Energy Procedia 38 ( 2013 ) 459 – 466
1876-6102 © 2013 The Authors. Published by Elsevier Ltd.
Selection and/or peer-review under responsibility of the scientifi c committee of the SiliconPV 2013 conference doi: 10.1016/j.egypro.2013.07.304
ScienceDirect
SiliconPV: March 25-27, 2013, Hamelin, Germany
Record efficiency of PhosTop solar cells from n-type Cz UMG silicon wafers
Yvonne Schiele
*, Svenja Wilking, Felix Book, Thomas Wiedenmann, Giso Hahn
University of Konstanz, Department of Physics, P.O. Box X915, D-78457 Konstanz, Germany
Abstract
Highly purified n-type Upgraded Metallurgical Grade (UMG) silicon carries a large potential for high efficiency low
cost solar cells. In this study, the industrially producib o manufacture
large-area n-type rear junction solar cells from such a Si material with a screen-printed Al-alloyed full-area emitter featuring a selective phosphorous front surface field (FSF) and a SiO2/SiNx:H passivation on the front.
Since resistivity at the seed end is about seven times as high as at the tail end of the UMG Si ingot and carrier lifetime decreases from seed to tail end, a clear
corresponding wafers in the UMG Si ingot is observable. Maximum conversion efficiency is reached (on a wafer = 19.0% being, to the so far reported on industrial type solar cells manufactured from 100%
UMG Si.
© 2013 The Authors. Published by Elsevier Ltd.
Selection and/or peer-review under responsibility of the scientific committee of the SiliconPV 2013 conference
Keywords: n-type; UMG silicon; Al emitter; selective
1. Introduction
Replacing p-type Si wafers predominantly utilized in current solar cell production by pure n-type Si material yields various benefits. n-type Si is in general more tolerant of residual metal impurities like, for instance, Fe allowing a higher carrier diffusion length for the same impurity concentration [1].
Additionally, in case of non-compensated material, no light-induced degradation (LID) of solar cell efficiency caused by B-O complexes occurs [2].
* Corresponding author. Tel.: +49-7531-88-4995; fax: +49-7531-88-3895.
E-mail address: yvonne.schiele@uni-konstanz.de.
Available online at www.sciencedirect.com
© 2013 The Authors. Published by Elsevier Ltd.
Selection and/or peer-review under responsibility of the scientifi c committee of the SiliconPV 2013 conference Open access under CC BY-NC-ND license.
Open access under CC BY-NC-ND license.
Konstanzer Online-Publikations-System (KOPS) URL: http://nbn-resolving.de/urn:nbn:de:bsz:352-250273
In recent years, applying the industrial standard processing sequence for p-type Si solar cells to n-type substrates, known ], has proven a promising candidate for a simple and economical fabrication of n-type Si solar cells. From this approach, a solar cell with a heavily POCl3- diffused and selectively etched-back FSF and a screen-printed Al-alloyed rear emitter emerges [4].
High purity UMG silicon features very low residual impurity concentrations [8, 9] and is used to produce high quality n-type mono-c-Si Cz ingots. Substituting the expensive Electronic Grade (EG) silicon by highly purified n-type UMG silicon to be employed in the PhosTop process provides a large potential for highly efficient low cost solar cells [5].
2. n-Type UMG silicon
UMG Si is compensated as it contains both polarity type dopants. For n-type UMG ingots, the dominant one is phosphorus featuring a lower segregation coefficient than boron which leads to a greater resistivity variation along the ingot. Thus, there is a critical amount of P required in the Si feedstock to make -type doped. Furthermore, the starting concentration of B, which needs to be overcompensated, plays a determining role for the resulting specific resistivity of an n-type UMG Si ingot [10]. In order to obtain n-type ingots with a high resistivity from UMG Si, which is especially relevant for rear emitter cell concepts demanding high diffusion lengths, the residual B concentration in the UMG Si feedstock should be as low as possible [10].
For this study, a 15 kg 6 inch mono-crystalline Cz silicon ingot is grown from 100% highly purified UMG silicon obtained by the PHOTOSIL process [5] with boron and phosphorus concentrations measured by Glow Discharge Mass Spectrometry (GDMS) of 0.15 ppmw and 0.8 ppmw respectively in the feedstock. The resulting ingot is n-type and fully mono-crystalline. Other impurities, especially metals, are not detectable with the analysis techniques applied (GDMS and Inductively Coupled Plasma Atomic Emission Spectroscopy, ICP-OES). The ingot is cut into 125×125 mm2 semi-square wafers of 180 μm thickness (after texturization).
3. Experimental details
The UMG material is firstly characterized in terms of specific resistivity and effective minority carrier lifetime eff being crucial parameters to the performance of the final device. Therefore, wafers from the seed and the tail end of the n-type Cz UMG Si ingot are extracted from the solar cell process before and after the different high temperature steps. Then, all layers which are already deposited on or diffused into the wafers are removed in diluted HF and a chemical polishing solution, whereupon resistivity is determined by four-point probe (4PP) at 10×10 measurement points distributed on the wafer area.
Subsequently, the wafers are chemically cleaned and passivated by 30 nm atomic layer deposited (ALD) Al2O3 on both sides. The passivation is activated by a 30 min anneal in N2 atmosphere at 420°C. By means of photoconductance decay measurement (PCD) [7] using a Sinton Lifetime Tester, eff is determined on five approximately 50×50 mm2 areas in the center and the corners of the 125×125 mm2 wafers n=1×1015 cm-3. In order to spatially resolve the recombination activity of the wafers, photoluminescence (PL) images are taken.
For the PhosTop solar cells, the emitter is formed by alloying the full-area screen-printed Al on the rear into the silicon wafer. The silver finger grid on the front is established by screen-printing, too. As cell performance of the n-type rear junction cell concept is limited mainly by its front surface recombination velocity [4], in particular with low ohmic substrates, an improved front surface passivation and antireflection coating (ARC) consisting of a SiO2/SiNx:H stack is applied. Additionally, a selectively etched-back phosphorous FSF is diffused into the Si surface. Complementarily to PhosTop solar cells
from the Cz UMG wafers, references are manufactured from standard n-type Cz material (125×125 mm2 semi-square, 220μm thick after texturization). The processing sequence and the resulting structure of the described PhosTop solar cell are depicted in Fig. 1.
Inkjnkj Inkj Inkj
Inkjkketetetetetetmaskmaskmaskmaskmaskaskkprinprinprinprinprinp tingtingtingtingtingting FSF FSF FSF FSF FSF FSFetchetchetchetchetchetchbackbackbackbackbackbackk Mask
Mask Mask Mask Mask Mask
MaskandandandandandandporSporSporSporSporSp iiiiremoremoremoremoremoe valvalvalvalvalva
Scre Scre Scre Scre Screeenenenenene-prinprinprinprinprinp tingtingtingtingtingting
Co Co Co Co Co--firfirifirifirifirif ngngngngngg PECV PECV PECV PECV PECV PECVDDDDD-SiNSiNSiNSiNSiNNNxxxxx:H:HHHHH Alka
Alka Alka Alka Alka
Alkaklinelinelinelinelineetextextextextextextureturetureturetureture/saw/saw/saw/sawsawawdamdamdamdamdamdamageage age ageageageremoremoremoremoremoremvalvalvalvalvalva
Ther Ther Ther Ther Ther
Themamalmalmalmalmaoxidoxidoxidoxidoxidxidatioatioatioatioatioat nnnnn
Edgedge Edge Edge Edge
Edge isolisolisolisolatiosolatioatioatioatioat nnnnn:sawsawisawisawisawiawngngngngngg POCl
POCl POCl POCl POCl
P 3333diffdiffdiffdiffdiffdiffffusiousiousiousiousionnnnn(35 (35 (35(35(355 //// ))))) AcidAcidAcidAcidAcidAcidicicicicicsurfsurfsurfsurfsurfsurfaceaceaceaceaceacepolipolipolipolipoliposhinshinshinshinshinsh gggggg RCA
RCA RCA RCA RCA
RCAsurfsurfsurfsurfsurfsurfaceaceaceaceaceacecleacleacleacleaclealeaningningningningningning ALD
ALD ALD ALD ALD
ALD---AAlAlAlAlA222OOOOO333+++++++anneanneanneanneanneanneaalalalala(30(30(30(30(303 min,min,min,minminm 420420420420420424 °°°C)C)C)C)C)C Remo
Remo Remo Remo Remo
RemvavalvalvalvalvaofofofofoffFSF,FSF,FSF,FSFFSFFSFPSGPSGPSGPSGPSGPSG, Si, Si, Si, SiSiSOOOOO2222, ,,,SiNSiNSiNSiNSiNSiNNxxxxx:H:HHHHH
nCz UMG Si p+Al alloyed emitter
Al-Si eutectic FSF passivation/ARC Ag front contact
n n+/ ++selective FSF
Al rear contact
Fig. 1. Processing sequence for solar cells (blue) and lifetime/specific resistivity (gray) samples; schematic representation of the PhosTop solar cell with selective FSF, SiO2/SiNx:H passivation/ARC on the front, Al-alloyed emitter on the rear and screen-printed
metal contacts.
4. Results and discussion
4.1. Lifetime and specific resistivity monitoring
The development of specific resistivity and minority carrier lifetime depending on the ingot position of the wafer employed in the PhosTop cell process and the beforehand conducted process steps is summarized in Table 1. is measured at the center of the wafer by the eddy current method using a Sinton Lifetime Tester. For spatial resolution of , the 10×10 measurement points distributed on the wafer area are displayed by smoothed contour plots in Fig. 2 exemplarily for the wafers taken from the process after oxidation. It has to be noticed, that due to measurement uncertainties the 4PP method leads to values approximately 12 to 25% smaller than the ones determined by the eddy current method. The eff
range listed in Table 1 results from the minimum and maximum value of the PCD measurements on the five different areas of the 125×125 mm2 n-type Cz UMG wafers. The spatially resolved recombination activity is exemplarily depicted in the PL images of the post-oxidation wafers (Fig. 3).
Table 1. Specific resistivity and effective minority carrier lifetime eff(with Al2O3passivation) of the n-type Cz UMG Si material after different PhosTop process steps. (**) Due to wafer breakage, only a small part of the post-firing sample consisting of one corner and the center was left for measurement, thus maximum effand the effrange is expected to be larger.
Previous process step Saw damage removal
POCl3
diffusion
Thermal oxidation
Firing seed end
tail end
5.2 0.72
5.9 0.83
5.8 0.77
5.9 0.79
eff(μs) seed end tail end
87 179 55 104
199 1813 200 425
245 2406 154 399
248 309(**) 74 444
Resistivity at the seed end is about seven times higher than at the tail end which is caused by the lower segregation coefficient of phosphorus. Specific resistivity of the wafers from both ends of the ingot increases during the first high temperature exposure (POCl3 diffusion) due to the dissolution of oxygen related thermal donors in the material. At the seed end, remains constant during the subsequent high temperature processes whereas it decreases again a little at the tail end.
4 5 6 7 8 9 10 11
0.82 0.84 0.86 0.88 0.90 0.92 0.94 0.96 0.98
Fig. 2. Smoothed contour plots of spatially resolved distribution measured by 4PP on post-oxidation wafers at the seed end (left) and tail end (right) of the n-type Cz UMG Si ingot.
Carrier lifetime distribution on the wafer area varies considerably. Particularly in the areas with higher lifetimes, eff rises during the POCl3 diffusion high temperature exposure which is explained by the POCl3
getter effect and the dissolution of thermal donors known to act as strong recombination centers. The relatively high thermal budget of the PhosTop solar cell process has no detrimental influence on the UMG Si material, as eff further increases during subsequent high temperature exposure. Especially the high lifetime areas of the wafers at the tail end exhibit a considerably lower eff than the ones at the seed end due to the higher concentration of (doping) impurities. Furthermore, the concentration of oxygen, thus of thermal donors, is higher there due to its apparent segregation coefficient >1. However, in PL images of the finished solar cells illuminated by an 808 nm laser, the inhomogeneities are not observable indicating that the low lifetime areas detected by the lifetime samples do not restrict solar cell efficiency.
The spatially resolved 4PP measurement at the seed end wafers exhibits a nearly radial resistivity distribution with maximal in the center, whereas at the tail end wafers, no explicit symmetry is observable. The and the eff spatial distribution do not show an obvious correlation.
PL intensity (counts/sec)
high
low
Fig. 3. Spatially resolved recombination activity determined by PL imaging of surface passivated post-oxidation wafers at the seed end (left) and tail end (right) of the n-type Cz UMG Si ingot.
The reference standard n-type Cz wafers exhibit a minority carrier lifetime of 3-5 ms (surface passivated by a-Si:H) and a specific resistivity of 10 (both values ascertained after the high temperature process) which corresponds to the doping density of the material typically chosen for the manufacturing of rear junction PhosTop solar cells.
4.2. IV characterization of solar cells
The characteristic IV parameters of the PhosTop solar cells exhibit a clear dependence on the original position of the wafer in the UMG Si ingot (see Fig. 4). The reference solar cell characteristics are close to the reported values achieved with this cell concept [4, 6]. The maximum efficiency of the UMG solar cells is reached on a wafer which has been extracted
the seed end with = 19.0%. According to the knowledge, this is the highest efficiency so far reported on industrial type solar cells manufactured from 100% UMG Si.
200 300 400 500 600
636 638 640 642 644 646 648 650
VOC (mV)
200 300 400 500 600
30 31 32 33 34 35 36 37
38 UMG solar cells
Reference solar cells
jSC (mA/cm2)
200 300 400 500 600
0.77 0.78 0.79 0.80 0.81 0.82
FF
Ingot position (a.u.) 15 200 300 400 500 600
16 17 18 19 20
(%)
Ingot position (a.u.)
Fig. 4. IV characteristics of PhosTop solar cells made from reference n-type Cz wafers (triangles) and n-type Cz UMG wafers (squares) as a function of the ingot position with seed end at ~200 a.u. and tail end at ~600 a.u.
Contrary to the conventional p-type solar cell, in the discussed n-type alternative cell design featuring the carrier collecting junction on the rear, the base contributes to lateral conductivity, which is injection level dependent. For n-type Si with high resistivity, the base is in high injection level condition under illumination which allows the employment of a wide range of low doping concentrations. Because of the base contribution to lateral conductivity, series resistance RS decreases with wafer resistivity towards the tail end of the ingot causing the growing fill factor FF (0.800 at ~230 a.u., 0.809 at ~600 a.u.). The pseudo fill factor PFF which disregards RS is, however, virtually identical at seed and tail end with PFF=0.825 or PFF=0.823, respectively. A two-diode model fit of the IV curve of a solar cell from the seed end (~230 a.u.) calculates RS at VOC to be 0.33 2, whereas at tail end (~600 a.u.) RS amounts to 0.29 2. Since series resistance of the solar cells from high-ohmic sub n and therefore on voltage, RS at seed end is even higher at maximum power point Vmpp.
The performance of the rear junction cell concept is limited mainly by its front surface recombination velocity [4]. Towards the tail end of the UMG ingot, FSF becomes less effective on the lower ohmic substrates, which causes a higher effective surface recombination velocity and thus a drastic drop in saturation current density jSC and open circuit voltage VOC. The reduced substrate carrier lifetime there adds to this effect by increasing the recombination activity in the Si bulk (cf. Table 1). PC1D [11]
simulations identify not to influence VOC, though eff changing with does cause the observed VOC
variation along the ingot. jSC is influenced by both and eff leading to the larger relative variation of jSC
with ingot height compared to the VOC variation. Saturation current density j01 determined by the two- diode model fit of the solar cell at 230 a.u. amounts to 3.9×10-13 A/cm2, whereas j01 =4.6×10-13 A/cm2 at 600 a.u. For a reference cell with standard n-type Cz substrate featuring higher base resistivity and higher carrier lifetime, j01 amounts to 2.8×10-13 A/cm2.
4.3. Light induced degradation
Since the n-type Cz UMG material is compensated, the solar cells suffer from LID caused by recombination active boron-oxygen related defects. For PhosTop solar cells, in particular, this is a significant object of investigation as this concept is largely affected by the resulting eff reduction.
LID of solar cells from seed end (~230 a.u.) and tail end (~600 a.u.) is compared. Therefore, the time resolved VOC degradation of these solar cells is measured until the degraded saturation value of VOC is reached (Fig. 5 left). As LID of compensated Si material is divided into a fast and a slow degradation, the initially occurring fast part is measured a second time with a higher time resolution (Fig. 5 right), assuming that the second annealing step after the first complete degradation has not modified the solar cells except for the B-O related defect concentration. The total VOC loss of the tail end solar cell is smaller VOC =4.2 mV) and the duration until the fast degradation saturates is longer (15-20 h) than the ones of the seed end solar cell VOC =10.6 mV, 1-2 h).
For p-type compensated Si, the B-O related degradation rate R is supposed to be dependent on the boron concentration and the net doping concentration p0 (R~p02) [12]. Based on lifetime samples of compensated n-type Si, R of lowly doped material has been observed to be higher, as it is the case for our investigation based on solar cells as well [12]. It is supposed that the initial lifetime values in the annealed state have an influence on the degradation amplitude. Furthermore, R might be proportional to the hole concentration in the material [12]. As the hole concentration decreases from seed to tail end assuming a nearly constant B but an increasing P concentration, this supposition is in accordance with the observed time resolved and ingot height dependent LID.
0 200 400 600 800 1000
-12 -10 -8 -6 -4 -2
0 600 a.u. 230 a.u.
VOC (mV)
Degradation duration (h)
0 50 100 150 200
-6 -5 -4 -3 -2 -1
0 600 a.u. 230 a.u.
VOC (mV)
Degradation duration (min)
Fig. 5. Complete (left) and fast part (right) of the time resolved degradation of solar cells from seed end (~230 a.u., left) and tail end (~600 a.u., right) under 0.5 suns illumination at 40±1°C.
eff of the post-oxidation lifetime samples diminishes after 20 h of degradation at 40±1°C and 0.2 suns illumination at seed end of the ingot to a range of 197-402 μs and at tail end to 133-178 μs (cf. Table 1).
The photoluminescence images reveal that all areas suffer from LID but eff in the high lifetime areas decreases more.
5. Conclusion
The industrial standard processing sequence for p-type Si solar cells applied to n-type substrates has been employed to manufacture PhosTop solar cells from high purity n-type Cz UMG silicon with very low residual impurity concentrations. and eff have been monitored during solar cell processing as being crucial parameters to the performance of the final device.
Since at the seed end is about seven times as high as at the tail end of the UMG Si ingot (due to the low segregation coefficient of P) and eff decreases from seed to tail end, a clear dependence of the solar ingot is observable. and eff ascend during the first high temperature exposure due to the dissolution of oxygen related thermal donors. eff distribution on the wafer area varies largely. However, the low lifetime areas detected on lifetime samples do not restrict solar cell efficiency. Maximum efficiency is reached on a wafer which has been extracted
= 19.0% so far reported on industrial type solar
cells manufactured from 100% UMG Si. RS increases with towards the seed end of the ingot causing a diminishing FF (0.800 at ~230 a.u., 0.809 at ~600 a.u.). Towards the tail end of the UMG ingot, the FSF becomes less effective on the lower ohmic substrates, which causes a higher effective surface recombination velocity and thus a drastic drop in jSC and VOC.
The PhosTop solar cells suffer from LID caused by recombination active B-O related defects. The total VOC loss of the tail end solar VOC =4.2 mV) and the duration until the fast degradation
saturates is longer (15-20 VOC =10.6 mV, 1-2 h).
Furthermore, degradation rate is supposed to be proportional to the hole concentration decreasing from seed to tail end, which is in accordance with the observed time resolved and ingot height dependent LID.
Acknowledgements
The authors would like to thank J. Ranzmeyer and L. Mahlstaedt for their processing support. The n- type Cz UMG Si material supply by Apollon Solar is gratefully acknowledged. The content of this publication is the responsibility of the authors.
References
[1] Macdonald D, Geerligs LJ. Recombination activity of interstitial iron and other transition metal point defects in p- and n-type crystalline silicon. Appl Phys Lett 2004;85:4061-3.
[2] Glunz SW, Rein S, Lee JY, Warta W. Minority carrier lifetime degradation in boron-doped Czochralski silicon. J Appl Phys 2001;90:2397-404.
[3] Meier DL, Davis HP, Garcia RA, Salami J, Rohatgi A, Ebong A, Doshi P. Aluminum alloy back p-n junction dendritic web silicon solar cell. Sol En Mat & Solar Cells 2001;65:621-7.
[4] Book F, Wiedenmann T, Schubert G, Plagwitz H, Hahn G. Influence of the front surface passivation quality on large area n- type silicon solar cells with Al-alloyed rear emitter. En Proc 2011;8:487-92.
[5] Einhaus R, Kraiem J, Degoulange J, Nichiporuk O, Foster M, Papet P, et al. 19% efficiency heterojunction solar cells on Cz wafers from non-blended upgraded metallurgical silicon. Proc 38th IEEE PVSC, Austin, Texas, USA; 2012, p.3234-7.
[6] Schiele Y, Book F, Seren S, Hahn G, Terheiden B. Screen-printed Al-alloyed rear junction solar cell concept applied to very thin (100 μm) large-area n-type Si wafers. En Proc 2012;27:460-6.
[7] Sinton RA, Cuevas A. Contactless determination of current voltage characteristics and minority carrier lifetimes in semiconductors from quasi steady state photoconductance data. Appl Phys Lett 1996;69:2510-2.
[8] Kishida Y, Okajima M, Kondo J, Okazawa K, Tokumaru S, Dohnomae H, Hiyoshi M. Production of compensation free SOG silicon feedstock by metallurgical refinement. Proc 37th IEEE PVSC, Seattle, Washington, USA; 2011, p.3425-8.
[9] Margaria T, Cocco F, Neulat L, Kraiem J, Einhaus R, Degoulange J, et al. UMG silicon from the PHOTOSIL project a status overview in 2011 on the way towards industrial production. Proc 26th EU PVSEC, Hamburg, Germany; 2011, p.1806-9.
[10] Kraiem J, Papet P, Degoulange J, Forster M, Nichiporuk O, Grosset-Bourbange D, et al. World class solar cell efficiency on n-type Cz UMG Silicon wafers by heterojunction technology. Proc 27th EU PVSEC, Frankfurt, Germany; 2012, p.657-60.
[11] Clugston DA, Basore PA. PC1D version 5: 32-bit solar cell modeling on personal computers. Proc 26th IEEE PVSC, Anaheim, California, USA; 1997, p.207-10.
[12] Lim B. Boron-oxygen-related recombination centers in crystalline silicon and the effects of dopant-compensation.
Dissertation, University of Hannover;2012.
