Silicon solar cell technology is dominating in the photovoltaic industry, however, a further increase in cost-effectiveness is needed. This can be achieved by increasing the light coupling in the active layer of the cell. Passivation layer of the commercial polycrystalline silicon solar cells was textured using the laser beam interference patterning technique to obtain enhanced optical properties due to the formation of periodic refractive index grating. As a consequence of laser irradiation, passivation layer of silicon nitride was partially oxidized at the intensity peaks of the interference distribution. Periodic distribution of oxidized areas was found by EDS. Investigation of optical and electrical properties of the laser treated solar cells shows increased light coupling and better photo-electrical performance. Simulations were carried out to evaluate the influence of refractive index grating formed in silicon oxy-nitride to optical properties of the patterned solar cells. Periodic oxide grating was found to be more efficient for light coupling in thinner film structures. The results of this work may find applications in other fields as it allows alteration of material composition in the well-defined periodic pattern.
© 2015 Optical Society of America
Crystalline silicon (c-Si) technology is dominant in photovoltaic (PV) industry and is expected to retain its market share due to maturity of the technology . An average PV module price is rapidly decreasing, however, the majority of cost reduction over the last years has been driven by declines in material prices . Additional means to increase the cost-effectiveness of silicon solar cells involves improved coupling of light in the active layer of the cell.
Current approaches for light management in c-Si photovoltaic elements are based on controlling light ray paths through geometrical optics. The predominant silicon cell technology utilizes pyramid texture, created by anisotropic etching and coated with anti-reflective film. Low reflection from the cell surface is ensured due to multiple reflections from the pyramid facets and the effect of anti-reflective coatings. In addition, the pyramid structure scatters incident light rays at various angles and optical path length in the material increases. Further improvement of solar cell efficiency can be achieved ensuring better light coupling in the active layer of the cell by further reduction of the front-side reflection.
The structures, based on wave optics, were shown to be capable, in principle, of outperforming all geometrical optics approaches. Light management using wave-optics includes enhancement of an optical path and reflectivity reduction via diffractive structures, i.e. periodic dielectric structures like diffraction gratings , photonic crystal structures , structures, that use plasmonic effects .
One of the widely investigated photonic structures is periodic hole arrays in the top passivation layer or silicon substrate. It was shown, that the hole arrays in the substrate provide better light coupling than the rod arrays . Effects of such arrays on properties of thin-film solar cells based on hydrogenated amorphous silicon (a-Si:H) were studied in . Silicon thin films with the periodic hole arrays showed increased short-wavelength absorption. Increased absorption was observed in the a-Si:H thin-film solar cell with a periodic hole array in both a-Si:H and the top ITO electrode . Manufacturing of plasmonic resonant structures in the passivation layer of the solar cell was also proposed. Simulations and experimental results showed an increase in photo-electrical conversion efficiency of the photovoltaic element with nanohole patterned antireflective layer .
Usually, periodic arrays in the passivation layer are formed utilizing interference lithography [7, 8, 10], combined with the reactive-ion etching. In this work, we present a patterning method for the top passivation layer using the direct laser beam interference patterning. The high pulse energy laser source provides sufficient energy to remove or modify the material for the creation of desired pattern without additional etching step after irradiation. Such method allows obtaining good quality structures with a single laser shot, is relatively simple and allows to reach industrial fabrication speeds. We proposed the direct laser-induced periodic structure formation in the passivation layer as a new approach to fabrication of photonic structures. In this paper, we investigate the effect of laser-induced silicon nitride/silicon oxy-nitride grating in the silicon nitride passivation layer on electrical and optical properties of commercially available polycrystalline silicon solar cells.
Silicon nitride is known to have good oxidation resistance. Si3N4 oxidation usually takes place at high temperatures and results in the formation of silicon oxy-nitride (SiOxNy) and silicon dioxide (SiO2), followed by the release of nitrogen . Thermal oxidation of silicon nitride was demonstrated in the temperature range from 800 to 1260 °C [12, 13].
Oxidation can also take place due to the impact of laser irradiation. The laser-induced temperature rise enhances the diffusion flux and the reaction rate of species within the irradiated area. This enhancement is based on a variety of different mechanisms: temperature dependence of ordinary diffusion, thermal generation of defects such as vacancies, and thermal excitation of electrons .
Laser ablation of silicon nitride layer also affects the lifetime of minority carriers. Ablation with nanosecond pulses was shown to reduce carrier lifetime . However, femtosecond pulse irradiation provided no significant impact to the lifetime . Our approach involves only modification of the silicon nitride layer and, therefore, allows minimizing any damage to the silicon layer.
2. Patterning results
Periodic selective changes in the silicon nitride layer were induced using the laser beam interference technique. The laser beam was split into four beams. Beamlets were collected on the sample using a system of lenses. Interfering laser beams provide periodic distribution of irradiation intensity, consisting of high-intensity spikes on the low-intensity background. Second harmonics (532 nm) of picosecond laser Atlantic HE (pulse duration 60 ps, repetition rate 4 kHz, maximum pulse energy 0.4 mJ) was used. Localized regions of increased temperature were induced by irradiating the silicon nitride layer with such intensity distribution. Ablation of pits in the polycrystalline silicon layer or selective removal of the silicon nitride layer can be obtained by varying pulse energy (Fig. 1). 2.5 μm period intensity distribution was selected, to ensure more straightforward sample analysis.
Removal of the passivation layer led to degradation of electrical properties of the solar cell. Therefore, the study was focused on the low-pulse energy irradiation regimes. Despite the lack of the visible alteration in the morphology of the silicon nitride layer, local changes in chemical composition of the sample were observed after laser irradiation (Fig. 2).
XPS analysis also provides that laser irradiation modify the element ratios in the sample (Fig. 3). Oxygen and nitrogen content changes after laser irradiation.
3. Optical measurements
Changes in optical properties of laser irradiated solar cell samples were evaluated experimentally by measuring diffusive reflection spectra (Fig. 4(a)). Measurements were performed using Shimadzu UV-VIS-NIR Spectrophotometer UV-3600 coupled with the MRC-3100 unit with integrating sphere. Integrating sphere collects reflected (R) and transmitted (T) light. Then absorbtance of the sample can be calculated from A = 1 – R – T. Solar cell samples are optically thick in the relevant spectral region. Therefore T = 0 absorbtance was calculated from: A = 1 – R (Fig. 4(b)). The absorption coefficient can be also evaluated from diffuse reflection spectrum using Kubelka-Munk theory. If Rd is the reflection coefficient, then the absorption coefficient K is: K = (1-Rd)2/(2Rd) [17–21]. Absorption spectra are presented in Fig. 4(c) and 4(d).
To understand an effect of laser-induced periodic modification in the antireflective SiNx layer, simulation of light propagation and related layer reflectivity was conducted. As localized presence of the areas with increased oxygen concentration was confirmed experimentally, structures with altered refractive index in the top passivation layer were simulated. We simulated gratings in the silicon nitride layer under two different terms: the large-period gratings (in comparison with the wavelength of visible light λ) with a period Λ>>λ and the small-period gratings with Λ~ = λ.
For the large period gratings, the grating structure itself does not affect light absorption, and reflectivity of the grating can be evaluated by a simulating structure with the uniformly oxidized top layer. The period of the experimentally obtained grating was 2.5 μm. This value is significantly larger than irradiation wavelength in the relevant spectral region. Therefore, such grating can be approximated as uniform silicon oxide layer. The only difference is in less significant change of reflectivity because only a part of the structure was coated with the modified refractive index layer in the experiment.
If grating period Λ and irradiation wavelength are comparable, an exact solution can be found only by solving Maxwell’s equations. Interaction of the small-period grating with the incident electromagnetic wave can no longer be described as interaction with uniform layers. For example, Fig. 6 shows that modified refractive index region concentrates electromagnetic power.
Rigorous Coupled Wave Analysis (RCWA) Matlab implementation by Pavel Kwiecien (rcwa-1d)  was used to simulate periodic gratings in SiNx layer on planar and pyramidal Si substrates. The rigorous coupled-wave analysis has been widely used for the analysis and the design of diffractive structures. It is an exact solution of Maxwell’s equations for the electromagnetic diffraction by grating structures. The grating is divided into a number of sufficiently thin planar slabs to approximate the grating profile to an arbitrary degree of accuracy. Electromagnetic fields in each grating layer are determined by the coupled-wave approach. The electromagnetic boundary conditions (continuity of the tangential electric- and magnetic-field components) are then applied in a sequence at the interfaces among the output region, the individual grating layers. Finally, the input region to yield the reflected and the transmitted diffracted field amplitudes and the diffraction efficiencies .
Simulation of the fully periodic structure is much less computationally expensive than random structure. Therefore, silicon substrate with the constant period upright-pyramid structure, instead of a random pyramid structure, used in the OPAL2 simulation, was utilized. The simulated dependence of reflectivity on the period Λ of the silicon oxide/oxy-nitride grating on a regular pyramidal silicon substrate (period of the pyramid pattern P was constant (P = 4 μm, α = 54.74 deg) in the simulation) is presented in Fig. 7. The refractive index of silicon nitride layer was taken from . The refractive index of silicon was given by . Oxidation of silicon nitride layer was counted in by introducing an effective refractive index ne: ne = x∙nSi3N4 + y∙nSiO2, where x + y = 1 . TM polarized wave (electric field vector is in the plane of incidence) was impinging on the top of the structure at 0 deg angle of incidence. Only the thin upper silicon nitride layer was modified in the simulation. Therefore, geometry changes of the structure are small compared to the irradiation wavelength. For this reason, the effect of the grating period on the reflectivity is modest. The effect of grating period Λ is apparent only in the short wavelength region. In the remaining region, reflectivity curve retains its shape.
The standard AM1.5 Global spectrum was multiplied by the simulated values of light absorbtance A (A = 1 - R) and integrated in the 280 nm – 1000 nm spectral range. Performance of the ne refractive index gratings with the periods of 1 μm, 0.75 μm, 0.5 μm and 0.375 μm in the 70 nm-thick SiNx layer on the silicon pyramid substrate was compared to the performance of the uniform SiNx layer on silicon pyramids (Table 1). Effective refractive index of the oxidized silicon nitride layer ne = 0.5 nSi3N4 + 0.5 nSiO2 was used in the calculation.
Refractive index grating in the silicon nitride layer provides more significant effect in the thin film device. Reflectivity and absorption spectra of the 2 μm-thick silicon layer with 75 nm-thick silicon nitride layer on top are presented in Fig. 8.
5. Electrical measurement
Commercial polycrystalline silicon solar cell samples were textured using the laser beam interference patterning technique. The 6” size poly-Si cell with average 11% PV efficiency was divided into smaller segments. Several lines were scanned between the contact grid to pattern uniformly the passivation layer in the 20.8 cm2 total area sample. Approximately 70% of the area between the contact grid was patterned. Patterning regimes with modifications in the SiNx layer only were selected by varying the laser pulse energy, energy density and pulse number. Volt-ampere characteristics of the samples were measured before and after the laser treatment using dual-channel system Source Meter Keithley 2602A and a solar simulator made in our laboratory.
The volt-ampere characteristics of the solar cell sample with the laser-induced 2.5 μm period grating were measured. For comparison V-A characteristic of not treated sample is presented as the dashed line. About 1% increase in PV efficiency was found after the laser treatment. Results are presented in Table 2.
Direct laser beam interference patterning allowed producing two types of periodic structure in the silicon nitride passivation layer of the poly-Si solar cell. Patterning using the higher laser fluencies led to the removal of this layer in the periodically arranged local regions (SiNx/air grating). The periodic modification of silicon nitride layer was observed using reduced laser fluence. EDS and XPS analysis of the samples shows increased oxygen content in the irradiated samples.
EDS element maps show that carbon, nitrogen and oxygen elements are distributed almost uniformly in the control sample (with no laser treatment). There are only some irregularities that can be explained by the geometry of the pyramidal surface. However, EDS analysis of laser irradiated sample shows increased local oxygen concentration after irradiation (Fig. 2 (a), (b)). Periodic oxygen distribution with the same period as the interference pattern was obtained. Therefore, the conclusion can be made that silicon nitride layer was oxidized in the regions with the highest irradiation intensity. In the nitrogen EDS map, more localized regions with low nitrogen content can be observed than in the untreated sample. However, no evident correlation between interference intensity pattern and nitrogen distribution is apparent. High carbon content in the EDS map of laser treated sample is caused by the contamination with organic debris, but not by the laser irradiation.
XPS analysis of samples treated using different pulse energies shows that oxygen and nitrogen content decreases for the high pulse energy (150 μJ) patterning regime, when passivation layer is ablated at the intensity peak regions. For lower pulse energies, when no visible damage to passivation layer takes place, relative oxygen content O/Si is increased, and relative content of nitrogen N/Si remains as in the untreated sample. This suggests that low energy pulse induces oxidation process but does not ablate material.
In the optical properties measurements, two modification regimes are apparent. For low laser fluencies, the reflectivity decreased, and absorption increased in the 400 – 1000 nm region almost uniformly. For higher fluencies, the reflectivity reduction in the 300 – 400 nm region and its increase in the remaining region was observed.
Simulation of a random pyramid structure with the oxidized top part of silicon nitride layer shows a good agreement of the experimental data with simulation results for the higher laser fluence regime. After the laser treatment, the reflectivity is decreased in the short wavelength region and slightly increased for longer wavelengths. Such effect was observed in reflectivity experiments for high laser fluency (Fig. 4(a)) as well as reproduced in simulation (Fig. 5(a)). In optical coatings technology, one of the conditions for reflection elimination is to ensure that amplitudes of the waves reflected from different interfaces are equal, to achieve complete destructive interference. This condition is satisfied if refractive index ratio is the same at both interfaces : na/nf = nf/ns. na is the air refractive index, nf is the film refractive index, ns is the substrate refractive index. The silicon oxide/oxy-nitride layer on top of the silicon nitride passivation layer helps to fulfill this condition better in the short wavelength region.
However, for lower laser fluence, the absorption increased in the wide 500 nm – 1000 nm spectral region. Such broad range of decreased reflectivity is not apparent from the simulation. This variance may arise due to the discrepancy between the refractive index values of real device and values used in the simulation. The refractive index values of the SiNx coating strongly depends on the deposition conditions.
Simulations of structure with oxidation pattern show that the small-period grating with lower effective refractive index in the silicon nitride passivation layer provides decreased reflectivity only in a narrow spectral region (~350-520 nm). However, such grating still provides increased sunlight collection (Table 1). The simulated periodic gratings in the SiNx layer provide increased light coupling, despite the narrow region of decreased reflectivity. However, for thin film silicon substrate influence of grating period was more evident, and the effect was stronger than for the thick silicon substrate. Therefore, the formation of SiNx/SiNxOy structure would be more suitable for the thin film devices.
Photo-electrical measurements of the patterned solar cell sample showed an increase of device efficiency above 1% after the formation of SiNx/SiNxOy structure. Since absorption was increased about 10 percent after the laser irradiation and PV efficiency of the cells are 11%, this result shows that SiNx/SiNxOy grating in the antireflective layer of the PV device was functioning as an antireflective structure.
The relevant aspect of the proposed laser-induced SiNx/SiNxOy grating structure is in its applicability to the industrial process. Fabrication of a periodic oxide fabrication in the silicon nitride layer requires a low irradiation fluence, and a single laser shot is sufficient for the process. The time tfab required to texture an area S using the interfering laser beams depends on the laser pulse repetition rate f and the diameter of the laser spot W: tfab = S/(W2f) . In the experiment, the 0.3 mm spot size and 4 kHz repetition rate were used. Therefore, patterning of a 6” solar cell takes about one minute. However, since the process requires a low irradiation fluence, the spot size and repetition rate can be increased. A laser source with the 1 mm diameter spot size and 25 kHz repetition rate would allow to reduce the patterning time of the 6” solar cell to approximately 1 second.
Modification of the passivation layer of complete commercial poly-Si solar cells was performed using the laser interference patterning. Using low laser fluencies, the formation of periodic silicon nitride – silicon oxide gratings in the passivation layer was observed. Such grating provided increased optical absorption in the 400 - 1000 nm spectral regions. Photo-electrical characteristics of solar cells were improved after the formation of the 2.5 μm period SiNx/SiNxOy grating in the top passivation layer. However, this improvement was modest, due to the low initial efficiency of solar cell samples. Simulations show that reduction of the grating period affects reflectivity only marginally in the thick silicon substrate structures. Nevertheless, the period of the grating makes significant influence if the silicon substrate is thin. Therefore, SiNx/SiNxOy grating structure could find application for the thin film PV devices. Results of this work may find applications in other fields as it allows alteration of material composition in the well-defined periodic pattern.
This research was funded from the Research Council of Lithuania by the grant No. ATE-11/2012.
References and links
1. EPIA, “Global Market Outlook for Photovoltaics 2013–2017,” Global Market Outlook for Photovoltaics 2013–2017 (2013).
2. “GTM Research. 2013. PV technology and cost outlook, 2013–2017. Boston.”
3. A. Gombert, K. Rose, A. Heinzel, W. Horbelt, C. Zanke, B. Bläsi, and V. Wittwer, “Antireflective submicrometer surface-relief gratings for solar applications,” Sol. Energy Mater. Sol. Cells 54(1-4), 333–342 (1998). [CrossRef]
4. Y. J. Hung, S. L. Lee, and L. A. Coldren, “Deep and tapered silicon photonic crystals for achieving anti-reflection and enhanced absorption,” Opt. Express 18(7), 6841–6852 (2010). [CrossRef] [PubMed]
5. S. Pillai, K. R. Catchpole, T. Trupke, and M. A. Green, “Surface plasmon enhanced silicon solar cells,” J. Appl. Phys. 101(9), 093105 (2007). [CrossRef]
7. G. Gomard, R. Peretti, S. Callard, X. Meng, R. Artinyan, T. Deschamps, P. Roca i Cabarrocas, E. Drouard, and C. Seassal, “Blue light absorption enhancement based on vertically channelling modes in nano-holes arrays,” Appl. Phys. Lett. 104(5), 051119 (2014). [CrossRef]
8. G. Gomard, R. Peretti, E. Drouard, X. Meng, and C. Seassal, “Photonic crystals and optical mode engineering for thin film photovoltaics,” Opt. Express 21(S3Suppl 3), A515–A527 (2013). [CrossRef] [PubMed]
9. N. Wang, Y. Zhu, W. Wei, J. Chen, P. Li, and Y. Wen, “Conversion efficiency enhanced photovoltaic device with nanohole arrays in antireflection coating layer,” Opt. Commun. 284(19), 4773–4777 (2011). [CrossRef]
10. V. Auzelyte, H. Sigg, B. Schmitt, and H. H. Solak, “Direct formation of ZnO nanostructures by chemical solution deposition and EUV exposure,” Nanotechnology 21(21), 215302 (2010). [CrossRef] [PubMed]
11. H. Du, R. E. Tressler, K. E. Spear, and C. G. Pantano, “Oxidation Studies of Crystalline CVD Silicon Nitride,” J. Electrochem. Soc. 136(5), 1527–1536 (1989). [CrossRef]
12. S. I. Raider, R. Flitsch, J. A. Aboaf, and W. A. Pliskin, “Surface Oxidation of Silicon Nitride Films,” J. Electrochem. Soc. 123(4), 560–565 (1976). [CrossRef]
13. I. Fränz and W. Langheinrich, “Conversion of silicon nitride into silicon dioxide through the influence of oxygen,” Solid-State Electron. 14(6), 499–505 (1971). [CrossRef]
14. D. W. Bäuerle, Laser Processing and Chemistry (Springer, 2011).
15. W. Pfleging, A. Ludwig, K. Seemann, R. Preu, H. Mäckel, and S. W. Glunz, “Laser micromachining for applications in thin film technology,” Appl. Surf. Sci. 154–155, 633–639 (2000). [CrossRef]
16. C. S. K. Neckermann, G. Andrä, M. Bähr, J. Lossen, and E. Ose, “Local structuring of dielectric layers on silicon for improved solar cell metallization,” in 22th European photovoltaic solar energy conference, 2007)
17. P. Kubelka and F. Munk, “Ein Beitrag zur Optik der Farbanstriche,” Z. Tech. Phys. 12, 593–601 (1931).
20. L. Yang, B. Kruse, and S. J. Miklavcic, “Revised Kubelka-Munk theory. II. Unified framework for homogeneous and inhomogeneous optical media,” J. Opt. Soc. Am. A 21(10), 1942–1952 (2004). [CrossRef] [PubMed]
21. L. Yang and S. J. Miklavcic, “Revised Kubelka-Munk theory. III. A general theory of light propagation in scattering and absorptive media,” J. Opt. Soc. Am. A 22(9), 1866–1873 (2005). [CrossRef] [PubMed]
22. K. R. McIntosh and S. C. Baker-Finch, “OPAL 2: Rapid optical simulation of silicon solar cells,” in Photovoltaic Specialists Conference (PVSC),201238th IEEE, 2012), 000265–000271. [CrossRef]
23. P. Kwiecien, “rcwa-1d”, SourceForge, 2014, http://sourceforge.net/projects/rcwa-1d/.
24. M. G. Moharam, D. A. Pommet, E. B. Grann, and T. K. Gaylord, “Stable implementation of the rigorous coupled-wave analysis for surface-relief gratings: enhanced transmittance matrix approach,” J. Opt. Soc. Am. A 12(5), 1077–1086 (1995). [CrossRef]
26. E. D. Palik, Handbook of Optical Constants of Solids (Academic Press, Boston, 1985).
27. E. Hecht, Optics (Addison-Wesley Longman, Inc., San Francisco, 2002).
28. S. Indrišiūnas, B. Voisiat, A. Žukauskas, and G. Račiukaitis, “Direct laser beam patterning technique for fast high aspect ratio surface structuring,” Proc. SPIE 9350, 935003 (2015).