Romain Scaffidi1,2,3,Guy Brammertz3,1,Jessica de Wild3,1,Denis Flandre2,Bart Vermang3,1
Hasselt University1,UCLouvain2,imec3
Romain Scaffidi1,2,3,Guy Brammertz3,1,Jessica de Wild3,1,Denis Flandre2,Bart Vermang3,1
Hasselt University1,UCLouvain2,imec3
CZTSSe kesterite materials constitute a promising candidate for thin-film photovoltaics, based on abundant and non-toxic elements. Despite their high absorption coefficient and tunable bandgap for either single or tandem junctions, they suffer from significant V<sub>oc</sub> losses impeding their efficiency to level with competing technologies such as CIGS, CdTe and Perovskites. Two main culprits for this large V<sub>oc</sub> deficit are electrostatic fluctuations and high density of bulk defects, highly dependent on the absorber composition typically considered as ideal within well-established Cu-poor and Zn-rich ranges. An increasingly popular solution to tackle these two loss mechanisms and boost V<sub>oc</sub> is the alloying of Ge to substitute Sn, leading here to Cu<sub>2</sub>Zn(Sn<sub>1-x</sub>,Ge<sub>x</sub>)Se<sub>4</sub> (CZTGSe) compounds, the bandgap of which is tunable through conduction band level following the x=Ge/Ge+Sn ratio. This work focuses precisely on the opto-electrical characterization of Sn-Ge bandgap-graded CZTGSe solar cells at both the absorber and complete device levels.<br/><br/>A sequential process is implemented and optimized, involving first the physical evaporation of a CZTG metallic precursor stack, then pre-annealed in N<sub>2</sub> and finally selenized. We obtain CZTGSe absorbers with low surface roughness and acceptable grain size within a Cu-poor and Zn-rich composition range. They exhibit an average x ratio around 45%, with a clear Sn-Ge segregation appearing naturally during the annealing step that leads to x=70% at the back surface and x=20% at the front surface. This corresponds to a back bandgap gradient which should simultaneously improve carrier collection and limit interface recombination, similarly to state-of-the-art CIGS. Yet, the low carrier lifetime is likely limited by poor top surface quality with observed surface defects and ZnSe secondary phases.<br/><br/>Advanced electrical characterization is performed on complete CZTGSe solar cell samples to gauge the impact of this Sn-Ge back bandgap grading approach and detect potential loss mechanisms. An efficiency of 7.2% is attained, along with values of 484 mV, 34.5 mA/cm<sup>2</sup> and 43.3 % for V<sub>oc</sub>, J<sub>sc</sub> and FF, respectively. Even though the performance is lower than record kesterite devices, the J<sub>sc</sub> value is close to the Shockley-Queisser (SQ) limit for a confirmed minimum bandgap of 1.18eV, probably due to the largely enhanced carrier collection by our back bandgap gradient design. However, our devices are affected by important V<sub>oc</sub> and FF deficits, i.e. about 50% of their SQ limits, for which the analysis of dark and light IV curves suggest are mainly the consequence of high-density deep defects as well as dominant electrical parasitics. This is investigated more deeply via temperature-dependent admittance spectroscopy, enabling the identification of a relatively deep defect level at the CZTGSe/CdS interface with activation energy of 260 meV and capture cross section of the order of 10<sup>-16</sup> cm<sup>-2</sup>. Minority carrier recombination via this defect state is likely the explanation for the poor V<sub>oc</sub> and FF values observed. SCAPS-1D simulations are well in agreement with these experimental findings and allow to highlight this interface loss mechanism as the main limiting factor in our samples.<br/><br/>Eventually, Ge inclusion into kesterites effectively permits bandgap gradient to boost carrier collection and approach SQ limits in terms of J<sub>sc</sub>. However, our study reveals that performance is still largely restrained by V<sub>oc</sub> and FF losses at the CZTGSe/CdS interface which should be here the first area of improvement. This is preliminarily attempted via front surface sulfurization, but further investigations are required whether it concerns the Sn-Ge bandgap gradient design or the top interface quality using chemical treatments or alternative buffer materials.<br/><br/>This work has received funding from the European Union's Horizon 2020 research and innovation program under grant agreement 952982 (CUSTOM-ART project).