Date Published: January 29, 2018
Publisher: John Wiley and Sons Inc.
Author(s): Jianjun Li, Dongxiao Wang, Xiuling Li, Yu Zeng, Yi Zhang.
As a promising candidate for low‐cost and environmentally friendly thin‐film photovoltaics, the emerging kesterite‐based Cu2ZnSn(S,Se)4 (CZTSSe) solar cells have experienced rapid advances over the past decade. However, the record efficiency of CZTSSe solar cells (12.6%) is still significantly lower than those of its predecessors Cu(In,Ga)Se2 (CIGS) and CdTe thin‐film solar cells. This record has remained for several years. The main obstacle for this stagnation is unanimously attributed to the large open‐circuit voltage (VOC) deficit. In addition to cation disordering and the associated band tailing, unpassivated interface defects and undesirable energy band alignment are two other culprits that account for the large VOC deficit in kesterite solar cells. To capture the great potential of kesterite solar cells as prospective earth‐abundant photovoltaic technology, current research focuses on cation substitution for CZTSSe‐based materials. The aim here is to examine recent efforts to overcome the VOC limit of kesterite solar cells by cation substitution and to further illuminate several emerging prospective strategies, including: i) suppressing the cation disordering by distant isoelectronic cation substitution, ii) optimizing the junction band alignment and constructing a graded bandgap in absorber, and iii) engineering the interface defects and enhancing the junction band bending.
Concerning the rapid increase of global energy demand and the growing severe environmental issues caused by the consumption of traditional fossil fuels, clean, safe, and renewable energy sources are in urgent need. Photovoltaic (PV) technology that can directly delivers the inexhaustible solar energy to clean electricity is considered to be an attractive solution.1 Taking advantages of less source material cost and adapted to flexible substrates,2 thin film PV technology is an important branch to make solar electricity more cost‐effective and prevalent in various application situations, namely building integrated photovoltaics (BIPV), unmanned aircraft systems,3 wearable power supply, and so on.4 Cu(In,Ga)Se2 (CIGS) and CdTe thin‐film solar cells have demonstrated over 20% power conversion efficiency (PCE).5, 6, 7, 8, 9, 10 However, the scarcity of In and Te elements and environmentally hazardous Cd constrain the large‐scale commercialization and the reduction in production costs.11, 12, 13, 14 In the last decade, the new emerging thin‐film photovoltaic devices based on kesterite structural semiconductors Cu2ZnSn(S, Se)4 (CZTSSe) have attracted considerable attention, owing to its large potential as a candidate for high‐performance photovoltaic technology with earth‐abundant source materials.11, 12, 15, 16, 17 Inheriting the device structure from the predecessor chalcopyrite CIGS solar cells,18, 19 and benefitting from the advanced first‐principle material calculations which give clear directions of chemical potential window, energy band, and defects of kesterite material family,11, 20, 21, 22 CZTSSe thin‐film solar cells have experienced a significant increase in power conversion efficiency from about 5% in 2004 to the record 12.6%.15 However, the current record efficiency of CZTSSe devices has been pinned at 12.6% for four years. In other word, the development of CZTSSe solar cells is experiencing a stagnation. However, this kind of stagnation is not rare in the 40 years’ history of well‐developed CIGS and CdTe thin‐film solar cells.23
The multinary feature of CZTSSe system introduces a large possibility of abundant point defects in these materials.11 As shown in Figure 1, predicted by the density functional theory (DFT) calculation, various point defects including antisite defects and vacancies may exist in the CZTSSe system. In addition, the formation energy of the charge‐compensated defect complexes is expected to be smaller than that of individual defects because of the charge transfer and attractive Coulomb interaction between positive and negative charged defects.21 For instance, the formation energy of CuZn+ZnCu clusters is particularly low, which makes these defect clusters rather prevalent in CZTSSe materials. By forming charge‐compensated defect clusters, the point defect density could be reduced effectively. However, this self‐compensation effect also induces an adverse factor: the formation energy of 2CuZn+SnZn clusters is quite small and these clusters can be in high concentration even in stoichiometric samples with Cu/(Zn+Sn) and Zn/Sn ratios near 1, though the formation energy of individual deep defects SnZn is relatively large. As a result, high concentration deep trap states are introduced and the electronic band gap is reduced.43
The bandgap of absorber material is another important factor, which has significant influence on the VOC deficit and related device performance. Theoretically, the bandgap of CZTSSe and CIGS should be adjusted to about 1.4–1.5 eV, which is the optimal value that expected to achieve highest theoretical efficiency according to the S–Q limit.28, 73 However, the current practical optimal bandgap of CZTSSe devices is much lower (1.1–1.2 eV).7, 8, 74 CZTSSe solar cells with high S content and large bandgap close to 1.5 eV usually output lower efficiency than the CZTSSe solar cells with small bandgap. One issue should be considered is the degradation of electronic properties of absorber as the bandgap is enlarged by tuning the ratio of cation or anion. For example, the transition energy level of GaCu is deeper than that of InCu, which dramatically deteriorates the device performance if the ratio of Ga/(Ga+In) is too high (Ga/(Ga+In) > 0.4).75 Analogously, the transition energy level of point defects (including the dominant CuZn antisite defects) in CZTS is also deeper than that in CZTSe with small bandgap,23 which leads to more nonradiative recombination loss in the bulk and grain boundaries. Another issue that limits the practical optimal bandgap of the CZTSSe absorber is the undesirable CBO at the buffer/absorber interface, especially for the pure sulfide CZTS solar cells.11 Despite these challenges, preparing absorber layer with large bandgap (1.3–1.5 eV) and excellent electronic property is a promising direction to explore more potential of kesterite solar cells. Besides, with respect to the engineering of bandgap, another prospective strategy to improve the photovoltaic performance of kesterite solar cells is constructing a graded bandgap similar to the “V” type bandgap grading of CIGS solar cells,6, 8, 9 which is regarded as the optimal solution to balance the trade‐off between Voc and Jsc. In this section, we discuss the recent efforts paid to optimize the bandgap of the absorber layer and improve the VOC of devices by cation substitution.
Besides the abundant bulk defects and associated disordering and band tailing, the defects type at the interface and related pining of Fermi level plays another critical role in the charge transport, separation, and extraction, which have a significant effect on the overall photovoltaic performance including VOC, JSC, and fill factor.34, 65, 154, 155 For the typical n+–p single‐side abrupt junction‐based thin film solar cells without interface defects, the band bending extent of the p‐type absorber layer is mainly determined by the carrier concentration of absorber.88, 156 In this case, the interface Fermi level locates close to the CBM of the absorber layer. However, the interface defects are inevitable and play an essential role in device operation in practice.157, 158, 159
In summary, this review provides an overview of recent advances of the efforts to break through the VOC limit of low cost and environment‐friendly kesterite‐based quaternary chalcogenide solar cells, with a particular focus on the strategies based on cation substitution. Though the kesterite thin‐film solar cells have experienced a rapid progress in the last decades, there is still a long journey to catch up with its predecessors CIGS and CdTe solar cells which have demonstrated more than 20% PCE. The main obstacle in the way is unanimously attributed as the large VOC deficit (generally > 0.6 V), which is now commonly attributed to the band and potential fluctuation caused by the substantial band tailing, as verified by commonly observed PL redshifting. The prevailing CuZn and ZnCu antisite defects and associated disordering with relatively low formation energy are recognized as the primary culprit that accounts for the band tailing and related large VOC deficit. Besides the cation disordering and associated band tailing, another issue needs to be addressed is the nonoptimized band alignment at the junction interface and the absorber bandgap in order to reduce Voc deficit. It is necessary to control over a delicate balance among the absorber bandgap matching with solar spectrum, the bandgap‐related electronic properties, and the conduction band offset at the junction interface. One more critical issue may be responsible for the large VOC deficit is the pinning of interface Fermi level to a low energy level by a large population of CuZn acceptor‐like defects, which reduces the band bending in the absorber and produces a weak junction, deteriorating both VOC and fill factor.
The authors declare no conflict of interest.