PCCP View Article Online
Published on 17 February 2015. Downloaded by UNIVERSITY OF OTAGO on 28/07/2015 04:37:54.
PAPER
Cite this: Phys. Chem. Chem. Phys., 2015, 17, 8182
View Journal | View Issue
The role of MoS2 as an interfacial layer in graphene/silicon solar cells† Kejia Jiao, Chunyang Duan, Xiaofeng Wu, Jiayuan Chen, Yu Wang* and Yunfa Chen* The role of MoS2 as an effective interfacial layer in graphene/silicon solar cells is systematically investigated by varying MoS2 film annealing temperature and thickness. It is found that the power conversion efficiency (PCE) is increased by B100% from B2.3% to B4.4% with 80 1C annealed MoS2 film whereas it drops significantly to B0.6% with 200 1C annealed MoS2 film. The results are well explained based on the device energy band diagram. That is, the incorporation of MoS2(80) films leads to the formation of type II structure, facilitating hole transport; while valence band mismatch is formed
Received 18th January 2015, Accepted 13th February 2015
with MoS2(200) films due to the increase in the work function of MoS2. Besides, the PCE increases gradually with decreasing MoS2 film thickness, and ‘‘saturates’’ at about 2 nm. The PCE can be further
DOI: 10.1039/c5cp00321k
enhanced to B6.6% with the aid of silicon surface passivation. Our work demonstrates that MoS2 is an excellent interfacial layer to improve the PCE with low-temperature annealing (80 1C in air), which may
www.rsc.org/pccp
be helpful in developing efficient and low-cost G/Si solar cells.
Introduction Graphene/n-silicon (G/n-Si) Schottky-barrier solar cells have attracted substantial attention due to their advantages of lowtemperature processability, low cost and potential of high efficiency.1,2 However, the power conversion efficiency (PCE) of pristine G/n-Si solar cells is usually low (typicallyo3%, here pristine refers to G/n-Si solar cells without any treatments), and chemical doping of graphene is the most commonly used method to improve PCE.3,4 Recently, we have demonstrated that interface tailoring is also a powerful method to enhance PCE and graphene oxide (GO) is selected as a model interfacial material.5 However, GO only works after high-temperature annealing (i.e., 400 1C under protection atmosphere). Exploring the alternative material of GO, which can improve the PCE effectively using a low-temperature procedure (o100 1C), is desirable. Molybdenum disulfide (MoS2) is a 2D material composed of S–Mo–S planes held together by van der Waals forces. It has drawn increasing attention due to its fascinating properties that are acceptable for a wide range of applications, such as field-effect transistors, light-emitting diodes and catalysis.6–9 Recently, the applications of MoS2 in photovoltaic devices are also demonstrated. For example, by employing MoS2 as a light State Key Laboratory of Multiphase Complex Systems, Institute of Process Engineering, Chinese Academy of Sciences, Beijing 100190, P.R. China. E-mail: [email protected], [email protected] † Electronic supplementary information (ESI) available. See DOI: 10.1039/ c5cp00321k
8182 | Phys. Chem. Chem. Phys., 2015, 17, 8182--8186
harvesting material, a PCE of 1.8% can be achieved for MoS2/Au Schottky junction solar cells.9 In addition, MoS2 has also been used as hole/electron transport layers in organic photovoltaic devices.10,11 The advantages of MoS2 over GO as an interfacial material are that on one hand, it is chemically inert with fewer traps; on the other hand, MoS2 has a higher hole mobility than GO, which may reduce annealing temperature. However, to date, there is no report on MoS2 as an interfacial layer in G/n-Si solar cells. Herein, MoS2 is integrated into G/n-Si solar cells as an interfacial layer. The device based on MoS2 shows a PCE of B4.4%, comparable to that of a GO-based device (B4.6%), yet only low-temperature annealing (i.e., 80 1C in air) is required. The role of MoS2 is investigated by varying MoS2 film annealing temperature and thickness. It is found that the incorporation of MoS2 can improve the PCE by forming a type-II heterojunction, which facilitates carrier transport between MoS2 and n-Si. The PCE can be further enhanced to B6.6% with the aid of silicon surface passivation.
Experimental section Synthesis of multilayer graphene (MLG) As reported previously,12 MLG is synthesized using a lowpressure chemical vapor deposition (LPCVD) method. Briefly, the system is first heated to 1000 1C for 10 min under the protection of hydrogen gas to clean the copper surface, then methane is introduced for 20 min.
This journal is © the Owner Societies 2015
View Article Online
Paper
PCCP
MoS2 film preparation
Published on 17 February 2015. Downloaded by UNIVERSITY OF OTAGO on 28/07/2015 04:37:54.
MoS2 nanoflake solution is purchased from Graphene Supermarket. It is spin-coated onto a silicon surface and annealed under two different conditions, i.e., at 80 1C in air (denoted as MoS2(80)) or at 200 1C under the protection of argon (denoted as MoS2(200)) for a fixed 30 min. The devices incorporating MoS2 are then labeled as G/MoS2(80)/n-Si and G/MoS2(200)/n-Si. Fabrication of solar cell devices n-type Si (2–4 O cm 1, with a square window of 3 mm 3 mm surrounded by 300 nm thick SiO2) is used as substrates. MoS2 is spin-coated onto Si substrates at a certain speed and then annealed at a different temperature. Subsequently MLG is transferred onto the top of MoS2 using the PMMA-assisted wet transfer method.13 Ga–In liquid (99.999%) and Ag paste are applied as the rear electrode and the front electrode, respectively. Characterizations The solar cells are tested under Air Mass 1.5 illumination (100 mW cm 2, the light density is calibrated using a standard solar reference cell, SRC-1000-TC-QZ, VLSI Standards S/N: 10510-0305). The J–V data and incident photon-to-electron conversion efficiency (IPCE) spectra are recorded using a Keithley 2400 SourceMeter and QTEST STATION 500AD, respectively. Raman Spectra are performed using Renishaw inVia-Reflex with 532 nm wavelength incident laser light. Ultraviolet photoelectron spectroscopy (UPS) measurements are made using an unfiltered HeI (21.22 eV) gas discharge lamp to determine the work function of graphene and MoS2. The UV-Vis spectrum is acquired using a Shimadzu UV-VIS-NIR Spectrophotometer UV-3600. X-ray photoelectron spectroscopy (XPS) measurements were performed using a monochromatic Al Ka source (1486.6 eV). The charge-transfer resistance is measured by electrochemical Impedance Spectroscopy (EIS), which is measured using CHI 760E, Shanghai Chenhua Instrument Co., Ltd, the measurements are carried out under illumination (AM 1.5G condition) by applying a bias that is equal to Voc. The MoS2 film thicknesses are obtained using an ellipsometer, SENTECH SE850, and the aperture diameter is 0.2 mm.
Fig. 1
The effects of MoS2 film annealing temperature on PCE.
Here 200 1C is chosen because it is high enough to vary the electronic properties of MoS2.14 As shown in Fig. S2 (ESI†), the Si substrates are fully covered by MoS2 and there are little changes in the morphologies of MoS2 films annealed at 80 and 200 1C, which can rule out the effects of morphology changes on PCE. The changes in PCE with MoS2 film annealing temperature are displayed in Fig. 1, and the corresponding photovoltaic parameters are listed in Table 1 (data highlighted in red). As seen, the PCE of the G/MoS2(80)/n-Si device is increased by B100% compared with the G/n-Si device (from 2.31% to 4.43%), while the G/MoS2(200)/n-Si device shows a decrease of B86% in PCE (compared to G/MoS2(80)/n-Si device) as a result of the decrease in Jsc, Voc and FF from 27.7 to 8.8 mA cm 2, 0.51 to 0.36 V and 0.32 to 0.21, respectively. To obtain reliable results on photovoltaic parameters, each set of solar cells is fabricated with three parallel devices and the statistical results are given in Fig. 2 and Table 1. Compared with the reference G/n-Si device, all the photovoltaic parameters of the G/MoS2(80)/n-Si device show B25% increase; while the main factor that results in poor performance of the G/MoS2(200)/n-Si device is Jsc, which shows a decrease of B60% from 21.9 to 8.8 mA cm 2. Firstly the effects of the SiOx interlayer (which may be formed since the annealing at 80 1C is done in air, it has been reported that the existence of a thin SiOx interlayer can greatly increase Voc15) on device performance is excluded by annealing the MoS2 films at 80 1C under the protection of Ar (denoted as
Results and discussion The graphene used here (B7 layers) is synthesized by LPCVD as reported elsewhere (for detailed characterizations on graphene, see ref. 5). The reference G/n-Si solar cells have a stable PCE of B2.3%. A typical device has Jsc = 21.9 mA cm 2, Voc = 0.46 V, FF = 0.23 and PCE = 2.31% (Fig. 1, black curve). The sizes of MoS2 nanoflakes are in the range of 100–300 nm (Fig. S1a, ESI†) and the thickness is in the range of 1–8 layers (Fig. S1b, ESI†). The out-of-plane A1g and in-plane E12g modes of MoS2 are observed using the Raman spectrum (Fig. S1c, ESI†). The MoS2 solution is spin-coated onto Si substrates and then baked at 80 1C in air for 15 min to evaporate the solvent (ethanol). Some MoS2 films are further annealed at 200 1C under the protection of argon to explore their role in G/MoS2/Si devices.
This journal is © the Owner Societies 2015
Table 1
Photovoltaic parameters of parallel solar cells
Ann. temp. (1C)
Jsc (mA cm 2)
Voc (V)
FF
PCE (%)
Ref.
21.9 18.4 21.3
0.46 0.37 0.42
0.23 0.28 0.29
2.31 1.88 2.65
Average (Ref.) 80
20.7 27.7 27.2 24 26.1
0.42 0.51 0.48 0.55 0.51
0.26 0.32 0.33 0.31 0.32
2.28 4.43 4.35 4.15 4.31
8.8 9.4 10.3 9.5
0.35 0.35 0.37 0.36
0.21 0.17 0.16 0.18
0.65 0.57 0.60 0.61
Average (80) 200 Average (200)
Phys. Chem. Chem. Phys., 2015, 17, 8182--8186 | 8183
View Article Online
Published on 17 February 2015. Downloaded by UNIVERSITY OF OTAGO on 28/07/2015 04:37:54.
PCCP
Fig. 2
Paper
Statistical results of photovoltaic parameters.
G/MoS2 (80, Ar)/n-Si), in contrast to the annealing at 80 1C in air. As shown in Fig. S3 and Table S1 (ESI†), the PCE of the two devices is almost the same. In addition, the Voc of G/MoS2 (80, Ar)/n-Si (0.55 V) is very close to that of G/MoS2(80)/n-Si (0.51 V). These results indicate that the enhancement in PCE has nothing to do with the SiOx interlayer, but solely due to the insertion of MoS2 interfacial layers. The effects of the SiOx interlayer are further excluded by valuing the performance of G/MoS2/n-Si devices without thermal treatment (see Fig. S4 for details, ESI†). In Schottky-junction solar cells, Voc is mainly related to the Schottky barrier height (SBH).16 The SBH is then extracted from the dark I–V curves using the Cheung method (Fig. 3 and Table S3, ESI†) to evaluate the changes in Voc.17 The series resistance are obtained both from dV/d(ln J) vs. J plots (denote as Rs1) and H( J) vs. J plots (denote as Rs2) to see the consistency of this method. As seen, Rs1 and Rs2 show good agreement within 5%. The calculated SBH of the reference G/n-Si device (SBHRef) is 0.64 eV, higher than the experimental value (0.46 V), implying that the interfacial recombination reduces Voc significantly.18 After the insertion of MoS2(80) films, the increase in SBH, DSBH, is 0.08 eV, consistent with the change in Voc (DVoc = 0.05 eV). However, SBHMoS2(200) is only lowered by 0.07 eV than SBHMoS2(80), smaller than the decrease in Voc (DVoc = 0.16 eV). The divergence can be explained by the higher interfacial recombination in G/MoS2(200)/n-Si
Fig. 3 The dark I–V curves of solar cells in Fig. 1. The device parameters, including SBH, n and Rs are calculated from the dark I–V curves using the dV/d(ln J) vs. J plots (the upper inset) and H(J) vs. J plots (the lower inset).
8184 | Phys. Chem. Chem. Phys., 2015, 17, 8182--8186
Fig. 4 Corresponding IPCE curves of devices in Fig. 1. Clearly IPCE80 4 IPCERef. 4 IPCE200, indicating that the G/MoS2(200)/n-Si device has the highest interfacial recombination.
devices, which leads to a higher DVoc than expected from DSBH. This is supported by the IPCE spectra, as shown in Fig. 4, IPCE80 4 IPCERef 4 IPCE200, higher IPCE is indicative of lower interfacial recombination (Note that the optical absorbance of MoS2(80) and MoS2(200) films made from the same MoS2 film is almost the same, see Fig. S5 for details, ESI†).19 The Jsc of the G/MoS2(200)/n-Si device decreases dramatically compared with that of G/MoS2(80)/n-Si devices, which means that the charge transport in these two devices varies tremendously. So, in the next step, the charge-transfer resistances (Zf) are measured using electrochemical impedance spectroscopy (EIS) to probe the charge transport.5 The measurements are conducted under AM 1.5G by applying a bias that is equal to Voc. As shown in Fig. 5, the Zf of G/n-Si, G/MoS2(80)/n-Si and G/MoS2(200)/n-Si devices are respectively B363 O, B148 O and B575 O, indicating that in the G/MoS2(80)/n-Si device the charge (hole) transport is prompted while it is impeded in the G/MoS2(200)/n-Si device. In order to further elucidate the charge transport in G/MoS2/n-Si devices, the energy band diagrams (EBD) are required. As can be seen in Fig. 6a, WFMoS2 is 4.41 and 4.52 eV for MoS2(80) and MoS2(200) films, respectively, as determined from UPS. Since the Fermi level (EF) of MoS2 falls between the conduction band (Ec) and the valence band (Ev), the
Fig. 5 The Electrochemical Impedance Spectroscopy of G/n-Si, G/MoS2(80)/ n-Si and G/MoS2(200)/n-Si devices. Their Zf are B363 O, B148 O and B575 O, respectively.
This journal is © the Owner Societies 2015
View Article Online
Published on 17 February 2015. Downloaded by UNIVERSITY OF OTAGO on 28/07/2015 04:37:54.
Paper
Fig. 6 (a) and (b) UPS spectra of MoS2, showing its Fermi level and the gap between the Fermi level and the conduction band, respectively. (c) UPS spectrum of graphene. (d) The absorbance spectra of MoS2 films, since the absorbance spectra of MoS2(80) and MoS2(200) films are almost the same, only one curve is given. The inset shows the calculation of the band gap using the (ahg)0.5 vs. hg plot.
distance between Ev and EF of MoS2 can be obtained from the onset energy. The distance is determined to be 0.69 eV as obtained from Fig. 6b. WFG is determined to be 4.7 eV (Fig. 6c). Since the optical transmittance spectra of MoS2(80) and MoS2(200) films are almost the same, only one curve is shown in Fig. 6d. The BG of MoS2 is determined to be 1.3 eV. Indeed, the BG of MoS2 is not important because it lies in the range of 1.2–1.9 eV (the BG of bulk and monlayer MoS2 is 1.2 eV and 1.9 eV, respectively). In all cases, the Ec of MoS2 is always higher than that of n-Si. The Fermi level of n-Si is 4.21 eV (EF Ei = 0.4 eV at 300 K), its electron affinity (conduction band) and BG is respectively 4.05 eV and 1.12 eV.16 Then the EBD of G/MoS2(80)/ n-Si and G/MoS2(200)/n-Si devices are constructed based on these data and presented in Fig. 7. As seen, the insertion of
PCCP
MoS2(80) films leads to the formation of type II heterojunction (Fig. 7a, the red circle). It has been demonstrated that such a structure is favorable for charge separation.20 To be specific, the hole transport from n-Si to MoS2 is facilitated. In contrast, the charge transport in the G/MoS2(200)/n-Si device (Fig. 7b) is blocked because of the valence band mismatch (the blue circle) triggered by the increase in WFMoS2. The changes in WFMoS2 can be explained by the phase transformation between the 1T and 2H phase.14 As shown in Fig. 7c, about 19% of the 1T phase is detected at 80 1C, while at 200 1C no noticeable 1T phase is detected (Fig. 7d), indicating that the material is predominantly in the 2H phase. This is in accordance with the previous report, which shows that at 200 1C the content of the 2H phase is higher than 95%.14 The thickness (d) of MoS2 films can also affect PCE. d is measured at three different points using an ellipsometer and fitted using a Tauc–Lorentz model, and the data are summarized in Table S4 (ESI†). The changes in PCE with d are presented in Fig. 8a (see Table S5 for the corresponding photovoltaic parameters, ESI†). As seen, the Jsc and FF increase with decreasing d. The variation trend of Jsc is consistent with that of IPCE (Fig. 8b), indicating that the interfacial recombination is the main factor determining Jsc. The variations in FF can be visualized by Zf (Fig. 8c). It is obvious that Zf decreases with decreasing d. In addition, the photovoltaic parameters change little at d = B2 nm, or the PCE ‘‘saturates’’ at d = B2 nm. Though the PCE of G/MoS2/n-Si solar cells can reach 4.4%, it is clear that the main factor that limits PCE is FF (FF = 0.32, note that the FF of traditional silicon solar cells is 40.7). Here
Fig. 8 (a) The effects of MoS2 film thickness on PCE. (b) The corresponding IPCE curves. (c) The corresponding Zf.
Fig. 7 (a) The EBD of the G/MoS2(80)/n-Si device, a type-II structure is formed, as indicated by the red circle. (b) The EBD of the G/MoS2(200)/n-Si device, valence band mismatch is observed, as indicated by the blue circle. (c) Mo 2d peak of the MoS2(80) film. (d) Mo 2d peak of the MoS2(200) film.
This journal is © the Owner Societies 2015
Fig. 9 (a) The effects of surface passivation on PCE. (b) Corresponding IPCE curves. (c) Corresponding Zf.
Phys. Chem. Chem. Phys., 2015, 17, 8182--8186 | 8185
View Article Online
PCCP Table 2
Published on 17 February 2015. Downloaded by UNIVERSITY OF OTAGO on 28/07/2015 04:37:54.
Pristine Modified
Paper
Notes and references
Photovoltaic parameters of devices in Fig. 9a 2
Voc (V)
Jsc (mA cm )
FF
PCE (%)
0.51 0.50
27.7 28.1
0.32 0.47
4.43 6.56
the hydroquinone–methanol surface passivation method is introduced to improve FF (for detailed preparation and characterizations, see ref. 21). Fig. 9a and Table 2 give respectively the J–V curves and corresponding photovoltaic parameters. After surface passivation, there are little changes in Jsc and Voc, but the FF is increased by B50% from 0.32 to 0.47. This is further confirmed by the IPCE (Fig. 9b) and Zf spectra (Fig. 9c): the IPCE spectra show little changes while Zf decreases from B148 to B70 O after surface modification. As a result, a PCE of 6.56% is obtained.
Conclusions In summary, the role of MoS2 is carefully investigated by varying the MoS2 film annealing temperature. It is shown that the insertion of MoS2(80) films results in the formation of a type II structure, facilitating hole transport; while the MoS2(200) film leads to valence band mismatch, which blocks hole transport. The effects of MoS2 film thickness on device performance are also investigated and it is found that the PCE gradually increases with decreasing film thickness and ‘‘saturates’’ at about 2 nm. A PCE of 6.56% is achieved with silicon surface passivation. Our work indicates that MoS2 is competitive as an effective interfacial layer in G/Si solar cells. It is believed that the PCE can be further enhanced using MoS2 nanoflakes with higher quality (e.g., larger flake sizes). Note: in the preparation of this manuscript, we notice that Tsai et al. demonstrate MoS2/p-Si devices having a PCE of B5.3%.22 Here we claim that the main focus of this work is on the interface of G/n-Si solar cells. It is believed that the efficiency of G/n-Si solar cells can be further enhanced by exploring a more efficient interfacial material.
Acknowledgements The authors gratefully acknowledge the support from National Natural Science Foundation of China (No. 51272253), Strategic Leading Science & Technology Programme (Grant No. XDB0505000), the 863 Hi-tech Research and Development Program of China (2010AA064903), and the Hundred Talents Program of the Chinese Academy of Sciences (CAS).
8186 | Phys. Chem. Chem. Phys., 2015, 17, 8182--8186
1 X. M. Li, H. W. Zhu, K. L. Wang, A. Y. Cao, J. Q. Wei, C. Y. Li, Y. Jia, Z. Li, X. Li and D. H. Wu, Adv. Mater., 2010, 22, 2743. 2 E. Shi, H. Li, L. Yang, L. Zhang, Z. Li, P. Li, Y. Shang, S. Wu, X. Li, J. Wei, K. Wang, H. Zhu, D. Wu, Y. Fang and A. Cao, Nano Lett., 2013, 13, 1776. 3 X. C. Miao, S. Tongay, M. K. Petterson, K. Berke, A. G. Rinzler, B. R. Appleton and A. F. Hebard, Nano Lett., 2012, 12, 2745. 4 T. X. Cui, R. T. Lv, Z. H. Huang, S. X. Chen, Z. X. Zhang, X. Gan, Y. Jia, X. M. Li, K. L. Wang, D. H. Wu and F. Y. Kang, J. Mater. Chem. A, 2013, 1, 5736. 5 K. J. Jiao, X. L. Wang, Y. Wang and Y. F. Chen, J. Mater. Chem. C, 2014, 2, 7715. 6 G. L. Frey, K. J. Reynolds, R. H. Friend, H. Cohen and Y. Feldman, J. Am. Chem. Soc., 2003, 125, 5998. 7 B. Radisavljevic, A. Radenovic, J. Brivio, V. Giacometti and A. Kis, Nat. Nanotechnol., 2011, 6, 147. 8 M. A. Lukowski, A. S. Daniel, F. Meng, A. Forticaux, L. S. Li and S. Jin, J. Am. Chem. Soc., 2013, 135, 10274. 9 M. Shanmugam, C. A. Durcan and B. Yu, Nanoscale, 2012, 4, 7399. 10 J. M. Yun, Y. J. Noh, J. S. Yeo, Y. J. Go, S. I. Na, H. G. Jeong, J. Kim, S. Lee, S. S. Kim, H. Y. Koo, T. W. Kim and D. Y. Kim, J. Mater. Chem. C, 2013, 1, 3777. 11 X. Gu, W. Cui, H. Li, Z. W. Wu, Z. Y. Zeng, S. T. Lee, H. Zhang and B. Q. Sun, Adv. Energy Mater., 2013, 3, 1262. 12 Z. Z. Sun, A. R. O. Raji, Y. Zhu, C. S. Xiang, Z. Yan, C. Kittrel, E. L. G. Samuel and J. M. Tour, ACS Nano, 2012, 6, 9790. 13 X. S. Li, W. W. Cai, J. H. An, S. Kim, J. Nah, D. X. Yang, R. Piner, A. Velamakanni, I. Jung, E. Tutuc, S. K. Banerjee, L. Colombo and R. S. Ruoff, Science, 2009, 324, 1312. 14 G. Eda, H. Yamaguchi, D. Voiry, T. Fujita, M. W. Chen and M. Chhowalla, Nano Lett., 2011, 11, 5111. 15 H. C. Card and E. S. Yang, Appl. Phys. Lett., 1976, 29, 51. 16 M. Sze and K. K. Ng, Physics of Semiconductor Devices, Wiley, Hoboken, 3rd edn, 2007. 17 S. K. Cheung and N. W. Cheung, Appl. Phys. Lett., 1986, 49, 85. 18 W. Shockley and H. J. Queisser, J. Appl. Phys., 1961, 32, 510. 19 X. Li, D. Xie, H. Park, T. H. Zeng, K. Wang, J. Wei, M. Zhong, D. Wu, J. Kong and H. Zhu, Adv. Energy Mater., 2013, 3, 1029. 20 D. Gross, I. Mora-Sero, T. Dittrich, A. Belaidi, C. Mauser, A. J. Houtepen, E. Da Como, A. L. Rogach and J. Feldmann, J. Am. Chem. Soc., 2010, 132, 5981. 21 K. Jiao, D. Zhang and Y. Chen, RSC Adv., 2014, 4, 55300. 22 M. L. Tsai, S. H. Su, J. K. Chang, D. S. Tsai, C. H. Chen, C. I. Wu, L. J. Li, L. J. Chen and J. H. He, ACS Nano, 2014, 8, 8317.
This journal is © the Owner Societies 2015
Published on 17 February 2015. Downloaded by UNIVERSITY OF OTAGO on 28/07/2015 04:37:54.
PAPER
Cite this: Phys. Chem. Chem. Phys., 2015, 17, 8182
View Journal | View Issue
The role of MoS2 as an interfacial layer in graphene/silicon solar cells† Kejia Jiao, Chunyang Duan, Xiaofeng Wu, Jiayuan Chen, Yu Wang* and Yunfa Chen* The role of MoS2 as an effective interfacial layer in graphene/silicon solar cells is systematically investigated by varying MoS2 film annealing temperature and thickness. It is found that the power conversion efficiency (PCE) is increased by B100% from B2.3% to B4.4% with 80 1C annealed MoS2 film whereas it drops significantly to B0.6% with 200 1C annealed MoS2 film. The results are well explained based on the device energy band diagram. That is, the incorporation of MoS2(80) films leads to the formation of type II structure, facilitating hole transport; while valence band mismatch is formed
Received 18th January 2015, Accepted 13th February 2015
with MoS2(200) films due to the increase in the work function of MoS2. Besides, the PCE increases gradually with decreasing MoS2 film thickness, and ‘‘saturates’’ at about 2 nm. The PCE can be further
DOI: 10.1039/c5cp00321k
enhanced to B6.6% with the aid of silicon surface passivation. Our work demonstrates that MoS2 is an excellent interfacial layer to improve the PCE with low-temperature annealing (80 1C in air), which may
www.rsc.org/pccp
be helpful in developing efficient and low-cost G/Si solar cells.
Introduction Graphene/n-silicon (G/n-Si) Schottky-barrier solar cells have attracted substantial attention due to their advantages of lowtemperature processability, low cost and potential of high efficiency.1,2 However, the power conversion efficiency (PCE) of pristine G/n-Si solar cells is usually low (typicallyo3%, here pristine refers to G/n-Si solar cells without any treatments), and chemical doping of graphene is the most commonly used method to improve PCE.3,4 Recently, we have demonstrated that interface tailoring is also a powerful method to enhance PCE and graphene oxide (GO) is selected as a model interfacial material.5 However, GO only works after high-temperature annealing (i.e., 400 1C under protection atmosphere). Exploring the alternative material of GO, which can improve the PCE effectively using a low-temperature procedure (o100 1C), is desirable. Molybdenum disulfide (MoS2) is a 2D material composed of S–Mo–S planes held together by van der Waals forces. It has drawn increasing attention due to its fascinating properties that are acceptable for a wide range of applications, such as field-effect transistors, light-emitting diodes and catalysis.6–9 Recently, the applications of MoS2 in photovoltaic devices are also demonstrated. For example, by employing MoS2 as a light State Key Laboratory of Multiphase Complex Systems, Institute of Process Engineering, Chinese Academy of Sciences, Beijing 100190, P.R. China. E-mail: [email protected], [email protected] † Electronic supplementary information (ESI) available. See DOI: 10.1039/ c5cp00321k
8182 | Phys. Chem. Chem. Phys., 2015, 17, 8182--8186
harvesting material, a PCE of 1.8% can be achieved for MoS2/Au Schottky junction solar cells.9 In addition, MoS2 has also been used as hole/electron transport layers in organic photovoltaic devices.10,11 The advantages of MoS2 over GO as an interfacial material are that on one hand, it is chemically inert with fewer traps; on the other hand, MoS2 has a higher hole mobility than GO, which may reduce annealing temperature. However, to date, there is no report on MoS2 as an interfacial layer in G/n-Si solar cells. Herein, MoS2 is integrated into G/n-Si solar cells as an interfacial layer. The device based on MoS2 shows a PCE of B4.4%, comparable to that of a GO-based device (B4.6%), yet only low-temperature annealing (i.e., 80 1C in air) is required. The role of MoS2 is investigated by varying MoS2 film annealing temperature and thickness. It is found that the incorporation of MoS2 can improve the PCE by forming a type-II heterojunction, which facilitates carrier transport between MoS2 and n-Si. The PCE can be further enhanced to B6.6% with the aid of silicon surface passivation.
Experimental section Synthesis of multilayer graphene (MLG) As reported previously,12 MLG is synthesized using a lowpressure chemical vapor deposition (LPCVD) method. Briefly, the system is first heated to 1000 1C for 10 min under the protection of hydrogen gas to clean the copper surface, then methane is introduced for 20 min.
This journal is © the Owner Societies 2015
View Article Online
Paper
PCCP
MoS2 film preparation
Published on 17 February 2015. Downloaded by UNIVERSITY OF OTAGO on 28/07/2015 04:37:54.
MoS2 nanoflake solution is purchased from Graphene Supermarket. It is spin-coated onto a silicon surface and annealed under two different conditions, i.e., at 80 1C in air (denoted as MoS2(80)) or at 200 1C under the protection of argon (denoted as MoS2(200)) for a fixed 30 min. The devices incorporating MoS2 are then labeled as G/MoS2(80)/n-Si and G/MoS2(200)/n-Si. Fabrication of solar cell devices n-type Si (2–4 O cm 1, with a square window of 3 mm 3 mm surrounded by 300 nm thick SiO2) is used as substrates. MoS2 is spin-coated onto Si substrates at a certain speed and then annealed at a different temperature. Subsequently MLG is transferred onto the top of MoS2 using the PMMA-assisted wet transfer method.13 Ga–In liquid (99.999%) and Ag paste are applied as the rear electrode and the front electrode, respectively. Characterizations The solar cells are tested under Air Mass 1.5 illumination (100 mW cm 2, the light density is calibrated using a standard solar reference cell, SRC-1000-TC-QZ, VLSI Standards S/N: 10510-0305). The J–V data and incident photon-to-electron conversion efficiency (IPCE) spectra are recorded using a Keithley 2400 SourceMeter and QTEST STATION 500AD, respectively. Raman Spectra are performed using Renishaw inVia-Reflex with 532 nm wavelength incident laser light. Ultraviolet photoelectron spectroscopy (UPS) measurements are made using an unfiltered HeI (21.22 eV) gas discharge lamp to determine the work function of graphene and MoS2. The UV-Vis spectrum is acquired using a Shimadzu UV-VIS-NIR Spectrophotometer UV-3600. X-ray photoelectron spectroscopy (XPS) measurements were performed using a monochromatic Al Ka source (1486.6 eV). The charge-transfer resistance is measured by electrochemical Impedance Spectroscopy (EIS), which is measured using CHI 760E, Shanghai Chenhua Instrument Co., Ltd, the measurements are carried out under illumination (AM 1.5G condition) by applying a bias that is equal to Voc. The MoS2 film thicknesses are obtained using an ellipsometer, SENTECH SE850, and the aperture diameter is 0.2 mm.
Fig. 1
The effects of MoS2 film annealing temperature on PCE.
Here 200 1C is chosen because it is high enough to vary the electronic properties of MoS2.14 As shown in Fig. S2 (ESI†), the Si substrates are fully covered by MoS2 and there are little changes in the morphologies of MoS2 films annealed at 80 and 200 1C, which can rule out the effects of morphology changes on PCE. The changes in PCE with MoS2 film annealing temperature are displayed in Fig. 1, and the corresponding photovoltaic parameters are listed in Table 1 (data highlighted in red). As seen, the PCE of the G/MoS2(80)/n-Si device is increased by B100% compared with the G/n-Si device (from 2.31% to 4.43%), while the G/MoS2(200)/n-Si device shows a decrease of B86% in PCE (compared to G/MoS2(80)/n-Si device) as a result of the decrease in Jsc, Voc and FF from 27.7 to 8.8 mA cm 2, 0.51 to 0.36 V and 0.32 to 0.21, respectively. To obtain reliable results on photovoltaic parameters, each set of solar cells is fabricated with three parallel devices and the statistical results are given in Fig. 2 and Table 1. Compared with the reference G/n-Si device, all the photovoltaic parameters of the G/MoS2(80)/n-Si device show B25% increase; while the main factor that results in poor performance of the G/MoS2(200)/n-Si device is Jsc, which shows a decrease of B60% from 21.9 to 8.8 mA cm 2. Firstly the effects of the SiOx interlayer (which may be formed since the annealing at 80 1C is done in air, it has been reported that the existence of a thin SiOx interlayer can greatly increase Voc15) on device performance is excluded by annealing the MoS2 films at 80 1C under the protection of Ar (denoted as
Results and discussion The graphene used here (B7 layers) is synthesized by LPCVD as reported elsewhere (for detailed characterizations on graphene, see ref. 5). The reference G/n-Si solar cells have a stable PCE of B2.3%. A typical device has Jsc = 21.9 mA cm 2, Voc = 0.46 V, FF = 0.23 and PCE = 2.31% (Fig. 1, black curve). The sizes of MoS2 nanoflakes are in the range of 100–300 nm (Fig. S1a, ESI†) and the thickness is in the range of 1–8 layers (Fig. S1b, ESI†). The out-of-plane A1g and in-plane E12g modes of MoS2 are observed using the Raman spectrum (Fig. S1c, ESI†). The MoS2 solution is spin-coated onto Si substrates and then baked at 80 1C in air for 15 min to evaporate the solvent (ethanol). Some MoS2 films are further annealed at 200 1C under the protection of argon to explore their role in G/MoS2/Si devices.
This journal is © the Owner Societies 2015
Table 1
Photovoltaic parameters of parallel solar cells
Ann. temp. (1C)
Jsc (mA cm 2)
Voc (V)
FF
PCE (%)
Ref.
21.9 18.4 21.3
0.46 0.37 0.42
0.23 0.28 0.29
2.31 1.88 2.65
Average (Ref.) 80
20.7 27.7 27.2 24 26.1
0.42 0.51 0.48 0.55 0.51
0.26 0.32 0.33 0.31 0.32
2.28 4.43 4.35 4.15 4.31
8.8 9.4 10.3 9.5
0.35 0.35 0.37 0.36
0.21 0.17 0.16 0.18
0.65 0.57 0.60 0.61
Average (80) 200 Average (200)
Phys. Chem. Chem. Phys., 2015, 17, 8182--8186 | 8183
View Article Online
Published on 17 February 2015. Downloaded by UNIVERSITY OF OTAGO on 28/07/2015 04:37:54.
PCCP
Fig. 2
Paper
Statistical results of photovoltaic parameters.
G/MoS2 (80, Ar)/n-Si), in contrast to the annealing at 80 1C in air. As shown in Fig. S3 and Table S1 (ESI†), the PCE of the two devices is almost the same. In addition, the Voc of G/MoS2 (80, Ar)/n-Si (0.55 V) is very close to that of G/MoS2(80)/n-Si (0.51 V). These results indicate that the enhancement in PCE has nothing to do with the SiOx interlayer, but solely due to the insertion of MoS2 interfacial layers. The effects of the SiOx interlayer are further excluded by valuing the performance of G/MoS2/n-Si devices without thermal treatment (see Fig. S4 for details, ESI†). In Schottky-junction solar cells, Voc is mainly related to the Schottky barrier height (SBH).16 The SBH is then extracted from the dark I–V curves using the Cheung method (Fig. 3 and Table S3, ESI†) to evaluate the changes in Voc.17 The series resistance are obtained both from dV/d(ln J) vs. J plots (denote as Rs1) and H( J) vs. J plots (denote as Rs2) to see the consistency of this method. As seen, Rs1 and Rs2 show good agreement within 5%. The calculated SBH of the reference G/n-Si device (SBHRef) is 0.64 eV, higher than the experimental value (0.46 V), implying that the interfacial recombination reduces Voc significantly.18 After the insertion of MoS2(80) films, the increase in SBH, DSBH, is 0.08 eV, consistent with the change in Voc (DVoc = 0.05 eV). However, SBHMoS2(200) is only lowered by 0.07 eV than SBHMoS2(80), smaller than the decrease in Voc (DVoc = 0.16 eV). The divergence can be explained by the higher interfacial recombination in G/MoS2(200)/n-Si
Fig. 3 The dark I–V curves of solar cells in Fig. 1. The device parameters, including SBH, n and Rs are calculated from the dark I–V curves using the dV/d(ln J) vs. J plots (the upper inset) and H(J) vs. J plots (the lower inset).
8184 | Phys. Chem. Chem. Phys., 2015, 17, 8182--8186
Fig. 4 Corresponding IPCE curves of devices in Fig. 1. Clearly IPCE80 4 IPCERef. 4 IPCE200, indicating that the G/MoS2(200)/n-Si device has the highest interfacial recombination.
devices, which leads to a higher DVoc than expected from DSBH. This is supported by the IPCE spectra, as shown in Fig. 4, IPCE80 4 IPCERef 4 IPCE200, higher IPCE is indicative of lower interfacial recombination (Note that the optical absorbance of MoS2(80) and MoS2(200) films made from the same MoS2 film is almost the same, see Fig. S5 for details, ESI†).19 The Jsc of the G/MoS2(200)/n-Si device decreases dramatically compared with that of G/MoS2(80)/n-Si devices, which means that the charge transport in these two devices varies tremendously. So, in the next step, the charge-transfer resistances (Zf) are measured using electrochemical impedance spectroscopy (EIS) to probe the charge transport.5 The measurements are conducted under AM 1.5G by applying a bias that is equal to Voc. As shown in Fig. 5, the Zf of G/n-Si, G/MoS2(80)/n-Si and G/MoS2(200)/n-Si devices are respectively B363 O, B148 O and B575 O, indicating that in the G/MoS2(80)/n-Si device the charge (hole) transport is prompted while it is impeded in the G/MoS2(200)/n-Si device. In order to further elucidate the charge transport in G/MoS2/n-Si devices, the energy band diagrams (EBD) are required. As can be seen in Fig. 6a, WFMoS2 is 4.41 and 4.52 eV for MoS2(80) and MoS2(200) films, respectively, as determined from UPS. Since the Fermi level (EF) of MoS2 falls between the conduction band (Ec) and the valence band (Ev), the
Fig. 5 The Electrochemical Impedance Spectroscopy of G/n-Si, G/MoS2(80)/ n-Si and G/MoS2(200)/n-Si devices. Their Zf are B363 O, B148 O and B575 O, respectively.
This journal is © the Owner Societies 2015
View Article Online
Published on 17 February 2015. Downloaded by UNIVERSITY OF OTAGO on 28/07/2015 04:37:54.
Paper
Fig. 6 (a) and (b) UPS spectra of MoS2, showing its Fermi level and the gap between the Fermi level and the conduction band, respectively. (c) UPS spectrum of graphene. (d) The absorbance spectra of MoS2 films, since the absorbance spectra of MoS2(80) and MoS2(200) films are almost the same, only one curve is given. The inset shows the calculation of the band gap using the (ahg)0.5 vs. hg plot.
distance between Ev and EF of MoS2 can be obtained from the onset energy. The distance is determined to be 0.69 eV as obtained from Fig. 6b. WFG is determined to be 4.7 eV (Fig. 6c). Since the optical transmittance spectra of MoS2(80) and MoS2(200) films are almost the same, only one curve is shown in Fig. 6d. The BG of MoS2 is determined to be 1.3 eV. Indeed, the BG of MoS2 is not important because it lies in the range of 1.2–1.9 eV (the BG of bulk and monlayer MoS2 is 1.2 eV and 1.9 eV, respectively). In all cases, the Ec of MoS2 is always higher than that of n-Si. The Fermi level of n-Si is 4.21 eV (EF Ei = 0.4 eV at 300 K), its electron affinity (conduction band) and BG is respectively 4.05 eV and 1.12 eV.16 Then the EBD of G/MoS2(80)/ n-Si and G/MoS2(200)/n-Si devices are constructed based on these data and presented in Fig. 7. As seen, the insertion of
PCCP
MoS2(80) films leads to the formation of type II heterojunction (Fig. 7a, the red circle). It has been demonstrated that such a structure is favorable for charge separation.20 To be specific, the hole transport from n-Si to MoS2 is facilitated. In contrast, the charge transport in the G/MoS2(200)/n-Si device (Fig. 7b) is blocked because of the valence band mismatch (the blue circle) triggered by the increase in WFMoS2. The changes in WFMoS2 can be explained by the phase transformation between the 1T and 2H phase.14 As shown in Fig. 7c, about 19% of the 1T phase is detected at 80 1C, while at 200 1C no noticeable 1T phase is detected (Fig. 7d), indicating that the material is predominantly in the 2H phase. This is in accordance with the previous report, which shows that at 200 1C the content of the 2H phase is higher than 95%.14 The thickness (d) of MoS2 films can also affect PCE. d is measured at three different points using an ellipsometer and fitted using a Tauc–Lorentz model, and the data are summarized in Table S4 (ESI†). The changes in PCE with d are presented in Fig. 8a (see Table S5 for the corresponding photovoltaic parameters, ESI†). As seen, the Jsc and FF increase with decreasing d. The variation trend of Jsc is consistent with that of IPCE (Fig. 8b), indicating that the interfacial recombination is the main factor determining Jsc. The variations in FF can be visualized by Zf (Fig. 8c). It is obvious that Zf decreases with decreasing d. In addition, the photovoltaic parameters change little at d = B2 nm, or the PCE ‘‘saturates’’ at d = B2 nm. Though the PCE of G/MoS2/n-Si solar cells can reach 4.4%, it is clear that the main factor that limits PCE is FF (FF = 0.32, note that the FF of traditional silicon solar cells is 40.7). Here
Fig. 8 (a) The effects of MoS2 film thickness on PCE. (b) The corresponding IPCE curves. (c) The corresponding Zf.
Fig. 7 (a) The EBD of the G/MoS2(80)/n-Si device, a type-II structure is formed, as indicated by the red circle. (b) The EBD of the G/MoS2(200)/n-Si device, valence band mismatch is observed, as indicated by the blue circle. (c) Mo 2d peak of the MoS2(80) film. (d) Mo 2d peak of the MoS2(200) film.
This journal is © the Owner Societies 2015
Fig. 9 (a) The effects of surface passivation on PCE. (b) Corresponding IPCE curves. (c) Corresponding Zf.
Phys. Chem. Chem. Phys., 2015, 17, 8182--8186 | 8185
View Article Online
PCCP Table 2
Published on 17 February 2015. Downloaded by UNIVERSITY OF OTAGO on 28/07/2015 04:37:54.
Pristine Modified
Paper
Notes and references
Photovoltaic parameters of devices in Fig. 9a 2
Voc (V)
Jsc (mA cm )
FF
PCE (%)
0.51 0.50
27.7 28.1
0.32 0.47
4.43 6.56
the hydroquinone–methanol surface passivation method is introduced to improve FF (for detailed preparation and characterizations, see ref. 21). Fig. 9a and Table 2 give respectively the J–V curves and corresponding photovoltaic parameters. After surface passivation, there are little changes in Jsc and Voc, but the FF is increased by B50% from 0.32 to 0.47. This is further confirmed by the IPCE (Fig. 9b) and Zf spectra (Fig. 9c): the IPCE spectra show little changes while Zf decreases from B148 to B70 O after surface modification. As a result, a PCE of 6.56% is obtained.
Conclusions In summary, the role of MoS2 is carefully investigated by varying the MoS2 film annealing temperature. It is shown that the insertion of MoS2(80) films results in the formation of a type II structure, facilitating hole transport; while the MoS2(200) film leads to valence band mismatch, which blocks hole transport. The effects of MoS2 film thickness on device performance are also investigated and it is found that the PCE gradually increases with decreasing film thickness and ‘‘saturates’’ at about 2 nm. A PCE of 6.56% is achieved with silicon surface passivation. Our work indicates that MoS2 is competitive as an effective interfacial layer in G/Si solar cells. It is believed that the PCE can be further enhanced using MoS2 nanoflakes with higher quality (e.g., larger flake sizes). Note: in the preparation of this manuscript, we notice that Tsai et al. demonstrate MoS2/p-Si devices having a PCE of B5.3%.22 Here we claim that the main focus of this work is on the interface of G/n-Si solar cells. It is believed that the efficiency of G/n-Si solar cells can be further enhanced by exploring a more efficient interfacial material.
Acknowledgements The authors gratefully acknowledge the support from National Natural Science Foundation of China (No. 51272253), Strategic Leading Science & Technology Programme (Grant No. XDB0505000), the 863 Hi-tech Research and Development Program of China (2010AA064903), and the Hundred Talents Program of the Chinese Academy of Sciences (CAS).
8186 | Phys. Chem. Chem. Phys., 2015, 17, 8182--8186
1 X. M. Li, H. W. Zhu, K. L. Wang, A. Y. Cao, J. Q. Wei, C. Y. Li, Y. Jia, Z. Li, X. Li and D. H. Wu, Adv. Mater., 2010, 22, 2743. 2 E. Shi, H. Li, L. Yang, L. Zhang, Z. Li, P. Li, Y. Shang, S. Wu, X. Li, J. Wei, K. Wang, H. Zhu, D. Wu, Y. Fang and A. Cao, Nano Lett., 2013, 13, 1776. 3 X. C. Miao, S. Tongay, M. K. Petterson, K. Berke, A. G. Rinzler, B. R. Appleton and A. F. Hebard, Nano Lett., 2012, 12, 2745. 4 T. X. Cui, R. T. Lv, Z. H. Huang, S. X. Chen, Z. X. Zhang, X. Gan, Y. Jia, X. M. Li, K. L. Wang, D. H. Wu and F. Y. Kang, J. Mater. Chem. A, 2013, 1, 5736. 5 K. J. Jiao, X. L. Wang, Y. Wang and Y. F. Chen, J. Mater. Chem. C, 2014, 2, 7715. 6 G. L. Frey, K. J. Reynolds, R. H. Friend, H. Cohen and Y. Feldman, J. Am. Chem. Soc., 2003, 125, 5998. 7 B. Radisavljevic, A. Radenovic, J. Brivio, V. Giacometti and A. Kis, Nat. Nanotechnol., 2011, 6, 147. 8 M. A. Lukowski, A. S. Daniel, F. Meng, A. Forticaux, L. S. Li and S. Jin, J. Am. Chem. Soc., 2013, 135, 10274. 9 M. Shanmugam, C. A. Durcan and B. Yu, Nanoscale, 2012, 4, 7399. 10 J. M. Yun, Y. J. Noh, J. S. Yeo, Y. J. Go, S. I. Na, H. G. Jeong, J. Kim, S. Lee, S. S. Kim, H. Y. Koo, T. W. Kim and D. Y. Kim, J. Mater. Chem. C, 2013, 1, 3777. 11 X. Gu, W. Cui, H. Li, Z. W. Wu, Z. Y. Zeng, S. T. Lee, H. Zhang and B. Q. Sun, Adv. Energy Mater., 2013, 3, 1262. 12 Z. Z. Sun, A. R. O. Raji, Y. Zhu, C. S. Xiang, Z. Yan, C. Kittrel, E. L. G. Samuel and J. M. Tour, ACS Nano, 2012, 6, 9790. 13 X. S. Li, W. W. Cai, J. H. An, S. Kim, J. Nah, D. X. Yang, R. Piner, A. Velamakanni, I. Jung, E. Tutuc, S. K. Banerjee, L. Colombo and R. S. Ruoff, Science, 2009, 324, 1312. 14 G. Eda, H. Yamaguchi, D. Voiry, T. Fujita, M. W. Chen and M. Chhowalla, Nano Lett., 2011, 11, 5111. 15 H. C. Card and E. S. Yang, Appl. Phys. Lett., 1976, 29, 51. 16 M. Sze and K. K. Ng, Physics of Semiconductor Devices, Wiley, Hoboken, 3rd edn, 2007. 17 S. K. Cheung and N. W. Cheung, Appl. Phys. Lett., 1986, 49, 85. 18 W. Shockley and H. J. Queisser, J. Appl. Phys., 1961, 32, 510. 19 X. Li, D. Xie, H. Park, T. H. Zeng, K. Wang, J. Wei, M. Zhong, D. Wu, J. Kong and H. Zhu, Adv. Energy Mater., 2013, 3, 1029. 20 D. Gross, I. Mora-Sero, T. Dittrich, A. Belaidi, C. Mauser, A. J. Houtepen, E. Da Como, A. L. Rogach and J. Feldmann, J. Am. Chem. Soc., 2010, 132, 5981. 21 K. Jiao, D. Zhang and Y. Chen, RSC Adv., 2014, 4, 55300. 22 M. L. Tsai, S. H. Su, J. K. Chang, D. S. Tsai, C. H. Chen, C. I. Wu, L. J. Li, L. J. Chen and J. H. He, ACS Nano, 2014, 8, 8317.
This journal is © the Owner Societies 2015
