Grain boundaries (GBs) in semiconductor solar cells are conventionally considered detrimental to cell performance because they create deep gap states that act as effective nonradiative electron-hole recombination centers. This has been proven in many conventional solar cells, such as GaAs and Si. However, other thin-film solar cell absorber, such as CdTe and Cu(In,Ga)Se2 (CIGS), have exhibited greater efficiency on polycrystalline phase than that on monocrystalline and the GBs in those semiconductors are believed to exhibit unusual physics. The physics of GBs in CdTe and CIGS has attracted great attention. However, owing to the complexity of CIS GB, the origin of benign behavior has not been sufficiently understood. Moreover, with recent development, the efficiency of earth-abundant Cu2ZnSnSe4 (CZTSe)-based solar cells has achieved 11.1% in 2012. However, this efficiency is still much lower than the record efficiencies of other thin-film solar cells, such as 18.3% for CdTe and 20.3% for CIGS. It is unclear whether GBs in CZTSe will behave similarly as GBs in CIGS and CdTe. If GBs were indeed detrimental in CZTSe, would there be any methods to engineer GBs in CZTSe to make it benign?

In this presentation, we will present our results of density functional theory (DFT) study on the GB properties of CdTe, CIS and CZTSe. The proper structural models for GBs are constructed. We have found that the intrinsic GBs for all those semiconductors are detrimental, probably due to structural similarity. The wrong bonds at intrinsic GBs result in deep defect states in the band gap, indicating that intrinsic GBs act as the Shockley-Read-Hall recombination centers and thus reduce both the open circuit voltage (Voc) and short circuit current (Jsc). However, the intrinsic defects or some extrinsic impurities have the tendency to spontaneously segregate around the grain boundaries. For example, the Cl/Cu (Na/O) has much lower formation energies at GBs of CdTe (CIS and CZTS) than those in the bulk. The defect segregations have two major effects: (1) passivating the deep defect states in the band gap by breaking or weakening the wrong bonds at GBs and (2) creating neutral hole barrier. The existence of Na+i further induces the band bending and increases the hole barrier at CIS and CZTS.
In conclusion, we used first-principle calculations to extensively study the GB properties of thin-film solar cell, including CdTe, CIS and CZTSe. We unraveled the origin of benign GB properties at CdTe and CIS and proposed the defect engineering methods for improving the performance of CZTSe-based solar cell.

Defects structure can be readily obtained from ab initio calculation, such as DFT, regardless of whether one is considering point defects or extended defects. Most of such calculations deal with isolated point defects and its is well known and documented that these have important impact on semiconductors electronic properties, and hence on their functionality in photovoltaic devices. The presentation will discuss the important case where defects are not isolated, but rather interact and organize into 1D, 2D or 3D structures. To be more specific, examples will be taken from various fields of thin film PV, e.g. Si, chalcogenides and III -V.

In the search for clean, sustainable energy sources, organic photovoltaics (OPVs) have emerged as a promising technology. These devices represent a lightweight, flexible, low-cost and environmentally benign power source, which has great potential in many applications. However, before large-scale commercialization is possible, these devices must first be efficient, reproducible and stable solar converters. In any microelectronic device, if such aspects are to be optimized, fundamental physical parameters must be well understood. One such prominent parameter is material defects and the energetic trap states they create, which are well-known to plague amorphous or otherwise impure semiconducting materials. Organic semiconductors are no strangers to such states and their electronic properties, reproducibility and stability are evidently tied to these defects. Thereby, the identification, characterization and eventual mitigation of these states remain important areas of interest for OPV devices. Herein, we discuss our work on the former two - identification and characterization - presenting our recent advancements on the identification of deeper, unknown defects in polymer solar cells as well as the characterization of the prefactor of trap emission in these same cells. In the former, low frequency (<1Hz) capacitance measurements were undertaken with the models of Walter et al. as well as Cohen and Lang to probe previously uncharted deep defect levels in polymer based bulk heterojunction cells. In the latter, capacitance versus frequency measurements were repeated as a function temperature to observe the Arrhenius behavior of trap emission and characterize the kinetics of trap occupancy. The prefactor of trap emission - commonly referred to as the ‘attempt to escape frequency&’ - for several polymers is discussed. The presented methods are generally applicable to doped organic materials, helping to guide future works in the identification and characterization of electroactive defects in OPV devices.

CuInSe2 (CIS) and Cu(In,Ga)Se2 (CIGS) are chalcopyrite-type semiconductor and one of the most promising materials for low cost photovoltaic solar-cell. Due to the self-regeneration mechanism [1], it is known that the processing of CIS is rather easy compared to Si-based photovoltaic solar cells. It is also known that the electrically benign nature of this material allows large off-stoichiometry, which leads a series of ordered defect compounds (ODCs) [1].
Recently, Yan et al., observed inhomogeneous distribution of Ga in CIGS by using TEM [2]. In this paper, based on first-principles calculations, we show that the spinodal nano-decomposition causes inhomogeneity in CIGS and will enhance the conversion efficiency in CIGS. Our calculations are based on the KKR-CPA-LDA [3] with the self-interaction correction [4]. From the calculated mixing energy of CIGS, it is found that the system favors the spinodal decomposition. We also perform Monte Carlo simulations and find that quasi-one-dimensional nano-structures with high concentration of impurities are formed under the layer-by-layer crystal growth condition in CIGS [5]. It is expected that the photo-generated electron-hole pairs are efficiently separated by the type-2 interface and then effectively transferred along the quasi-one-dimensional structures in CIGS.
Next, we apply the same idea to CIS and show that ODC is also useful to enhance the efficiency of CIS based photovoltaic materials. Due to the compensation, defect pair of InCu and two Cu vacancies (2VCu) is very much stable in CIS. According to our calculation of the mixing energy delta-E(x) = E(Cu1-3x/2VxInx/2InSe2) - [(1-3x/2)E(CuInSe2) - (3x/2)E(In2Se3)], where x is the concentration of Cu vacancy and E is the total energy of each system, this system favors the phase separation to CIS and In2Se3 (which is the end member of ODCs). Due to the phase separation, interfaces between CIS and In2Se3 are generated spontaneously. Due to the suppression of hybridization between Cu-d states and Se-p states, energy gap of ODC, Eg(ODC), is larger than Eg(CIS). Since, the predicted band alignment between CIS and In2Se3 is type 2, we can expect efficient electron-hole separation at the interfaces between CIS and In2Se3.
References:
[1] S. B. Zhang et al., PRB57 (1998) 9642.
[2] Y. Yan and M. Al-Jassim, Current Opinion in Solid State and Materials Science 16 (2012) 39.
[3] H. Akai, http://sham.phys.sci.osaka-u.ac.jp/kkr/
[4] A. Filippetti and N. A. Spaldin, Phys. Rev. B 67 (2003) 125109.
[5] Y. Tani et al., APEX 3 (2010) 101202, JJAP 51 (2012) 050202.

Cu is an abundant element that exists in many materials, such as high-temperature superconductors, dilute-magnetic semiconductors, luminescent nanomaterials, and solar cells. Experimentally, it is well-known that Cu exhibits fast diffusion in many semiconductors. However, the origin of this unusual diffusion behavior in semiconductors is not clear. Because the diffusion of Cu in semiconductors plays an important role in many device applications, it is imperative to understand the underlying physics behind this puzzling phenomenon. Using first-principles calculations and taking the CdTe:Cu system as an example, we have compared the diffusion behaviors between Cu, Ag and group-IA atoms. We find that the existence of highly active d electrons in Cu is the main reason for the novel diffusion behavior of Cu in semiconductors. The strong coupling between Cu d states and unoccupied s states alters the stable doping site, diffusion pathway, and diffusion energy curve of Cu, compared to those of group-IA atoms that have no active d electrons. The s-d coupling lowers the diffusion barrier for interstitial Cu and contributes to the fast diffusion. Because our results and analysis are based mainly on the symmetry argument, we expect that the same conclusion can be applied for Cu diffusion in other zinc-blende semiconductors.

Polycrystalline CdTe is a prominent photovoltaic material with proven industry success. To develop the next generation of thin film CdTe solar cells, higher open-circuit voltages and longer minority carrier lifetimes must be achieved. Playing a major role in doping, defect migration, recombination, and current transport are grain boundaries and other extended defects within grains of poly-crystalline CdTe. A further understanding of these defects is needed to develop methods to mitigate their effects and promote higher doping with less recombination for device engineering.
Modern STEM techniques now allow for the direct determination of the chemical species in an image, allowing for ready extraction of the two-dimensional projections of atomic configurations. Commonly observed with STEM in CdTe are twins and stacking faults that extend throughout the entire grain. These twins can appear as lamellar repeating twins, or as single column stacking faults occurring in repetition near that of a Wurtzite structure.
In this talk, we will use first principles density functional theory (DFT) to investigate the thermodynamics and electronic structures of micro- and lamellar twins observed in STEM. The interaction energetics between adjacent twins and sets of twins are investigated.
We will also investigate the likelihood of formation of neutral and charged native point defects in and near these extended defect structures. Binding energies of multiple point defects near such structures are also revealed. Implications towards PV efficiencies are discussed.

Grain boundaries (GBs) in polycrystalline silicon thin film solar cells are frequently found to be detrimental for device performance. One way to control GB defects is to develop biaxially-textured silicon solar cells which have well-aligned grains in-plane and out-of-plane. These well-aligned grains form low-angle GBs, which could potentially have fewer dangling bonds and thus lower carrier recombination. This beneficial property of biaxially-textured silicon has stimulated the research in this emerging technology.
In this work, we use TCAD Sentaurus device simulator and known experimental work to investigate and quantify the potential performance gains of biaxially-textured silicon. For the first time, the conditions under which biaxially-textured films would lead to higher efficiency solar cells has been identified and quantified. The simulation reveals there can be performance gain from well-aligned grains when GB defects dominate carrier recombination or when grains are small. On the other hand, when intra-grain defects dominate recombination and grains are large, well-aligned grains do not lead to much performance gain.
Another important result from the simulation is when intra-grain and GB defects are low, Jsc is shown to be almost independent of grain size while Voc drops with decreasing grain size. The almost constant Jsc shows that when a film has high minority carrier diffusion length, carrier collection is not affected by grain size. However, a higher density of GBs and thus GB junction area in small-grained films can lead to an increase in dark current and decrease in Voc.

The commercial prominence in the optoelectronics industry of tunable semiconductor alloy materials based on nitride semiconductor devices, specifically In1-xGaxN, motivates the search for earth-abundant alternatives for use in efficient, high-quality optoelectronic devices and solar energy-conversion systems. In this presentation we will describe theoretical studies of the optoelectronic behavior and defect physics of the ZnSnN2 alloy series, as well as experimental investigations via X-ray absorption and emission spectroscopy of the alloys that probe the conduction and valence band partial density of states, that are in excellent agreement with first principles theory. We found that these materials have crystal structure and electronic structure similar to the In1-xGaxN alloys, but with significantly better miscibility (component-uniformity), i.e., the alloyed elements are fully miscible across the composition range without evidence for phase separation. Therefore, optical band-gaps of the ZnSn1-xGexN2 alloys that range from ~2.0 to ~3.1 eV can be tuned almost linearly as a function of the composition x. Based on the calculated formation energy of defects, we show that SnZn (Sn on Zn antisite) is the dominant intrinsic defect with a high concentration, and ON (O on N antisite) is another possible dopant. Both defects are donors, and the high concentration of donor states form a defect band below the conduction band, making the ZnSnN2 sample degenerately n-type. Resonant inelastic scattering observations to provide understanding of the role of Ge when incorporated into the ZnSnN2 lattice will be presented along with carrier dynamics of ZnSnxGe1-xN2 alloys as elucidated from photoluminescence (room and low temperature) and pump-probe spectroscopy.
This work was performed by the JCAP, a DOE Energy Innovation Hub, supported through the office of Science of the U.S. Department of Energy under Award No. DE-SC0004993.

Photovoltaic materials contain a plethora of defects, both point and extended defects such as intra-grain dislocations and grain boundaries (GBs). As the efficiency of polycrystalline CdTe exceeds that of single crystal cells, it has been suggested that GBs must be beneficial to carrier collection. A direct correlation from atomic structure to electrical activity has been achieved by a combination of scanning transmission electron microscopy (STEM), electron energy loss spectroscopy (EELS), electron-beam induced current (EBIC), cathodoluminescence (CL) and density-functional theory (DFT) calculations.
EBIC observations directly image the collection efficiency of the GBs, showing greatly improved collection after either a CdCl2 or a Cu heat treatment. Each treatment improves the cell performance independently. CL observations show reduced emission from GBs. STEM/EELS observations show that after CdCl2 treatment a substantial Cl concentration lies on Te sites within a few atomic spacings of the GBs. DFT calculations show that this Cl is sufficient to invert the GBs to n-type, explaining the EBIC and CL observations. Surprisingly, we find that intra-grain partial dislocation pairs do not create gap states but cause band bending, hence assisting the separation of electron-hole carriers and reducing recombination [1, 2]. Possible strategies for improving cell efficiency will be discussed.
References:
[1] C. Li, J. Poplawsky, Y. Wu, A. R. Lupini, A. Mouti, D. N. Leonard, N. Paudel, K. Jones, W. Yin, M. Al-Jassim, Y. Yan and S. J. Pennycook, Ultramicroscopy 134, 113 (2013)
[2] C. Li, Y. Wu, T. J. Pennycook, A. R. Lupini, D. N. Leonard, W. Yin, N. Paudel, M. Al-Jassim, Y. Yan and S. J. Pennycook, Phys. Rev. Lett. 111, 096403 (2013).
Acknowledgment:
This research was supported by the US DOE Office of Energy Efficiency and Renewable Energy, Foundational Program to Advance Cell Efficiency (F-PACE), (CL, YW, JP, AM, NP, WY, MAJ, JM, YY, SJP), the Office of BES, Materials Science and Engineering Division (ARL), and through a user project supported by ORNL&’s Center for Nanophase Materials Sciences, which is also sponsored by DOE-BES. This research used resources of the National Energy Research Scientific Computing Center, which is supported by the DOE Office of Science under Contract No. DE-AC02-05CH11231.

Deep-level transient spectroscopy (DLTS) is a general technique to study the properties of the electrically active defects in materials: energy level, concentration, and capture cross-section. We present recent DLTS studies of several thin-film solar cell materials. In close-space sublimation grown CdTe, DLTS was used to study the engineering of intrinsic defects through stoichiometry. In co-evaporated Cu(In,Ga)Se2, we use DLTS to reveal information about both the shallow levels responsible for doping and deep levels responsible for recombination. In co-evaporated Cu2ZnSnSe4, we use DLTS to investigate the influence of external impurities such as sodium on the electrical properties of the absorber.
The overall application of DLTS to thin-film solar cell materials is perhaps not yet as satisfactory and effective as what has been achieved for the more developed materials such as single-crystalline Si and GaAs. We will discuss possible reasons why DLTS is more challenging in the thin-film materials. On one hand, we will identify certain experimental artifacts that are not truly due to defects. On the other hand, we will discuss certain intrinsic complications due to the nature of the thin-film photovoltaic materials, which at the same time serve as opportunities for new physics.
To validate and clarify DLTS results, we will discuss corroborative experiments such as admittance spectroscopy, capacitance-voltage, drive-level capacitance profiling, temperature-dependent current-voltage, and transient photo-capacitance. We will also present recent improvements to instrumentation and analysis methods for overcoming the challenges facing the DLTS measurements in thin-film photovoltaic materials.

Post-growth processing is essential in order to achieve high performance in thin film CdTe solar cells. It is almost universally achieved by heating in the presence of CdCl2. We will review the evidence as to how it operates, and give new experimental results from alternative processes.
The performance of as-grown CdTe solar cells is improved very substantially by CdCl2 treatment, which causes increases in Voc, Jsc and FF. We will review the evidence for grain boundary electrical passivation and for metallurgical and chemical change in the devices resulting from the treatment. This will include transport measurements, direct studies of grain boundaries, deep level analysis, chemical profiling and structural investigations.
The results of comparing six alternatives to ‘standard&’ CdCl2 processing will be presented for the case devices fabricated in our laboratory from close-sublimation grown CdTe on sputtered CdS films. Details of the processes will be given. Characterisation has included performance (AM1.5), external quantum efficiency, current transport mechanism (J-V-T), carrier concentration (C-V), chemical analysis (SIMS) and microstructure (SEM). While most of the processes tested increase Jsc, not all cause increases in Voc. One treatment in particular gave results that were comparable to the device performance achievable with CdCl2 in our laboratory. At the time of writing this was Voc = 820 mV and conversion efficiencies of 13.5% for 5 x 5 mm2 contacts in a nominally Cu-free process.

Cadmium telluride (CdTe) has great potentials for the applications in solar cells and photo-detectors due to its low price and direct band gap (~1.5 eV) for ideal photovoltaic conversion efficiency. Graphene has been considered as a one of the replacements of ITO-based transparent electrodes due to its excellent properties such as high conductivity, high thermal conductivity and high mechanical strength. Also, graphene shows higher transmittance from ultra-violet to infra-red region than the ITO thin film. The application of graphene in CdTe solar cells has been rarely investigated. In our experiments, we demonstrate well-oriented CdTe thin film structures on modified graphene. In addition, we observed the effects of CdCl₂ treatment which is an essential process to achieve a high efficiency CdTe solar cell.
We prepared bare SiO₂/Si, pristine 2-layer grapehene/SiO₂/Si and modified 2-layer graphene/SiO₂/Si. Graphene was grown on the copper foil using chemical vapor deposition method. To create defects intentionally on graphene, UV-ozone treatments were performed. The formation of defects on graphene was monitored by micro-Raman spectroscopy. In Raman spectra, D-peak, which is related to the defects in graphene, was increased after UV-ozone treatments.
Close-spaced sublimation system was used to grow CdTe structures. We controled the morphologies of CdTe structures through adjusting the growth temperature gradients and growth time. Scanning electron microscope was used to observe the morphology of CdTe structures on different substrate conditions. In case of pristine graphene-based seed layer, CdTe structures did not grow on entire graphene layer. CdTe structures covered more area on top of the UV-treated graphene compared with pristine graphene. The difference in the coverage at same deposition conditions can be explained by the amount of defects on graphene. Since the growth of CdTe structures might initiate at defect sites. Also, the surface morphology of CdTe changed after the CdCl₂ treatments. To characterize the crystallinity of CdTe structures, X-ray diffraction (XRD) was used. Every sample shows the same behavior regardless of different substrate conditions. These XRD data confirm that high quality of our CdTe on the graphene layer. The recrystallization of CdTe structures was observed after the CdCl₂ treatments. Detail experiment process, results and discussion will be presented.

In recent years, scientific developments in renewable energy sources have been remarkable. The large-scale use of wind, solar, geothermal and hydropower energy sources to obtain electrical energy is a clear example of the possibility to obtain usable energy from clean and renewable sources. In particular, CdS/CdTe-based solar cells have attracted a lot of attention in recent years with efficiencies comparable to those of silicon-based solar cells. However, better efficiencies for CdS/CdTe can be expected with better control of the CdS/CdTe interface to achieve a defect-free p-n junction.
In this work, we used pulsed laser deposition (PLD) methods to grow in-situ CdS/CdTe heterojunctions to minimize oxygen and impurities contaminating resulting when exposing the CdS films to air before depositing the CdTe films. Our method allows obtaining higher quality p-n interface, when compared to other fabrication process techniques. In particular, we studied and correlated the effect of deposition pressure and temperature of the CdS and CdTe films with the electrical characteristics of the resulting p-n heterojunction diode behavior. A bilayer Cu/Au is used as Ohmic contact for CdTe and indium tin oxide (ITO) for the CdS. Optimized diodes showed Ion/Ioff ratios up to 10E6 with ideality factor of approximately 2. Carrier concentrations in the CdTe film were calculated using capacitance-voltage curves with values in the range of 10E15 cm-3. The impact of CdS doping concentration in the diode depletion width is also investigated. Finally, we study the effect of depositing a subsequent thick layer of CdTe (~ 3 µm) by closed space sublimation (CSS) on top of the already formed CdS/CdTe interface to demonstrate the benefits of creating the heterojunction in-situ.

The I-III-VI chalcopyrites show especially for the Cu compounds a pronounced tendency for a Cu depletion at the surface /1/. This depletion was recently /2/ experimentally observed for epitaxial prepared (112) and (100) surfaces of CuInSe2. Despite this reconstruction of the clean surface an additional depletion in Cu is observed on junction formation. Here we report on the change in the Cu to III concentration for CuInSe2 and CuGaSe2 (112) epitaxial surfaces as a result of the exposure to diethylzinc (DEZ) and water. DEZ and water form at 350° C clean ZnO films in a MOMBE deposition process /3/. In this study photo emission, LEED and ISS data will be presented to clarify the junction formation under the influence of the Cu depletion and the implications on the electronic structure.
/1/ J.E. Jaffe, A. Zunger, Phys. Rev. B 64, 241304, 2001
S.B. Zhang, S.-H. Wei, Phys. Rev. B 65, 81402, 2002
/2/ A. Hofmann, C. Pettenkofer, Surf. Sci. 606, 1180, 2012
/3/ S. Andres, C. Pettenkofer, F. Speck, T. Seyller, J. Appl. Phys. 103, 103720, 2008

The optical and structural properties of oxygenated CdS films were studied as a function of substrate temperature and oxygen concentration in the sputtering ambient. CdS films deposited at room temperature (RT) exhibit nanocrystalline phases and increased band gaps. This increase in oxygen concentration leads to a decrease in grain size and therefore enhanced quantum confinement effects. X-ray diffraction shows strong attenuation of CdS peaks with a significant broadening effect for ge; 2% ambient oxygen fraction. In contrast, when CdS films were deposited at 270 ^oC, the band gap decreases as the oxygen concentration increases. Polycrystalline CdS:O films deposited at this higher temperature have a hexagonal wurtzite structure. Despite the red-shift in absorption edge, solar cell devices made using the higher temperature deposited CdS:O films exhibited better performance than those using the room temperature deposited CdS:O films. With the application of CSS processed CdTe absorber layers, our small-area cells on commercial SnO₂:F/SnO₂-coated soda-lime glass substrates have shown efficiencies of 13.8% and 13.1% respectively for high temperature and RT sputter deposited CdS window layers. The observed difference in cell efficiencies was due to the variation in short-circuit current. The open-circuit voltages and fill factors are comparable for cells using both types of window layers.
This work was partially supported by the DOE-FPACE program under contract No: 0492-1662 and the Ohio Research Scholar Program.

The highest efficiency CuIn1-xGaxSe2 (CIGS) based solar cells have been produced from films with x~0.3 which gives a value of Eg around 1.1-1.2eV. Increasing the Ga content of the CIGS absorber allows tuning of the band gap which can enhance performance under actual operating conditions and make it possible to use CIGS film in multi-junction devices. However, champion cells have not yet been produced for values of x significantly greater than 0.3. As grain boundaries have a strong influence on device performance, this work focuses on how increased Ga content in CIGS films affects the recombination behavior of grain boundaries. Performing cathodoluminescence spectral imaging (CLSI) and electron beam induced current (EBIC) measurements on the same region has allowed us to directly compare the collection of carriers generated near a particular grain boundary with the dominant recombination transition associated with the boundary. Our results suggest that grain boundaries at different distances from the junction have different recombination behavior. In particular, the red shift in CL emission observed at grain boundaries near the surface is less pronounced or non-existent for boundaries farther from the junction. This difference in degree of red shift was more evident in higher efficiency cells with x~0.3 than in lower efficiency cells with larger values of x. Additionally, the red shift at grain boundaries near the junction was accompanied by strong collection in the EBIC images, but this correlation did not persist at boundaries deeper in the film. In order to definitively correlate the CLSI and EBIC observations with the Ga content we have performed nano-scale compositional analysis on multiple boundaries analyzed with both techniques. Our results show that it is necessary to formulate alternative models for recombination at grain boundaries in CIGS solar cells based on their character and position in the film.

Formation of CuIn3Se5 and CuIn5Se8 phases via In at Cu antisite defects (InCu), known as ordered vacancy compounds (OVC), in Cu(In1-x,Gax)Se2 (CIGS) active layers are believed to play an important role in the performance of CIGS photovoltaics. However the spatial non-uniformity of CIGS composition , both in the substrate normal direction due to varied Ga concentration as well as laterally, could potentially cause spatial non-uniformity in the OVC layer formation leading to drastic spatially non-uniformity in the overall performance of the cell. In this work we utilize aberration corrected transmission electron microscopy (TEM), Grazing-incidence X-ray scattering (GIWAXS), Raman spectroscopy, diffraction limited spatially and wavelength resolved laser beam induced current (LBIC) and temperature dependent admittance spectroscopy with capacitance-voltage (CV) profiling on CIGS cells having power-conversion efficiencies from 14%-19% and Ga fractions ranging from 0.3 to 0.5, in order to better understand the spatial variation of OVC and its role played in achieving high conversion efficiency in CIGS.
From TEM measurements we obtain electron-diffraction patterns of CuIn3Se5 and CuIn5Se8 OVC phases as well as the local stoichiometry from EELS measurements, and investigate the spatial non-uniformity of these phases and its effect on device performance. GIWAXS measurements are used to quantify the vertical extent of the OVC layer over a larger lateral area. Finally we use admittance spectroscopy and CV profiling to understand the relationship between the OVC layer characteristics and formation of shunt pathways and defects in the cells, and determine with high resolution the spatial non-uniformities in the performance using LBIC.

Cadmium telluride (CdTe) is a leading material used in thin-film photovoltaics. Limitations in performance due to crystal defects, including point defects, dislocations, and grain boundaries, need better fundamental understanding to push the state-of-the-art. Often, such defects are deleterious by contributing to reduced carrier lifetimes and increased recombination through the introduction of deep traps. Understanding the electronic structure of the defects therefore is critical to optimizing material and device performance.
Here we describe a theoretical study of the impact of dislocations on the electronic properties of CdTe. First-principles calculations based on density functional theory are used to compute the atomic and electronic structure of prototypical screw and edge dislocation cores in CdTe. The accuracy of the description of long-range strain around the dislocation cores is studied by carefully examining finite-size and edge termination effects in the atomic models, including long-range strain from very large-scale semi-empirical models. The spectra of gap electronic states are then computed for a number of different core structures, allowing direct assessment of the impact on carrier trapping and device electrical performance. In particular, comparison of Te-core and Cd-core dislocations is made, as well as various core symmetry structures.
Prepared by LLNL under Contract DE-AC52-07NA27344 and funded by the National Nuclear Security Administration Office of Nonproliferation and Verification Research and Development (NA-22).

Micropatterned CdTe-CdS islands have been prepared through SiO2 windows to enhance photovoltaic performance by relieving lattice mismatch stress. To efficiently measure their local response down to the nanoscale, a new measurement scheme is presented for mapping PV performance. Sequences of images are acquired with contact mode conductive atomic force microscopy (cAFM) for a single area each with incremented DC applied biases during simultaneous illumination from below through an inverted optical microscope. The local I-V response for any given pixel position is easily determined in each distinct image (voltage) which then used to construct drift-free property maps (photoconduction, Voc, Ish, fill factor etc.). This new measurement scheme provides ~5nm spatial resolution over 65,536 pixels with orders of magnitude shorter acquisition times than conventional cAFM based methods. The PV performance of continuous and microstructured films as well as different growth regions within individual films will be compared and related to crystallographic orientations based on EBSD results. Enhanced photoconductivity is clearly observed for certain grains even revealing twin boundaries, confirming the microstructural control on polycrystalline PV performance.

Copper-indium-gallium-selenium (CIGS) is one of possible candidates for high conversion efficiency solar cells with direct band gap of 1~1.7 eV and high absorption coefficient of 10^4~10^5 cm-1. To obtain better device performance, enlarged band gap on the surface of CIGS can not only enhance open circuit voltage due to junction characteristics improvement, but also increase short circuit current because of smaller positive conduction band offset at CIGS/buffer layer interface. The latter one results in easier transport for carriers within solar cells. Sulfur is often used for expanding band gap by change compositions of CIGS. Generally, toxic H2S gas is used to sulfurize the CIGS at high temperature. Here, a facile wet sulfurization process is achieved by dipping CIGS thin film into InCI3 and thioacetamide aqueous solution at 80 oC. From Raman spectroscopy, peaks are observed with a blue-shift after surface treatment, which is attributed by increased S/Se ratio. For photoluminescence spectroscopy results, blue-shifted peaks indicate an enlarged surface band gap. The results can be applied to zinc sulfide (ZnS) buffer layer, which is potential to replace toxic CdS layer of CIGS solar cell. Finally, a high performance solar cell can be demonstrated.

Polycrystalline p-type Cu(In,Ga)Se₂ semiconductors represent the absorber material in thin film solar cells currently reaching the highest power conversion efficiency. Efficiencies above 20% are surprising considering the high density of grain boundaries in these films. Their role in the solar cell as well as their electronic structure are largely investigated and discussed. Recently, a number of scanning probe microscopy studies has contributed to the understanding of the grain boundary properties. In this talk we will present our contributions to this research field by employing Kelvin probe force microscopy (KPFM) and KPFM in conjunction with other techniques.
Several KPFM investigations of polycrystalline Cu(In,Ga)Se₂ thin films have shown potential variations at grain boundaries [1], mainly exhibiting a lower work function around the grain boundary. However, recently also grain boundaries without potential variation and with higher work function have been found. These potential variations are typically assigned to a local band bending due to the presence of charges [2]. A study of polycrystalline samples with a variation in the In-to-Ga ratio is presented [3]. From a combination of KPFM with electron backscatter diffraction (EBSD), we could identify twin grain boundaries as predominantly neutral, whereas higher disorder grain boundaries are predominantly charged [4].
Additional understanding was gained from model samples consisting of large bicrystals on which, in addition to the KPFM characterization, also regular electrical characterization techniques and transmission electron microscopy were applied. These studies reveal a charge neutral barrier to majority transport for twin boundaries [4], while higher disorder Σ9 grain boundaries are charged and present a thin and high (~500 meV) electronic barrier for hole electrical transport [6], which can be attributed to the lower atomic density in the grain boundary core by means of comparison to density-functional theory calculations [7].
References:
[1] S. Sadewasser, Thin Solid Films 515, 6136 (2007).
[2] D. Fuertes Marroacute;n et al., Phys. Rev. B 71, 033306 (2005).
[3] R. Baier et al., Sol. Energy Mat. Sol. Cells 103, 86 (2012).
[4] R. Baier et al., Appl. Phys. Lett. 99, 172102 (2011).
[5] S. Siebentritt et al., Phys. Rev. Lett. 97, 146601 (2006).
[6] M. Hafemeister et al., Phys. Rev. Lett. 104, 196602 (2010).
[7] S.S. Schmidt et al., Phys. Rev. Lett. 109, 095506 (2012).

Yttria (Y2O3) is a good host material of rare earth elements sensitizers and activators thanks to its large band gap and strong mechanical strength. It has been known that absorption and emission spectra of Bi-doped yttria are dependent much on their fabrication procedures such as annealing condition or fuel in combustion, as well as population of Bi at C2 and C6 sites of yttrium. Despite many years of researches, unrevealed questions are still there about origins of photoluminescence dependency on fabrication procedures. We have investigated the geometric and electronic structures of Bi-doped Y2O3 including intrinsic point defects, using density functional theory (DFT) calculations and thermodynamic formalism. The formation of point defects was thoroughly investigated under varying oxygen partial pressure and temperature, and their effect in electronic structures. We found that, under usual annealing temperature, the concentration of defects is not ignorable, and defects might degrade emission properties in certain conditions. We explain roles of defects in light absorption and emission.

We report electron and hole photocarrier drift mobility measurements in high-efficiency, thin-film polycrystalline CdTe solar cells. The magnitude of these drift mobilities is typically 1 cm²/Vs for both carriers, which is vastly lower than the values in bulk crystals (typically 100 cm²/Vs for holes and 800 cm²/Vs for electrons). Since little has been known about mobilities in thin-film CdTe solar cells, device modelers have provisionally used values that are fairly close to single-crystal values. The new measurements may require a reconsideration of device modeling results for the cells.
The drift-mobilities are measured using the time-of-flight technique with a pulsed laser. Special bifacial samples were prepared to simplify the electron mobility measurements. The samples were prepared at First Solar, Inc. with a variety of treatments so as to explore the range of mobilities. We measured drift-mobilities over the temperature range 100 K - 350 K.
As a broad summary, hole drift mobilities at room-temperature varied about a factor 5 between ten coupons prepared using different deposition procedures. The variation in electron drift-mobilities is substantially greater. Neither the electron nor the hole drift-mobilities are strongly temperature dependent; a typically variation in mobility over the range 100 - 350 K was about a factor three.
We have not identified the microscopic mechanism underlying the reduction of the mobility in these CdTe thin-films compared to bulk single-crystals. Trapping of carriers onto deep levels is ruled out from the time-of-flight technique. Carrier transport is through the thickness of these thin-films, and we don&’t have evidence for traversal of grain boundaries as the mechanism. While the mobility magnitudes are low enough to make comparisons with nanocrystalline materials reasonable, the temperature-dependence in CdTe is much weaker. More speculative mechanisms such as disorder-induced polaron formation may need to be considered.

To achieve high-efficiency polycrystalline CdTe-based thin-film solar cells, the CdTe absorbers must go through a post-deposition CdCl2 heat treatment followed by a Cu diffusion step, but the electronic property changes induced by these treatments are not well understood. To this end, CdTe solar cells with and without Cu diffusion and CdCl2 heat treatments are investigated using electron beam induced current (EBIC) and electron backscatter diffraction (EBSD) techniques in order to understand the separate role of each treatment. The results reveal how the CdCl2 treatment and Cu diffusion increase device efficiency. In particular, the microscopic property changes that cause an increase in short circuit current due to a CdCl2 heat treatment and high open-circuit voltage achieved by a Cu diffusion treatment are identified. Also, a drastic change in the carrier collection properties of the grain boundaries due to the inclusion of Cu and Cl is revealed, in which an inversion from a decrease to an enhancement in the grain boundary short circuit current occurs. The increase in efficiency at the grain boundaries is shown to be due to the presence of a space charge region. Also, each treatment improves the overall carrier collection efficiency of separate regions of the solar cell, and, therefore, the benefits realized by CdCl2 and Cu treatments are shown to be completely independent of each other.
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 sponsored by the Scientific User Facilities Division, Office of Basic Energy Sciences, U.S. Department of Energy.

Indium-antimonide (InSb) has the narrowest bandgap energy and high electron mobility among the III-V compound semiconductors. It has high potential for device applications involving infrared detectors, high frequency electronics, and magnetic field sensors. In order to fabricate InSb devices with high performance, developing an optimized etching process that can reduce dark currents is one of the most important issues. Ar ion beam etching (IBE) which can make smooth etched surface without any by-products is widely used for the fabrication of InSb devices, but even in this case, Ar ion bombardment can make defects near the surface. Such damaged region, undetectable even by atomic force microscopy (AFM), can act as surface trap sites and dark current sources. To develop a high performance InSb device, research on the measuring method is required to evaluate the damaged surface for the optimization of the etching process.
In this work, Ar ion-induced damages on the etched InSb (100) surfaces were examined by resonant Raman scattering. Raman scattering measurements for the etched InSb surface show that the splitting of LO and TO phonons is drastically enhanced as the RF power increases. According to the Raman selection rule for InSb (100) with zinc-blende crystal structure, the LO phonon scattering is allowed and the TO phonon scattering is forbidden. Sb vacancies caused by preferential sputtering can account for the increase of the number of defects near the etched surface. The structural disordering due to induced-defects breaks the symmetry of the crystal and would lead to breaking of selection rule. These scattering are shown to be very sensitive to the change in the geometry and the lattice perfection. The ion-induced structural and morphological deformation makes the forbidden TO phonon mode to reveal. Thus, the enhancement of TO scattering represents surface defects due to ion bombardment.
Both the integrated area ratio of A(TO)/A(LO) and the FWHM of LO peaks are proportional to the applied RF power. The A(TO)/A(LO) increases from 0.05 to 0.23 and the FWHM of LO peak increases from 5.71 cm-1 to 7.46 cm-1 as the RF power increases from 50 to 200 W. The integrated area ratio of A(TO)/A(LO) and the FWHM of LO peak was observed to increase with increasing the energy of Ar ions, which is proportional to the RF power. Moreover, amorphous phase is observed in the Ar ion beam etched InSb sample by Raman spectroscopy. The presence of these phases is a strong evidence for structural damages of after Ar IBE process.

We report a combined atomic resolution and X-ray diffraction study of strain and composition in semiconductor nanostructure and III-V superlattices. The atomic resolution study is based on advances in aberration corrected scanning transmission electron microscopy, which now allow recording of atomic columns at 0.1 nm or less resolution at high fidelity. Quantitative composition and lattice analysis can be carried out in real space using atomic resolution Z-contrast images, in combination with electron spectroscopy. Using information obtained from electron imaging, we have developed a model for X-ray diffraction which allows a quantitative measurement of superlattice properties. Together, we show electron imaging and X-ray diffraction offer unprecedented details about the structure and defects in MBE grown III-V superlattices. This work is supported by U.S. Army Research Office (Grant No. Army W911NF-10-1-0524 and monitored by Dr. William Clark) through the MURI program.

Carrier lifetimes vary throughout a photovoltaic material due to defects such as grain boundaries, traps and surface defects that act as local recombination centers. Mapping and quantifying the effect of these defects is critical to improving the performance of devices where defects limit efficiencies. Traditional optical techniques are inherently surface sensitive due to light-absorption that predominately occurs near the surface, thus making determination of bulk or sub-surface properties difficult. Here we show that two-photon absorption permits sub-surface optical excitation, enabling us to map optoelectronic properties in three-dimensions. One example use of this technique is two-photon time-resolved PL measurements (2P-TRPL). With 2P-TRPL we can now measure the true bulk minority carrier lifetimes of materials even when the sample has high surface recombination velocity. We demonstrate how the traditional one-photon TRPL technique can underestimate the bulk lifetime by 10x in CdTe crystals with high SRV while showing that two-photon excitation accurately measures the bulk lifetime. Finally, we generate three-dimensional spatial maps of optoelectronic properties (locally-excited photocurrent and carrier lifetime) at the surface and in the bulk of photovoltaic materials using two-photon excitation.

Transparent conductors play an increasingly important role in a number of semiconductor technologies. Of special interest are transparent conductors that can be used as contacts in solar cells where there are strict requirements on a high conductivity and a good transparency in the usable part of the solar spectrum ranging from near infrared to ultraviolet. In this presentation I will discuss our recent work on optimization of properties of CdO based transparent conductors. We have shown that intentionally doped CdO can have electron density higher than 10^21 cm^-3 and electron mobility in excess of 100 cm^2/Vs leading to record high electron conductivity exceeding 10^4 S/cm. The high electron mobility can be attributed to a large static dielectric constant and efficient screening of the ionized impurity scattering. The weak scattering is responsible for very low free carrier absorption with the infrared transparency window extending to 1500 nm making this material suitable for solar cells that rely on the infrared part of the solar spectrum. The short wavelength transparency is determined by the Burstein-Moss shift moving the absorption edge from 2.2 eV in lightly to 3.2 eV in most heavily doped material. The unusual electrical and optical properties stem from interplay between electronic band structure and defect properties of CdO. We have performed systematic experimental studies of the doping behavior of Ga and In donors. The results are explained in terms of the amphoteric defect model. The model is then used as guidance in search for transparent conductors with improved ultraviolet transparency. We have synthesized Cd-rich CdMgO and CdZnO alloys. The latter alloys offer a potential for an increased band gap without compromising maximum doping level and electron mobility. Supported by US DOE

Semiconducting nanowires have been recognized as promising materials for high-performance electronics and optics where optical and electrical properties can be tuned individually. The feasibility of III-V nanowire (NW) integration with existing silicon processing technology due to the small footprint between the silicon substrate and the nanowire material has further sparked that interest. Especially, InP NW PV has been grown epitaxially on Si substrates [1]. For NWs to provide the new architecture for next generation photovoltaics there is a strong need to take complete control over synthesis. By optimizing growth conditions with respect to tapering we created nanowire-InP nanowire based solar cells using Au seed particles for growth.
We will report on the growth, processing and characterization of nanowire array-based solar cells with 13.8 % efficiency [2]. First, gold particles were patterned on InP substrates with a 500 nm pitch, using nanoimprint lithography. Then, about 1.5 mu;m long InP nanowires were grown using DEZn and TESn as doping precursors, to create an axially defined p-i-n junction. HCl was used to prevent radial overgrowth. The nanowires were processed as-grown with a transparent conductive oxide as top contact to create 1x1 mm2 solar cells, with 4 million nanowires per cell.
The solar cells were investigated using a sun simulator at Fraunhofer ISE CalLab reference setup. Although the 180 nm-diameter NWs only covered 12 % of the surface, the photocurrents were 71 % of the theoretical maximum for an InP solar cell. This is six times the limit in a simple ray optics description, and comparable to the record planar InP cell. To understand the absorption, we used three-dimensional electromagnetic optical modeling [3, 4]. We find excellent agreement between the spectra of modeled absorption and the experimentally measured external quantum efficiency.
[1] Magnus T. Borgström et al., IEEE J. Select. Topics Quantum Electr. 17 (2011), 1050
[2] Jesper Wallentin et al. Science, 339 (2013) 1057
[3] Nicklas Anttu et al., Phys. Rev. B 83 (2011), 165431
[4] Jan Kupec et al., Opt. Express 18 (2010), 27589

Intermediate band photovoltaics have been proposed to enhance efficiencies of solar cells by harvesting additional energy from sub-bandgap photons. To fabricate an intermediate band material, we introduce deep-level dopants at high concentrations. Sulfur hyperdoped silicon, a material that exhibits strong sub-bandgap light absorption, is a good model system for understanding properties of an intermediate band. Previous work observed metallic-like conduction and identified insulator-to-metal transition at sulfur concentration above 2 × 10^20 1/cm^3. In addition to identifying the critical concentration for band formation, understanding transport properties and the energetics of the intermediate band and are important. In this work, we use femtosecond-laser hyperdoping to introduce non-equilibrium concentrations of sulfur into silicon and study the nature of the resulting intermediate band. By changing the pressure of the dopant precursor, we fabricate samples doped with different concentrations of sulfur. To better understand the dopant energetics, we perform temperature-dependent Hall and resistivity measurements. We analyze the energetics of the intermediate band by using a two-band model. The temperature-dependence of the carrier concentration and resistivity suggests that the dopant concentration is below the insulator-to-metal transition and that the samples have a localized intermediate band at 70 meV below the conduction band edge. The two-band model and analysis technique presented in our work enable studying intermediate band energetics from temperature-dependent transport measurements.

Hybrid solar cells, typically comprised of a p-type highly conducting polymer, poly(3,4-ethylenedioxythiophene):poly(styrenesulfonate) (PEDOT:PSS) on a n-type Si substrate, have been intensively studied recently for the realization of cost-effective, high-efficiency energy-generation devices. PEDOT:PSS is also a technologically important polymer used for the fabrication of organic electronic devices. Many efforts have been made recently to improve the properties of PEDOT:PSS and to fabricate high-efficiency solar cells. However, most of the high power conversion efficiencies (PCE) have thus far been obtained from solar cells fabricated on surface structured Si substrates. High performance planar single-junction solar cells have considerable advantages in terms of processing and cost, because they do not require the complex surface texturing processes. Here, we report that the addition of an appropriate wt% of ethylene glycol and fluorosurfactant can dramatically enhance the solar cell performance, especially when compared to the commonly used co-solvent, dimethyl sulfoxide (DMSO), and non-ionic surfactant, triton-X100. In addition, the interface of these single-junction solar cells can critically influence the performance. We provide morphological models to explain the influence of the co-solvent in PEDOT:PSS, in which the co-solvent modifies the internal crystalline ordering of individual PEDOT nanocrystals that increases the crystal size and forms closely packed nanocrystals. The co-solvent also facilitates rearrangement of PSS that reduces its surface chain networks and enhances the polymer conductivity. Using time-of-flight secondary ion mass spectroscopy (TOF-SIMS), we obtain the first three-dimensional chemical image of the interface, which shows the presence of micropore defects limiting the device efficiency. We further demonstrate that these defects can be controlled by optimizing the amount of co-solvent and surfactant in PEDOT:PSS. By minimizing the defects using this approach, we fabricate a solar cell with an efficiency exceeding 13% on a planar Si substrate for the first time. This method could be further exploited to enhance the performance of other diverse organic electronic devices based on PEDOT:PSS, such as polymer light-emitting diodes or polymer solar cells.

Transition metal oxides are often employed as buffer layers in conventional organic solar cells. Take MoO₃ for example, it has the advantage in enhancing cell stability when compared to PEDOT:PSS. In this study, we use a solution process to synthesize the MoO₃ layer doped with vanadium. The doped vanadium will either form V₂O₅ or introduce defects in molybdenum oxide, which may modify the position of valence band to match that of P3HT more suitably. Chemical composition of the V-doped MoO₃ layer is studied by X-ray photoelectron spectroscopy (XPS), and the energy level of valence band is investigated by ultraviolet photoelectron spectroscopy (UPS). It is demonstrated that the solution process is an effective method to modify the V:Mo ratio of the buffer layer. With that, an appropriate V:Mo ratio is obtained to expedite the hole transportation successfully and the performance of solar cells is enhanced.

For high efficiency thin-film silicon solar cells in single amorphous silicon (a-Si:H) or micromorph tandem configuration, a low band gap of the intrinsic a-Si:H absorber layer is desirable for better utilization of the illumination spectrum. Typically, the band gap of the absorber layer is varied by using different hydrogen dilutions of the silane gas during the plasma-enhanced chemical vapor deposition (PECVD): lower hydrogen dilution leads to smaller band gaps. This method comes to its limit at last with no hydrogen dilution, i.e. depositing from pure silane. Moreover, stability against light-induced degradation (Staebler-Wronski effect) deteriorates with lower hydrogen dilutions.
Another efficient possibility to decrease the bandgap is the increase of deposition temperature. In case of using the superstrate configuration the window layers are subject to an annealing step since they are deposited before the absorber layer. Issues related to annealing of the window layer are the subject of the present study.
With increasing deposition temperature of the absorber layer, we observe a systematic decrease of the band gap, leading to more absorption at long wavelengths. However, the short-circuit current assumes a maximum and drops at high deposition temperature due to bad charge collection at short wavelengths. This experimental signature is typically associated with the so-called boron-tailing. Possible mechanisms are investigated by SIMS measurements of dedicated i-p-i layer stacks to identify boron diffusion, by optical measurements (photothermal deflection spectroscopy, ellipsometry, transmittance, FTIR) of p-layers annealed at different temperatures, and by conductivity measurements of the annealed p-layers.
In addition to changes of the p-layers, also the low-pressure chemical vapor deposition (LPCVD) grown zinc oxide is affected by high temperatures. ZnO layers, annealed at different temperatures, were investigated by photospectroscopy, Hall, and conductivity measurements. Further, the change of ZnO by light soaking has been investigated.
A third issue with increased deposition temperatures concerns the voltage provided by the solar cell: with a decrease of the bandgap of the absorber layer, the voltage decreases much stronger than expected. Possible reasons such as a change of the activation energy of the p-layer are investigated.
Although not all mechanisms that limit deposition temperature of the absorber layer are yet understood, they seem at least partially to be linked to boron as a dopant in a-Si:H. Therefore, we investigated trimethylgallium instead of trimethylboron as gas precursor for the deposition of p-doped a-Si:H. Preliminary results of this test will be presented.

High efficiency silicon solar cells have been a topic of research for many years. Initially this research consisted of exploring the limits of the applied material and investigating the cell design and fabrication process steps. Subsequently, in a second phase transferring the obtained knowledge to produce solar cells at a competitive cost and at increased scale became the focus. Significantly reducing the cell&’s thickness, while increasing the efficiency, represents a path toward cost competitiveness with fossil fuel-based energy sources. Up-to-date the PV market is yet dominated by Wafer-based crystalline silicon c-Si solar cells, having 90% of market share. The efficiency record of 25% was achieved at UNSW where the PERL concept (Passivated Emitter and Rear Locally diffused) has been employed [1]. Heterojunction a-Si:H/c-Si solar cell is a promising candidate for low-cost and high-efficient PV technology. Specifically, the HIT (Heterojunction with intrinsic thin-layer) cell concept developed by Sanyo Electric Corp, holds to date the world record for industrial cells with an efficiency exceeding 22% at the manufacturing level and close to 23% at R&D level [2]. Recently, using a 98 µm thick silicon wafer with HIT concept, Panasonic PV achieved the highest R&D efficiency of 24.9% , and this on larger cell area of 100 cm2 [3,4].
Up to date, top-research laboratories failed to achieve similar results as Sanyo&’s. The large variation of HIT cells efficiencies obtained by several research and industrial groups is due to the fact that the performances of these cells critically depend on the properties of the 20 nm amorphous silicon (intrinsic, p- and n- doped) that act as emitter, BSF (Back Surface Field) and passivation layers. In this work, the a-Si/c-Si HIT cell is modeled on both n-type and p-type c-Si substrates using AFORS-HET software. We note that the density of localized states in the band gap of amorphous silicon has been modeled by exponential band tails (Urbach tails) and Gaussian mid-gap states (associated to silicon dangling bonds). The effect of different factors such as the amorphous layer thickness, flat and textured surface shape and more importantly, the doping level of p and n type layers and densities of interface states (Dit) on the efficiency was investigated. This later parameter plays a major role in the surface/interface recombination rate and hence dictates the passivation quality. Finally, the effect of the work function of the TCO layers on the interface quality of a-Si/TCO and the heterojunction interface a-Si/c-Si is discussed. The influence of these parameters on the cell efficiency is discussed and the key factors to achieve a promising cell design that leads to high efficiency of 25 % are identified.
[1] Zhao et al., APL 1998; 73: 1991-1993.
[2] Mishima, T. et al., Sol. Energ. Mat. Sol. 95, 18 (2010).
[3] Kinoshita, T et al., 26th EU PVSC Proceedings, 2011; 871-874.
[4] http://www.pv-magazine.com

This talk will highlight and review successful interface and bulk defect-engineering approaches to increase the conversion efficiency of novel thin-film materials. The impact of bulk defect engineering — extended defects, grain size, and carrier concentration — to improve short-circuit current will be explored from device simulation and experimental points of view. Strategies for interface engineering to improve open-circuit voltage and minimize series resistance will also be discussed, with attention to controlling charge-carrier concentrations at absorber surfaces. The talk will highlight the role of laboratory-scale and synchrotron-based characterization techniques to pin-point specific materials and device loss mechanisms, toward defining a brachistochrone path to novel materials development.

The Cu(In,Ga)Se2 chalcopyrite compound (also called CIGS) is an established material for high performance thin film solar cells. On rigid soda-lime glass substrates photovoltaic conversion efficiencies up to 20.3% are reported, which represents the highest thin film solar cell performance on lab-scale. The CIGS technology is also used for flexible substrates such as plastic or metal foils, which can enable low cost production owing to their applicability to roll-to-roll manufacturing systems. Certified efficiencies of 18.7% on polyimide and 17.7% on stainless steel (SS) foils were achieved on lab-scale so far.
In comparison to polyimide substrates, stainless steel foils show higher temperature stability and tensile strength, where CIGS process temperatures up to 600 omicron;C are possible to apply. However, a major disadvantage of steel substrates is that iron as the main component of stainless steel can diffuse through the back contact into the CIGS absorber, where Fe impurities are known to reduce the solar cell performance. To prevent iron diffusion, oxide (e.g. Al2O3,SiO2) or nitride10 (e.g., TiN, Si3N4) diffusion barrier layers are conventionally used. These barrier layers are deposited with different techniques, such as magnetron sputtering, plasma-enhanced chemical vapor deposition, thin film anodization, and atomic layer deposition (ALD). Slow ALD method is hard to implement on industrial scale. Beside the diffusion barrier blocking performance and production costs, residual film stress must be considered for thin flexible substrates, as stress can lead to structural deformation (foil bending) and thin film micro-cracking.
In this study, we form the iron oxide layer as diffusion barrier from SS substrate annealing at 600 omicron;C for 1 min. It has the advantage that this method is simple and need not to use other materials. Diffusion barrier was measured by Raman spectroscopy and X-ray diffraction. It demonstrates that surface of SS substrate oxidized by annealing process is iron oxide layer. Mo back contact and CIGS layer was deposited by co-sputtering system and selenization process in order to investigate iron diffusion through the diffusion barrier into the absorber layer. Fe impurity diffused in CIGS was measured by secondary ion mass spectroscopy. As a result of SIMS, iron oxide diffusion barrier had reduced the diffusion of Fe from SS substrate. Also, performance of CIGS solar cell was improved with iron oxide layer because Fe impurity concentration in CIGS was reduced which influence the efficiency of CIGS solar cell.

Thin films including CdTe and Cu(In,Ga)Se₂ (CIGS) have been considered as a promising solar cell due to their low cost and flexibility. In particular, CIGS solar cell has attracted since its large optical absorption coefficient (α > ~ 10⁵ /cm) is suitable for high conversion efficiency. Recently, the CIGS solar cell with 20.3 % efficiency was achieved by co-evaporation method. However, vacuum process exhibited a limitation in cost reduction coming from wasted materials and high initial investment. Thus, electrodeposition method which can provide high quality film, high rate process and very low initial cost has been suggested as breakthrough. However, electrodeposition method shows the lower conversion efficiency than vacuum process, since optimization is not sufficient yet. Even though nucleation and growth mechanism is effective in uniformity and performance of solar cell, the fundamental research was not proceeded. In addition, additives, such as PEG, saccharin, thiourea and etc. could affect the reduction of copper during the electrodeposition. In this study, copper layer was electrodeposited in electrolyte consists of CuSO₄, H₂SO₄, and Na₂SO₄ was used. After that, indium layer was deposited in solution composed of InCl₃ and NaCl. Finally, rapid-thermal-process was employed to form the CuInSe₂ absorber at 650 ^oC and selenium atmosphere. SEM, XRD and I-V measurement was used to characterize the films.
REFERENCES
[1] R.N. Bhattacharya., A.M. Fernandez, Solar Energy Materials and Solar Cells 76 (2003) 331.

CIGS thin film solar cells have recently demonstrated a record efficiency of 20.3%, the absolute record for all thin film photovoltaics [1]. Chalcopyrite based technologies have already entered the stage of industrial mass production, with commercial modules that provide stable efficiencies in the range of 12 - 13% [2]. However, attaining higher efficiency is challenging due to the difficulty of controlling the production processes on large area substrates. In addition, the fundamental studies, such as growth and characterization of CIGS single crystal are very little, recently. First of CIGS solar cell was CIS single crystal solar cell and achieved 12% [3]. Therefore, return to the start line, or investigate single crystal is necessary to achieve higher efficiency.
It is generally difficult to grow high-quality single crystals of the I-III-VI2 chalcopyrite compounds, because most of the compounds grow through a peritectic reaction or a solid state transition during the cooling process. CIGS single crystals are grown by traveling heater method (THM), which is one of the solution growth techniques. Advantages of the THM growth are following that growth temperature is low compared with that of the other melt growth and larger crystals can be grown compared with a conventional solution growth. The principle of THM is shown in our previous studies [4, 5].
In this paper, large size CIGS single crystal was obtained by THM and characterized from using X-ray diffraction (XRD), Raman spectroscopy, X-ray rocking curve (XRC), electron probe microanalysis (EPMA) and Hall effect measurement.
The purpose of this study is to investigate growth and characterization of single crystal to achieve higher efficiency. Large size CIGS single crystal was successfully grown by THM. No secondary phases are observed from XRD and Raman spectroscopy measurements. The FWHM of the XRC for the (112) oriented CIGS single crystal is 103 arcsec. The results show that good-quality single-phase CZTSe single crystals can be obtained by THM growth. The composition of the CIGS single crystal was homogeneous and the stoichiometric ratio of CIGS was found to be slightly Cu-poor, In-rich, Ga-rich and Se-poor. Therefore, it is assumed that the In on Cu antisite (In_Cu) and the Se vacancy are dominant, leading to n-type conduction. The electron carrier concentration, n = ~10^(16-17) cm^-3, the hole mobility, mu;n = 550-720 cm^2 V^-1 s^-1, and the resistivity, ρ =0.1~10 Omega;cm by Hall effect measurement.
[1] P. Jackson, D. Hariskos, E. Lotter, S. Paetel, R. Wuerz, R. Menner, W. Wischmann, and M. Powalla, Prog. Photovol. 19 (2011) 894.
[2] M. A. Green, K. Emery, Y. Hishikawa, W. Warta, and E. D. Dunlop, Prog. Photovol. 20 (2012) 606.
[3] J. L. Shay, S. Wagner, and H. M. Kasper, Appl. Phys. Lett. 27 (1975) 89.
[4] H. Miyake, and K. Sugiyama, J. Crystal Growth, 125 (1992) 548.
[5] H. Miyake, T. Haginoya, and K. Sugiyama, Jpn. J. Appl. Phys., 36 (1997) 785.

Solar energy conversion is a highly attractive process for clean and renewable power for the future. Cu(In,Ga)(Se,S)₂/CdS/ZnO has proven itself to be a worthwhile photovoltaic combination achieving record efficiencies upwards of 20%. These record efficiencies have been obtained through vacuum-based approaches. However, the high cost of vacuum-based fabrication processes become a barrier to affordable solar cells as an energy source. As a result, solution-based approaches have become the focus of many studies which have achieved record efficiencies of 15% for copper-indium-gallium-di sulfide (CIGS) materials.
In this study, we have investigated various approaches to improve CIGS solar cells after thin film deposition. CIGS devices have been fabricated by a hydrazine solution based process. CIGS was deposited in a Cu poor-rich-poor sequence with a Ga rich layer at the Mo surface to improve contact resistance. After thin film deposition, post-deposition treatments of sulfurization and wet chemical etching were studied with focuses on the change of material structures and physical properties. Sulfurization has shown to increase grain size and band gap of the absorber layers at higher temperatures or prolonged treatment times. This property change has shown a direct impact on open circuit voltage of the solar cell devices. Wet chemical etching eliminates remaining small crystals of CIGS which act as recombination centers and improves the charge carrier collection. Through these post-deposition processes, improved quality of CIGS materials can be obtained and therefore the associated solar cell devices show better performance.

Both CIGS and CZTS,Se based PV devices have voltage deficits that grow with increasing band gap (Eg). In all transient photo-capacitance (TPC) spectra published on CIGSe and CZTS devices a transition is observed from a defect with a Gaussian distribution centered at 0.8+/-0.05eV above the valence band. This commonality indicates that either the signal originates somewhere other than the bulk absorber in these otherwise similar devices, or that a very similar defect may be found in these distinct but related materials. In this study, TPC spectra are obtained for CIGS and CZTS devices with varied device architectures. In particular, devices with ZnO,S buffers in place of CdS are examined to determine whether the deposition of these layers affects the formation of the 0.8eV defect.
Recently, temperature and intensity dependent open-circuit voltage measurements (Voc(T, G)) have indicated that the increased voltage deficit in large-gap CIGS is dominated by interface recombination, while in low-gap CIGS bulk recombination dominates [Appl. Phys. Lett. 103, 093502 (2013)]. Meanwhile, a primary difference between the highest performance CZTS and CIGS devices is that the voltage deficit in CZTS grows more slowly with increasing Eg than in CIGS but has a much larger initial value. Taken together, these observations indicate that identifying this deep defect and its location within the device may help increase Voc in low-gap CIGS and perhaps all CZTS devices. By comparing the TPC signal in devices with identical absorbers but alternate buffer and emitter layers, we will determine whether this signal originates in the bulk of the absorber. Comparison of TPC spectra with other techniques such as Voc(T,G) and photo-luminescence will provide a better understanding of what role the 0.8eV defect plays in limiting Voc.

Fabrication of CIGS thin film based on a precursor solution has been attracted because it is possible to make low cost and tunable band gap thin film. In this study, in particular, multiple coatings with two different precursor solution pastes were applied to achieve more efficient fabrication of CIGS thin films (fewer coating and drying cycles) and an improved solar cell performance. Paste A using polyvinyl acetate as a binder material is enable to make a thin and dense layer for high efficiency solar cell performance but required several coating and drying cycles for the proper film thickness. On the other hand, paste B using ethylcellulose lead to form relatively thick films in spite of less number of spin-coating process but the porosity of CIGS films limited the solar cell performance. Multiple coating with the two pastes conducted three type of configurations of CIGS films (A + B, B + A, and A + B + A) for a combination of the advantage of each paste. The A + B configuration showed the best solar cell efficiency (4.66%) compared with other configurations. The details will be discussed in the presentation.

Chalcopyrite semiconductors such as Cu(In1-x,Gax)Se2 (CIGS), which are promising active material for highly-efficient thin-film photovoltaics (PV), have significantly non-uniform optical and electrical properties both as a function of depth (in the substrate-normal) and in the in-plane direction. These variations in the refractive index, absorption coefficient, carrier mobility, lifetime, trap-density and recombination-velocities determine the absolute power conversion efficiency as well as the performance variation of individual cells. In this work, we present a rigorous experimental and theoretical approach that accurately measures, with high-spatial resolution, both vertical and lateral charge collection probabilities via wavelength-specific, diffraction-limited excitation of CIGS PVs. The photocurrent and consequently the external quantum efficiency generated from the above excitations are measured under various bias and cell temperatures. Afterwards, we utilize robust optical models calculated using the Finite-Difference Time-Domain (FDTD) method, with experimentally measured optical constants obtained from spectroscopic ellipsometry, to calculate the optically generated charge-carrier density profile. This profile, together with the experimentally measured quantum efficiency, yields the charge collection probability as a function of depth within the CIGS cell. Furthermore, this depth analysis is carried out laterally in high resolution by rastering the cell with micron-sized steps, yielding the three-dimensional charge collection probability of the cell. We present the dependence of the spatially resolved collection probability on the individual grain-specific vertical and lateral variation in the CIGS absorber layer.

It is well known that CdTe thin films are successfully used in solar cells¹. We study CdTe thin films deposited by conventional CSS technique and annealed under gas mixtures based on CHClF₂ gas. The CdTe films annealed by using Ar-CHClF₂-O₂ show incorporation of Cl in the bulk . Also, they have a smooth surface, with a compact structure and thin grain boundaries. CdCl₂ grains are formed at grain boundaries after annealing in mixtures with CHClF₂; the CdTe surface of as-grown and annealed in Ar samples is composed by CdTe, TeO₂ and CdO; in samples annealed in Ar-CHClF₂ and Ar-CHClF₂-O₂ the surface is composed by CdTe, TeO₂, CdO, TeCl₂O and CdCl₂. However, oxidized compounds significantly decrease after annealing in the mixture with O₂.
We show in photovoltaic parameters of solar cells based in CdS/CdTe how the passivation treatment is related to its better performance. Additionally, for analysis, were used photoelectron X-ray spectroscopy and scanning electron microscopy with high- resolution.
1.www.firstsolar.com/innovation/advanced-thin-film-modules, (2013).
sect;.-This work has been supported by CONACYT-México and CONCIYTEY-Yucatan under contracts FORDECYT-116157 and FOMIX-170098. Measurements were performed at LANNBIO CINVESTAV-Mérida,under support from projects FOMIX-Yucatán 2008-108160,CONACYT LAB-2009-01 No. 123913 and CB2012/178947. The authors wish acknowledges O. Goacute;mez, W. Cauich, D. Aguilar and D. Huerta for technical support and L. Pinelo for its secretarial assistance.

Cu(In,Ga)(S,Se)2 (CIGS) photovoltaic absorber of having high absorption coefficient of about 105/cm has been proved to give the highest conversion efficiency of 20.8% among thin film and multicrystalline Si solar cell [1]. In commercial manufacrtuing of CIGS photovoltaic module, the absorber films are usually produced by sputtering and selenization/sulfurization process. During th igh temperature selenization above 450 oC, the surface of Mo back contact is easily transformed to Mo-selenide (MoSex), which is beneficial for ohmic contact of Mo/CIGS. However, an excessive MoSex of being thicker than 200 nm may induce a high device resistance and therefore, controlling of its thickness is necessary.
The electrical properties of CIGS as an intrinsic semiconductor depend heavily on its defect chemistry. Although a high temperature selenization process is favorable for suppressing of harmful donor formation, it may lead to formation of excessive MoSex layer and decrease the solar cell efficiency.
In this study, for the purpose of strengthening the advantage of high temperature selenization and controlling of the MoSex thickness, a noble multilayer of Mo/MoNy/Mo was used for the back contact of CIGS solar cell. The conducting compound of MoNy as an diffusion barrier against Se was deposited by a reactive sputtering. The thickness and N content of MoNy was found to be linearly proportional to N2 gas flow rate, which suggests that MoSex could be easily adjusted by a simple modification of Mo formation process. The CIGS absorber films were obtained by solid state selenization, where a sputtered CuInGa alloy film on Mo/MoNy/Mo/glass was subsequently selenizated at 450~600 oC by using Se vapor. The formation behavior of MoSey and the performance of ZnO:Al/ZnO/CdS/CIGS/Mo/MoNy/Mo/glass solar cell, will be discussed in terms of The dependence of the thickness and N content of MoNy film.
Acknowledgements
The authors appreciate the financial support from "Development of 25% Efficiency Grade Tandem CIGS Thin Film Solar Cell Core Technology" of MKE (Ministry of Knowledge Economy) and ISTK (Korea Research Council for Industrial Science and Technology) of Republic of Korea. This work was supported by the New & Renewable Energy of the Korea Institute of Energy Technology Evaluation and Planning (KETEP) grant funded by the Korea government Ministry of Knowledge Economy.(No. 20123010010130)
Reference
[1] http://www.nrel.gov/ncpv/images/efficiency_chart.jpg

In CdTe solar cells the absorption in the CdS layer reduces the flux of high- energy photons to the CdTe absorber. One solution to this problem is to use semiconductor with wider bandgap as window material. ZnS with 3.7 eV bandgap can be suitable for this purpose, but it is highly resistive, and the lattice mismatch with CdTe is about 16%, which causes an increase of defects at the interface. This difficulty can be compensated by i) the formation of ternary compound ZnCdTe at the ZnS/ CdTe interface, or ii) using ZnS/CdS bilayer as window layer [1] among others methods.
In this work, we have studied the solar cells structures ZnS/CdTe and ZnS/CdS/CdTe by low temperature photoluminescence (PL) technique and room temperature photoacoustic (PA) spectroscopy. We analyze PL spectra in the sub- band- gap regime, which reflects defect- associated recombination processes, in order to investigate the quality of the interfaces. The results were correlated with the PA spectra, which give information about non-radiative recombination processes in these structures. Both, ZnS and CdS thin films were deposited using rf- sputtering and the CdTe thin film was deposited using closed space sublimation technique. The PL and PA studies were carried out on as-deposited as well as annealed ZnS/ CdTe and ZnS/CdS/CdTe structures in chlorine atmosphere. We report the influence of this treatment on defect behavior and device efficiency.
[1] Junfeng Han et al, Journal of Physics and Chemistry of Solids 74 (2013) 1879

In the aim to reduce the costs, the photovoltaic industry must attempt to extend the mean lifetime of solar panels. Polycrystalline CdS/CdTe solar cells are considered an excellent option for PV conversion at low cost; however, device stability of this solar cell has been given little attention. Several causes have been proposed to explain cell degradation, such as junction degradation, degradation of the electrical contact to the CdTe and shunting. A variety of degradation mechanisms may be occurring in CdS/CdTe solar cells. The most suspected cause for cell instability is the Cu diffusion from the back contact into the CdS region.
In this work we investigated the degradation of CdS/CdTe solar cells by mean of electrical pulses. The cell condition was monitored using I-V and spectral quantum efficiency techniques. Voc values remained constants, consistent with no PV junction degradation, while a decrease in Isc was observed associated with an increase in the series resistance.

Developments in have been meteoric over the last 2 years, with small-area efficiencies surpassing 15%. We address the fundamental issue of how organic-inorganic lead halide-based perovskite solar cells operate by applying a scanning electron microscopy-based technique, Electron Beam-Induced Current, EBIC, to cell cross sections, obtained by careful fracturing, so as to minimize to eliminate damage from the process. EBIC is used to map the variation in efficiency of charge separation and collection in pristine areas of cross-sections, using low beam voltages (~ 1.5 keV) and the electro-voltaic effect. The cells achieve their performances because of the high electronic quality of the absorber, which functions as (“i”) in a “p-i-n” junction configuration, with high efficiency of charge separation at/near the absorber/hole-blocking-layer, and at/near the absorber/electron-blocking-layer interfaces, with the former more pronounced for the CH3NH3PbI3-xClx (i.e., prepared with a mixed PbI3-PbCl3 solution; the role of the Cl is the subject of a separate study). For the CH3NH3PbI3 the junction efficiencies are reversed. Proper energy level alignment at the interfaces is an obvious sine qua non. If the electron blocker is replaced by a gold contact, only a heterojunction at the absorber/hole-blocking interface remains. This mechanisms immediately explains the high voltages (relative to the optical bandgap of the absorber) that these systems yield and provide guidelines for further development, based on understanding and to find new materials. Scanning probe microscopy measurements served to complement and confirm the conclusions from the EBIC studies.

Thin film CdTe solar cells are a successful photovoltaic (PV) technology in the market today owing to their effective optical absorption and inexpensive fabrication processes. Currently, the module efficiency is asymp;13 %, well below the theoretically estimated maximum efficiency (asymp;30 %). Inhomogeneity of the PV materials plays a key role in determining the power conversion efficiency, and thus a firm understanding of the correlation between microstructural properties and the overall PV properties is required to optimize the cell performance. In this work, we investigate local PV properties of polycrystalline CdTe solar cells focusing on grain bulk, grain surface, and grain boundaries using electron beams (3 keV - 30 keV). Real-time spectroscopic ellipsometry was applied to study the evolution of the thin film optical structure during sputter deposition of n-CdS / p-CdTe layers on a TCO (transparent conductive oxide) coated glass substrate. Following light and dark current-voltage measurements, we performed local characterizations using electron beams for high (> 13 %) and low efficiency (< 6 %) devices that were fabricated under the same processes. Electron beam induced current (EBIC) was used to measure local carrier collection efficiency with a spatial resolution as high as asymp;20 nm both on the top surface and in cross-section of the device. High-resolution EBIC data (acceleration voltage of 5 keV) reveals that the highest carrier collection occurs near the p-n junction in the high efficiency device, whereas in the middle of CdTe layer for the low efficiency device. The EBIC contrasts at grains/grain boundaries in these devices are also compared. The measured local electronic properties are correlated to microstructural morphology, orientation (Electron Back Scattered Diffraction characterization), and chemical composition (Energy Dispersive X-ray spectroscopy). Furthermore, we perform 2D model drift-diffusion simulations to obtain the magnitude of downward band-bending near grain boundaries (with typical magnitude of 0.2 eV).

We investigate the photo-stability and thermal-stability in inert atmosphere of polymer solar cells(PSCs) based on the blend of poly(4,8-bis-alkyloxybenzo(1,2-b:4,5-bprime;) dithiophene-2,6-diyl- alt-(alkyl thieno(3,4-b) thiophene-2-carboxylate)-2,6-diyl) (PBDTTT-C) and [6,6]-phenyl C70- butyric acid methyl ester (PC70BM). The increase of ideality factors and the decrease of built-in potential under sustained-illumination indicate the trap-assisted recombination plays a dominant role in the device photo-degradation. The variation of the phase separation in the active layer under the thermal stress results in the decrease of the absorption spectra and inefficient exciton dissociation and charge-carrier transport, leading to thermal-degradation.

III-V compound multi-junction (MJ) solar cells have high efficiency potential of more than 50% due to wide photo response, while limiting efficiencies of single-junction solar cells are 31-32%. In order to realize high efficiency III-V compound MJ solar cells, understanding and controlling imperfections (defects) are very important. This paper reviews fundamentals of defects and defect management for III-V MJ, space and concentrator solar cells.
There are several loss mechanisms to be solved for realizing high-efficiency III-V MJ solar cells. 1) bulk recombination loss, 2) surface recombination loss, 3) interface recombination loss, 4) voltage loss, 5) fill factor loss, 6) optical loss, 7) insufficient photon energy loss. The origins of those losses and key technologies for reducing those losses are presented. Key technologies for reducing the above losses are high quality epitaxial growth, reduction in density of defects, optimization of carrier concentration in base and emitter layers, double-hetero (DH) junction structure, lattice-matching of active layers and substrate, surface and interface passivation, reduction in series resistance and leakage current, anti-reflection coating, photon recycling and so forth. In order to solve energy loss problems due to thermalization or lack of absorption, multi-junction (tandem) structure is essential.
The key issues for realizing super-high-efficiency MJ solar cells are 1) sub cell material selection, 2) tunnel junction for sub cell interconnection, 3) lattice-matching, 4) carrier confinement, 5) photon confinement, 6) anti-reflection in wide wavelength region and so forth. Especially, DH structure have been found to effectively prevent from impurity diffusion from tunnel junction and high tunnel peak current density has been obtained by the authors.
InP and InP-related materials such as InGaP and InGaAsP have been found to be superior radiation resistant compared to GaAs and Si and have unique light-illumination defect annealing phenomena by the authors. As a result, InGaP-based MJ solar cells have been industrialized for space use even in Japan. Origins of radiation-induced defects in InP and InP-related materials, and radiation damages to solar cells are discussed.
For concentrator applications by using MJ cells, the cell contact grid structure should be designed in order to reduce the energy loss due to series resistance, and tunnel junction with high tunnel peak current density is necessary.
The conversion efficiency of inverted epitaxially grown InGaP/GaAs/InGaAs triple-junction solar cells has been improved to 37.9% (1-sun, AM1.5G) and 44.4% (250- 300 suns) as a result of proposing double-hetero structure wide-band-gap tunnel junctions, and inverted epitaxial growth. Concentrator photovoltaics (PV) is expected to contribute as one of major PV as well as the first crystalline Si PV and the second thin-film PV.

Measurement of the bulk minority carrier lifetime (tau;B) requires a high-quality surface passivation to minimize the contribution from the surface recombination velocity (SRV). Passivation schemes used to date are: iodine ethanol/methanol (I-E), quinhydrone/methanol (QH), SiO2, a-Si, SiN:H, and Al2O3. There is a great deal of information in the literature on each of these methods. However, they have been studied primarily to make a single measurement on the wafer. However, in many cases, it is required to have spatial maps of tau;B. As an example, a recent need has arisen to study defect/impurity distributions and their influence on tau;B, in 156 mm x 156 mm, N-type, CZ wafers. Because of their high life time (3-5 ms) and thinness (180 mu;m), this requires that the passivation be very high quality and uniform over the entire wafer. We have done a detailed study of each of these passivation methods on high lifetime (range 3 to 24 ms) wafers to evaluate their passivation quality, stability, and uniformity. This paper will discuss the results of this study. The salient results are:
(a) Iodine-ethanol gives the lowest SRV both on P and N-type wafers, but the measurement window is only a few minutes, which shrinks with higher lifetime. Highest lifetime we have measured is 26 ms on FZ wafers.
(b) Quinhydrone is capable of producing low SRV. However, it has an “incubation time” of about 30 min. Because the degree of passivation also depends on the thickness of the QH layers on the wafer surfaces, it is difficult to attain uniform passivation.
(c) Aluminum oxide gives a superb passivation on N-type wafers. The uniformity is limited by the deposition equipment. We have also observed some instabilities over time and as a result of illumination. The quality of passivation is as good as IE.
(d) We have not been able to get a-Si passivation to be as good as IE, QH, or Al2O3.
(e) Thermal SiO2 has given us the best uniformity and passivation. Because SiO2 growth is a high temperature process, it is necessary to minimize the high temperature process time. We have tested passivation characteristics of different thicknesses of the oxide and found that about 200 Å of the oxide can give stable passivation to map tau;B of N-type wafers. We have used photoluminescence mapping to establish that this process does not produce measurable changes in the wafer structure itself.
Wafer preparation and oxidation process are very important in attaining uniform passivation. We will describe these in detail and also discuss application of using passivation to generate tau;B maps and to relate them to the structure of high-tau;B, N-type wafers.

Engineering and optimizing high-efficiency thin-film solar cells requires an intimate knowledge of each layer in the device structure. One common concern for thin-film solar cells is interfacial intermixing between layers, particularly for Cu(In,Ga)(Se,S)₂ (CIGS) and Cu₂ZnSn(Se,S)₄ (CZTS) absorbers that contain mobile atoms like Cu. Therefore, understanding the interfacial defects and mixed phases is critical to achieving the best performance for a given photovoltaic device. Here we present a first principles study to investigate the thermodynamic, electrical, and optical characteristics of absorber-related defects in CdS and ZnS, two prototypical buffer layer materials. We use density functional theory to calculate the structures, formation energies, migration barriers, and energy levels of each defect, from which we identify their role(s) in carrier recombination and device metastabilities. The extent of interfacial point-defect intermixing is also correlated with the buffer layer growth conditions in terms of film stoichiometry.
This work performed under the auspices of the U.S. Department of Energy by Lawrence Livermore National Laboratory under Contract DE-AC52-07NA27344.

GaAs is a promising material for high efficiency photovoltaic devices but traditional MOCVD or MBE methods of growth have not proven cost-effective for large-scale unconcentrated energy conversion. The close space vapor transport (CSVT) method of film growth has potential to be a low-cost alternative that does not rely on toxic or pyrophoric precursors.(1-2) We report for the first time solid-state solar cells fabricated using CSVT to deposit n-GaAs emitters on p-type substrates and p-type CSVT-grown films.
Initial results give a maximum open circuit voltage of 770 mV under one sun illumination, which is similar to the open circuit voltages achieved using photoelectrochemical measurements on liquid junction cells. Internal quantum efficiencies for solid-state devices are much smaller than liquid junction cells, however, and may be due to defects at the electronic junction or to surface recombination. We report on devices grown under conditions designed to minimize oxygen at the junction between the CSVT film and substrate as well as the results of surface passivation. Characterization of the junction and bulk material is performed using capacitance-voltage profiling, admittance spectroscopy, drive-level capacitance profiling, and transient photocapacitance spectroscopy. (3)
References:
(1) Ritenour, A. J.; Cramer, R. C.; Levinrad, S.; Boettcher, S. W. Efficient n-GaAs Photoelectrodes Grown by Close-Spaced Vapor Transport from a Solid Source. ACS Appl. Mater. Interfaces 2012, 4, 69-73.
(2) Ritenour, A. J.; Boettcher, S. W. Towards High-Efficiency GaAs Thin-Film Solar Cells Grown via Close Space Vapor Transport from a Solid Source. IEEE PVSC 38, 2012.
(3) Boucher, J. W.; Miller, D. W.; Warren, C. W.; Cohen, J. D.; McCandless, B. E.; Heath, J. T.; Lonergan, M. C.; Boettcher, S. W. Optical Response of Deep Defects as Revealed by Transient Photocapacitance and Photocurrent Spectroscopy in CdTe/CdS solar cells. Submitted to Solar Energy Materials and Solar Cells.

Ge wafers have been employed as substrates for high efficiency III-V multi-junction solar cells because of the almost ideal lattice match. However, the cost of Ge wafers is high and the Ge bandgap is too small which limits its contribution to the overall voltage output. Since Si has much higher bandgap and lower cost than Ge, using epitaxial Ge on Si as virtual substrates for III-V material growth is a promising solution. Si could be employed as the bottom cell and therefore boost the overall voltage output. In this work we deposit the Ge films on Si by sputtering which is an inexpensive process without the requirement of ultra-high vacuum. Sputtering also avoids the use of toxic gas and is potentially capable for fabricating large area films. Only very thin layers of Ge can be grown defect free on Si substrates owing to the 4.2% lattice mismatch between Si and Ge. Such layers are biaxially strained to adapt its lattice constant to that of the underlying Si substrate. When layer thickness is above the critical thickness, dislocations will nucleate with misfit segments near the interface and threading segments running through the layer to the surface. The threading dislocations density might be very high for a fully relaxed Ge layer directly grown on top of Si substrates. In this work, the grown Ge layers have to be beyond the critical thickness for fully strain relaxation to work as virtual substrate. Defect blocking and thermal annealing are investigated to reduce the dislocation density of the epitaxial Ge layers. Defect blocking by epitaxial necking could be used to obtain a defect-free top Ge surface on Si substrate. Selective epitaxial growth and defect crystallography are important to force defects to the oxide sidewall, resulting in a perfect top layer. Since threading segments have a 45^o angle to the underlying Si substrate, if the aspect ratio of the holes in the oxide mask is greater than 1, threading segments will be blocked by the oxide sidewall. This goal could be achieved in a relative small film thickness by using ultrathin porous alumina membranes as mask for substrate patterning. In addition, thermal annealing could be introduced to reduce the threading dislocation density. After a 780^oC/900^oC-10 min cyclic thermal annealing the threading dislocation density of the Ge layer could be reduced by a factor of 10. X-ray diffraction(XRD) and Transmission Electron Microscopy (TEM) measurements are employed to investigate the dislocation density reduction of the epitaxial Ge layers.

To date, high efficiency multijunction solar cells have been developed on Ge or GaAs substrates for space applications, and terrestrial applications are hampered by high fabrication costs. In order to reduce this cost, we propose a breakthrough technique of III-V compound heteroepitaxy on Si substrates without generation of defects harmful for PV applications.
The integration technique is based on the epitaxial lateral overgrowth of microscale GaAs crystals on a 0.6 nm thick SiO2 layer from nanoscale Si seeds. The nucleation from nanoscale openings through the SiO2 was found to avoid the emission of misfit dislocations and the formation of antiphase domains. Consequently, the interface between the GaAs crystal and the SiO2 layer remains perfectly sharp and free of defects. On Si(001) substrate, the only defects found by transmission electron microscopy in GaAs islands were pairs of twins. Each island contains a pair of {111} twin planes sectioning its volume into a central part in direct epitaxy with the Si susbtrate and two twinned parts above the SiO2 layer. A simple model based on the anisotropy of zinc blende crystal with respect to the [110] and [ 110] directions was proposed to explain their formation. On the one hand, each As atom in the (1-10) plane is bonded with Ga atoms at its top right and left sides and with two Ga bottom atoms in the (110) plane. Consequently, as the growing crystal meets the edge of the SiO2 layer, the As (001) plane can be directly linked to the upper Ga (001) plane and the growing GaAs crystal can wet the SiO2 layer without perturbation regarding the chemical bonds. On the other hand, each As atom in the (110) plane is bonded with two Ga at its bottom right and left side, and with two Ga top atoms in the (-110) plane. If we consider now the lateral growth of GaAs crystal along the [-110] direction, one can notice that an As atom at the edge of the SiO2 layer cannot bind with a fourth bottom Ga atom, and thus exhibits a dangling bond. In this case, this As atom may "attract" a neighbour Ga atom situated on the SiO2 layer in order to insure the neutrality along the As [110] column at the edge of the SiO2 layer. The resulting bond distorsion provokes the anticlockwise rotation of the GaAs lattice which leads to the {111} twin plane formation.
A way to avoid twinning is to overcome the above mentioned effect of anisotropy of the zinc blende structure by starting the growth on a Si (111) substrate. In this case, perfect integration of heterogeneous GaAs on oxidized Si (111) substrates was achieved, and will be presented. Complementary analyses such as micro-Raman and micro-photoluminescence spectra were also performed and confirm the very good quality of microscale GaAs grown on Si.

GaAs is an attractive material for high-efficiency photovoltaics, but large-scale production is impeded in part by the cost and toxicity of gas-phase precursors (e.g. arsine and trimethylgallium) employed by metal-organic chemical vapor deposition (MOCVD). Close-spaced vapor transport (CSVT) of GaAs is a plausibly scalable process similar to commercial CdTe deposition which uses water vapor to generate gas-phase As₂ and Ga₂O in-situ at atmospheric pressure with high (~1 µm/min) growth rates and ~95% overall transport efficiency.^{1, 2} We report epitaxial GaAs films with controllable p- and n-type doping deposited via CSVT from potentially inexpensive GaAs powder sources. We studied the films' defects using a combination of impedance spectroscopy and secondary ion mass spectrometry (SIMS) to show that the doping of CSVT GaAs films can be controlled via addition of dopant powders to the source powder.³
Photovoltaic measurements of n- and p-GaAs films were performed using liquid junction photoelectrochemical (PEC) cells. The 1-sun efficiency for CSVT GaAs photoanodes in ferrocene/ferrocenium was eta; ~ 10% with no correction for resistive and optical losses. Spectral response in the PEC configuration was analyzed to determine the minority carrier diffusion lengths, which were typically ~3 µm and ~5 µm for n- and p-GaAs films, which are significantly longer than those for control n- and p-GaAs wafers (0.6 µm and 1.1 µm) and competitive with MOCVD-grown films.
Device-physics modeling of a CSVT GaAs pn junction solar cell based on the measured diffusion lengths and Hall mobilities predicts that devices with eta; ~25% should be achievable given our current material quality. Our first CSVT GaAs pn junctions exhibited 1-sun V_OC nearly 800 mV despite the high (~ 10¹⁸ cm^-3) concentration of oxygen defects at the pn junction interface, which was profiled using SIMS. Efforts to understand and reduce the defects present in solid-state CSVT pn junctions are underway and will also be discussed as they progress.
1. Ritenour, A. J.; Cramer, R. C.; Levinrad, S.; Boettcher, S. W., Efficient n-GaAs Photoelectrodes Grown by Close-Spaced Vapor Transport from a Solid Source. ACS Appl. Mater. Interfaces 2011, 4, 69-73.
2. Ritenour, A. J.; Boettcher, S. W. In Towards High-Efficiency GaAs Thin-Film Solar Cells Grown via Close Space Vapor Transport from a Solid Source, Proc. 38th IEEE Photovoltaic Spec. Conf., 2012; 913-917.
3. Ritenour, A. J.; Boucher, J. W.; Greenaway, A. L.; Boettcher, S. W., Control of p/n doping of GaAs Films deposited via Vapor Transport from Powder Sources. Manuscript in preparation.