- Email: [email protected]
A comparison study of boron emitter passivation by silicon oxide and a PECVD silicon nitride stack
Available online at www.sciencedirect.com Available online at www.sciencedirect.com
ScienceDirect ScienceDirect
Energy Procedia 00 (2017) 000–000 Available online www.sciencedirect.com Available online atatwww.sciencedirect.com Energy Procedia 00 (2017) 000–000
ScienceDirect ScienceDirect
www.elsevier.com/locate/procedia www.elsevier.com/locate/procedia
Energy (2017) 000–000 288–294 EnergyProcedia Procedia124 00 (2017) www.elsevier.com/locate/procedia
7th International Conference on Silicon Photovoltaics, SiliconPV 2017 7th International Conference on Silicon Photovoltaics, SiliconPV 2017
A comparison study of boron emitter passivation by silicon oxide A comparison study of boron emitter passivation by silicon oxide a PECVD silicon nitrideHeating stack The 15thand International Symposium on District and a PECVD silicon nitride stackand Cooling b b Barbora Mojrováaa*, the Haifeng Chubb, Christop Peterbbthe , Pirmin Preis b, Jan Lossenb, Valentin Assessing feasibility of using heat demand-outdoor b b Barbora Mojrová *, Haifeng Chu , Christop Peter , Pirmin Preis , Jan Lossen , Valentin D. Mihailetchi b, Radovan Kopecekb temperature function for a long-term district D. Mihailetchi , Radovan Kopecekheat demand forecast a
Departement of Microelectronics The Faculty of Electrical Engineering and Communication Brno University of Technology Departement ofa,b,c Microelectronics The Faculty of3058/10, Electrical Engineering andb Communication Brno University of Technologyc 616 00, Czech Republic aTechnicka a Brno c b Technicka 3058/10, Brno 616 15, 00,Konstanz Czech Republic ISC Konstanz e. V., Rudolf-Diesel-Str. 78467, Germany b ISC Konstanz e. V., Rudolf-Diesel-Str. 15, Konstanz 78467, Germany a IN+ Center for Innovation, Technology and Policy Research - Instituto Superior Técnico, Av. Rovisco Pais 1, 1049-001 Lisbon, Portugal b Veolia Recherche & Innovation, 291 Avenue Dreyfous Daniel, 78520 Limay, France c Département Systèmes Énergétiques et Environnement - IMT Atlantique, 4 rue Alfred Kastler, 44300 Nantes, France a
I. Andrić
*, A. Pina , P. Ferrão , J. Fournier ., B. Lacarrière , O. Le Corre
Abstract Abstract Surface recombination of minority charge carriers is a significant loss mechanism in crystalline-Si (c-Si) solar cells. For + Surface recombination minority charge carriers is a significant loss mechanism crystalline-Si (c-Si) solar cells. For (phosphorous) doped of regions a sufficient surface passivation is achieved by using a in silicon nitride (SiN nAbstract X) film deposited by n+ (phosphorous) regions a sufficient surface passivation is achieved using a silicon nitride ) film deposited the by X Plasma Enhanced doped Chemical Vapour Deposition (PECVD). However, on p+ by boron doped regions, SiN(SiN does not passivate X boron doped regions, SiN does nota passivate the Plasma Enhanced Chemical Vapour Deposition (PECVD). However, on p+2), X or aluminium oxide (Al O ), and SiN film is surface. Instead, a stack of layers comprising of a thin silicon dioxide (SiO 2 3 X District heating networks are commonly addressed in the literature as one of the most effective solutions for decreasing the surface. Instead, a passivate stack of layers comprising of a thin silicon dioxide (SiO2), or aluminium oxide (Al2O3), and a SiNX film is commonly boron regions. greenhouseused gastoemissions from thedoped building sector. These systems require high investments which are returned through the heat commonly used toinvestigated passivate boron doped regions. Insales. this study the passivation quality of boron renovation doped emitters by varying the composition of SiO 2/SiN X stack Due towethe changed climate conditions and building policies, heat demand in the future could decrease, In this For studythis wepurpose, investigated the (passivated passivation emitter, quality of boron doped emitters bycells varying the composition of SiO 2/SiN X stack layers. n-PERT rear totally-diffused) solar with boron doped front side emitter and prolonging the investment return period. layers. For this purpose, n-PERT (passivated emitter, rearsymmetrical totally-diffused) solar cells with boron doped front on side6-inch emitter and phosphorous doped back-surface-field (BSF), as well as boron doped structures, were fabricated n-type The main scope of this paper is to assess the feasibility of using the heat demand – outdoor temperature function for heat demand phosphorous doped back-surface-field (BSF), as well as symmetrical boron doped structures, were fabricated on 6-inch n-type wafers. forecast. The district of Alvalade, located in Lisbon (Portugal), was used as a case study. The district is consisted of 665 wafers. a thickness at least nm, The results that showvary thatintheboth optimum passivation is achieved by a stack of a thermal 2, with buildings construction period and typology. Threelayer weather scenariosSiO (low, medium, high)ofand three10 district , with a thickness of at least 10 nm, The aresults show that the optimum passivation is achieved by a stack layer of a thermal SiO 2 layer with a low refractive index. The chemical composition of SiN capping layer plays an important role for surface and SiN X scenarios were developed (shallow, intermediate, deep). To estimate X renovation the error, obtained heat demand values were and a SiNX layer withSi-rich a low SiN refractive index. The chemical composition of SiNXpassivation capping layer plays an important role for surface passivation. A more layer show significant degradation in surface of the stack, due to the increase in the X compared with results from a dynamic heat demand model, previously developed and validated by the authors. show significant degradation in surface passivation of the stack, due to the increase in the passivation. A morestates Si-rich SiN X layer ) and fixed positive charges (Q ) at the interface. density of interface (D it tot The results showed that when only weather change is considered, the margin of error could be acceptable for some applications density of interface states (Dit) and fixed positive charges (Qtot) at the interface. (the error in annual demand was lower than 20% for all weather scenarios considered). However, after introducing renovation © 2017 Thethe Authors. Published by Elsevier Ltd. error value increased up to 59.5% ©scenarios, 2017 The Authors. Published by Elsevier Ltd. (depending on the weather and renovation scenarios combination considered). © 2017 The by Authors. Published by Elsevier Ltd. of SiliconPV 2017 under responsibility of PSE AG. Peer review the scientific conference committee Thereview value of coefficient increased on average within the range of responsibility 3.8% up to 8% per decade, that corresponds to the Peer by slope the scientific conference committee of SiliconPV 2017 under of PSE AG. Peer review by the scientific conference committee of SiliconPV 2017 under responsibility of PSE AG. decrease in the number of heating hours of 22-139h during the heating season (depending on the combination of weather and Keywords: Solar cell; boron emitter; n-PERT; passivation; NAOS renovation scenarios considered). On the other hand, function intercept increased for 7.8-12.7% per decade (depending on the Keywords: Solar cell; boron emitter; n-PERT; passivation; NAOS coupled scenarios). The values suggested could be used to modify the function parameters for the scenarios considered, and improve the accuracy of heat demand estimations. © 2017 The Authors. Published by Elsevier Ltd. * Corresponding author. Tel.: +420 54114 fax: +420 54114 6298. Peer-review under responsibility of the6107; Scientific Committee of The 15th International Symposium on District Heating and * Corresponding Tel.: +420 54114 6107; fax: +420 54114 6298. E-mail address:author. [email protected] Cooling. E-mail address: [email protected]
1876-6102 2017demand; The Authors. Published Elsevier Ltd. Keywords:©Heat Forecast; Climatebychange 1876-6102 The Authors. Published by Elsevier Ltd. Peer review©by2017 the scientific conference committee of SiliconPV 2017 under responsibility of PSE AG. Peer review by the scientific conference committee of SiliconPV 2017 under responsibility of PSE AG.
1876-6102 © 2017 The Authors. Published by Elsevier Ltd. Peer-review under responsibility of the Scientific Committee of The 15th International Symposium on District Heating and Cooling.
1876-6102 © 2017 The Authors. Published by Elsevier Ltd. Peer review by the scientific conference committee of SiliconPV 2017 under responsibility of PSE AG. 10.1016/j.egypro.2017.09.301
Barbora Mojrová et al. / Energy Procedia 124 (2017) 288–294 Barbora Mojrová et al. / Energy Procedia 00 (2017) 000–000
289
1. Introduction Surface passivation of heavily doped p+ emitters is required in order to realize cost-effective industrial highefficiency solar cells based on n-type silicon wafers. The boron emitter can be passivated by a dielectric having a low density of interface states (Dit), potentially in combination with a high fixed negative charge density. However conventional SiNX is not suitable for p+ surface passivation [1]. One of the reasons for the relatively poor surface passivation of p+ surfaces by SiNX is related to the very high fixed charge density in PECVD SiNX. It is attributed to the band gap defects that SiNX form during deposition, which are more detrimental to p+ Si [2, 3]. It was observed that the fixed charge density can be reduced by adding a thin silicon oxide layer between the crystalline silicon and the SiNX film. Another approach to passivate p+ layer is based on using thin Al2O3 dielectric with the negative charge as a surface passivation layer and SiNX as an antireflection coating (ARC) layer [4-7]. 2. Experimental part 2.1. Sample preparation In this work we investigated the passivation of boron doped emitters with different SiO2/SiNX stack layers. For this purpose we used, on the one hand, an n-PERT cell with a homogeneous diffused front boron emitter and a phosphorous back-surface-field (BSF) and on the other hand, symmetrical boron diffused test structure. Both device structures were fabricated on 6-inch n-type monocrystalline Si wafers with base resistivity of 2.5 cm. Standard industrial processes were used, which included wet chemical alkaline texturization, cleaning by HCl, HF, and HF/O3 solutions, diffusion in quartz tube furnaces containing phosphoryl chloride (POCl3; for n+ BSF) or boron tribromide (BBr3; for p+ emitter). For the solar cell devices the back side was chemically polished and the n+ BSF diffusion was passivated by a thermal SiO2/ SiNX stack. Screen printing and firing through of commercial silver containing pastes were applied to both sides for metallization. A schematic cross-section of the studied solar cells is presented in the Fig. 1.
Fig. 1. Schematic cross-section of the investigated n-PERT cells.
2.2. Passivation stacks All solar cells were processed identically, with the exception of boron emitter passivation. For the boron emitter 5 different passivation stacks were investigated and compared, as detailed in Table I. First, the stacks G1 and G2 aimed to investigate the effect of SiO2 interfacial layer thickness on the passivation quality of the SiO2/SiNX stack. A symmetrical boron diffused test structure was used, in addition to the solar cells, to find the optimum thickness of the SiO2 in the stack. Here, a thermal SiO2 was grown in situ during the BBr3 diffusion on boron emitter surface with a thickness of about 40 nm. Subsequently the thick SiO2 layer was partially etched (thinned) in HF (2%) solution to yield various thicknesses, followed by a 62 nm PECVD SiNX antireflection coating (ARC; n ≈ 2) deposition of the capping layer to complete the passivation stack. Then, the stacks G1 and G3 aimed to compare different SiO2 interface passivation layers, whereas the SiNX capping layer was kept the same (ARC; n ≈ 2). Stack G3 contained a 1.5 nm chemical SiO2 interfacial layer grown
Barbora Mojrová et al./ Energy Procedia 00 (2017) 000–000 Barbora Mojrová et al. / Energy Procedia 124 (2017) 288–294
290
by nitric acid oxidation of silicon (NAOS), as it is known from literature that such stacks can passivate boron emitters [1]. Finally, in the stacks G1, G4, and G5, the interfacial layer was kept constant while the chemical composition of the SiNX capping layer was varied. Layers with higher Si- and hydrogen-content were achieved by changing the NH3:SiH4 gas ratios during deposition. Table 1. The overview of investigated passivation layers stacks for boron emitter. Passivation
Interfacial layer
Capping layer
Stack
(thickness)
(thickness / refractive index)
G1
Thermal SiO2 (10 nm)
SiNX (62 nm / n ≈ 2)
G2
Thermal SiO2 (30 nm)
SiNX (62 nm / n ≈ 2)
G3
NAOS (1.5 nm)
SiNX (62 nm / n ≈ 2)
G4
Thermal SiO2 (10 nm)
SiNX (54 nm / n ≈ 2.08)
G5
Thermal SiO2 (10 nm)
SiNX (47 nm / n ≈ 2.22)
3. Results Quasi-Steady-State Photoconductance (QSSPC), current-voltage (IV), and internal quantum efficiency (IQE) measurements were done to compare the passivation quality of different investigated layer stacks. Before printing the metallization grid, but after a firing step, we measured the iVOC of the precursor cells using the QSSPC. Values of iVOC were obtained at 1 sun illumination after annealing. The iVOC values for each passivation stack are compared in the Fig. 2. The metalized solar cells were investigated by IV measurements to determine the values of short-circuit current (JSC) and VOC (see the Fig. 3), and by IQE measurements (see the Fig. 4) to see the spectral response of different ARC layers. 690 680 670 660 650 640 630
G1
G2
G3
G4
G5
Passivation Stack Fig. 2. iVOC of the cell precursor (before metallization) obtained at 1 sun illumination.
Barbora Mojrová al. / Energy Procedia 124 000–000 (2017) 288–294 Barbora Mojrová et al. / et Energy Procedia 00 (2017)
291
39,0 38,0 37,0 36,0 35,0 650 640 630 620 610
G1
G2
G3
G4
G5
Passivation Stack Fig. 3. VOC and JSC measured on metallized solar cells were obtained by IV measurement at 25 °C and under AM1.5 spectrum with illumination intensity of 1000 W/m2.
From iVOC values it is obvious that the best passivation quality on the cell precursor level was achieved with G2, where the passivation stack was composed of 30 nm SiO2 and 62 nm SiNX with refractive index 2.0. Slightly lower iVOC values were measured on samples from groups G1 and G4, where the ARC layer was made up of 10 nm SiO2 and SiNX with low refractive index (62 nm with n = 2.0 or 54 nm with 2.08). The passivation stacks G1 (JSC = 39,0 mA/cm2; VOC = 651 mV) and G2 (JSC = 38,4 mA/cm2; VOC = 654 mV) showed the best passivation quality while there was no decrease at short wavelengths in contrast with the samples from G4 which had JSC and VOC values almost on the same level. The absorption in SiNX due to higher refractive index is not corrected in the IQE curve, so the reduction in IQE for G4 compared with G1 should be caused by the additional absorption in SiNX and not necessary by a poorer passivation. The difference in VOC and JSC values between G1 and G2 comes from different optical properties of passivation stack. The G2 gives slightly better passivation (higher VOC) but the JSC value was lowered due to higher optical losses caused by worse light coupling. The SiNX thickness was optimized to give the best ARC stack in combination with ~ 10 nm SiO2 (G1). For G2 the SiO2 thickness is too thick when the same SiNX layer is deposited on top. Therefore the JSC will be lower for G2 due to higher reflection, but the passivation quality remains the same, as seen in IQE, which is corrected against reflection. The chemical SiO2 (NAOS) did not show good passivation, as seen from results of QSSPC, IV, and IQE measurements (very low value of iVOC, VOC, and poor IQE response at short wavelengths). The reason may be that the thin SiO2 layer was partially damaged during the PECVD SiNX deposition.
292
Barbora Mojrová et al./ Energy Procedia 00 (2017) 000–000 Barbora Mojrová et al. / Energy Procedia 124 (2017) 288–294
Fig. 4. Internal quantum efficiency (IQE) of the solar cells with different front boron emitter passivation stacks, as detailed in Table I.
3.1. Effect of SiO2 interfacial layer thickness A symmetrical boron diffused test structure was used to find the optimum thickness of the SiO2 in the stack. The minimum thickness of 10 nm (Fig. 5) seems to be necessary to achieve good passivation of boron emitter in this cell concept.
Fig. 5. iVOC of symmetrical p+/n/p+ test structures with thermal SiO2/SiNX passivation stacks as a function of SiO2 interfacial layer thickness.
3.2. Effect of SiNX capping layer variation From the IQE results in the Fig. 3 it is obvious that the passivation quality was worse in case of SiNX capping layers with higher refractive index (n = 2.08 or 2.22) despite the fact that these layers might be better for good passivation because of higher content of hydrogen. The reason is that the deposition SiNX with higher refractive index also increased the density of interface states and positive interface charge (see the Fig. 6), which both had a negative effect on boron emitter passivation.
Barbora Mojrová al. / Energy Procedia 124000–000 (2017) 288–294 Barbora Mojrová et al. / et Energy Procedia 00 (2017)
293
Fig. 6. Density of interface traps (Dit) and total fix interface charge (Qtot) for different composition of SiNX capping layer determined by corona charge measurements. The measurements were performed using Semilab PV2000 tool on un-diffused p-type c-Si wafers passivated by similar stack structures of G1, G4, and G5.
4. Conclusion Within this work we presented how the passivation quality of boron doped emitters depends on the composition of SiO2/SiNX passivation and ARC stack layers. The results from QSSPC, IV, and IQE measurements show that the optimum passivation is achieved by a stack layer of a thermal SiO2, with a thickness of at least 10 nm, and a PECVD SiNX layer with a low refractive index (n = 2.0). From the passivation results it is observed that the chemical SiO2 deposited by NAOS method does not show good passivation quality. This could be caused by PECVD deposition of SiNX, which could have partially damaged the thin SiO2 layer. The passivation degrades if the chemical composition of SiNX changes towards more Si-rich (higher refractive index), despite the fact that the layer contains more hydrogen which would have been expected to be helpful for a good passivation. The reason seems to be that, at the same time, the deposition increases the density of interface states as well as the positive interface fix charges. Acknowledgement The article was supported by project no. FEKT-S-17-3934, Utilization of novel findings in micro and nanotechnologies for complex electronic circuits and sensor applications. References [1] Mihailetchi, V. D., Y. Komatsu and L. J. Geerligs. Nitric acid pretreatment for the passivation of boron emitters for n-type base silicon solar cells. Applied Physics Letters. 2008, 92(6), 063510-. DOI: 10.1063/1.2870202. ISSN 00036951. [2] Schmidt, J., F. M. Schuurmans, W. C. Sinke, S. W. Glunz a A. G. Aberle. Observation of multiple defect states at silicon–silicon nitride interfaces fabricated by low-frequency plasma-enhanced chemical vapor deposition. Applied Physics Letters [online]. 1997, 71(2), 252-254. DOI: 10.1063/1.119512. ISSN 0003-6951. [3] Schmidt, J., A. G. Aberle. Carrier recombination at silicon–silicon nitride interfaces fabricated by plasma-enhanced chemical vapor deposition. Journal of Applied Physics [online]. 1999, 85(7), 3626-3633. DOI: 10.1063/1.369725.
294
Barbora Mojrová et al./ Energy Procedia 00 (2017) 000–000 Barbora Mojrová et al. / Energy Procedia 124 (2017) 288–294
[4] Hoex, B., J. Schmidt, R. Bock, P. P. Altermatt, M. C. M. Van De Sanden and W. M. M. Kessels. Excellent passivation of highly doped p-type Si surfaces by the negative-charge-dielectric Al2O3. Applied Physics Letters [online]. 2007, 91(11), 107-112. DOI: 10.1063/1.2784168. ISSN 00036951. [5] Duttagupta, S., F. Ma, B. Hoex and A. G. Aberle. Extremely low surface recombination velocities on heavily doped planar and textured p silicon using low-temperature positively-charged PECVD SiOx/SiNx dielectric stacks with optimised antireflective properties. In: 2013 IEEE 39th Photovoltaic Specialists Conference (PVSC). IEEE, 2013, p. 1776-1780. DOI: 10.1109/PVSC.2013.6744487. ISBN 978-1-47993299-3. [6] Kalio, A., A. Richter, C. Schmiga, M. Glatthaar and J. Wilde. Study of Dielectric Layers for Bifacial n-type Silicon Solar Cells with Regard to Optical Properties, Surface Passivation Quality, and Contact Formation. IEEE Journal of Photovoltaics. 2014, 4(2), 575-580. DOI: 10.1109/JPHOTOV.2013.2296374. ISSN 2156-3381. [7] Duttagupta, S., F. Lin, K. Shetty, M. Wilson, F. Ma, J. Lin, A. G. Aberle and B. Hoex. State-of-the-art surface passivation of boron emitters using inline PECVD AlOx/SiNx stacks for industrial high-efficiency silicon wafer solar cells. In: 2012 38th IEEE Photovoltaic Specialists Conference. IEEE, 2012, p. 001036-001039. DOI: 10.1109/PVSC.2012.6317780. ISBN 978-1-4673-0066-7.
ScienceDirect ScienceDirect
Energy Procedia 00 (2017) 000–000 Available online www.sciencedirect.com Available online atatwww.sciencedirect.com Energy Procedia 00 (2017) 000–000
ScienceDirect ScienceDirect
www.elsevier.com/locate/procedia www.elsevier.com/locate/procedia
Energy (2017) 000–000 288–294 EnergyProcedia Procedia124 00 (2017) www.elsevier.com/locate/procedia
7th International Conference on Silicon Photovoltaics, SiliconPV 2017 7th International Conference on Silicon Photovoltaics, SiliconPV 2017
A comparison study of boron emitter passivation by silicon oxide A comparison study of boron emitter passivation by silicon oxide a PECVD silicon nitrideHeating stack The 15thand International Symposium on District and a PECVD silicon nitride stackand Cooling b b Barbora Mojrováaa*, the Haifeng Chubb, Christop Peterbbthe , Pirmin Preis b, Jan Lossenb, Valentin Assessing feasibility of using heat demand-outdoor b b Barbora Mojrová *, Haifeng Chu , Christop Peter , Pirmin Preis , Jan Lossen , Valentin D. Mihailetchi b, Radovan Kopecekb temperature function for a long-term district D. Mihailetchi , Radovan Kopecekheat demand forecast a
Departement of Microelectronics The Faculty of Electrical Engineering and Communication Brno University of Technology Departement ofa,b,c Microelectronics The Faculty of3058/10, Electrical Engineering andb Communication Brno University of Technologyc 616 00, Czech Republic aTechnicka a Brno c b Technicka 3058/10, Brno 616 15, 00,Konstanz Czech Republic ISC Konstanz e. V., Rudolf-Diesel-Str. 78467, Germany b ISC Konstanz e. V., Rudolf-Diesel-Str. 15, Konstanz 78467, Germany a IN+ Center for Innovation, Technology and Policy Research - Instituto Superior Técnico, Av. Rovisco Pais 1, 1049-001 Lisbon, Portugal b Veolia Recherche & Innovation, 291 Avenue Dreyfous Daniel, 78520 Limay, France c Département Systèmes Énergétiques et Environnement - IMT Atlantique, 4 rue Alfred Kastler, 44300 Nantes, France a
I. Andrić
*, A. Pina , P. Ferrão , J. Fournier ., B. Lacarrière , O. Le Corre
Abstract Abstract Surface recombination of minority charge carriers is a significant loss mechanism in crystalline-Si (c-Si) solar cells. For + Surface recombination minority charge carriers is a significant loss mechanism crystalline-Si (c-Si) solar cells. For (phosphorous) doped of regions a sufficient surface passivation is achieved by using a in silicon nitride (SiN nAbstract X) film deposited by n+ (phosphorous) regions a sufficient surface passivation is achieved using a silicon nitride ) film deposited the by X Plasma Enhanced doped Chemical Vapour Deposition (PECVD). However, on p+ by boron doped regions, SiN(SiN does not passivate X boron doped regions, SiN does nota passivate the Plasma Enhanced Chemical Vapour Deposition (PECVD). However, on p+2), X or aluminium oxide (Al O ), and SiN film is surface. Instead, a stack of layers comprising of a thin silicon dioxide (SiO 2 3 X District heating networks are commonly addressed in the literature as one of the most effective solutions for decreasing the surface. Instead, a passivate stack of layers comprising of a thin silicon dioxide (SiO2), or aluminium oxide (Al2O3), and a SiNX film is commonly boron regions. greenhouseused gastoemissions from thedoped building sector. These systems require high investments which are returned through the heat commonly used toinvestigated passivate boron doped regions. Insales. this study the passivation quality of boron renovation doped emitters by varying the composition of SiO 2/SiN X stack Due towethe changed climate conditions and building policies, heat demand in the future could decrease, In this For studythis wepurpose, investigated the (passivated passivation emitter, quality of boron doped emitters bycells varying the composition of SiO 2/SiN X stack layers. n-PERT rear totally-diffused) solar with boron doped front side emitter and prolonging the investment return period. layers. For this purpose, n-PERT (passivated emitter, rearsymmetrical totally-diffused) solar cells with boron doped front on side6-inch emitter and phosphorous doped back-surface-field (BSF), as well as boron doped structures, were fabricated n-type The main scope of this paper is to assess the feasibility of using the heat demand – outdoor temperature function for heat demand phosphorous doped back-surface-field (BSF), as well as symmetrical boron doped structures, were fabricated on 6-inch n-type wafers. forecast. The district of Alvalade, located in Lisbon (Portugal), was used as a case study. The district is consisted of 665 wafers. a thickness at least nm, The results that showvary thatintheboth optimum passivation is achieved by a stack of a thermal 2, with buildings construction period and typology. Threelayer weather scenariosSiO (low, medium, high)ofand three10 district , with a thickness of at least 10 nm, The aresults show that the optimum passivation is achieved by a stack layer of a thermal SiO 2 layer with a low refractive index. The chemical composition of SiN capping layer plays an important role for surface and SiN X scenarios were developed (shallow, intermediate, deep). To estimate X renovation the error, obtained heat demand values were and a SiNX layer withSi-rich a low SiN refractive index. The chemical composition of SiNXpassivation capping layer plays an important role for surface passivation. A more layer show significant degradation in surface of the stack, due to the increase in the X compared with results from a dynamic heat demand model, previously developed and validated by the authors. show significant degradation in surface passivation of the stack, due to the increase in the passivation. A morestates Si-rich SiN X layer ) and fixed positive charges (Q ) at the interface. density of interface (D it tot The results showed that when only weather change is considered, the margin of error could be acceptable for some applications density of interface states (Dit) and fixed positive charges (Qtot) at the interface. (the error in annual demand was lower than 20% for all weather scenarios considered). However, after introducing renovation © 2017 Thethe Authors. Published by Elsevier Ltd. error value increased up to 59.5% ©scenarios, 2017 The Authors. Published by Elsevier Ltd. (depending on the weather and renovation scenarios combination considered). © 2017 The by Authors. Published by Elsevier Ltd. of SiliconPV 2017 under responsibility of PSE AG. Peer review the scientific conference committee Thereview value of coefficient increased on average within the range of responsibility 3.8% up to 8% per decade, that corresponds to the Peer by slope the scientific conference committee of SiliconPV 2017 under of PSE AG. Peer review by the scientific conference committee of SiliconPV 2017 under responsibility of PSE AG. decrease in the number of heating hours of 22-139h during the heating season (depending on the combination of weather and Keywords: Solar cell; boron emitter; n-PERT; passivation; NAOS renovation scenarios considered). On the other hand, function intercept increased for 7.8-12.7% per decade (depending on the Keywords: Solar cell; boron emitter; n-PERT; passivation; NAOS coupled scenarios). The values suggested could be used to modify the function parameters for the scenarios considered, and improve the accuracy of heat demand estimations. © 2017 The Authors. Published by Elsevier Ltd. * Corresponding author. Tel.: +420 54114 fax: +420 54114 6298. Peer-review under responsibility of the6107; Scientific Committee of The 15th International Symposium on District Heating and * Corresponding Tel.: +420 54114 6107; fax: +420 54114 6298. E-mail address:author. [email protected] Cooling. E-mail address: [email protected]
1876-6102 2017demand; The Authors. Published Elsevier Ltd. Keywords:©Heat Forecast; Climatebychange 1876-6102 The Authors. Published by Elsevier Ltd. Peer review©by2017 the scientific conference committee of SiliconPV 2017 under responsibility of PSE AG. Peer review by the scientific conference committee of SiliconPV 2017 under responsibility of PSE AG.
1876-6102 © 2017 The Authors. Published by Elsevier Ltd. Peer-review under responsibility of the Scientific Committee of The 15th International Symposium on District Heating and Cooling.
1876-6102 © 2017 The Authors. Published by Elsevier Ltd. Peer review by the scientific conference committee of SiliconPV 2017 under responsibility of PSE AG. 10.1016/j.egypro.2017.09.301
Barbora Mojrová et al. / Energy Procedia 124 (2017) 288–294 Barbora Mojrová et al. / Energy Procedia 00 (2017) 000–000
289
1. Introduction Surface passivation of heavily doped p+ emitters is required in order to realize cost-effective industrial highefficiency solar cells based on n-type silicon wafers. The boron emitter can be passivated by a dielectric having a low density of interface states (Dit), potentially in combination with a high fixed negative charge density. However conventional SiNX is not suitable for p+ surface passivation [1]. One of the reasons for the relatively poor surface passivation of p+ surfaces by SiNX is related to the very high fixed charge density in PECVD SiNX. It is attributed to the band gap defects that SiNX form during deposition, which are more detrimental to p+ Si [2, 3]. It was observed that the fixed charge density can be reduced by adding a thin silicon oxide layer between the crystalline silicon and the SiNX film. Another approach to passivate p+ layer is based on using thin Al2O3 dielectric with the negative charge as a surface passivation layer and SiNX as an antireflection coating (ARC) layer [4-7]. 2. Experimental part 2.1. Sample preparation In this work we investigated the passivation of boron doped emitters with different SiO2/SiNX stack layers. For this purpose we used, on the one hand, an n-PERT cell with a homogeneous diffused front boron emitter and a phosphorous back-surface-field (BSF) and on the other hand, symmetrical boron diffused test structure. Both device structures were fabricated on 6-inch n-type monocrystalline Si wafers with base resistivity of 2.5 cm. Standard industrial processes were used, which included wet chemical alkaline texturization, cleaning by HCl, HF, and HF/O3 solutions, diffusion in quartz tube furnaces containing phosphoryl chloride (POCl3; for n+ BSF) or boron tribromide (BBr3; for p+ emitter). For the solar cell devices the back side was chemically polished and the n+ BSF diffusion was passivated by a thermal SiO2/ SiNX stack. Screen printing and firing through of commercial silver containing pastes were applied to both sides for metallization. A schematic cross-section of the studied solar cells is presented in the Fig. 1.
Fig. 1. Schematic cross-section of the investigated n-PERT cells.
2.2. Passivation stacks All solar cells were processed identically, with the exception of boron emitter passivation. For the boron emitter 5 different passivation stacks were investigated and compared, as detailed in Table I. First, the stacks G1 and G2 aimed to investigate the effect of SiO2 interfacial layer thickness on the passivation quality of the SiO2/SiNX stack. A symmetrical boron diffused test structure was used, in addition to the solar cells, to find the optimum thickness of the SiO2 in the stack. Here, a thermal SiO2 was grown in situ during the BBr3 diffusion on boron emitter surface with a thickness of about 40 nm. Subsequently the thick SiO2 layer was partially etched (thinned) in HF (2%) solution to yield various thicknesses, followed by a 62 nm PECVD SiNX antireflection coating (ARC; n ≈ 2) deposition of the capping layer to complete the passivation stack. Then, the stacks G1 and G3 aimed to compare different SiO2 interface passivation layers, whereas the SiNX capping layer was kept the same (ARC; n ≈ 2). Stack G3 contained a 1.5 nm chemical SiO2 interfacial layer grown
Barbora Mojrová et al./ Energy Procedia 00 (2017) 000–000 Barbora Mojrová et al. / Energy Procedia 124 (2017) 288–294
290
by nitric acid oxidation of silicon (NAOS), as it is known from literature that such stacks can passivate boron emitters [1]. Finally, in the stacks G1, G4, and G5, the interfacial layer was kept constant while the chemical composition of the SiNX capping layer was varied. Layers with higher Si- and hydrogen-content were achieved by changing the NH3:SiH4 gas ratios during deposition. Table 1. The overview of investigated passivation layers stacks for boron emitter. Passivation
Interfacial layer
Capping layer
Stack
(thickness)
(thickness / refractive index)
G1
Thermal SiO2 (10 nm)
SiNX (62 nm / n ≈ 2)
G2
Thermal SiO2 (30 nm)
SiNX (62 nm / n ≈ 2)
G3
NAOS (1.5 nm)
SiNX (62 nm / n ≈ 2)
G4
Thermal SiO2 (10 nm)
SiNX (54 nm / n ≈ 2.08)
G5
Thermal SiO2 (10 nm)
SiNX (47 nm / n ≈ 2.22)
3. Results Quasi-Steady-State Photoconductance (QSSPC), current-voltage (IV), and internal quantum efficiency (IQE) measurements were done to compare the passivation quality of different investigated layer stacks. Before printing the metallization grid, but after a firing step, we measured the iVOC of the precursor cells using the QSSPC. Values of iVOC were obtained at 1 sun illumination after annealing. The iVOC values for each passivation stack are compared in the Fig. 2. The metalized solar cells were investigated by IV measurements to determine the values of short-circuit current (JSC) and VOC (see the Fig. 3), and by IQE measurements (see the Fig. 4) to see the spectral response of different ARC layers. 690 680 670 660 650 640 630
G1
G2
G3
G4
G5
Passivation Stack Fig. 2. iVOC of the cell precursor (before metallization) obtained at 1 sun illumination.
Barbora Mojrová al. / Energy Procedia 124 000–000 (2017) 288–294 Barbora Mojrová et al. / et Energy Procedia 00 (2017)
291
39,0 38,0 37,0 36,0 35,0 650 640 630 620 610
G1
G2
G3
G4
G5
Passivation Stack Fig. 3. VOC and JSC measured on metallized solar cells were obtained by IV measurement at 25 °C and under AM1.5 spectrum with illumination intensity of 1000 W/m2.
From iVOC values it is obvious that the best passivation quality on the cell precursor level was achieved with G2, where the passivation stack was composed of 30 nm SiO2 and 62 nm SiNX with refractive index 2.0. Slightly lower iVOC values were measured on samples from groups G1 and G4, where the ARC layer was made up of 10 nm SiO2 and SiNX with low refractive index (62 nm with n = 2.0 or 54 nm with 2.08). The passivation stacks G1 (JSC = 39,0 mA/cm2; VOC = 651 mV) and G2 (JSC = 38,4 mA/cm2; VOC = 654 mV) showed the best passivation quality while there was no decrease at short wavelengths in contrast with the samples from G4 which had JSC and VOC values almost on the same level. The absorption in SiNX due to higher refractive index is not corrected in the IQE curve, so the reduction in IQE for G4 compared with G1 should be caused by the additional absorption in SiNX and not necessary by a poorer passivation. The difference in VOC and JSC values between G1 and G2 comes from different optical properties of passivation stack. The G2 gives slightly better passivation (higher VOC) but the JSC value was lowered due to higher optical losses caused by worse light coupling. The SiNX thickness was optimized to give the best ARC stack in combination with ~ 10 nm SiO2 (G1). For G2 the SiO2 thickness is too thick when the same SiNX layer is deposited on top. Therefore the JSC will be lower for G2 due to higher reflection, but the passivation quality remains the same, as seen in IQE, which is corrected against reflection. The chemical SiO2 (NAOS) did not show good passivation, as seen from results of QSSPC, IV, and IQE measurements (very low value of iVOC, VOC, and poor IQE response at short wavelengths). The reason may be that the thin SiO2 layer was partially damaged during the PECVD SiNX deposition.
292
Barbora Mojrová et al./ Energy Procedia 00 (2017) 000–000 Barbora Mojrová et al. / Energy Procedia 124 (2017) 288–294
Fig. 4. Internal quantum efficiency (IQE) of the solar cells with different front boron emitter passivation stacks, as detailed in Table I.
3.1. Effect of SiO2 interfacial layer thickness A symmetrical boron diffused test structure was used to find the optimum thickness of the SiO2 in the stack. The minimum thickness of 10 nm (Fig. 5) seems to be necessary to achieve good passivation of boron emitter in this cell concept.
Fig. 5. iVOC of symmetrical p+/n/p+ test structures with thermal SiO2/SiNX passivation stacks as a function of SiO2 interfacial layer thickness.
3.2. Effect of SiNX capping layer variation From the IQE results in the Fig. 3 it is obvious that the passivation quality was worse in case of SiNX capping layers with higher refractive index (n = 2.08 or 2.22) despite the fact that these layers might be better for good passivation because of higher content of hydrogen. The reason is that the deposition SiNX with higher refractive index also increased the density of interface states and positive interface charge (see the Fig. 6), which both had a negative effect on boron emitter passivation.
Barbora Mojrová al. / Energy Procedia 124000–000 (2017) 288–294 Barbora Mojrová et al. / et Energy Procedia 00 (2017)
293
Fig. 6. Density of interface traps (Dit) and total fix interface charge (Qtot) for different composition of SiNX capping layer determined by corona charge measurements. The measurements were performed using Semilab PV2000 tool on un-diffused p-type c-Si wafers passivated by similar stack structures of G1, G4, and G5.
4. Conclusion Within this work we presented how the passivation quality of boron doped emitters depends on the composition of SiO2/SiNX passivation and ARC stack layers. The results from QSSPC, IV, and IQE measurements show that the optimum passivation is achieved by a stack layer of a thermal SiO2, with a thickness of at least 10 nm, and a PECVD SiNX layer with a low refractive index (n = 2.0). From the passivation results it is observed that the chemical SiO2 deposited by NAOS method does not show good passivation quality. This could be caused by PECVD deposition of SiNX, which could have partially damaged the thin SiO2 layer. The passivation degrades if the chemical composition of SiNX changes towards more Si-rich (higher refractive index), despite the fact that the layer contains more hydrogen which would have been expected to be helpful for a good passivation. The reason seems to be that, at the same time, the deposition increases the density of interface states as well as the positive interface fix charges. Acknowledgement The article was supported by project no. FEKT-S-17-3934, Utilization of novel findings in micro and nanotechnologies for complex electronic circuits and sensor applications. References [1] Mihailetchi, V. D., Y. Komatsu and L. J. Geerligs. Nitric acid pretreatment for the passivation of boron emitters for n-type base silicon solar cells. Applied Physics Letters. 2008, 92(6), 063510-. DOI: 10.1063/1.2870202. ISSN 00036951. [2] Schmidt, J., F. M. Schuurmans, W. C. Sinke, S. W. Glunz a A. G. Aberle. Observation of multiple defect states at silicon–silicon nitride interfaces fabricated by low-frequency plasma-enhanced chemical vapor deposition. Applied Physics Letters [online]. 1997, 71(2), 252-254. DOI: 10.1063/1.119512. ISSN 0003-6951. [3] Schmidt, J., A. G. Aberle. Carrier recombination at silicon–silicon nitride interfaces fabricated by plasma-enhanced chemical vapor deposition. Journal of Applied Physics [online]. 1999, 85(7), 3626-3633. DOI: 10.1063/1.369725.
294
Barbora Mojrová et al./ Energy Procedia 00 (2017) 000–000 Barbora Mojrová et al. / Energy Procedia 124 (2017) 288–294
[4] Hoex, B., J. Schmidt, R. Bock, P. P. Altermatt, M. C. M. Van De Sanden and W. M. M. Kessels. Excellent passivation of highly doped p-type Si surfaces by the negative-charge-dielectric Al2O3. Applied Physics Letters [online]. 2007, 91(11), 107-112. DOI: 10.1063/1.2784168. ISSN 00036951. [5] Duttagupta, S., F. Ma, B. Hoex and A. G. Aberle. Extremely low surface recombination velocities on heavily doped planar and textured p silicon using low-temperature positively-charged PECVD SiOx/SiNx dielectric stacks with optimised antireflective properties. In: 2013 IEEE 39th Photovoltaic Specialists Conference (PVSC). IEEE, 2013, p. 1776-1780. DOI: 10.1109/PVSC.2013.6744487. ISBN 978-1-47993299-3. [6] Kalio, A., A. Richter, C. Schmiga, M. Glatthaar and J. Wilde. Study of Dielectric Layers for Bifacial n-type Silicon Solar Cells with Regard to Optical Properties, Surface Passivation Quality, and Contact Formation. IEEE Journal of Photovoltaics. 2014, 4(2), 575-580. DOI: 10.1109/JPHOTOV.2013.2296374. ISSN 2156-3381. [7] Duttagupta, S., F. Lin, K. Shetty, M. Wilson, F. Ma, J. Lin, A. G. Aberle and B. Hoex. State-of-the-art surface passivation of boron emitters using inline PECVD AlOx/SiNx stacks for industrial high-efficiency silicon wafer solar cells. In: 2012 38th IEEE Photovoltaic Specialists Conference. IEEE, 2012, p. 001036-001039. DOI: 10.1109/PVSC.2012.6317780. ISBN 978-1-4673-0066-7.