Alina
Senina
a,
Anatol
Prudnikau
a,
Angelika
Wrzesińska-Lashkova
ab,
Yana
Vaynzof
abc and
Fabian
Paulus
*ac
aLeibniz Institute for Solid State and Materials Research (IFW) Dresden, Helmholtzstraße 20, 01069 Dresden, Germany. E-mail: f.paulus@ifw-dresden.de
bChair for Emerging Electronic Technologies, Technische Universität Dresden, Nöthnitzer Str. 61, 01187 Dresden, Germany
cCenter for Advancing Electronics Dresden (cfaed), Technische Universität Dresden, Helmholtzstraße 18, 01069 Dresden, Germany
First published on 6th March 2024
Silver bismuth sulfide (AgBiS2) nanocrystals have emerged as a promising eco-friendly, low-cost solar cell absorber material. Their direct synthesis often relies on the hot-injection method, requiring the application of high temperatures and vacuum for prolonged times. Here, we demonstrate an alternative synthetic approach via a cation exchange reaction. In the first-step, bis(stearoyl)sulfide is used as an air-stable sulfur precursor for the synthesis of small, monodisperse Ag2S nanocrystals at room-temperature. In a second step, bismuth cations are incorporated into the nanocrystal lattice to form ternary AgBiS2 nanocrystals, without altering their size and shape. When implemented into photovoltaic devices, AgBiS2 nanocrystals obtained by cation exchange reach power conversion efficiencies of up to 7.35%, demonstrating the efficacy of the new synthetic approach for the formation of high-quality, ternary semiconducting nanocrystals.
The first attempts to fabricate AgBiS2 based solar cells were made using the sequential ionic layer adsorption reaction (SILAR) technique, demonstrating that AgBiS2 can be used as a potential solar sensitizer.21 The synthesis of AgBiS2 QDs by hot injection method was initially proposed in 2013 by Chao Chen et al.22 In 2016, Bernechea et al. presented the first AgBiS2-based solar cells synthesized by this method with a PCE of 6.3%.23 Further research was focused on optimizing the synthesis by employing various precursor ratios, changing the injection temperature, and using different capping ligands.24–26 These attempts in combination with optimized hole and extraction layers resulted in a record result PCE of 9.17%, which further solidified AgBiS2 as a highly promising eco-friendly alternative to toxic chalcogenides.27 However, despite all the advantages of the hot injection method, it requires both vacuum and high temperatures (typically 100 °C) for several hours, which remains a challenge not only in terms of the thermal budget for QD synthesis, but also the ability to reproducibly synthesize large volumes for commercial applications.
Several studies have explored more inexpensive alternative routes for the AgBiS2 synthesis. For instance, Wu et al. proposed a simple synthesis of AgBiS2 nanorods by spin-coating the precursor solutions and subsequent annealing of the films at 300 °C in air.28 Gu et al. demonstrated the synthesis of compact AgBiS2 thin films via the formation of Ag-Bi-thiourea-DMSO molecular inks followed by decomposition upon thermal annealing.29 In addition, Akgul et al. presented the first solar cells with AgBiS2 QDs fabricated at room temperature by simple mixing of metal halides with separately dissolved sulfur in oleylamine, resulting in a PCE of 5.55%.26 These promising attempts highlight the importance of low-cost and low-temperature synthetic routes for AgBiS2 QDs in order to pave the way for their large scale commercial application.
Although direct synthesis allows the rapid preparation of ternary AgBiS2 QDs, this approach is fraught with difficulties due to the presence of precursors of two cations with different reactivities in the reaction system. The different reactivities limit the simultaneous control over size, shape, and composition of the resulting QDs. Cation exchange can be a suitable alternative for a low-temperature synthesis. The cation exchange (CE) reaction is a well-studied method for various material systems in which a guest cation replaces the host cation with subsequent incorporation into the nanocrystal structure.30 Organic ligands such as phosphines or carboxylates act as soft bases that interact with the softer cation and facilitate its removal from the nanocrystal.31,32 Modifying the composition of nanocrystals without changing their size and shape makes cation exchange an extremely versatile tool for the synthesis of various materials.
The cation exchange approach has been successfully employed in the synthesis of ternary material compounds.30,33,34 In the case of CE, easily obtainable binary chalcogenides such as Cu2S, Ag2S, and CdSe are used as a template to further incorporate guest cations (e.g., In3+, Hg2+) and result in the formation of ternary compounds with precise compositions.35–37 The outcome of a CE reaction is dictated by subtle thermodynamic and kinetic factors enabling the control over the structure and composition of the desired product.33 The choice of size, shape, and structure of the parent nanocrystals as well as the choice of solvent, additional ligands, and temperature of the reaction allows not only to tune the morphology and composition of the final products,38,39 but also to prepare various heterostructures such as Cu–In–Se/ZnS core–shell nanoparticles or CdS/Cu2S binary nanorods.40,41 Despite the widespread use of CE for synthesizing ternary nanoparticles, the synthesis of AgBiS2 QDs using this approach has not yet been reported.
Here, we demonstrate a facile synthesis of small Ag2S nanocrystals at room temperature and subsequent cation exchange at mild temperatures to form AgBiS2 QDs. The use of bis(stearoyl)sulfide (St2S) as air stable and highly reactive sulfur precursor enables the synthesis of small, almost monodisperse Ag2S nanoparticles at room temperature under ambient conditions. Bismuth neodecanoate (Bi(neo)3) in combination with the soft Lewis base trioctylphosphine (TOP) facilitates the cation exchange of Ag+ to Bi3+ at only 50 °C in air. We show that the cation exchange progresses rapidly under these conditions and the size and shape of the nanocrystals does not change throughout the reaction. The AgBiS2 QDs obtained by cation exchange were utilized in planar solar cells that reach a PCE >7%, which is the highest PCE for AgBiS2 QDs synthesized under ambient conditions.
![]() | ||
Fig. 1 Schematic representation of Ag2S QDs synthesis and formation of AgBiS2via cation exchange at mild temperature. |
The kinetics of the cation exchange reaction was monitored using absorbance spectroscopy (UV-VIS), X-ray diffraction (XRD), and transmission electron microscopy (TEM). Fig. 2 shows the TEM images of Ag2S QDs before and throughout the exchange, taken at different times of the cation exchange reaction. The average diameter of St2S-based Ag2S QDs is approximately ≈2.8 nm and exhibits a rather narrow size distribution (see Fig. S1†). Upon the incorporation of Bi3+ the average size and shape of the QDs remain the same throughout the reaction. It is also noteworthy that the introduction of TOP into the reaction system does not lead to the full or partial dissolution of the quantum dots, but rather leads to a reduction of larger agglomerates that are present at earlier times of the reaction probably due to a better colloidal stability of the QDs after CE. Consequently, with progressing time of the cation exchange reaction the size distribution becomes narrower (see Fig. 2 and Fig. S1, ESI†).
Fig. 3a shows the evolution of the optical properties of the QDs throughout the cation exchange process. The absorbance of QDs in solution slightly red shifts already after 5 minutes of cation exchange and undergoes only minute changes after 15 minutes. While an absorption shift usually implies a change in the size of the quantum dots for a given material, in this case, it could also be an indication of a change in the band gap. Since bulk, stoichiometric Ag2S exhibits a bandgap of 0.9 eV, while bulk AgBiS2 exhibits a slightly smaller bandgap of only 0.8 eV, a red shift, as observed, suggests a change in QD composition.43 However, the shift in bandgap will also be affected by variations of the individual QD stoichiometry and the corresponding exciton Bohr radius of each material, which complicates a full understanding of the observed red-shift. Continuous absorbance measurements at a wavelength of 300 nm throughout the cation exchange process in Fig. 3b support the aforementioned observation that the cation exchange process to AgBiS2 QDs progresses continuously with time and is not based on a full dissolution and new nucleation of nanocrystals. XRD measurements (Fig. 3c) confirm that the transformation from the monoclinic β-Ag2S,44 to cubic AgBiS2 occurs already within minutes. Subsequent growth in intensity and decrease in peak width for reaction times up to 30 minutes could be an indication of an improvement in the crystallinity of the QDs.
![]() | ||
Fig. 3 (a) Absorbance spectra of Ag2S-based QDs throughout the cation exchange reaction to AgBiS2 at different reaction times, normalized to the absorbance at 300 nm, (b) comparison of absorbance at 300 nm, and (c) XRD patterns of original Ag2S QDs and AgBiS2 QDs at different times of CE reaction. Ag2S and AgBiS2 reference spectra in panel c correspond to crystallographic information obtained from ref. 44 and 45 (Ag2S CCDC 1692248 and AgBiS2 CCDC 1612237), respectively. |
To examine the compositional changes of the QDs upon cation exchange, X-ray photoemission spectroscopy (XPS) measurements were performed. The Ag 3d spectra (Fig. 4a) show the presence of a single doublet for both types of QDs, which confirm the absence of metallic silver in the QD films. Comparison of the Ag 3d spectra of Ag2S and AgBiS2 nanocrystals also reveals the absence of a clear shift in the binding energy upon cation exchange. In the spectrum of S 2p (Fig. 4b), on the other hand, a slight change in the peak position to lower binding energies upon cation exchange is observed, which may indicate a weaker interaction between silver and sulfur after bismuth incorporation. The Bi 4f spectrum consists of two sets of doublets, one associated with AgBiS2 and an additional high binding energy doublet, which is attributed to the formation of bismuth oxide or hydroxylated bismuth species at the QDs surface.46 The binding energies of the Ag 3d (367.8 and 373.8 eV), S 2p (160.8 and 162.1 eV) and Bi 4f (158.0 and 163.3 eV) doublets that originate from AgBiS2 QDs are in agreement with previously reported data in literature.23,25,47
The calculated atomic ratios of Ag, Bi, and S in Ag2S and AgBiS2 QDs are shown in Fig. 4c. We observe a sharp decrease in Ag content and an increase in Bi content within the first 30 min of cation exchange, which corroborates a successful exchange of cations. This is followed by a stabilization in the Ag and Bi composition, with a small decrease in bismuth content for very long cation exchange times. The final stoichiometry after 120 minutes is Ag:
Bi
:
S = 0.8
:
0.6
:
1 (Table S1, ESI†). While this composition is clearly non-stoichiometric, AgBiS2 QDs synthesized using the hot-injection method using either St2S or TMS2S precursors result in an even larger deviation from the theoretical stoichiometry (Table S2, ESI†). The larger deviation in material composition in the hot-injection method is most likely a result from the formation of non-negligible amounts of silver oleates – and in the case of St2S, silver stearates – in these synthetic procedures, which cannot be easily removed from QDs through the employed redispersion/precipitation procedures.42 These findings highlight, that synthesis via the cation exchange route provides a significantly higher control over the desired QD composition with potentially higher purity.
The synthesis of Ag2S quantum dots using elemental sulfur S and TMS2S as precursors under identical synthetic conditions resulted in the formation of larger Ag2S QDs with sizes of 3.2 ± 1.0 nm and 3.5 ± 1.9 nm, respectively. Both precursors result in more polydisperse nanocrystals (see Fig. S2 and S3†). Attempts to perform a cation exchange for those Ag2S QDs under identical conditions did not result in the formation of pure, high-quality AgBiS2 QDs. For the largest Ag2S QDs no cation exchange could be achieved, indicated by no changes in the diffraction pattern of these nanocrystals (Fig. S2†). While the diffraction pattern changes in the case of Ag2S obtained by elemental sulfur, the absorbance spectrum differs from the one of AgBiS2 obtained from St2S-based Ag2S QDs. Consequently, the integration of QDs based on TMS2S and elemental sulfur as precursors resulted in very low photovoltaic (PV) performance (Fig. S4†). These results could indicate that the Ag2S quantum dot size is crucial for a successful cation exchange with Bi3+. To corroborate the size dependence, we synthesized larger Ag2S dots based on St2S. Quantum dots with an initial average size larger than 3 nm do not undergo the cation exchange reaction within 120 min under the same conditions (Fig. S5†). This suggests that only small Ag2S quantum dots (<3 nm) can be converted into AgBiS2 QDs sufficiently fast through cation exchange with Bi3+. We note that also other factors may influence the efficacy of cation exchange, such as the choice of ligands, the quantum dot stoichiometry and the presence of defects, however the investigation of these goes beyond the scope of the current study and is subject to future research.
To explore the photovoltaic performance of the QDs formed by cation exchange, they were integrated into solar cells whose architecture is illustrated in Fig. 5a. In short, a thin SnO2 layer was deposited as electron extraction layer (ETL) on top of prepatterned ITO substrates. QDs were deposited in a layer-by-layer process employing a solid phase ligand exchange until a final layer thickness of ≈35 nm was achieved. Next, an ultra-thin poly[bis(4-phenyl)(2,4,6-trimethylphenyl)amine] (PTAA) hole extraction layer (HTL) was deposited.27 The devices were completed with a thin layer of thermally evaporated MoO3 (that increases the conductivity of the PTAA HTL by p-doping48,49) and a Ag electrode. The PCEs and current density – voltage (J–V) curves of solar cells based on Ag2S (0 min of CE) and AgBiS2 (30 and 120 min of CE time) are shown in the Fig. 5b and c. It can be seen that while Ag2S (0 min) demonstrates a very low efficiency, probably due to surface recombination and poor alignment between the ETL and the energy levels of Ag2S. AgBiS2 QDs, on the other hand, exhibit a significant improvement reaching an average performance of 3.31% (best 4.27%) after 30 min of reaction. Interestingly, a further increase in the cation exchange reaction time leads to a stark enhancement in solar cell performance, reaching to an average PCE of 6.76% (best 7.35%). While the open-circuit voltage (Voc) is similar for both devices, the fill factor and especially the short-circuit current density are noticeably higher for the QDs after 120 minutes of CE compared to only 30 minutes of CE time (Table S3, ESI†).
Extending the reaction time of the cation exchange beyond 120 min causes the solar cell performance to drop significantly (Fig. S6†). While the exact origin of this decrease in performance remains elusive, the reaction conditions of the cation exchange and the presence of oxidation-sensitive TOP and Ag+-species might affect the AgBiS2 quality, composition or purity for such prolonged reaction times.
We would like to emphasize that the reported solar cell performance represents the performance level of the QD solar cell after one week. Our QD solar cells based on AgBiS2 obtained via the cation exchange process exhibit a similar increase in performance within the first few days of storage in ambient as in the case of QDs synthesised via the hot-injection synthesis (Fig. S7†).27 While the exact origin of this improvement remains an open question in the field, our findings suggest that the performance evolution of AgBiS2 solar cell seems to be a universal observation for this device architecture and independent of the synthetic route. Similar observations were also reported for other types of quantum dot based solar cells,50,51 with the origin of this behavior remaining under debate.
The increase in current density for solar cells of 30 min and 120 min of reaction time is in agreement with the external quantum efficiency (EQE) spectra shown in Fig. 5d, where a maximum EQE of 80% is achieved at 500 nm for the 120 minutes CE devices accompanied by a clear red shift of the EQE onset. Similar changes in the EQE spectra have been attributed in literature to be a consequence of increased cation homogenization within AgBiS2 quantum dots upon thermal annealing.27,44 We believe that a similar process gradually occurs as the duration of the cation exchange is increased. Cation exchange processes occur through the quantum dot factettes at the surface of the QDs. Within the first minutes of the cation exchange process Bi3+ is incorporated into the quantum dot crystal lattice resulting in separated Ag- and Bi-rich domains, while the latter are most likely be in the proximity of the nanocrystal's surface. As the reaction time increases, the overall cation distribution becomes more homogenous across the QD, resulting in a red-shifted EQE, higher short-circuit current and an overall better photovoltaic performance, similar to the observations made by Konstantatos and coworkers,27 albeit without the need for additional annealing. This simplified model is supported by a detailed examination of the XPS data for bismuth after different times of the cation exchange reaction (Fig. S8†). Higher amounts of bismuth at the nanocrystal surface would result in a higher fraction of hydroxylated or oxidic bismuth species after these QDs have undergone the normal washing procedure and film formation processing, while less of such species are to be expected for bismuth homogenously distributed throughout the entire AgBiS2 QD lattice. Fig. 6 depicts the process of homogenization of the cation disorder schematically and the evolution of the two bismuth species with respect to the reaction time. Our XPS studies show that the ratio between high-binding energy and low binding energy bismuth dramatically decreases by more than 50% with increasing cation exchange duration, while (as discussed above) the overall ratio of the total amount of bismuth to silver remains quasi stable for prolonged reaction times. The hydroxylated or oxidic bismuth species exhibit a doublet at higher binding energies of 158.6 and 163.9 eV as compared to the bismuth of the nanocrystal lattice (158.0 eV and 163.3 eV). These values correspond well to Bi2O3 and do not stem from the potentially unreacted Bi precursor (Fig. S9†). These data also correlate with previously published XPS spectra of the oxygen-induced degradation of AgBiS2, indicating the formation of Bi2O3.46
After cation exchange, AgBiS2 was purified by precipitation and redispersion with acetone and toluene 2 times. Then, QDs were dissolved in hexane and stirred overnight. The solution was centrifuged for 5 min at 10000 rpm. The precipitate was discarded, the solution was washed with acetone and centrifuged one time. Purified AgBiS2 QDs were dissolved in dry toluene.
Footnote |
† Electronic supplementary information (ESI) available. See DOI: https://doi.org/10.1039/d3nr06128k |
This journal is © The Royal Society of Chemistry 2024 |