Interface engineering and efficiency improvement of monolayer graphene–silicon solar cells by inserting an ultra-thin LiF interlayer

Abstract

Graphene–silicon (Gr–Si) Schottky junction solar cells have recently attracted intensive attention as candidates for low-cost photovoltaic devices. However, the efficiency of Gr–Si solar cells still needs further improvement. In this study, we have introduced an ultra-thin LiF layer between the Si and aluminum (Al) back electrode of Gr–Si solar cells. It is found that carrier recombination at the back surface is significantly suppressed, resulting directly in the improvement of external quantum efficiency (EQE) of devices in the long wavelength range of 800–1100 nm. Moreover, the back contact resistance is greatly reduced, and therefore the fill factor (FF) of devices is greatly improved. As a result, the highest power conversion efficiency (PCE) of 6.25% has been obtained for a pristine Gr–Si solar cell, which is further improved to 10.61% after chemical doping. These results pave a new way to the fabrication of high efficiency Gr–Si solar cells.

---

1. Introduction

Tremendous interest has been focused on photovoltaic (PV) electricity based on solar cells as power sources due to worldwide energy issues. Commonly-used commercial solar cells are fabricated from Si material since it is earth-abundant, stable and suitable for the photovoltaic industry. However, the fabrication of traditional crystalline Si solar cells based on the formation of a p–n junction is complicated. Compared to p–n junctions, Schottky junction solar cells have merits of low-cost and ease of fabrication. However, the thick metal layer usually used for the formation of a Schottky junction with Si has poor transparency, which limits the sunlight absorption of solar cells. Therefore, in the past few years transparent carbon nanotube (CNT) films have been used to form solar cells with Si, which yield a power conversion efficiency (PCE) of 10–15%.^1–4 Graphene (Gr), a fascinating two-dimensional (2D) material with high transparency, high conductivity and unique quantum properties, has triggered great research enthusiasm in PV applications.^5–8 Since the first Gr–Si solar cell achieved a PCE of 1.5%, many efforts have been made to improve the solar cell efficiency by different routes.⁹ It has been widely accepted that the PCE of Gr–Si Schottky solar cells can be improved by increasing the electrical conductivity of Gr film with chemical or electrical field doping, which results in the highest efficiency of 8.9%. If combined with the application of antireflective film, the efficiency of a chemically doped Gr–Si solar cell can reach 14.5%.¹⁰ However, the chemical doping is usually not stable and the corresponding Gr–Si solar cell has to suffer some efficiency degradation. For the pristine Gr–Si solar cell, the difficulty in efficiency improvement arises from the interface recombination of carriers, including the front and back surfaces. Recently, a Gr–Si solar cell free from chemical doping with a record PCE of 6.18% was obtained by inserting a graphene oxide (GO) interlayer between Si and Gr to suppress the front surface recombination of solar cells, which has been reported by our group.¹¹ In order to further improve the performance of Gr–Si solar cells, it is also necessary to reduce the recombination at the back surface of Gr–Si solar cells.

Herein, we have introduced a lithium fluoride (LiF) layer with nominal thickness of 0.5 nm between the Si and Al back electrode to suppress the back surface recombination of Gr–Si solar cells. The external quantum efficiency (EQE) of solar cells is significantly improved; moreover, the dark current and the back contact resistance decrease. A PCE of up to 6.25% can be obtained for the pristine Gr–Si solar cell. This value can be further improved to 10.61% after HNO₃ chemical doping. To the best of our knowledge, such efficiencies are record values for the pristine and chemically-doped Gr–Si solar cells without the application of an antireflection layer.¹⁰

2. Experimental

A. Graphene preparation

The large-area monolayer Gr films were grown on a copper foil by a low pressure chemical vapor deposition (LPCVD) method at 1000 °C using CH₄ (20 sccm) as the carbon source and H₂ (40 sccm) as the reduction gas. The sheet resistances of Gr films are in the range of 400–1000 Ω sq⁻¹ determined by a four probe method. A well-defined polymethylmethacrylate-assisted (PMMA) wet transfer process was used for graphene transfer.¹² Briefly, graphene was spin-coated with PMMA in ethyl lactate. The PMMA was dried on a hot plate at 150 °C for 15 min. Gr on the backside of the Cu foil was etched away by an ultraviolet ozone cleaning process for 10 min. Furthermore, the underlying Cu foil was etched by aqueous ammonium persulfate ((NH₄)₂S₂O₈, Aladdin®) solution (1 M). The resultant Gr films were cleaned with deionized water several times.

B. Device fabrication

N-type <100> c-Si wafer (1.2 × 1.2 cm², resistivity 1–10 Ω cm⁻¹ and thickness 300 μm) was used to construct Gr–Si Schottky junction solar cells. The Si substrates were first soaked in a diluted HF solution for several minutes to remove the native SiO₂ and then rinsed by deionized water followed by N₂ gas drying. The cleaned Si substrates were set aside for about 2 hours permitting a thin native SiO₂ passivation layer to grow. The PMMA-supported Gr films were directly transferred onto the Si substrates. After drying at room temperature, residual PMMA on the Gr films was removed by annealing at 400 °C for 1 h in an N₂ atmosphere. Ag paste was coated along the outline of the Gr films. The device active area was determined using black tapes in the range of 0.12–0.18 mm². The Ag paste and the outside area were completely covered. A 0.5 nm thick LiF layer and a 150 nm thick Al layer were successively thermally deposited on the back side of the Si substrate as a back contact for devices. The thickness of the LiF layer was determined by a film thickness meter (INFICON, SQM-160). The schematic of the device structure is shown in Fig. 1(a). Reference solar cells without the LiF interlayer were fabricated in the same way. Furthermore, the HNO₃ doping process was carried out by exposing Gr films to HNO₃ fumes for 1 min. The Ag electrodes gradually turn yellow during prolonged HNO₃ doping process so the doping process should be carefully controlled to avoid severe oxidation of the Ag front electrodes.

Fig. 1 (a) Schematic of a Gr–Si Schottky solar cell; (b) the Raman spectrum of a Gr film.

The performance of solar cells were measured by a Keithley 2400 source meter and a solar simulator (94022A, Newport®) under AM 1.5G conditions at an illumination intensity of 100 mW cm⁻², calibrated by a standard Si solar cell (PVM937, Newport®). The external quantum efficiency (EQE) of solar cells was measured by an EQE measurement system (QEX10, PV Measurements, Inc.) across a wavelength range of 400–1100 nm.

3. Results and discussion

Fig. 1(b) shows the Raman spectrum of a Gr film. A very weak D peak at ∼1350 cm⁻¹ indicates that the Gr film is of high quality. The intensity ratio of the 2D and the G peak is 2.2, suggesting that the Gr film is a monolayer. Fig. 2 shows the scanning electron microscopy (SEM) image of a nominally 0.5 nm thick LiF layer deposited on the Si wafer. The LiF layer as grown is formed by close-packed flakes with diameters of around 20 nm. Since the deposited LiF layers are not very uniform, all the thickness of the LiF interlayer mentioned in following discussions are nominal thicknesses.

Fig. 2 Scanning electron microscopy (SEM) image of 0.5 nm thick LiF layer deposited on Si.

Fig. 3(a) displays the current density–voltage (J–V) curves of our best graphene–Si solar cell with and without a LiF interlayer under AM 1.5G conditions and illumination of 100 mW cm⁻². Moreover, the photovoltaic parameters of solar cells with and without a LiF interlayer, including short-circuit current density (J_SC), open-circuit voltage (V_OC), fill factor (FF) and PCE, are summarized in Table 1. It can be noted that an obvious increase in V_OC can be observed, with a maximum value up to ∼424 mV, for the solar cells with a LiF interlayer. The J_SC for the devices with and without LiF interlayers were in the ranges of 24.83–27.18 mA cm⁻² and 21.39–26.59 mA cm⁻², respectively. The values of FF for the reference cells were in the range of 44.21–52.25%, which increased to 52.78–54.50% for solar cells with a LiF interlayer. As a result, the PCE of devices increased from 3.81–4.67% to 5.48–6.25% after the application of LiF interlayers. Such high efficiencies, for our solar cells with LiF interlayers are comparable or even higher than those for Gr–GO–Si MIS solar cells reported by Yang et al., which hold the current record for Gr–Si Schottky solar cells without chemical doping.¹¹ After the HNO₃ doping process, the V_OC, J_SC, FF and PCE of the Gr–Si solar cells with LiF interlayers were dramatically improved to 540 mV, 29.07 mA cm⁻², 67.54% and 10.61%, respectively. To the best of our knowledge, this efficiency is the highest value to date for Gr–Si Schottky solar cells without an antireflection layer.

Fig. 3 (a) J–V curves under the illumination of pristine and HNO₃-doped devices with and without a LiF interlayer; inset shows the image of the device; (b) external quantum efficiency (EQE) spectra of devices with and without a LiF interlayer; (c) J–V curves of devices with and without a LiF interlayer under dark conditions; (d) series resistances (R_s) extracted from dV/d(ln J) versus J curves for our best Gr–Si Schottky junction solar cells with and without a LiF interlayer.

Table 1 Summary of the photovoltaic parameters of devices with and without a LiF interlayer fabricated under the same conditions

Back electrode	VOC (mV)	JSC (mA cm^−2)	FF (%)	PCE (%)	Rs (Ω cm^2)
Al	336.21	25.63	44.21	3.81	2.43
	411.03	21.39	52.25	4.59	1.15
	361.55	26.59	45.18	4.34	0.95
	397.70	23.23	50.55	4.67	0.93
LiF/Al	424.83	27.11	54.23	6.25	0.85
	400.42	26.29	53.06	5.58	0.80
	405.34	24.83	54.50	5.48	0.92
	417.26	27.18	52.78	5.99	0.87

---

Fig. 3(b) compares the EQE spectra of Gr–Si solar cells with and without a LiF interlayer. It is clearly seen that the EQE of devices with a LiF interlayer was significantly improved in the long wavelength (800–1100 nm) range. The integrated current densities for the devices with and without a LiF interlayer were 26.26 mA cm⁻² and 24.66 mA cm⁻², respectively, which are consistent with the J_SC improvement observed in J–V curves. It should be noted that the EQE of the Gr–Si solar cell with a LiF interlayer has the highest value of ∼65% in the range of 750–850 nm, which is in agreement with the reflectivity rate of Si.^13,14 It indicates that our solar cell has high carrier collection efficiency.

To further understand the improvement provided by the LiF interlayer, we compare the J–V curves of Gr–Si solar cells with and without a LiF interlayer under dark conditions, as shown in Fig. 3(c). The reversed saturation current (J_s) can be extracted by fitting the J–V curves based on the thermionic emission model as follows,

where J is current density; V is applied voltage; T is the absolute temperature; k is the Boltzmann's constant and q is the electronic charge. The J_s of devices with a LiF interlayer was 2.50 × 10⁻⁴ mA cm⁻², which is one order of magnitude lower than that of devices without a LiF interlayer, (6.34 × 10⁻³ mA cm⁻²). This indicates that the carrier recombination is significantly suppressed due to the introduction of the LiF interlayer. Fig. 3(d) shows the dV/d(ln J) versus J curves for the solar cells with and without a LiF interlayer. The series contact (R_s) of devices were extracted from the dV/d(ln J) versus J curves (see details in ESI†).^15,16 It can be seen that the values of R_s for the devices with a LiF interlayer are in the range of 0.80–0.92 Ω cm², while those for the devices without a LiF interlayer are in the range of 0.93–2.43 Ω cm². This reduction in R_s will improve the FF of devices.

Fig. 4 compares the contact resistances between Al and n-Si before and after introducing the LiF interlayer, based on the transmission line measurement (TLM) method.¹⁷ It can be seen that the resistance is 10.5 Ω for the direct Al/Si contact and 3.49 Ω for the Al/LiF/Si contact. Further calculations (see details in ESI†) indicate that the contact resistance decreases from 1.9 × 10⁻¹ Ω cm² to 6.3 × 10⁻² Ω cm² upon the introduction of a LiF interlayer.

Fig. 4 Contact resistance (R_c) extracted by the transmission line method (TLM) for the direct Al/Si contact and the Al/LiF/Si contact. The inset shows the schematic of a device for measurement.

In principle, the reduction of carrier recombination leads to the improvement of EQE. Long wavelength photons tend to be absorbed in the deep region of the Si substrate. In the Gr–Si solar cells, holes generated in such deep regions can easily migrate towards the back surface, wherein carrier recombination takes place.¹⁸ Therefore, we argue the observed higher EQE at long wavelengths can be ascribed to less recombination and more efficient electron collection at the back surface of solar cells.

Based on the analysis of the J–V, EQE and contact resistance data for the samples with and without LiF interlayers, we propose that the LiF interlayer has beneficial effects on band alignment of Gr–Si solar cells (Fig. 5). It is well-accepted that a Schottky barrier always exists in the case of metal directly contacting with Si, due to Fermi level pinning by interface states. For Al/n-Si Schottky junction, the potential barrier height (ϕ_B) was reported to be ∼0.5 eV.¹⁹ In this case, the electrons need to cross through such an energy barrier to be collected by the Al back electrode and therefore reduce the electron current (see the red solid line in Fig. 5(a)). However, these holes will increase the recombination probability of the electron current. As a result, the decrease of current and the drop of EQE in the long wavelength range occur for the reference Gr–Si solar cell, which are consistent with our observations. In the presence of a LiF ultra-thin interlayer (see Fig. 5(b)), the inserted LiF can dramatically reduce the work function of Al to about 3.3 eV,²⁰ due to the existence of a large dipole moment.²¹ This finally reduces the Schottky barrier between Al and Si. In some extreme cases, the large work function difference generates band bending towards the opposite direction, which could repel the holes back and therefore reduce the carrier recombination.¹⁶ The suppressed recombination and enhanced holes blocking effect should lead to improvement of J_SC and V_OC of devices. It should be noted that the LiF in this experiment, as an insulator, is thin enough for electrons to tunnel through. Therefore, the reduction in contact resistance, as well as series resistance of Gr–Si solar cells, can be ascribed to the lowering of the Al work function by the introduction of LiF, which has been widely observed in OLED^20–23 and OPV^24,25 devices.

Fig. 5 Scheme of band structure of devices (a) without and (b) with a LiF interlayer.

The adjustment of band alignment of Gr–Si Schottky solar cells offered by the LiF interlayer could help explain the improvement of J_SC, V_OC and FF of devices. We should mention that the optimal thickness of the LiF interlayer reported in photovoltaic devices was in the range of 0.5–1.2 nm.^26–29 However, in our experiments, the value 0.5 nm is preferred. We compared six Gr–Si solar cells in the same batch with 0.5 and 1.2 nm thick LiF interlayers. Relatively poorer performance for devices with a 1.2 nm LiF interlayer is obtained (see J–V characteristics under illumination shown in Fig. S2(a)† and the photovoltaic parameters of all six devices summarized in Table S1†). We notice that the J_SC of solar cells with different thicknesses of LiF interlayers are almost identical, which is consistent with the EQE spectrum (see Fig. S2(b)†). The electron transport could become more difficult for a thicker LiF interlayer. Therefore, lower FF can be obtained for solar cells with a thicker LiF interlayer, leading to poorer photovoltaic performance.

Herein, special attention has been placed on the stability of Gr–Si solar cells with a LiF interlayer after storing for several days. Table S2† summarizes the evolution of the photovoltaic parameters for pristine Gr–Si solar cells with LiF interlayers during the storing time. A dramatic increase of V_OC and PCE is obtained after one day of storing, which can be attributed to the natural doping effect induced by O₂ and H₂O absorbed on annealed Gr films.^30,31 Furthermore, the performance of devices remains almost constant after one week, indicating that the Gr–Si solar cells with a LiF interlayer have a good stability. However, the performance of HNO₃-doped Gr–Si solar cells degrades rapidly, which is consistent with previous reports.³² Much more effort should be paid on developing stable doping method for Gr films.

4. Conclusions

In summary, we have introduced a LiF ultra-thin interlayer between the Si and Al back contact to improve the performance of Gr–Si Schottky junction solar cells. High PCEs of 6.25% and 10.61% for pristine and chemically-doped Gr–Si solar cells, respectively, have been achieved. The improved performance was ascribed to the reduction in carrier barrier height, recombination, and contact resistance at the back side of the solar cell. Such device architecture has potential to be further developed if antireflection techniques and multilayer Gr are applied. These results pave a way to the fabrication of high efficiency Gr–Si solar cells, which are of great interest for low-cost photovoltaics.

Acknowledgements

This study is supported by the National Natural Science Foundation of China (nos 61422404 and 51472219), the Central basic scientific research in colleges and universities operating expenses, the Program for Innovative Research Team in University of Ministry of Education of China (IRT13R54), the State Key Laboratory of Optoelectronic Materials and Technologies (Sun Yatsen University) and the Zhejiang University K.P.Chao's High Technology Development Foundation.

References

1. Y. Jia, J. Wei, K. Wang, A. Cao, Q. Shu, X. Gui, Y. Zhu, D. Zhuang, G. Zhang and B. Ma, Adv. Mater., 2008, 20, 4594–4598.
2. D. D. Tune, B. S. Flavel, R. Krupke and J. G. Shapter, Adv. Energy Mater., 2012, 2, 1043–1055.
3. J. Wei, Y. Jia, Q. Shu, Z. Gu, K. Wang, D. Zhuang, G. Zhang, Z. Wang, J. Luo, A. Cao and D. Wu, Nano Lett., 2007, 7, 2317–2321.
4. X. Li, Y. Jung, K. Sakimoto, T.-H. Goh, M. A. Reed and A. D. Taylor, Energy Environ. Sci., 2013, 6, 879.
5. A. Ferrari, J. Meyer, V. Scardaci, C. Casiraghi, M. Lazzeri, F. Mauri, S. Piscanec, D. Jiang, K. Novoselov and S. Roth, Phys. Rev. Lett., 2006, 97, 187401.
6. A. K. Geim and K. S. Novoselov, Nat. Mater., 2007, 6, 183–191.
7. Y. Zhang, Y.-W. Tan, H. L. Stormer and P. Kim, Nature, 2005, 438, 201–204.
8. K. Novoselov, A. K. Geim, S. Morozov, D. Jiang, M. K. I. Grigorieva, S. Dubonos and A. Firsov, Nature, 2005, 438, 197–200.
9. X. Li, H. Zhu, K. Wang, A. Cao, J. Wei, C. Li, Y. Jia, Z. Li, X. Li and D. Wu, Adv. Mater., 2010, 22, 2743–2748.
10. 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, 1776–1781.
11. L. F. Yang, X. G. Yu, M. S. Xu, H. Z. Chen and D. R. Yang, J. Mater. Chem. A, 2014, 2, 16877–16883.
12. X. Li, W. Cai, J. An, S. Kim, J. Nah, D. Yang, R. Piner, A. Velamakanni, I. Jung, E. Tutuc, S. K. Banerjee, L. Colombo and R. S. Ruoff, Science, 2009, 324, 1312–1314.
13. H. M. Branz, V. E. Yost, S. Ward, K. M. Jones, B. To and P. Stradins, Appl. Phys. Lett., 2009, 94, 231121.
14. Y.-F. Huang, S. Chattopadhyay, Y.-J. Jen, C.-Y. Peng, T.-A. Liu, Y.-K. Hsu, C.-L. Pan, H.-C. Lo, C.-H. Hsu and Y.-H. Chang, Nat. Nanotechnol., 2007, 2, 770–774.
15. A. Tataroğlu and Ş. Altındal, Microelectron. Eng., 2008, 85, 233–237.
16. Y. Zhang, W. Cui, Y. Zhu, F. Zu, L. Liao, S.-T. Lee and B. Sun, Energy Environ. Sci., 2015, 8, 297–302.
17. G. K. Reeves and H. B. Harrison, Electron Device Lett., 1982, 3, 111–113.
18. J. Britt and C. Ferekides, Appl. Phys. Lett., 1993, 62, 2851.
19. B. Carr, E. Friedland and J. Malherbe, J. Appl. Phys., 1988, 64, 4775–4777.
20. S. E. Shaheen, G. E. Jabbour, M. M. Morrell, Y. Kawabe, B. Kippelen, N. Peyghambarian, M. F. Nabor, R. Schlaf, E. A. Mash and N. R. Armstrong, J. Appl. Phys., 1998, 84, 2324.
21. X. J. Wang, J. Appl. Phys., 2004, 95, 3828.
22. T. M. Brown, R. H. Friend, I. S. Millard, D. J. Lacey, J. H. Burroughes and F. Cacialli, Appl. Phys. Lett., 2000, 77, 3096.
23. T. Mori, H. Fujikawa, S. Tokito and Y. Taga, Appl. Phys. Lett., 1998, 73, 2763–2765.
24. V. Mihailetchi, P. Blom, J. Hummelen and M. Rispens, J. Appl. Phys., 2003, 94, 6849–6854.
25. C. J. Brabec, S. E. Shaheen, C. Winder, N. S. Sariciftci and P. Denk, Appl. Phys. Lett., 2002, 80, 1288–1290.
26. E. Ahlswede, J. Hanisch and M. Powalla, Appl. Phys. Lett., 2007, 90, 163504.
27. G. Li, C. W. Chu, V. Shrotriya, J. Huang and Y. Yang, Appl. Phys. Lett., 2006, 88, 253503.
28. S. E. Shaheen, C. J. Brabec, N. S. Sariciftci, F. Padinger, T. Fromherz and J. C. Hummelen, Appl. Phys. Lett., 2001, 78, 841.
29. Y. Zhang, R. Liu, S.-T. Lee and B. Sun, Appl. Phys. Lett., 2014, 104, 083514.
30. Z. H. Ni, H. M. Wang, Z. Q. Luo, Y. Y. Wang, T. Yu, Y. H. Wu and Z. X. Shen, J. Raman Spectrosc., 2009, 41, 479–483.
31. S. Ryu, L. Liu, S. Berciaud, Y. J. Yu, H. Liu, P. Kim, G. W. Flynn and L. E. Brus, Nano Lett., 2010, 10, 4944–4951.
32. L. Yang, X. Yu, W. Hu, X. Wu, Y. Zhao and D. Yang, ACS Appl. Mater. Interfaces, 2015, 7, 4135–4141.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c5ra05619e

---