards the Mo back contact, lowering the band gap. EDX measurements near the surface gave a Ge/(Sn+Ge) ratio of 0.87, corresponding to a bandgap of 1.5 eV assuming that there is a linear relationship between high Ge/(Sn+Ge) ratio and bandgap [3]. Additionally, CuxS secondary phases have been identified by GI-XRD, which can be removed with KCN etching. Aqueous KCN etching is extensively investigated for CIGSe and known to improve the device performance. However, KCN etching of the CTGS absorbers before device completion (Mo/CTGS/CdS/Al:ZnO/i:ZnO/Ni:Al grids) led to lower device device efficiencies (< 1%) than devices made without KCN etching. To further investigate the etching process, absorbers were analysed photoelectrochemically with and without KCN etching, in order to study the charge carrier generation and transport, which can be done without buffer and window layers in a Shottky diode configuration. KCN etching was found to reduce the dark current in reverse bias. Surprisingly, compared to what is known for CuInGaSe2, KCN etching also reduced the photocurrent and hence does not seem to benefit the final device performance.
Based upon this evidence, new solar cell devices were built from absorbers which did not undergo the KCN etching, leading to device efficiencies between 1 and 1.5% with open circuit voltages above 200 mV. External quantum efficiency (EQE) measurement on those devices gives an extrapolated bandgap of 1.4 eV, lower than what was expected from the EDX. We attribute this to the gradient in the absorber layer that has a lower band gap at the back. With the EQE the lowest band gap is determined. Additionally, the EQE does not show the double optical transition in the EQE, as observed usually for this material [3], which is interpreted as the presence of different CTGS band gaps, i.e the slope of the EQE is a superposition of the several bandgaps in the material and hence the double transition smoothens out.
Further process modifications allowed us to reach efficiencies around 3% on 0.5 cm2 surface areas, revealing the potential of CTGS and influence of etching process prior to device preparation.
[1] M. Umehara et al, Applied Physics Express, 9 (2016) 072301.
[2] J. de Wild et al, Solar Energy Materials and Solar Cells, 157 (2016) 259-265.
[3] E.V.C. Robert et al, Proc. SPIE 9936 , Thin Films for Solar and Energy Technology VIII (2016) 993607.
10:00 AM - ES14.15.04
Metastable Cubic SnS, SnSe and Their Alloys as Photovoltaic Materials
Stephan Lany 1 , Aaron Holder 1 , Rachel Kurchin 2 , Enue Barrios Salgado 3 , M.T. Santhamma Nair 3 , P.Karunakaran Nair 3
1 , National Renewable Energy Laboratory, Golden, Colorado, United States, 2 , Massachusetts Institute of Technology, Cambridge, Massachusetts, United States, 3 , Universidad Nacional Autónoma de México, Mexico City Mexico
Show AbstractThe observation of cubic polymorphs of SnS in nanostructures and thin-film form have created interest, since a zinc-blende type structure would have seamlessly fit into the paradigm of tetrahedrally coordinated semiconductors, a materials class encompassing Si, GaAs, CdTe, CIGS, and CZTS. However, the structure assignment as cubic zinc-blende or rocksalt phases proved wrong, and the cubic SnS phase was refined as an unusually large 64 atom simple cubic (SC) cell [1] with spacegroup (SG) 198. Indeed, first principles calculations show that the zinc-blende and rocksalt structures are dynamically unstable, whereas the SG 198 is a metastable low energy structure, only 6 meV/at higher in energy than the simple orthorhombic (ORC) ground state. SnSe can assume a similar structure with 13 meV/at above the ground state. We report a computational characterization of simple cubic SnS and SnSe, including electronic structure, defects, and the phase diagram for cubic SnS-SnSe alloys. We compare the computational predictions with recent experimental work on synthesis and characterization of cubic SnS, SnSe, and their alloys [2]. Compared to the ORC ground state, the SC phases have a larger and direct band gap with a sharper absorption onset. In particular, cubic SnSe or Se rich SnSSe alloys could prove very interesting materials for photovolatics.
[1] A. Rabkin, S. Samuha, R.E. Abutbul, V. Ezersky, L. Meshi, and Yuval Golan, Nano Lett. 15, 2174 (2015).
[2] E. Barrios-Salgado, L.A. Rodríguez-Guadarrama, A.R. Garcia-Angelmo, J. Campos Álvarez, M.T.S. Nair, and P.K. Nair, Thin Solid Films 615, 415 (2016).
10:15 AM - ES14.15.05
Cu2BaSn(S,Se)4—Earth-Abundant Chalcogenides for Solar Energy Conversion Applications
Donghyeop Shin 1 , Tong Zhu 1 , Jon-Paul Sun 2 , William Huhn 1 , Xuan Huang 1 , Edgard Ngaboyamahina 1 , Yihao Zhou 1 , Ian Hill 2 , Volker Blum 1 , Jeffrey Glass 1 , David Mitzi 1
1 , Duke University, Durham, North Carolina, United States, 2 , Dalhousie University, Halifax, Nova Scotia, Canada
Show AbstractChalcogenides such as CdTe, Cu(In,Ga)(S,Se)2 (CIGSSe), and Cu2ZnSn(S,Se)4 (CZTSSe) have enabled remarkable progresses in thin-film photovoltaic performance, but concerns remain regarding (i) the toxicity (CdTe), (ii) scarcity (CIGSSe/CdTe) of the constituent elements, and (iii) the unavoidable anti-site disordering and band tailing that limit further efficiency improvement (CZTSSe). Recently, Cu2BaSnS4-xSex materials have been proposed as an emerging PV absorber to address these issues, while using earth-abundant metals with an ionic size mismatch to reduce anti-site disordering.1 These compounds exhibit a tunable band gap in the range of 1.5~2.0 eV, spanning relevant values for single- or multiple-junction photovoltaic or photocatalytic applications. A relatively sharp cutoff in external quantum efficiency and sharp photoluminescence feature support the notion that band tailing may not be as much of an issue in this material systems relative to CZTSSe. As a proof of concept, pure sulfide (Cu2BaSnS4)-based PV devices have been successfully demonstrated, yielding a power conversion efficiency of 1.62% with Voc= 713 mV, Jsc= 4.11 mA/cm2, and FF= 55.32.1 Given that the pure sulfide has a relatively large band gap (2.0 eV), a feasible route to enhance the device performance is to reduce the band gap by incorporating Se. In this work, we have developed processes to produce high-quality Cu2BaSnS4-xSex films with smooth surface and large grains. Using structural, optical, and electrical analysis tools, various properties of Cu2BaSnS4-xSex materials have been investigated. Furthermore, the obtained knowledge can be used to optimize device architectures for photovoltaic and photoelectrochemical applications and enhance performances of Cu2BaSnS4-xSex-based devices.
1D. Shin, B. Saparove, T. Zhu, W. P. Huhn, V. Blum, and D. B. Mitzi, Chem. Mater. 2016, 28, 4771
10:30 AM - ES14.15.06
Reduced Secondary Phase Formation and Doping in Cu2SnS3 by Alloying with Ag to be Used for Solar Cell Applications
Jessica de Wild 1 , Erika Robert 1 , Alex Redinger 2 , Phillip Dale 1
1 , Luxembourg Univ, Belvaux Luxembourg, 2 Department Structure and Dynamics of Energy Materials, Helmholtz-Zentrum Berlin für Materialien und Energie, Berlin Germany
Show AbstractCu2SnS3 (CTS) is a new p-type semiconductor used as an absorber layer in solar cells of which the monoclinic polymorph shows the best device properties [1]. Best device power conversion efficiency of 4.6% is obtained with absorber layers made from evaporated NaF/Cu/Sn precursors, showing the potential of this material [2]. However, CTS seems too highly doped for a standard p/n+ junction configuration [3]. Additionally, a detrimental secondary phase NaxCuSnS3 was found [1]. We show that both problems are tackled by doping CTS with Ag.
Silver replaces Cu in the CTS lattice making it less p-type, since the compound Ag2SnS3 is n-type [4]. A slight increase in band gap can be expected as well due to lower shifting of the valance band. CTS layers doped with Ag (CATS) were made on glass and Mo. Precursors were deposited on glass via wet chemical technique [1] with Ag added separately by ink. Additionally Cu/Sn precursors were sputtered on Mo and Ag deposited via ion beam evaporation. All precursors are annealed at 550 °C in a S and SnS atmosphere. The wet chemical precursors have an initial Ag/Cu ratio varying from 0.4 to 1.2. However, it seems that only a small amount of Ag is incorporated into the CTS lattice, since most silver forms Ag2S. Additionally, negligible change in band gap was measured. Hence, less silver can be used. With ion beam evaporation smaller amounts of Ag can be deposited than by ink and several thicknesses of silver are deposited corresponding to concentrations between 0.1 and 2 at%. XRD shows no signs of Ag2S for Ag between 0.1 and 1 at%, while for higher Ag concentrations Ag2S forms again. None of the samples has the NaxCuSnS3 phase found in pure CTS [1].
Optical properties were measured with photoluminescence (PL). The PL peak maximum of the wet chemical CATS layers show a blue shift of 7 meV, independent on the initial Ag/Cu ratio of the precursor. This follows the finding that Ag uptake into the CTS host is limited and too much Ag results in Ag2S. The PL yield of CATS layers made with ion beam evaporated Ag is always higher than pure CTS, which steadily increases up to 4 times for the Ag doping of 0.7 at%. No shift in PL peak position was measured for the ion beam samples.
Solar cells were made from the wet chemical CATS layers with the Mo/CATS/CdS/iZnO/AZO configuration with Ni/Al grids and external quantum efficiency was measured, to subtract the depletion width as shown in ref [3]. Fitting the EQE indicates that the depletion width is larger than any standard CTS solar cell i.e 200 nm vs 90 nm, which is attributed to the decreased doping. These results show the potential of doping CTS with only very small amount of Ag to reduce the p-type doping.
[1] J. de Wild, et al. Sol. Energy Mater. Sol. Cells, 157, 259 (2016)
[2] M. Nakashima, et al. Applied Physics Express 8, 042303 (2015)
[3] J. de Wild et al. IEEE J. of Photo. 10.1109/JPHOTOV.2016.2612359 (2016)
[4] A. Fedorchuk, et al. Mater. Chem. and Phys. 135, 249 (2012)
ES14.16: Kesterite Absorber Growth and Devices
Session Chairs
Raquel Caballero
Katharina Rudisch
Friday PM, April 21, 2017
PCC North, 200 Level, Room 229 B
11:15 AM - *ES14.16.01
11.5% CZTSSe Devices Spray Coated from a Water-Ethanol Ink—Current Limitations and Ways Forward
Gilles Dennler 1
1 , IMRA, Sophia Antipolis, MA, France
Show AbstractOur spraying process of water-ethanol based CZTS colloid followed by an annealing in Selenium vapor allowed us recently to reach efficiencies up to 11.5%. This achievement was made possible by a thorough optimization of the stoichiometry of the active layer as well as a proper management of the incoming light. Furthermore, this industrially applicable and easily up-scalable process allowed us to produce 1.1 cm2 devices certified at 10%, which entered recently the Table I of the Efficiency Tables published by Progress in Photovoltaics. Besides, preliminary lifetime tests indicated that these solar cells may succeed to pass the standard accelerated lifetime tests without major difficulties: Indeed, non-encapsulated cells left under simulated AM1.5G irradiation and open circuit voltage condition for more than 5000 hours were observed to lose only about 10% of their initial efficiency.
In spite of the fact that these achievements appear encouraging, especially in view of the simplicity of our process, these devices still suffer from major losses that reduce severely their photovoltaic performances. During this talk, we will present our latest results, discuss various potential source of losses like the anionic and cationic disorders, and propose some ways forward.
11:45 AM - ES14.16.02
Cu2ZnSnS4 Solar Cell with over 700 mV Open Circuit Voltage from High Sulfur Partial Pressure during the Annealing Process
Yi Ren 1 , Nils Ross 1 , Jes Larsen 1 , Katharina Rudisch 1 , Jonathan Scragg 1 , Charlotte Platzer-Bjorkman 1
1 Solid State Electronics, Uppsala University, Uppsala Sweden
Show AbstractHigh temperature annealing is a critical step in the fabrication of Cu2ZnSnS4 (CZTS) based solar cells. In this process, it is difficult to control the sulfur partial pressure (PS2) under the non-equilibrium condition. Here we tackled this issue by characterization of a monitoring material placed together with the CZTS films in the same annealing process. The examined process includes annealing in a semi-closed graphite box with a sulfur excess environment at approximately 580 °C under static 465 mbar Ar. We discovered that the PS2 declined with increasing annealing time. The PS2 began to decrease after about 3min annealing and could possibly be lower than 0.1 mbar (the decomposition PS2 of CZTS) after 13min annealing. Furthermore, the estimated PS2 can be nicely correlated to variations in the CZTS material quality. The results indicate that the PS2, rather than the dwell time, is in fact the most essential parameter during annealing.
In this work, identical Cu-Zn-Sn-S precursors were prepared from co-sputtering of CuS, ZnS and SnS targets, and the resulting compositions of the precursors were Cu/Sn ~1.88 and Zn/Sn ~1.02. XRD and multi-wavelength Raman scattering indicated the successive evolution of ZnS and Sn-S secondary phases as the annealing proceeded. In addition, it could be inferred that a gradual modification of the dominant defect type in CZTS occurred with the increased annealing time. We deduced that the possible evolution of defects, besides the Cu/Zn cation disorder, could start from the mixed A-type (VCu+ZnCu) and B-type (2ZnCu+ZnSn) defects to the pure B-type defect. This is on basis that: 1) EDS revealed alterations in the local composition of the annealed CZTS film even though the overall composition of all CZTS films was unchanged according to XRF. 2) Different secondary phases started to form as annealing advanced. 3) Near-resonant Raman showed that the value of the ordering parameter (the intensity ratio of the peaks at 287 cm-1 and 304 cm-1) was initially constant but started to reduce with increased annealing time. 4) The PL peak position was decreased from 1.39 eV to 1.30 eV as the annealing time continued from 1 min to 40 min. The defect variation in CZTS as function of the annealing time became more apparent after adopting an optimized 35h slow cooling process to reduce the Cu/Zn disorder.
Lastly, it was demonstrated that high PS2, i.e. 1 min annealing, could improve the Voc of the solar cell above 700 mV with a non-optimized CdS buffer layer. Several repeated experiments confirmed that the Voc of the solar cell is between 700 and 720 mV, with the overall device performance comparable to our references (~ 6% efficiency). The Voc can be further improved to 783 mV with the optimized slow cooling. It is important to note from this work how precise control of the PS2 is linked to bulk defects in CZTS and the Voc of CZTS solar cells.
12:00 PM - ES14.16.03
Thermal Annealing Effect on Layer Morphology and Performance of Kesterite Solar Cells
Samira Khelifi 1 , Maria Batuk 2 , Leo Choubrac 3 , Joke Hadermann 2 , Nicolas Barreau 3 , Bart Vermang 4 , Guy Brammertz 5 , Marc Meuris 5 , Jeroen Beeckman 1 , Johan Lauwaert 1 , Kristiaan Neyts 1
1 ELIS, Gent University, Gent Belgium, 2 EMAT, University of Antwerp, Antwerp Belgium, 3 CNRS, Institut des matériaux Jean Rouxel, Université de Nantes, Nantes France, 4 , imec – Partner in Solliance, Leuven Belgium, 5 , imec - Division IMOMEC, Hasselt Belgium
Show AbstractIn this work, the effect of low temperature post-annealing of complete CZTSe/CdS solar cells has been extensively investigated using electrical characterization techniques: current-voltage, capacitance-voltage and EQE. The samples were annealed at temperatures ~ 200 °C under air atmosphere, and the annealing time was varied between 5 min to 14 hours. An increase of the efficiency from 2 % to 8 % was reported after the annealing, mainly due to an improvement in the short-circuit current and the open-circuit voltage. In order to find a correlation between the performance and morphology of the solar cells, detailed transmission electron microscopy analysis was performed to characterize the microstructure of the solar cells.
The capacitance-voltage measurements show a slight increase in the doping profile in accordance with the increase reported in the size of some CZTSe grains after the annealing. From the orientation mapping the grains have close orientation (<110> orientation dominates), while after annealing the grains have more random orientation.
Segregation of CuxS secondary phases were detected at the grain boundaries of the CZTSe before thermal annealing and appeared as inclusions after annealing. Sn-rich phases were detected within the absorber and also in the CdS layer.
12:15 PM - ES14.16.04
High Voltage CZTS,Se Devices with High Efficiencies
Richard Haight 1 , Priscilla Antunez 1 , Douglas Bishop 1 , Yu Luo 1 , Suarabh Singh 1 , Teodor Todorov 1 , James Hannon 1
1 , IBM T.J. Watson Research Ctr, Yorktown Heights, New York, United States
Show AbstractIncreasing the power conversion efficiency of CZTS,Se based devices has been the primary focus of research in earth abundant thin film photovoltaics. But the ability to tune the band gap of CZTS,Se by modulating the concentration of S and in turn adjusting the open circuit voltage Voc is a powerful property of this material. Careful choice and placement of the amount of S utilized during the brief high temperature “hard bake” anneal of solution deposited CZTS,Se films has resulted in both high voltage and high efficiency devices. In particular we have fabricated devices with a Voc of 550 mV and 12% efficiency. Even higher voltage devices with Voc in the range of 650-670 mV with device efficiencies as high as 11.5% were fabricated. This was achieved through the deposition of a unique back contact stack on standard high efficiency devices suitably modified. This modification involves the exfoliation of an active high efficiency device from the Mo/soda-lime-glass substrate followed by the deposition of a back contact stack consisting of a layer of MoO3 and Au. This process resulted in devices exhibiting increased Voc, fill factor, short circuit current and efficiency relative to pre-exfoliated devices. Exfoliation followed by back contact deposition is essentially a room temperature process that side-steps concerns of contact degradation that occurs following typical high temperature anneals. Details of this process will be described. In addition to the fabrication of these high voltage, high performing devices, series connection of multiple devices in a monolithic array has produced voltages of >5.5V and Jsc~25mA/cm2 under 1 sun illumination and 1.7V under typical office light conditions of 10-3 suns. This performance indicates that CZTS,Se based devices can be utilized for energy harvesting even under low light conditions for powering sensors, microprocessors and charging batteries, key elements to autonomous devices that might be deployed for “Internet-of-things” applications.
12:30 PM - ES14.16.05
Cu2ZnSnSe4 Solar Cell with 11.5% Efficiency Achieved by Sputtering from Quaternary Target
Rujun Sun 1 , Daming Zhuang 1
1 Materials Science and Engineering, Tsinghua University, Beijing China
Show AbstractUp to now, the record efficiency of CZTSSe solar cells has reached 12.6%, which is still far low than 22.6% of CIGS solar cells. The efficiency of CZTSSe solar cell is sensitive to composition of absorber, thus the composition should be strictly controlled in Cu-poor and Zn-rich region. The typical fabrication of absorber consists of the deposition of precursor, followed by heat treatment. During heat treatment, SnS is one of intermediate phases before that CZTSSe is formed. SnS is volatile and possesses high saturated vapor pressure at typical annealing temperature. Hence, the actual composition of absorber would differ from the designed one, thus greatly affecting the performance of solar cell. This deviation in composition usually varies in different processing condition, leading to poor repeatability and stability of production. On the other hand, CZTS is predicted to decompose to form Cu2-xS and ZnS in the absence of S and SnS in the gas phase. This suggests that CZTS is in fact in equilibrium with two gas phase species. Attention must be taken to avoid significant loss of Sn. There are two approaches to avoid substantial tin loss. One is to supply or compensate for the loss of volatile species of SnS and S by adding excess sulfur and tin sulfide powders inside chamber. When sufficient sulfur partial pressure is applied, a thick MoS2 interface layer is consequently formed, leading to high series resistance. The other is to use high background pressure. This high pressure slows the movement of volatile species, thus reducing rate of loss and increasing the partial pressure of volatile species above the film.
Annealing for short time is also be adopted. Though short time anneal is effective to reduce the loss of tin, a lot of defects caused by insufficient grain growth will form in the absorber. Etching off secondary phases before CdS deposition is alternative to eliminate the harmful impact of resulting secondary phases. KCN, which etches Cu2-xS, is highly toxic and acid solution to etch ZnS is corrosive. Moreover, additional procedure will increase cost.
In this work, the loss of tin is successfully suppressed by supplying H2Se prior to CZTS decomposition, to convert CZTS precursor to CZTSe absorber. Firstly, CZTS precursor by sputtering from quaternary is adopted to avoid the formation of intermediate phase SnS. Secondly, the temperature Td that decomposition of CZTS is accelerated by significant volatilization of gaseous SnS, is determined by annealing CZTS precursor under Ar atmosphere. Thirdly, the possibility of converting CZTS into CZTSe below this temperature Td is confirmed. This enable tin loss to be avoided in subsequent heat treatment. Finally, a two-step hearting regime is used to restrain the excess growth of CZTSe grain. Accordingly, a CZTSe solar cell with efficiency of 11.5 % has been achieved using this strategy of selenization.
12:45 PM - ES14.16.06
Cu2ZnSnS4 Thin Films from a Single Precursor Solution—Effect of Na and Sb Doping on Device Performance
David Fermin 1 , Devendra Tiwari 1 , Rainer Klenk 2
1 , University of Bristol, Bristol United Kingdom, 2 , HZB, Berlin Germany
Show AbstractLow-cost solution processing of active layers can generate a profound impact on scalability of thin-film PV technology. However, controlling composition and phase purity of multicomponent chalcogenides remain a significant challenge. In this contribution, we explore a new route for depositing of Cu2ZnSnS4 (CZTS) thin film via a single molecular precursor solution, focusing on the effect of dopants such as Sb and Na on device performance. Homogeneous CZTS films with thickness of 1.2 mm are prepared by spin-coating of a single precursor solution containing metal chloride salts and thiourea onto Mo coated glass, followed by annealing at 560 °C. Quantitative analysis of X-ray diffraction and Raman spectroscopy reveal the nucleation of highly pure kesterite phase, with Na and Sb dopants significantly improving film crystallinity. These dopants also promote higher photoluminescence yield and narrowing of the band-to-band peak. Thin-film devices were investigated with the configuration Mo/CZTS/CdS/i-ZnO/Al:ZnO/Ni-Al with a 0.5 cm2 active area. Analysis of over 70 cells for each of the precursor composition show that introduction of Na and Sb leads to an increase of the average power conversion efficiency from 3.2±0.5 to 5.2±0.3%. The best cell was obtained upon Na and Sb doping, featuring 14.9 mA cm-2 short-circuit current, 610 mV open circuit voltage and 63% fill factor under AM 1.5 illumination (5.72% power conversion efficiency). We rationalise the increase in cell performance in terms of the formation of alkali antimony chalcogenides fluxes during the annealing step, while a decrease in the structural disorder is brought about by the interactions of Na and Sb with specific lattice sites.