Sunteți pe pagina 1din 7

Solar Energy Materials and Solar Cells 185 (2018) 97–103

Contents lists available at ScienceDirect

Solar Energy Materials and Solar Cells


journal homepage: www.elsevier.com/locate/solmat

Effect of basicity of glass frits on electrical properties of Si solar cells T


a b a,⁎
Yurian Kim , Jeungsoo Huh , Hyungsun Kim
a
Department of Materials Science and Engineering, Inha University, 100 Inharo, Namgu, Incheon 22212, Republic of Korea
b
School of Materials Science and Engineering, Kyungpook National University, 80 Daehakro, Bukgu, Daegu 41566, Republic of Korea

A R T I C LE I N FO A B S T R A C T

Keywords: The purpose of this study is to understand the effects of the basicity parameter of glass frits added to the front
Ag electrode contact of three-busbar Si solar cells on the removal of the anti-reflective coating (ARC) and the electrical
Glass frits characteristics of the cell. We investigated PbO-TeO2-SiO2 glass systems by adding oxides to improve the etching
Si solar cell of the interface between the Ag electrode and the Si wafer. The thermal properties and viscous flows of the frits
Tellurite glass
were characterized using differential thermal analysis and a hot-stage microscope. The microstructure of the
Basicity
glass layer at the interface of the Ag electrode and Si wafer was observed using scanning electron microscopy and
transmission electron microscopy. The results showed that the electrical properties were dependent on the
morphology of the interfacial structure. The addition of oxides to the glass system produced more recrystallites
on the n+ emitter with a greater penetration depth in the Si wafer, thereby increasing the efficiency of the solar
cell. The reactivity of the glass affected the removal of the ARC and the electrical properties of the Si solar cells
were influenced by the charge transport facilitated by the Ag crystallites.

1. Introduction Kim et al. [13] also proposed electron injection into the B emitter
during the firing process to increase the Ag crystallite formation and
Ag paste has been used in Cu2O solar cells and perovskite solar cells decrease the contact resistance. The size of the glass frits is an im-
[1,2], and is commonly used to form contacts with the n+ emitter in the portant parameter affecting solar cell performance, as nanosized glass
front electrodes of Si solar cells. The addition of glass frits to Ag paste frits prevent residual glass from accumulating at the Ag/Si interface
has important effects on the etching of the silicon nitride anti-reflective [14]. The density and distribution of Ag recrystallites and the thickness
coating (ARC) layers as insulators and, thereby, on the Ag crystallite of the glass layer vary as a function of the viscous flows of the frits
formation in the glass layer to accomplish electrical contact between during fast firing [15,16].
the Ag electrode and the Si wafer [3–5]. The current conduction paths The chemical composition of the glass frits affects the electrical
from the Si wafer to the Ag electrode are classified into different micro- performance during the etching of Si3N4, with or without an ARC layer.
contact types, depending on the manner of contact formation: (1) direct Watanabe et al. [17] reported that the basicity of the glass frit affected
connections between the Ag electrode and Si [6,7], (2) current trans- the removal of the ARC coating. Ag paste containing tellurite glass with
port via Ag crystallites grown on the n+ emitter [8], and (3) tunneling a high basicity produces a low binding energy for reaction with Si3N4,
via hopping through metal granules (nano-metal colloids) precipitated resulting in a low-contact-resistance interface between the Ag electrode
in the glass layer [9,10]. Direct connections between the bulk Ag and and Si wafer. Several studies have described the relationship between P
Ag crystallites are considered the main path of current transport in such doping and microstructural changes in solar cells [18,19]. It has been
cells. suggested that the doping profiles can increase recombination by in-
In order to achieve good Ohmic contact with the Si cell, a thin glass ducing crystallite formations that deeply penetrate the Si, thereby im-
layer between the Ag electrode and Si wafer and uniform Ag re- proving device performance. However, no fundamental studies have
crystallization over the Si emitter surface are required. Previous studies investigated the relationship between the thermal and chemical prop-
regarding the influence of the front-contact microstructure in Si solar erties of glass frits, the interface between the Ag electrodes and the Si
cells have focused on the effects of sheet resistances [11] and firing wafer, and Si solar cell performance.
atmospheres during sintering. Li et al. [12] reported an evolution of Ag This study investigated the effects of adding oxides to the PbO-
crystallite and Ag colloids in glass by microstructural comparisons with TeO2-SiO2 glass system on the contact formation between Ag and Si, the
varied firing conditions to explain the current transport mechanisms. corresponding contact resistance efficiency, and the performance of the


Corresponding author.
E-mail address: kimhs@inha.ac.kr (H. Kim).

https://doi.org/10.1016/j.solmat.2018.05.015
Received 13 November 2017; Received in revised form 25 April 2018; Accepted 6 May 2018
0927-0248/ © 2018 Elsevier B.V. All rights reserved.
Y. Kim et al. Solar Energy Materials and Solar Cells 185 (2018) 97–103

Si solar cell. Glass frits with high basicity were designed to increase Table 2
reactivity with SiNx; hence, this study focused on the effects of the Basicity and I–V parameters of each cell with Ag pastes with different glass frit
microstructures and phases of the precipitates formed in the Si wafer on compositions.
the cell performance. The vibration of recrystallites formed on and Cell Glass frit Basicity Voc (mV) Isc (A) FF (%) η (%) Rs (mΩ)
penetrating the n+ emitter was quantitatively measured, and the re-
lationship between current conduction via the crystallites and the PTS Glass matrix 0.72 638 8.97 63.83 15.28 14.21
4N 1L4N 0.78 639 8.97 74.58 17.89 7.99
electrical properties of the solar cell was investigated. Our results
4Z 1L4Z 0.72 629 7.17 18.22 3.45 110.41
showed that Ag crystallites formed on the emitter increased the number 4B 1L4B 0.69 635 9.03 49.67 11.92 21.56
of current paths between the Ag electrodes and Si wafer.

2. Materials and methods monocrystalline commercial Si wafers (15.6 cm × 15.6 cm) with the
sheet resistance of 95 Ω/sq. as emitters. The front Ag electrode and back
2.1. Fabrication and characterization of glass frits Al electrode were screen-printed using the corresponding metallic
pastes and then co-fired in a belt furnace at a peak firing temperature of
The PbO-TeO2-SiO2-Bi2O3-Li2O-RO system (R = Na, Zn, or B) with 800 °C.
various glass compositions was considered. PbO (99.9%; Sigma- The electrical performance of the solar cells, as characterized by the
Aldrich), TeO2 (99%, Sigma-Aldrich), SiO2 (99.5%; Sigma-Aldrich), open-circuit voltage (Voc), short-circuit current (Isc), fill factor (FF), and
Bi2O3 (99.9%, Sigma-Aldrich), Li2CO3 (99.0%; Sigma-Aldrich), Na2CO3 efficiency (η), were measured using a solar simulator (Oriel Sol3A,
(99.5%; Sigma-Aldrich), ZnO (99.9%; Sigma-Aldrich), and H3BO3 Newport Corp., USA), and is summarized in Table 2. A map of the
(99.97%, Sigma-Aldrich) powders were weighed, mixed in the desired contact resistance of the front side of the solar cell was determined
proportions using a mixer (Turbula T2 F, Glenmills Inc., Switzerland) using a Correscan instrument (SCRA 10 9001, SunLab, Netherlands).
for 30 min, and melted at 900 °C for 20 min in an Al2O3 crucible. After The cross-sections of the solar cells were characterized using high-re-
the molten glass was quenched on a ribbon roller, the glass cullet was solution scanning electron microscopy (HR-SEM, SU 8010, Hitachi,
crushed to a powder with an agate mortar and pestle by hand and then Japan). The average thickness of the glass layer was determined by
ball-milled using a planetary mill (Pulverisette-7, Fritsch, Germany) for using five cross-sectional SEM images. For transmission electron mi-
10 min. The mean D50 particle size of the glass frits was 4–5 µm. The croscopy (TEM) examinations, cross-sectional samples of the interfaces
amorphous state of the glass frits was characterized using X-ray dif- between the Ag electrodes and Si wafers were prepared by dual-beam
fraction (XRD; DMAX-2500, Rigaku, Japan) at 40 kV and 100 mA in the focused ion beam milling (Nova 600, FEI's NanoLab, USA).
2θ range between 5° and 80° at a scan speed of 1°/min with Cu Kα The phases of the recrystallites formed on the n+ emitter were ex-
radiation. amined using an energy dispersive X-ray spectroscopy (EDS) system
The glass transition temperature (Tg) was determined using ther- integrated with the TEM (Talos F200X, FEI, USA). To observe the Ag
mogravimetric differential thermal analysis (TG-DTA; STA409PC2, crystallites on the n+ emitter of the Si solar cells and the shape of the
Netzsch, Germany) at a heating rate of 10 °C/min in air from 30 °C to pits, the Si wafers were cut to a size of 7 mm × 7 mm. These samples
600 °C. The sintering behavior of the glass frits was analyzed using hot- were firstly dipped into a 5% HF buffer solution for 2 min to remove the
stage microscopy (HSM; Misura, Expert System Solution Inc., Italy). The Ag electrodes and glass layer, followed by immersion in a 68% HNO3
composition and thermal properties (Glass-transition temperature Tg, solution for 3 min to selectively remove Ag crystallites grown inside the
softening temperature Ts) of the glass frits are summarized in Table 1. Si wafer. The surface microstructure below the glass layer between the
The design tool adopted in the current work was the basicity parameter Ag electrode and Si wafer was examined using SEM (S-4300, Hitachi,
of the oxide glasses, which permitted the prediction of the basicity of Japan). In order to analyze the number of Ag recrystallites and crys-
the resulting glass using the formula [17,20]: tallite imprints (pits) quantitatively, we chose four images of each
sample at different analysis positions. The average crystallite size, total
β= ∑ ni βi (1) crystallite area (Acryst), imprint area (Apit), and their distribution over
where β is the basicity of the resultant glass, and ni and βi are the cation each imaged area (the dimensions of each sample area were 10.65 µm
fraction and basicity parameter of the glass component i in the multi- × 7.25 µm) was calculated using image analysis software (Image-Pro
component glass, respectively. Plus, Version 6.0, Media Cybernetics Inc., USA) and the selected SEM
micrographs.
2.2. Cell fabrication and characterization
3. Results and discussion
The Ag pastes were prepared by mixing 90 wt% Ag powder (Daejoo
Co. Ltd., Korea), 8 wt% of an organic medium, and 2 wt% of each glass The efficiency of the Si solar cells is proportional to the basicity
frit (glass matrix, 1L4N, 1L4Z, 1L4B) using a three-roll mill for 20 min. parameter of the different glass frits used in the Ag pastes, and sample
The organic medium consisted of ethyl cellulose, Texanol (C12H24O3), 4 N shows the highest efficiency (Fig. 1a). This result verifies that the
and DE acetate (C8H16O4). Crystalline Si solar cells were fabricated hypothesis of reacting with the SiNx to increase the basicity parameter
using the different Ag pastes (PTS, 4 N, 4Z, 4B) with p-type was valid. However, no significant correlation is observed between the

Table 1
The composition and thermal properties (Tg, Ts) of the glass frits.
Composition (mol%)

Glass frits PbO TeO2 SiO2 Bi2O3 Li2O Na2O ZnO B2O3 Tg (°C) Ts(°C)

glass matrix 52 30 15 4 – – – 245 553


1L4N 48 29 14 4 1 4 – – 232 482
1L4Z 48 29 14 4 1 4 275 454
1L4B 48 29 14 4 1 4 244 407

Tg; Glass-transition temperature, Ts, Softening temperature of glass.

98
Y. Kim et al. Solar Energy Materials and Solar Cells 185 (2018) 97–103

Fig. 1. (a) Relationship between the basicity parameter of Pb-Te-Si glass and the efficiency of corresponding Si solar cells. (b) Series resistance (Rs) of the Si solar cells
as a function of glass layer. SEM image of the cross-section of the interfaces between the Ag electrodes and Si wafers (c1) PTS, (c2) 4 N, (c3) 4Z, and (c4) 4B.

basicity of the sample with the ZnO glass (4Z) and the efficiency of the crystal structure of the recrystallites is identified as that of Ag. From
cell. This may be because the 4Z glass has a different reactivity for a TEM images and corresponding EDS mapping of the elemental dis-
single element, and the mutual reactivity between two and three ele- tribution, no ARC is observed in the 4 N cell and no N is detected in the
ments determine the calculated glass basicity in the multicomponent glass layer at the interface between the Ag electrode and the Si. The
glasses [21]. Ag2O dissolved in the glass begins to etch the SiNx layer via the redox
The contact resistance comprises the largest part of the series re- reactions shown below [24]:
sistance, and the series resistance (Rs) dominates the overall resistance −2
4Ag+
(in glass) + 2O(in glass)+ SiNx (S) → 4Ag(in glass) + SiO2(in glass) + x/
in the structure of the solar cell. Therefore, the contact resistance
2N2 (g) (2)
should be minimized through the appropriate etching of the glass layer
for a high-efficiency solar cell. The series resistance (Rs) values of the 4Ag+
(in glass) +2O−2
(in glass)+ Si(S) → 4Ag(into Si) + SiO2(in glass) (3)
solar cell decreases on increasing the average thickness of the glass
layer from 350 nm to 500 nm (Fig. 1b). Because the average ARC The glass with added Na2O showed evidence of Ag recrystallized on
thickness is ~80 nm and the n+ emitter thickness is ~500 nm, the glass the n+ emitter, where the ARC formed a good Ohmic contact.
layer of the 4 N and PTS frits penetrated and etched the n+ emitter, Nano-precipitates with diameters of 20 nm were formed in the glass
forming an Ohmic contact with low Rs. The microstructure between the layers between the Ag electrodes and Si wafers. The ARC layers are
Ag electrodes and the Si wafer interfaces is observed as a distinctly incompletely etched and Ag recrystallites do not penetrate the Si wafer
thicker glass layer (Fig. 1 c2). However, the SiNx layer of the 4Z cell in the 4Z sample (Fig. 4a). In order to determine the composition of
cannot be etched because the film was thin, demonstrating a suitable these precipitates in the glass layer, elemental mapping was performed.
contact (Fig. 1 c3). This result confirms that the improved etching The precipitate phase contained a Bi-Te alloy; both Bi and Te were
ability of the glass frits optimizes the interface structure between the Ag detected by EDS. Because of the Zn composition, the Bi-Te alloy was
electrodes and Si wafers, and reduces Rs of the solar cells. formed in addition to Ag recrystallization. Thus, the following chemical
The cross-sections of the interfaces between the Ag electrodes and Si reaction between Bi and Te in the glass is highly probable [25]:
wafers with different frits are observed using TEM (Fig. 2). Re- BiOx(in glass) + TeOx(in glass) + SiNx (S)→ Bi1–xTexO(3+x)/2(nanoparticles) +
crystallites penetrate the Si wafer and nano-precipitates in the glass SiO2(in glass) + x/2N2(g) (4)
layer are measured with higher efficiency in the 4 N sample. However,
the 4Z glass frit shows lower efficiency than the other samples because The reactions of frits with dissolved Ag in the glass layer have been
of the etching effect; no continuous glass layer between the Ag elec- described in previous reports. The reactions among Ag, bismate glass
trode and Si wafer is observed. Ag particles are observed in the glass frits, and Si wafers formed recrystallites such as Ag-Bi alloys or Bi na-
layer as well as in the glass phase trapped between the sintered Ag noparticles [26]. The crystallization behavior of Ag powder and PbO-
electrode and Si wafer, while Si is observed in the glass frit of the glass TeO2 glass frits induced the formation of recrystallites such as Bi-Te-O
matrix (PTS) and 4B. The 4B specimen has larger Ag precipitations than or Bi-B-O [27]. Because glass with lower efficiency contains Bi-Te alloys
the PTS frit because 4B glass frits with low Ts show higher viscous flow in the glass layer, we assumed that the other phases would contain Ag
[22,23]. precipitates because of the disturbance caused by tunneling via hopping
When analyzing the 4 N sample using TEM, it is seen that the re- through Ag nanocolloids in the glass layer as the dominant mode of
crystallites penetrating the Si comprise Ag from the paste (Fig. 3a). The current conduction.

99
Y. Kim et al. Solar Energy Materials and Solar Cells 185 (2018) 97–103

Fig. 2. Microstructures of the interfaces between the Ag electrodes and Si wafers. (a) PTS, (b) 4 N, (c) 4Z, and (d) 4B.

Fig. 3. (a) Indexed selected-area electron diffraction (SAED) pattern taken along the [001] zone axis (inset). Cross-sectional TEM image and corresponding EDS
compositional maps for (b) Ag, (c) Si, (d) O, (e) N, (f) Bi, and (g) Te within the Ag contact of textured Si solar cells produced using a 4 N frit.

To confirm the relationship between the basicity of the glass frits and the oxygen ion for glass components [20], causing an increase in
and the chemical reaction with the ARC, the microstructures of the the donor electrons from oxide ions and in the reactivity with SiNx [17];
samples prepared with different basicity parameters are shown in Fig. 5 this would enhance the formation of inverse pyramidal pits on the n+
a1–d1. Higher basicity parameters lead to higher Ag crystallite den- emitter, according to the redox reaction (2). With increases in the ba-
sities. The mean number of Ag recrystallites formed on the Si wafer is sicity parameter of the glass frits, the network modifiers in the glass are
decreased from 349 to 74 with the different frits (in the order of 4 N, decreased, and thus the dissolved O2− ions in the glass layer are in-
PTS, 4B, and 4Z) for an area of 77 µm2. It is expected that increasing the creased [20]. According to redox reactions (2) and (3), an increase in
basicity of glass would reduce the coulombic force between the cation O2− corresponds to an increase in Ag crystallite formation at the

100
Y. Kim et al. Solar Energy Materials and Solar Cells 185 (2018) 97–103

Fig. 4. (a) Cross-sectional TEM image and corresponding EDS compositional maps for (b) Ag, (c) Si, (d) O, (e) N, (f) Bi, and (g) Te within the Ag contact of textured Si
solar cells produced using the 4Z frit.

interface between the glass layer and Si [8,13]. Many pits (imprints of frits showed a lower density of Ag recrystallites and pits on the n+
Ag recrystallites) remain in the Si wafer after the embedded re- emitter due to the formation of the Bi-Te alloy, suggesting that the
crystallites are removed (Fig. 5 a2–d2). The higher basicity parameter number of current paths from the Ag bulk or Ag recrystallizes to the Si
of the glass frits induces the formation of more Ag recrystallites and pits wafer was reduced.
on the Si surface. Lohmüller et al. [30] and the analysis of their work by Rothhardt
The Ag crystallite imprints on the glass-free pyramid top were at- et al. [31] showed that increasing the size of Ag crystallites or pits
tributed to direct contact with the bulk Ag finger, as suggested by corresponded to enhanced direct contact between the Ag electrodes and
Cabrera et al. [28]. However, we considered that the Ag recrystallites the Si surface, thereby increasing the conduction channels; this corre-
on the n+ emitter also increased the number of conduction paths be- sponded to a reduction in the total contact resistance. It is proposed that
tween the Ag bulk and Si wafer. Therefore, it is assumed that the se- the current conduction is inversely proportional to the area of a single
parating influence of the Ag recrystallites formed conductive channels crystallite in direct contact, which is expressed by the following equa-
and covered imprints penetrating the n+ emitter, resulting in direct tion.
conduction paths [29]. A higher concentration of larger Ag crystallites
ρc−1 = Acryst or pit / ρc, crystal (5)
was observed for higher-efficiency cells. The fraction of the sample
covered by crystallites and pits on the Si wafer decreased in the order of where the parameter ρc, crystal represents the contact resistance of a
4 N, PTS, 4B, and 4Z frits (Fig. 6a). The etched Ag electrodes of 4 N single crystallite. To investigate the effect of pits on the Si emitter on
showed a higher area fraction of Ag crystallites (13.2%) and smaller the contact resistance of the cell, we quantitatively calculated the re-
(< 0.1 µm2) Ag crystallites compared to the 4Z sample (1.8%). The lative coverage area of the Ag crystallites and pits, and measured the
mean diameter of Ag crystallites of the 4 N sample is approximately contact resistance of the cell.
155 ± 4 nm, while that of the 4 N pit is 197 ± 37 nm (Fig. 6b). The 4Z To observe the effect of recrystallites and pits in the n+ emitter on

Fig. 5. Surface SEM images of recrystallites formed on the n+ emitter with (a1) PTS, (b1) 4 N, (c1) 4Z, and (d1) 4B after removing the Ag electrodes and glass layers.
Surface SEM image of the etching pits caused by Ag precipitates on Si wafers after HF and HNO3 etching for (a2) PTS, (b2) 4 N, (c2) 4Z, and (d2) 4B.

101
Y. Kim et al. Solar Energy Materials and Solar Cells 185 (2018) 97–103

Fig. 6. Relative number of (a) Ag recrystallites and (b) pit size depending on the glass basicity parameter.

the conductive paths between the Ag bulk and Si wafer, the correlation
between the inverse of contact resistance (σc−1) and the area fraction of
recrystallites and pits was investigated (Fig. 7). When the area fraction
of crystallites is increased from 1.8% to 13.2%, the inverse of contact
resistance is increased substantially, from 0.03 (mΩ·cm2)-1 to 0.78 (m
Ω·cm2)-1. The contact resistance is decreased linearly with increases in
the area fraction of pits. The increased density of pits on the n+ emitter
at the contact interface causes a decrease in the contact resistance; this
should correspond to an increase in the conductivity of the Ag crys-
tallites and pits, allowing improved current transport.
These mechanisms of formation of a contact model comprising the
Ag electrodes, glass phase, Ag recrystallites, and Si wafers during the
sintering process are shown schematically (Fig. 8). The glass frits began
to show viscous flow when fired at the softening temperature Ts and the
molten glass began to wet the SiNx layers. The dissolved Ag and oxygen
Fig. 7. Inverse of the specific contact resistance σc−1 as a function of the area ions in the glass layer etched the SiNx layer, and the pits formed on the
fraction of crystallite and pits with the Ag busbar. Si defects, according to the redox reaction (2). After fast cooling, Ag

Fig. 8. Schematic of glass frits with different basicity parameters on Ag-Si contacts, showing the formation of crystallites in the glass during firing: (a) 4 N, (b) 4Z, and
(c) 4B.

102
Y. Kim et al. Solar Energy Materials and Solar Cells 185 (2018) 97–103

recrystallites and Bi-Te alloy formed between Ag electrodes and Si, and contacts of crystalline Si solar cells–Review of existing models and recent devel-
then, Ag recrystallites grew through Ostwald ripening, according to the opments, Sol. Energy Mater. Sol. Cells 90 (2006) 3399–3406.
[7] E. Cabrera, S. Olibet, J. Glatz-Reichenbach, R. Kopecek, D. Reinke, G. Schubert,
redox reaction (3). The higher basicity parameter of 4 N relative to 4B Experimental evidence of direct contact formation for the current transport in silver
correlated to stronger etching behavior by the glass frits in the n+ thick film metalized silicon emitters, J. Appl. Phys. 110 (2011) 114511.
emitter (Fig. 8a). This increased the density of Ag crystallites present in [8] M. Eberstein, S. Körner, High performance front side pastes by Ag precursor
equipped glasses, Photovoltaic Specialist Conference (PVSC) 42nd IEEE, IEEE 14-19
the n+ emitter, which deepened the penetration of the Si wafer. It is June, New Orleans, LA, 2015.
expected that increasing the number of size of conductive channels via [9] C. Ballif, D.M. Huljić, G. Willeke, A. Hessler-Wyser, Silver thick-film contacts on
crystallites minimizes the series resistance and enhances the solar cell highly doped n-type silicon emitters: structural and electronic properties of the
interface, Appl. Phys. Lett. 82 (2003) 1878–1880.
efficiency. However, in the case of the 4Z sample, thin and incomplete [10] C. Lin, S. Tsai, S. Hsu, M. Hsieh, Investigation of Ag-bulk/glassy-phase/Si hetero-
etching of the SiNx layer was observed due to the Bi-Te alloy crystallites structures of printed Ag contacts on crystalline Si solar cells, Sol. Energy Mater. Sol.
formed in the glass matrix during sintering, thus the kinetics for the Cells 92 (2008) 1011–1015.
[11] D. Pysch, A. Mette, A. Filipovic, S.W. Glunz, Comprehensive analysis of advanced
reactions between the Ag, glass frits, SiNx, and Si were decreased
solar cell contacts consisting of printed fine-line seed layers thickened by silver
(Fig. 8b). This could correlate to smaller crystallites and lower crys- plating, Prog. Photovolt: Res. Appl. 17 (2009) 101–114.
tallinity, attributed to poor contact formation. Therefore, the fabricated [12] Z.G. Li, L. Liang, A.S. Ionkin, B.M. Fish, M.E. Lewittes, L.K. Cheng, K.R. Mikeska,
solar cell based on the 4Z glass frit achieved a higher contact resistance Microstructural comparison of silicon solar cells' front-side Ag contact and the
evolution of current conduction mechanisms, J. Appl. Phys. 110 (2011) 074304.
but lower efficiency. [13] C. Kim, J. Choi, S. Choi, S. Kim, H. Park, H. Song, S. Yoon, J. Huh, Y. Kang, H. Lee,
D. Kim, Effects of current-injection firing with Ag paste in a boron emitter, Sci. Rep.
4. Conclusions 6 (2016) 21553.
[14] Y. Shih, Y. Lin, J. You, F.G. Shi, Screen-printable silver pastes with nanosized glass
frits for silicon solar cells, J. Electron. Mater. 42 (2013) 410–416.
We investigated the correlations between the formation of contacts [15] D. Kim, S. Hwang, H. Kim, Effect of the thermal properties of frits on the electrical
between Ag paste electrodes with glass frits of different basicity and Si properties of screen-printed silicon solar cells, J. Korean Phys. Soc. J. 55 (2009)
1046–1050.
wafers and the electrical characteristics of corresponding three-busbar [16] Y. Zhang, Y. Yang, J. Zheng, W. Hua, G. Chen, Thermal properties of glass frit and
Si solar cells. Higher-basicity glass frits showed greater concentrations effects on Si solar cells, Mater. Chem. Phys. 114 (2009) 319–322.
of Ag recrystallites formed on the n+ emitter, where the penetration [17] S. Watanabe, T. Kodera, T. Ogihara, Influence of tellurite glass on reaction between
Si3N4 anti-reflection coating film and Ag paste for electrodes in Si solar cells, J.
depth of these recrystallites into the Si was positively correlated to the Ceram. Soc. Jpn. 124 (2016) 218–222.
efficiency of the solar cell. However, a solar cell with Ag electrodes [18] G. Kulushich, R. Zapf-Gottwick, V.X. Nguyen, J.H. Werner, Role of phosphorus in
containing a glass doped with ZnO showed the precipitation of a Bi-Te contact formation on silicon solar cells, Phys. Status Solidi RRL 6 (2012) 370–372.
[19] E. Cabrera, S. Olibet, D. Rudolph, P.E. Vullum, R. Kopecek, D. Reinke, C. Herzog,
alloy in the glass layer between the Ag electrodes and Si wafer, which
D. Schwaderer, G. Schubert, Impact of excess phosphorus doping and Si crystalline
was thought to prevent the formation of crystallites on the n+ emitter defects on Ag crystallite nucleation and growth in silver screen-printed Si solar
and decrease the efficiency of the cell because of the increased contact cells, Prog. Photovoltaics Res. Appl. 23 (2015) 367–375.
resistance at the interface. The measured contact resistance linearly [20] K. Morinaga, H. Yoshida, H. Takebe, Compositional dependence of absorption
spectra of Ti3+ in silicate, borate, and phosphate glasses, J. Am. Ceram. Soc. 77
increased with increasing crystallites and pits, corresponding to con- (1994) 3113–3118.
duction paths from the Si wafer to the Ag bulk. This study confirmed [21] W. Thiemsorn, K. Keowkamnerd, S. Phanichphant, P. Suwannathada,
that the growth of Ag crystallites and their penetration into the n+ H. Hessenkemper, Influence of glass basicity on rodex interactions of ir-
on–manganese–copper ion pairs in soda–lime–silica glass, Glass Phys. Chem. 34
emitters strongly governed the development of conduction paths, thus (2008) 19–29.
contributing to improvements in Si solar cell efficiency. [22] G. Zheng, Y. Tai, H. Wang, J. Bai, Effect of the Pb–Te–B–O system glass frits in the
front contact paste on the conversion efficiency of crystalline silicon solar cells, J.
Mater. Sci. - Mater. Electron. 25 (2014) 3779–3786.
Acknowledgements [23] J. Qin, W. Zhang, S. Bai, Z. Liu, Study on the sintering and contact formation
process of silver front side metallization pastes for crystalline silicon solar cells,
This research was supported by the World Class 300 Project Appl. Surf. Sci. 376 (2016) 52–61.
[24] K. Hong, S. Cho, J. You, J. Jeong, S. Bea, J. Huh, Mechanism for the formation of Ag
(WC300,Development of the Electrode Materials for High Efficiency
crystallites in the Ag thick-film contacts of crystalline Si solar cells, Sol. Energy
(21%) and Low Cost C-Si Solar Cell, S2340684) funded by The Ministry Mater. Sol. Cells 93 (2009) 898–904.
of SMEs and Startups(MSS), South Korea. [25] L. Pöppl, Z. Szaller, Reactions and phases within the TeO2-rich part of the Bi2O3-
TeO2 system, J. Therm. Anal. Calorim. 74 (2003) 375–386.
[26] D. Kim, H. Kim, Effect of bismate frits under fast firing on reactions between Ag
References electrodes and Si wafer for Si solar cells, J. Am. Ceram. Soc. 96 (2013) 774–780.
[27] J. Qin, W. Zhang, S. Bai, Z. Liu, Effect of Pb–Te–O glasses on Ag thick-film contact in
[1] X. Zeng, T. Zhou, C. Leng, Z. Zang, M. Wang, W. Hu, X. Tang, S. Lu, L. Fang, crystalline silicon solar cells, Sol. Energy Mater. Sol. Cells 144 (2016) 256–263.
M. Zhou, Performance improvement of perovskite solar cells by employing a CdSe [28] E. Cabrera, S. Olibet, J. Glatz-Reichenbach, R. Kopecek, D. Reinke, G. Schubert,
quantum dot/PCBM composite as an electron transport layer, J. Mater. Chem. A 5 Experimental evidence of direct contact formation for the current transport in silver
(2017) 17499–17505. thick film metallized silicon emitters, J. Appl. Phys. 110 (2011) 114511.
[2] Z. Zang, Efficiency enhancement of ZnO/Cu2O solar cells with well oriented and [29] M. Pfeffer, P. Kumar, O. Eibl, High-efficiency crystalline-Si solar cells with screen-
micrometer grain sized Cu2O films, Appl. Phys. Lett. 112 (2018) 042106. printed front-side metallization: a percolation model to explain the current path, J.
[3] G.C. Cheek, R.P. Mertens, R. Van Overstraeten, L. Frisson, Thick-film metallization Electron. Mater. 45 (2016) 5764–5772.
for solar cell applications, IEEE T. Electr. Dev. ED 31 (1984) 602–609. [30] E. Lohmüller, S. Werner, R. Hoenig, J. Greulich, F. Clement, Impact of boron doping
[4] M. Prudenziati, L. Moro, B. Morten, F. Sirotti, L. Sardi, Ag-based thick-film front profiles on the specific contact resistance of screen printed Ag–Al contacts on si-
metallization of silicon solar cells, Act. Passiv. Electron. Compon. 13 (1989) licon, Sol. Energy Mater. Sol. Cells 142 (2015) 2–11.
133–150. [31] P. Rothhardt, S. Meier, R. Hoenig, A. Wolf, D. Biro, Influence of surface near doping
[5] D. Neuhaus, A. Munzer, Industrial silicon wafer solar cells, Adv. OptoElectron. concentration on contact formation of silver thick film contacts, Sol. Energy Mater.
(2007) 24521. Sol. Cells 153 (2016) 25–30.
[6] G. Schubert, F. Huster, P. Fath, Physical understanding of printed thick-film front

103

S-ar putea să vă placă și