Synthesis and characterization of Ag2SxSe1−x nanocrystals and their photoelectrochemical property

Changyin Ji; Yu Zhang; Xiaoyu Zhang; Peng Wang; Hongzhi Shen; Wenzhu Gao; Yiding Wang; William W Yu

doi:10.1088/1361-6528/aa523c

1. Introduction

Infrared materials are of great importance in solar cell, detector and biological probe applications, because the infrared light (beyond 700 nm) occupies about 70% photon flux of the whole sun light energy and it can penetrate deep tissues without the influence of bioluminescence. Previous works about infrared materials are mainly concentrated on CdTe [1, 2], InAs [3, 4], PbS and PbSe [5–10] nanocrystals (NCs) and their variations [11, 12]. However, the intrinsic toxicity of these materials restricts their applications. Thus, it is urgent to find environmentally friendly materials [13]. I–VI semiconductor NCs (Ag₂X, X = S, Se, Te), as typical near infrared (NIR) materials, have some unique properties such as low toxic [14–16], high ionic and electronic mobility [17], and high electrical conductivity [18, 19]. Ag₂X NCs provide a broadband absorption in the NIR region, which is an intrinsic advantage to improve the efficiency of the photovoltaic devices. Although there are quite a few reports about photovoltaic device structures, they performed poor power conversion efficiencies (PCEs) [20–25].

Compared with binary NCs, ternary alloyed NCs have some superior properties. For example, bandgap tuning is crucial in photovoltaic devices. With the composition control of ternary alloyed NCs, one can obtain different bandgap materials without changing the particle size [26–28]. Alivisatos' group reported PbS_xSe_1−x NC photovoltaic devices and the results showed that the ternary alloys were more efficient than pure binary alloy NCs based devices [29–31]. Zhong and coworkers described CdSe_xTe_1−x alloyed NCs sensitized solar cells with greater than 6% efficiency and high stability [32]. Moreover, ternary alloys supply a unique way to investigate the micro interaction of exciton and hole or other electron−hole decoupling [33, 34].

In this work, monodisperse Ag₂S_xSe_1−x ternary NCs with successive size tunability were synthesized using a modified synthesis method. Different ratios of high reactivity precursors (TMS)₂S and (TMS)₂Se were used to adjust the band-gap of the NCs. High reactivity precursors can overcome the intrinsic lattice mismatch that play important roles in crystal growth orientation and final morphology. Furthermore, the photoelectrochemical performance, as an evaluation index, shows that composition doping bring about an impressive promotion to photocurrent. Benefited from composition regulation, we obtained different band-gap materials without size change, which is of great importance to the photovoltaic devices.

2. Experimental section

2.1. Materials

Bis(trimethylsilyl)sulfide ((TMS)₂S, 98%, Sigma Aldrich), bis(trimethylsily)selenide ((TMS)₂Se, 98%, Sigma Aldrich), silver acetate (AgAc, 99.99%, Alfa Aesar), myristic acid (MA, 98%, J&K Scientific), octylamine (OAm, 99%, J&K Scientific), trioctylphosphine (TOP, 90%, Sigma Aldrich), 1-octadecene (ODE, 90%, Sigma Aldrich), 20 nm and 400 nm TiO₂ nanoparticles (≥99.8%,Veking), titanium butoxide (98%, Sinopharm Chemical Reagent Co. Ltd), 1,2-ethanedithiol (EDT, 98 + %, Alfa Aesar), Zn(NO₃)₂, Na₂S, hexane, methanol, toluene, and acetone were used as received.

2.2. Preparation of (TMS)₂S-TOP and (TMS)₂Se–TOP precursors

The precursors of (TMS)₂S–TOP and (TMS)₂Se–TOP were prepared in a glove box. The total number of moles of S and Se was kept at 0.40 mmol and the two precursors with different atomic ratios (S/(S + Se) = 0, 0.25, 0.5, 0.75, 1) were mixed with 10 ml TOP in a glass vial with vigorous stirring, then the precursor solution was transferred into a 20 ml syringe.

2.3. Synthesis of Ag₂S_xSe_1−x NCs

We modified the synthesis route of Ag₂S_xSe_1−x NCs based on a method reported by Pang et al [35]. The synthetic process was performed by adding 0.4 mmol AgAc powder, 6.4 mmol MA, 9.6 mmol OAm and 25 ml ODE in a 100 ml three-neck flask under continuous stirring. Then the solution was degased by vacuum for at least 1 h. After this procedure, the system was protected by N₂ atmosphere throughout the reaction. Followed the purging progress, the solution was heated to 80 °C for 5 min, and 5 ml OAm was injected. When the AgAc powder completely dissolved, the solution turned to be transparent. The reaction system was then cooled down to 50 °C, and the (TMS)₂S–TOP and (TMS)₂Se–TOP mixture was injected into the flask swiftly. After the injection, the temperature was adjusted to 40 °C, the color of the solution immediately changed from transparent to dark, indicating the formation of the NCs. The temperature of the reaction system was set at 40 ± 0.3 °C for 5 min to guarantee the stable growth of the NCs.

2.4. Purification of Ag₂S_xSe_1−x NCs

The reaction was quenched by injecting 10 ml 4 °C anhydrous toluene and immersing the flask into an ice-water bath. The as synthesized NCs were purified by adding 25 ml of anhydrous methanol/acetone to make the NCs precipitate out and then centrifuged at 6000 r min⁻¹ for 20 min This step was repeated at least three times. The gained brown-black product was dissolved in 5 ml hexane and centrifuged again at 12 000 r min⁻¹ for 3 min to remove any big size of impurity. At last, the pure product was filtered by a 0.22 μm filter and transferred into a glove box for further use.

2.5. Characterizations

The microstructures of the samples were measured by using field emission scanning electron microscopy (FESEM, JEOL JSM-6700). A JEM 2100 F was used to measure the transmission electron microscopy (TEM) and high resolution transmission electron microscopy (HRTEM). The acceleration voltage was 200 kV. The samples of TEM were made by dropping the NCs (dissolved in hexane) onto carbon film coated copper grids and let it dry. The absorption and photoluminescence spectra were recorded by Shimadzu UV-3600 spectrophotometer and Omni-λ300 monochromotor/spectrograph, respectively. X-ray diffraction (XRD) patterns were obtained using a Bruker D8 Advance XPert diffractometer (Cu Kα: λ = 1.5406 Å), in the range of 10°–70° at a scan rate of 4 ° min⁻¹. Inductively coupled plasma mass spectrometry (ICP-MS) was performed on Agilent/7700. X-ray photoelectron spectroscopy (XPS) was conducted on an ESCALAB250 spectrometer. The highest occupied molecular orbital (HOMO) energy level was measured by using high resolution ultraviolet photoelectron spectroscopy (PREVAC XPS/UPS System, R3000/VUV5K/MX-650). The film thickness was measured by Profile-system (DEKTAK 150, Vecco). Each cycle of NC film consists of successive immersion of the clean glass in NC solution for 1 min and washing with 1,2-ethanedithiol (EDT) between each step. Ten cycles were applied and the thickness was measured to be about 100 nm, so the thickness of a single layer film was about 10 nm.

2.6. Photoanode fabrication

FTO glass (Nippon Sheet Glass, Japan, resistance 14 Ω/square, transmittance >90% from 400 to 700 nm) was cleaned using aqueous soap solution, deionized water, ethanol and acetone for 30 min, and finally immerged into ethanol for use. TiO₂ nanorods (NRs) were grown on FTO glass by a hydrothermal method reported in the literature [36]. Briefly, 60 ml concentrated hydrochloric acid (36.5%–36.8% by weight) was added to 60 ml double-distilled water under vigorous stirring for 5 min, then 1.2 ml titanium butoxide was added to the mixture. After continous stirring, 30 ml of the mixture was transferred into a 100 ml Teflon-lined stainless steel autoclave. Three pieces of FTO glass were placed against the wall of the Teflon-liner with a slope. After treated with hydrothermal method at 180 °C for 14 h, the FTO glass was rinsed with double-distilled water and ethanol, and immersed into 50 mM TiCl₄ aqueous solution for 30 min at a 50 °C water bath. The TiCl₄-treated sample was annealed at 500 °C for 2 h. Two layers (an active layer and a scattering layer) of TiO₂ nanoparticles (NPs) were then modified onto FTO/TiO₂ NRs. The TiO₂ paste (particle size ∼20 nm) was applied onto the FTO/TiO₂ NRs as an active layer by a blade. The film was dried in a muffle furnace at 80 °C for 1 h and followed by annealing at 500 °C for 2 h. The scattering TiO₂ layer (particle size ∼400 nm) was coated onto the active layer by a blade, then the electrode was dried in the muffle furnace at 80 °C for 1 h and annealed at 500 °C for 2 h.

2.7. Electrophoretic deposition (EPD)

EPD was employed to fabricate FTO/TiO₂ with Ag₂S_xSe_1−x in toluene [37]. A FTO/TiO₂ photoanode and a graphene electrode were placed in an electrochemical cell with a distance of 4 mm, then 150 V cm⁻¹ bias was used to supply a constant voltage. The thickness of the NC layer could be adjusted by the EPD time and the NC concentration. After the modification, the electrode was rinsed with ethanol and dried in N₂ flow.

After the EPD, we used the successive ionic layer adsorption and reaction method to coat the FTO/TiO₂/Ag₂S_xSe_1−x with a ZnS layer. Briefly, the FTO/TiO₂/Ag₂S_xSe_1−x photoanode was dipped into cationic sources (Zn(NO₃)₂ solution, 0.1 M in methanol) and anionic sources (Na₂S solution, 0.1 M in V_methanol:V_water = 1:1) for 1 min, respectively. This process was repeated for two cycles. Between each cycles, the photoanode was rinsed with deionized water.

2.8. Photoelectrochemical measurements

Photoelectrochemical measurements were performed in a quartz cell (20 mm × 20 mm) containing a sulfide/polysulfide electrolyte ( ${{\rm{S}}}^{2-}/{{{\rm{SO}}}_{3}}^{2-},$ 0.25 M Na₂S and 0.35 M Na₂SO₃ dissolved in deionized water). FTO/TiO₂/Ag₂S_xSe_1−x, Ag/AgCl (3 M KCl), and Pt foil electrode were employed as the photoanode, reference and photocathode, respectively. The photovoltaic performances of photoelectrochemical cells were evaluated using an electrochemical workstation (Corrtest CS150). An AM 1.5 light with a power of 100 mW cm⁻² (from Zolix SS150 solar simulator) was employed as the illumination source. The efficiency of the light source at different wavelengths was measured using (Newport) QE/IPCE measurement kit with a silicon detector.

3. Results and discussion

It is not easy to achieve monodisperse ternary NCs because many factors will influence the result in microscopic scale, such as the reactivity of passivation group, the temperature, and the concentration of precursors. Among these factors, the relative reactivity of two competitive precursors (S and Se in this case) is of great importance [26]. Precursors with different activities will lead to separated nucleation. So it is hard to obtain ternary alloyed NCs and the size homogeneity of the NCs usually deteriorates. Take this factor into consideration, one should choose the precursors with comparable reactivity to reduce the individual self-nucleation, and adjust the molar ratio of both precursors to make the desired composition in the NCs. S powder and Se powder dissolved in trioctylphosphine (TOP) were firstly selected as the precursors. The TEM images displayed bad size distribution, and serious self-nucleation of Ag₂S and Ag₂Se. (TMS)₂S and (TMS)₂Se were finally selected as the ideal precursors. Because trimethylsilyl terminated S and Se atoms have high and comparable reactivity, which would ensure the homogeneous nucleation of ternary NCs at low temperature. Figure 1 exhibits the TEM and element mapping images. To gain strong element signals, we exchanged the myristic acid and octylamine ligands to butylamine and washed the NCs more than 30 min The nucleation was a fast step in a few seconds, with growth step carrying on; the comparable reactivity of two precursors made them react in a similar ratio of feeding atoms. Finally, we gained nearly uniform sized NCs as shown in figure 1 with good crystallinity. Figure 1(a) is the TEM image of Ag₂S_0.66Se_0.34 NCs, and the white frame zone was selected to scan atom mapping image. HRTEM indicated the lattice spacing of 0.223 and 0.243 nm corresponding to the (103) and (013) planes of orthorhombic phase, respectively. The separated maps in figures (b)–(d), illuminate that Ag, S and Se were distributed evenly in the NCs.

**Figure 1.** (a) TEM image and HRTEM (inset) of Ag₂S_xSe_1−x NCs. Element maps of (b) Ag, (c) Se and (d) S in blue, red and green, respectively.
Download figure:
Standard image High-resolution image

ICP-MS was applied to analyze the actual atomic ratio of S and Se anions in the product. 3 ml aqua regia was used to dissolve 6 mg samples and then diluted by 25 ml deionized water. The result in figure 2(a) is the proportion of S atom (S/(S + Se)) in precursors and products collected from 9 samples. When the molar ratio of the precursor was set at 1:1, the S and Se found in the product was close to 1:1 (i.e., Ag₂S_0.48Se_0.52). It is attributed to the comparable reactivity of Se precursor and S precursor that the graph shows a better linearity than the previous reported result [31]. The S element analysis in figure 2(b) demonstrates the trend of S composition varied a little in the beginning of the reaction, and quickly reached a steady level of (S/(S + Se)) ≈ 0.5.

**Figure 2.** (a) ICP-MS data show the relative amount of S in the product versus the relative amount of S in the precursor solution. (b) S element in NCs taken from products reacted for 0–30 min.
Download figure:
Standard image High-resolution image

The absorbance and photoluminescence spectra of Ag₂S_xSe_1−x NCs are shown in figures 3(a) and (b) with a strong excitonic absorption peak. The photoluminescence peak positions are located at 1026 nm, 933 nm and 847 nm for Ag₂Se, Ag₂S_0.48Se_0.52 and Ag₂S NCs. The calculated Stokes shift is around 180 nm for the three NCs. The spectrum shifted to red with the increase of Se composition. As Vegard's law [38] says, this shift of alloyed materials can be described as E_alloy = χE_A + (1 − χ)E_B, where χ is the molar fraction of component A, E_A, E_B, and E_alloy are the band gap energies (or other physic properties) of pure A, pure B, and the alloyed material, respectively. Figures 3(a) and (b) reveal that both of the photoluminescence and absorption spectra follow the Vegard's law. Fortunately, the precursors' high reactivity ensures the smooth growth of the ternary NCs even at low temperature, thus we did not observe an 'optical bowing' effect as reported in CdSe_xTe_1−x alloys synthesis [39].

**Figure 3.** (a) Absorbance and photoluminescence of Ag₂Se, Ag₂S, and Ag₂S_0.48Se_0.52. (b) Absorbance spectra of alloyed NCs with gradient varied ratio of S and Se in precursors.
Download figure:
Standard image High-resolution image

Phase identification of Ag₂S, Ag₂S_0.48Se_0.52, and Ag₂Se NCs in figure 4 were carried out by XRD measurement. The XRD patterns show that monoclinic Ag₂S formed in our synthesis and all the characteristic diffraction peaks can match the JCPDS card very well (JCPDS, 14-0072). The lattice fringe of Ag₂S ₍₋₀₁₃₎ with spacing of 0.238 nm is a dominant orientation which is also observed in figure 6. The diffraction peaks of Ag₂Se NCs correspond to the orthorhombic Ag₂Se in the β form (JCPDS, 24-1041). The lattice fringe of Ag₂Se ₍₁₂₁₎ is corresponding to the fringe spacings of 0.258 nm in figure 6. The characteristic peaks of Ag₂S_0.48Se_0.52 NCs appear between the two pure materials (JCPDS, 27-0620), which is a gradient chang between Ag₂Se and Ag₂S NCs. Two dominant orientations of Ag₂S_0.48Se_{0.52 (013)} and Ag₂S_0.48Se_{0.52 (103)} with fringe spacings of 0.243 and 0.223 nm are also observed in figure 6. The broadenings of the peaks of NCs disclosed the nanocrystalline nature of the products, which primarily related to the size of NCs which are much smaller than their bulk materials.

XPS was measured to investigate the surface composition and valence condition of the NCs. Figure 5(a) is the low-resolution scan spectrum of Ag₂S_0.48Se_0.52 NCs. Figures 5(b)–(d) are the high-resolution scan spectra of the Ag 3d, S 2p, Se 3p, and Se 3d. The binding energy of Ag (3d_3/2, _5/2) emerges at 373.60 and 367.60 eV, respectively, suggesting that the oxidation state of silver is univalent Ag⁺ in the NCs [40]. As for Se element, the peak of Se (3d_{3/2, 5/2}) is observed at 53.51 eV. Because the peaks of S and Se elements have an overlap, it is essential to separate them by a fitting method. After fitting, Se 3p_1/2 and Se 3p_3/2 appear at 159.32 and 165.33 eV, respectively. So the resolved signals eliminate the existing of mutiple oxidation states of selenium within the peak range. In other words, the valence of selenide is Se²⁻ [41, 42]. S 2 P appears at 160.45 and 161.32 eV based on the possible peak fitting result, which demonstrates the existing of S²⁻. All the experiment data are consistent with the previous reports [41, 43].

The TEM and HRTEM images in figure 6 show the shape, structure, and size of the NCs. The size distributions of the NCs were nearly monodisperse. The sizes of the spherical NCs were 5.2, 6.8 and 9.2 nm, respectively. With the ratio of S/(S + Se) decreased, the lattice spacing changed from 0.238 to 0.243 and finally to 0.258 nm. The same trend was also observed from XRD results in figure 4. The uniform lattice structures and nanocrystal sizes demonstrate that there is no core/shell structure or stack faults. In other words, S and Se atoms grew within the ternary NCs.

The SEM images in figure 7(a) reveal that the FTO surface covered with compact TiO₂ crystal seeds. TiO₂ NRs grow above the seeds uniformly. The length of TiO₂ NR is about 1 μm and the diameter is between 80 and 120 nm, respectively. The planar density is about 8–12 wire μm^–2. Figure 7(a) gives the tilted cross view of the TiO₂ NRs treated with TiCl₄. In the image, we can observe that the TiO₂ NRs covered densely, uniformly and nearly vertical on the FTO substrate. The quadrangular prism morphology reveals that the TiO₂ NRs are tetragonal crystal structure. After being treated with TiCl₄, TiO₂ NCs with diameter ∼5 nm grow onto the surface of TiO₂ NRs. This procedure effectively enhances the specific surface area of NRs, which helps the large scale coating of Ag₂S_xSe_1−x NCs onto the TiO₂ NRs surface. Figures 7(b) and (c) demonstrate that Ag₂S_xSe_1−x NCs were deposited on the TiO₂ NRs. After the deposition process, the space between the NRs was compactly covered with Ag₂S_xSe_1−x NCs. The rough surface of the NRs became even rougher after overgrown with Ag₂S_xSe_1−x NCs. The HRTEM in figure 6(d) show the interface of the TiO₂ NRs and Ag₂S_xSe_1−x can be clearly distinguished for a high crystallinity. The lattice fringes of rutile TiO_{2 (101)} and Ag₂S_0.66Se_{0.34 (013)} with spacings of 0.246 and 0.243 nm are denoted in the figure 7(d). The approximate lattice parameter can ensure the stable growth of two crystals from dynamic aspect.

UPS was used to measure the kinetic energy of Ag₂S_xSe_1−x NCs. The NCs were coated on a piece of clean glass and the film thickness was controlled to be 10 nm. Figure 8 shows the Fermi energy levels. The peak absorption values taken from figure 3 are used as the band-gap (2.12 eV for Ag₂S, 1.97 eV for Ag₂S_0.66Se_0.34, 1.85 eV for Ag₂S_0.48Se_0.52, 1.61 eV for Ag₂S_0.21Se_0.79 and 1.39 eV for Ag₂Se). We calculated the HOMO and lowest unoccupied molecular orbital (LUMO) energies of Ag₂S_xSe_1−x NCs. Here, Ag₂S_xSe_1−x NC photoelectrochemical cells are N type solar cell structure. The LUMO energy level of NCs must higher than the LUMO energy level of TiO₂ NRs. Based on the precise energy level positions, we could design the structure of photoelectrochemical cell more reasonably. The LUMO and HOMO energy levels were −3.47 and −5.60 eV for Ag₂S, −3.49 and −5.46 eV for Ag₂S_0.66Se_0.34, −3.43 and −5.28 eV for Ag₂S_0.48Se_0.52, −3.45 and −5.06 eV for Ag₂S_0.21Se_0.79, and −3.44 and −4.84 eV for Ag₂Se, respectively.

**Figure 8.** UPS data of Ag₂S, Ag₂S_0.66Se_0.34, Ag₂S_0.48Se_0.52, Ag₂S_0.21Se_0.79 and Ag₂Se NCs. The left panel shows the amplified views of the high binding energy regions and the right panel displays the amplified views of the low binding energy regions.
Download figure:
Standard image High-resolution image

Finally, we tested the photoelectrochemical cell performance using Ag₂S_xSe_1−x NCs with different ratios of S and Se. Figure 9(a) shows that the Ag₂S_0.66Se_0.34 NCs have an excellent performance of 2.034% IPCE due to its highest photocurrent in figure S1(b) (supporting information), and ternary alloy NCs give higher PCEs than binary NCs in figure 9(b). The interpretation is that ternary alloy can broaden the absorption spectrum region of Ag₂S NCs and the photocurrent will be higher with the increase of photogeneration carrier concentration. At the same time, ternary alloy can balance the transfer of photo-excited electrons through both interfaces of the TiO₂/Ag₂S_xSe_1−x and the Ag₂S_xSe_1−x/electrolyte by minimizing electron–hole recombination. When the composition of Se increased, the band-gaps of the ternary alloy NCs decreased. The obtained highest photocurrent value is 5.11 mA cm⁻². With further increase of Se composition, we observed a decrease in photocurrent due to the electron–hole recombination in figure S1(b) (supporting information). Taking the absorption factor into consideration, we prolonged the reaction time of the NCs to gain relative bigger NCs, which is a good way to promote the photon absorption. During the composition adjusting process, the LUMO energy level stayed steady with the gradual variation of S and Se composition (figure 8), which is consistent with the variation trend of open voltage in figure S1(a) (supporting information). Benefited from the use of composition hybridization and bis(trimethylsilyl) group, the surface passivation ligand is stable to the ambient atmosphere and electrolyte solution. The interpretations is that the morphology change decreased the aggregation of the NCs [44]. It can also supply a stable photocurrent under continuous work in figure S2 (supporting information). On the basis of these results, composition adjustment helps us find the optimal value of ternary NCs that supplies the highest PCE. Alloying has been proved to be a simple and useful method to achieve target band-gap and maintain the binary advantages.

4. Conclusion

In summary, we have synthesized and characterized monodispersed Ag₂S_xSe_1−x NCs. HRTEM and XRD confirmed the lattice change; the lattice parameter gradually shifted from Ag₂S to Ag₂S_xSe_1−x and finally to Ag₂Se NCs. The HOMO and LUMO energies were also determined by UPS and absorption spectra for the precise design of photoelectrochemical cells. The preliminary photoelectrochemical cell study demonstrated that composition dependent of band-gap tuning of the ternary alloying helped to enhance the photocurrent and broaden the absorption spectrum region of Ag₂S_x NCs. It can also balance the transfer of photo-excited electrons through both the TiO₂/Ag₂S_xSe_1−x interface and Ag₂S_xSe_1−x/electrolyte by minimizing electron–hole recombination. This approach turned out to be an ideal strategy to design and tuning material characteristics.

Acknowledgments

This work was financially supported by the National Natural Science Foundation of China (51272084, 61306078, 61225018, 61475062, 61675086), the Jilin Province Key Fund (20140204079GX), and BORSF LA.

Synthesis and characterization of Ag₂S_xSe_1−x nanocrystals and their photoelectrochemical property

Article metrics

Submit

Permissions

Author e-mails

Author affiliations

Author notes

Dates

Peer review information

Abstract

1. Introduction