- Email: [email protected]
Identifying Screen Properties to Optimise Solar Cell Efficiency and Cost
Available online at www.sciencedirect.com
Energy Procedia 33 (2013) 84 – 90
PV Asia Pacific Conference 2012
Identifying Screen Properties to Optimise Solar Cell Efficiency and Cost Z. Fortu, L. X. Song* Heraeus Materials Singapore Pte Ltd, 9 Tuas Avenue 5, Singapore 639335, Singapore
Abstract With an increasing competition in the solar market, most cell manufacturers are moving towards aggressive reduction in direct material cost, while still committed to improve cell efficiency. Although front side silver metallisation has always been part of cell efficiency optimisation, recent focus has been heavily shifted to reducing the paste consumption as well while maintaining the contact quality in a vast range of emitters. Therefore, this challenge relies greatly on paste manufacturers to produce low resistivity materials and on choosing the right screen mesh, print layout and emulsion thickness to optimise efficiency at a reduced cost. The purpose of this study is to identify cell parameters trends with respect to changes in screen properties and consumption.
©2013 2013ThePublished by Elsevier Ltd. Ltd. Selection and/or peer-review under responsibility of Solar Energy © Authors. Published by Elsevier Research Institute of Singapore (SERIS) University ofInstitute Singapore (NUS). The PV –Asia Pacific Selection and peer-review under responsibility ofNational Solar Energy Research of Singapore (SERIS) National University of Singapore Theorganised PV Asia Pacifi Conference jointly organised byIndustry SERIS and the Asian Conference 2012 was(NUS). jointly by cSERIS and 2012 the was Asian Photovoltaic Association Photovoltaic (APVIA). Industry Association (APVIA) Keywords: Screen; emulsion; paste; contact resistance; line resistance
* Corresponding author. Tel.: +65 9137 2004; fax: +65 6571 7520. E-mail address: [email protected]; [email protected]
1876-6102 © 2013 The Authors. Published by Elsevier Ltd.
Selection and peer-review under responsibility of Solar Energy Research Institute of Singapore (SERIS) – National University of Singapore (NUS). The PV Asia Pacific Conference 2012 was jointly organised by SERIS and the Asian Photovoltaic Industry Association (APVIA) doi:10.1016/j.egypro.2013.05.043
Z. Fortu and L.X. Song / Energy Procedia 33 (2013) 84 – 90
1. Introduction Current industry trend showed three common efficiency and cost optimisation approaches that directly affect front side metallisation: Shift to fine line printing, from >70 um to <60 um where choice of screen meshes and design becomes increasingly important. Changes in the nominal emulsion thickness which is critical in maintaining aspect ratio Increasing sheet resistances of the emitters from the traditional 65 ohms/sq to a higher range of 80 to 100 ohms/sq, which critically affects the front side contact formation With the above mentioned cell design changes, the challenge is not up to the cell manufacturers only but also to the paste manufacturers who are confronted with producing low resistivity paste that can be finely printed in at least 10% reduction in consumption. Heraeus, for example, has recently developed it new front side silver paste, SOL 9600 series, which is distinguished from reference (previous generation paste) as having considerably better contact and line resistivity, refer to Fig. 1. Contact resistance defines the ability for the paste to contact to the emitter and line resistivity defines the conductivity of the material along the printed metal lines.
Boxplot of Rline
Rline
Rcontact
Boxplot of Rcontact
Reference
SOL 9600 Paste
Reference
SOL 9600 Paste
Fig. 1. Contact resistance (left) and line resistance (right) comparing reference and SOL 9600
SOL 9600 paste enables fine line printing, applicable to a wider range of emitter types, without affecting the contact quality. Therefore, this addresses issues from the material side and is contributing much in the reduction of silver consumption while efficiency improvements are still obvious. The remaining issue then is on choosing the correct screen properties to suit printing requirements. This paper discusses the trends related to EOM (emulsion over mesh) thickness, finger opening and screen meshes widely available in the market.
85
86
Z. Fortu and L.X. Song / Energy Procedia 33 (2013) 84 – 90
2. Experimental Studies have been made specifically on the printing behaviour of screens tabulated in Table 1 below, using standard 65 ohms per sq. wafers. Each wafer is printed with SOL 9600 paste at exactly the same printing parameters and fired at the optimum firing temperature. To understand the trends, finger width and height were evaluated using a Keyence microscope, Rcontact and Rline are measured using GP Solar Tester and electrical results were taken from Berger Cell Tester. Table 1. Screen experiment matrix Screen Mesh
Finger Opening (um)
EOM Thickness (um)
290/20
40 | 45 | 50 | 55 | 60 | 65 | 70
12 | 15 | 18 | 20 | 23
400/18
40 | 45 | 50 | 55 | 60 | 65 | 70
12 | 15 | 18 | 20 | 23
3. Results and discussion 3.1. Paste consumption As expected the consumption increases with EOM (emulsion over mesh), but the consumption between 290/20 and 400/18 mesh is different and is illustrated in Fig. 2. From the result, 400/18 mesh is more economic in terms of consumption and the differences between EOM thickness is more pronounced. The reason for this is that 400/18 mesh printing narrower than 290/20 mesh as shown in the line width measurements in Fig. 3. Mesh/EOM Vs Laydown 105
Mesh 290/20 400/18 REF-290/20
Mean of Laydown
100 95 90 85 80 75
Fig. 2. Mesh/EOM vs. Consumption
290/20 400/18 REF-290/20 20
290/20 400/18 REF-290/20 290/20 REF-20 400/18 290/20 REF-290/20
290/20 400/18 REF-290/20 18
23
290/20 400/18 REF-290/20 15
EOM Thk
290/20 400/18 REF-290/20
65 Mesh
12
70
87
Z. Fortu and L.X. Song / Energy Procedia 33 (2013) 84 – 90
3.2. Finger width As the cell efficiency is dependent on paste weight, it is needed to study the finger profile and aspect ratio of each group. Aside from the observation that 400/18 mesh produces narrower finger, there is also a general trend that increasing EOM within the same finger opening reduces the width and is more obvious at less than 60 um (refer to Fig. 3: finger width trends). 3.3. Finger height Increasing finger height with EOM is only true at opening greater than 60 um; at below 55 um the height drops with increasing EOM. Hence, there is a need to reduce EOM when using narrower finger design (refer to Fig. 4: finger height trends).
Boxplot of Finger Width
Screen
12 15 40 18 20 23 12 15 45 18 20 23 12 15 50 18 20 23 12 15 400/18 55 18 20 23 12 15 60 18 20 23 12 15 65 18 20 23 12 15 70 18 20 23
EOM Finger Opening
12 15 40 18 20 23 12 15 45 18 20 23 12 15 50 18 20 23 12 15 290/20 55 18 20 23 12 15 60 18 20 23 12 15 65 18 20 23 12 15 70 18 20 23
Finger Width
Finger Opening 40 45 50 55 60 65 70
Fig. 3. Finger width trends Boxplot of Finger Height
Screen
Fig. 4. Finger height trends
12 15 40 18 20 23 12 15 45 18 20 23 12 15 50 18 20 23 12 15 400/18 55 18 20 23 12 15 60 18 20 23 12 15 65 18 20 23 12 15 70 18 20 23
EOM Finger Opening
12 15 40 18 20 23 12 15 45 18 20 23 12 15 50 18 20 23 12 15 290/20 55 18 20 23 12 15 60 18 20 23 12 15 65 18 20 23 12 15 70 18 20 23
Finger Height
Finger Opening 40 45 50 55 60 65 70
88
Z. Fortu and L.X. Song / Energy Procedia 33 (2013) 84 – 90
3.4. Line resistance Finger height observations are reinforced by the Rline trend. Figure 5 below shows that the Rline trends lower with increasing EOM but becomes worst at very high EOM (23 um). Figure 5 can be used as a guideline for choosing EOM at certain finger opening. 3.5. Aspect ratio Considering the combined effect of finger width and height which is expressed in terms of aspect ratio (height divided by width), and with consumption considered, the optimum condition is between 50 to 60 um. Generally, trends become relatively flat comparing 50 to 70 um, such that printing on 70 um will just incur more paste consumption without much gain in EFF (refer to Fig. 6: Aspect ratio trends). Boxplot of RLine
Screen
12 15 40 18 20 23 12 15 45 18 20 23 12 15 50 18 20 23 12 15 400/18 55 18 20 23 12 15 60 18 20 23 12 15 65 18 20 23 12 15 70 18 20 23
EOM Finger Opening
12 15 40 18 20 23 12 15 45 18 20 23 12 15 50 18 20 23 12 15 290/20 55 18 20 23 12 15 60 18 20 23 12 15 65 18 20 23 12 15 70 18 20 23
RLine
Finger Opening 40 45 50 55 60 65 70
Fig. 5. Rline trends
Boxplot of Aspect Ratio
Screen
Fig. 6. Aspect ratio trends
12 15 40 18 20 23 12 15 45 18 20 23 12 15 50 18 20 23 12 15 400/18 55 18 20 23 12 15 60 18 20 23 12 15 65 18 20 23 12 15 70 18 20 23
EOM Finger Opening
12 15 40 18 20 23 12 15 45 18 20 23 12 15 50 18 20 23 12 15 290/20 55 18 20 23 12 15 60 18 20 23 12 15 65 18 20 23 12 15 70 18 20 23
Aspect Ratio
Finger Opening 40 45 50 55 60 65 70
89
Z. Fortu and L.X. Song / Energy Procedia 33 (2013) 84 – 90
3.6. Electrical trends EOM affects electrical trends as shown in Fig. 7. For 290/20, the optimum EOM is between 18 um to 20 um and for 400/18 is between 15 um to 18 um, where, for both case, Isc and FF are optimum. Analysis will have to be aligned with Rline trends in Fig. 4, to choose relative finger opening. To balance Rline, however, adding fingers, will help and will need relative Isc losses considerations, and is not discussed in this paper. For finger width, 400/18 offers advantage, hence Isc gain is seen. 290/20, on the other hand, will have advantage at wider opening. 400/18 will have FF advantage at narrower finger line as the Rline will be much better. 290/20 NCell-Isc Trend
290/20/12
290/20/15 EFF
Isc
290/20/18
290/20/20
Poly. (EFF)
400/18 NCell-Isc Trend
290/20/23
400/18/12
Poly. (Isc)
400/18/15 EFF
290/20 Rs-FF Trend
290/20/12
290/20/15 FF
Rs
290/20/18
290/20/20
Poly. (FF)
Isc
400/18/18
400/18/20
Poly. (EFF)
400/18/23
Poly. (Isc)
400/18 Rs-FF Trend
290/20/23
Poly. (Rs)
400/18/12
400/18/15 FF
Rs
400/18/18
400/18/20
Poly. (FF)
400/18/23
Poly. (Rs)
Fig. 7. Electrical trends with different mesh and EOM thickness
4. Conclusions Before changing the screen design, especially when going for finer line printing, it is important to understand the relative Rcontact and Rline of the pastes. Then, choosing the right screen that suits the emitter and the printability of the paste are the keys to optimising the cell efficiency and cost. With SOL9600, minimum paste consumption savings is 10% relative to reference and can save even further with reducing finger opening and EOM.
90
Z. Fortu and L.X. Song / Energy Procedia 33 (2013) 84 – 90
Between meshes, 400/18 mesh will always have advantage over 290/20 mesh in terms of cost with minimum changes in efficiencies. Trade off in FF and Isc happens in choosing between 290/20 and 400/18 meshes. Figure 5 above, can be used as a guide in choosing the correct EOM for a specific finger opening. To balance resistance losses, additional finger lines may need to be added. It is not always true that higher EOM will give higher finger or better aspect ratio. At finer finger, 40 um for example, the trend in aspect ratio is inversely proportional to EOM thickness. If the EOM is too high and the finger opening is narrow, the paste transfer will be affected.
Acknowledgements Thanks to the entire Heraeus Team, for giving their expertise and support for the experiment. Thanks to Brave C&H for providing the screens required for this experiment. References [1]
Top Contact Design; http://www.pveducation.org/pvcdrom/design
Energy Procedia 33 (2013) 84 – 90
PV Asia Pacific Conference 2012
Identifying Screen Properties to Optimise Solar Cell Efficiency and Cost Z. Fortu, L. X. Song* Heraeus Materials Singapore Pte Ltd, 9 Tuas Avenue 5, Singapore 639335, Singapore
Abstract With an increasing competition in the solar market, most cell manufacturers are moving towards aggressive reduction in direct material cost, while still committed to improve cell efficiency. Although front side silver metallisation has always been part of cell efficiency optimisation, recent focus has been heavily shifted to reducing the paste consumption as well while maintaining the contact quality in a vast range of emitters. Therefore, this challenge relies greatly on paste manufacturers to produce low resistivity materials and on choosing the right screen mesh, print layout and emulsion thickness to optimise efficiency at a reduced cost. The purpose of this study is to identify cell parameters trends with respect to changes in screen properties and consumption.
©2013 2013ThePublished by Elsevier Ltd. Ltd. Selection and/or peer-review under responsibility of Solar Energy © Authors. Published by Elsevier Research Institute of Singapore (SERIS) University ofInstitute Singapore (NUS). The PV –Asia Pacific Selection and peer-review under responsibility ofNational Solar Energy Research of Singapore (SERIS) National University of Singapore Theorganised PV Asia Pacifi Conference jointly organised byIndustry SERIS and the Asian Conference 2012 was(NUS). jointly by cSERIS and 2012 the was Asian Photovoltaic Association Photovoltaic (APVIA). Industry Association (APVIA) Keywords: Screen; emulsion; paste; contact resistance; line resistance
* Corresponding author. Tel.: +65 9137 2004; fax: +65 6571 7520. E-mail address: [email protected]; [email protected]
1876-6102 © 2013 The Authors. Published by Elsevier Ltd.
Selection and peer-review under responsibility of Solar Energy Research Institute of Singapore (SERIS) – National University of Singapore (NUS). The PV Asia Pacific Conference 2012 was jointly organised by SERIS and the Asian Photovoltaic Industry Association (APVIA) doi:10.1016/j.egypro.2013.05.043
Z. Fortu and L.X. Song / Energy Procedia 33 (2013) 84 – 90
1. Introduction Current industry trend showed three common efficiency and cost optimisation approaches that directly affect front side metallisation: Shift to fine line printing, from >70 um to <60 um where choice of screen meshes and design becomes increasingly important. Changes in the nominal emulsion thickness which is critical in maintaining aspect ratio Increasing sheet resistances of the emitters from the traditional 65 ohms/sq to a higher range of 80 to 100 ohms/sq, which critically affects the front side contact formation With the above mentioned cell design changes, the challenge is not up to the cell manufacturers only but also to the paste manufacturers who are confronted with producing low resistivity paste that can be finely printed in at least 10% reduction in consumption. Heraeus, for example, has recently developed it new front side silver paste, SOL 9600 series, which is distinguished from reference (previous generation paste) as having considerably better contact and line resistivity, refer to Fig. 1. Contact resistance defines the ability for the paste to contact to the emitter and line resistivity defines the conductivity of the material along the printed metal lines.
Boxplot of Rline
Rline
Rcontact
Boxplot of Rcontact
Reference
SOL 9600 Paste
Reference
SOL 9600 Paste
Fig. 1. Contact resistance (left) and line resistance (right) comparing reference and SOL 9600
SOL 9600 paste enables fine line printing, applicable to a wider range of emitter types, without affecting the contact quality. Therefore, this addresses issues from the material side and is contributing much in the reduction of silver consumption while efficiency improvements are still obvious. The remaining issue then is on choosing the correct screen properties to suit printing requirements. This paper discusses the trends related to EOM (emulsion over mesh) thickness, finger opening and screen meshes widely available in the market.
85
86
Z. Fortu and L.X. Song / Energy Procedia 33 (2013) 84 – 90
2. Experimental Studies have been made specifically on the printing behaviour of screens tabulated in Table 1 below, using standard 65 ohms per sq. wafers. Each wafer is printed with SOL 9600 paste at exactly the same printing parameters and fired at the optimum firing temperature. To understand the trends, finger width and height were evaluated using a Keyence microscope, Rcontact and Rline are measured using GP Solar Tester and electrical results were taken from Berger Cell Tester. Table 1. Screen experiment matrix Screen Mesh
Finger Opening (um)
EOM Thickness (um)
290/20
40 | 45 | 50 | 55 | 60 | 65 | 70
12 | 15 | 18 | 20 | 23
400/18
40 | 45 | 50 | 55 | 60 | 65 | 70
12 | 15 | 18 | 20 | 23
3. Results and discussion 3.1. Paste consumption As expected the consumption increases with EOM (emulsion over mesh), but the consumption between 290/20 and 400/18 mesh is different and is illustrated in Fig. 2. From the result, 400/18 mesh is more economic in terms of consumption and the differences between EOM thickness is more pronounced. The reason for this is that 400/18 mesh printing narrower than 290/20 mesh as shown in the line width measurements in Fig. 3. Mesh/EOM Vs Laydown 105
Mesh 290/20 400/18 REF-290/20
Mean of Laydown
100 95 90 85 80 75
Fig. 2. Mesh/EOM vs. Consumption
290/20 400/18 REF-290/20 20
290/20 400/18 REF-290/20 290/20 REF-20 400/18 290/20 REF-290/20
290/20 400/18 REF-290/20 18
23
290/20 400/18 REF-290/20 15
EOM Thk
290/20 400/18 REF-290/20
65 Mesh
12
70
87
Z. Fortu and L.X. Song / Energy Procedia 33 (2013) 84 – 90
3.2. Finger width As the cell efficiency is dependent on paste weight, it is needed to study the finger profile and aspect ratio of each group. Aside from the observation that 400/18 mesh produces narrower finger, there is also a general trend that increasing EOM within the same finger opening reduces the width and is more obvious at less than 60 um (refer to Fig. 3: finger width trends). 3.3. Finger height Increasing finger height with EOM is only true at opening greater than 60 um; at below 55 um the height drops with increasing EOM. Hence, there is a need to reduce EOM when using narrower finger design (refer to Fig. 4: finger height trends).
Boxplot of Finger Width
Screen
12 15 40 18 20 23 12 15 45 18 20 23 12 15 50 18 20 23 12 15 400/18 55 18 20 23 12 15 60 18 20 23 12 15 65 18 20 23 12 15 70 18 20 23
EOM Finger Opening
12 15 40 18 20 23 12 15 45 18 20 23 12 15 50 18 20 23 12 15 290/20 55 18 20 23 12 15 60 18 20 23 12 15 65 18 20 23 12 15 70 18 20 23
Finger Width
Finger Opening 40 45 50 55 60 65 70
Fig. 3. Finger width trends Boxplot of Finger Height
Screen
Fig. 4. Finger height trends
12 15 40 18 20 23 12 15 45 18 20 23 12 15 50 18 20 23 12 15 400/18 55 18 20 23 12 15 60 18 20 23 12 15 65 18 20 23 12 15 70 18 20 23
EOM Finger Opening
12 15 40 18 20 23 12 15 45 18 20 23 12 15 50 18 20 23 12 15 290/20 55 18 20 23 12 15 60 18 20 23 12 15 65 18 20 23 12 15 70 18 20 23
Finger Height
Finger Opening 40 45 50 55 60 65 70
88
Z. Fortu and L.X. Song / Energy Procedia 33 (2013) 84 – 90
3.4. Line resistance Finger height observations are reinforced by the Rline trend. Figure 5 below shows that the Rline trends lower with increasing EOM but becomes worst at very high EOM (23 um). Figure 5 can be used as a guideline for choosing EOM at certain finger opening. 3.5. Aspect ratio Considering the combined effect of finger width and height which is expressed in terms of aspect ratio (height divided by width), and with consumption considered, the optimum condition is between 50 to 60 um. Generally, trends become relatively flat comparing 50 to 70 um, such that printing on 70 um will just incur more paste consumption without much gain in EFF (refer to Fig. 6: Aspect ratio trends). Boxplot of RLine
Screen
12 15 40 18 20 23 12 15 45 18 20 23 12 15 50 18 20 23 12 15 400/18 55 18 20 23 12 15 60 18 20 23 12 15 65 18 20 23 12 15 70 18 20 23
EOM Finger Opening
12 15 40 18 20 23 12 15 45 18 20 23 12 15 50 18 20 23 12 15 290/20 55 18 20 23 12 15 60 18 20 23 12 15 65 18 20 23 12 15 70 18 20 23
RLine
Finger Opening 40 45 50 55 60 65 70
Fig. 5. Rline trends
Boxplot of Aspect Ratio
Screen
Fig. 6. Aspect ratio trends
12 15 40 18 20 23 12 15 45 18 20 23 12 15 50 18 20 23 12 15 400/18 55 18 20 23 12 15 60 18 20 23 12 15 65 18 20 23 12 15 70 18 20 23
EOM Finger Opening
12 15 40 18 20 23 12 15 45 18 20 23 12 15 50 18 20 23 12 15 290/20 55 18 20 23 12 15 60 18 20 23 12 15 65 18 20 23 12 15 70 18 20 23
Aspect Ratio
Finger Opening 40 45 50 55 60 65 70
89
Z. Fortu and L.X. Song / Energy Procedia 33 (2013) 84 – 90
3.6. Electrical trends EOM affects electrical trends as shown in Fig. 7. For 290/20, the optimum EOM is between 18 um to 20 um and for 400/18 is between 15 um to 18 um, where, for both case, Isc and FF are optimum. Analysis will have to be aligned with Rline trends in Fig. 4, to choose relative finger opening. To balance Rline, however, adding fingers, will help and will need relative Isc losses considerations, and is not discussed in this paper. For finger width, 400/18 offers advantage, hence Isc gain is seen. 290/20, on the other hand, will have advantage at wider opening. 400/18 will have FF advantage at narrower finger line as the Rline will be much better. 290/20 NCell-Isc Trend
290/20/12
290/20/15 EFF
Isc
290/20/18
290/20/20
Poly. (EFF)
400/18 NCell-Isc Trend
290/20/23
400/18/12
Poly. (Isc)
400/18/15 EFF
290/20 Rs-FF Trend
290/20/12
290/20/15 FF
Rs
290/20/18
290/20/20
Poly. (FF)
Isc
400/18/18
400/18/20
Poly. (EFF)
400/18/23
Poly. (Isc)
400/18 Rs-FF Trend
290/20/23
Poly. (Rs)
400/18/12
400/18/15 FF
Rs
400/18/18
400/18/20
Poly. (FF)
400/18/23
Poly. (Rs)
Fig. 7. Electrical trends with different mesh and EOM thickness
4. Conclusions Before changing the screen design, especially when going for finer line printing, it is important to understand the relative Rcontact and Rline of the pastes. Then, choosing the right screen that suits the emitter and the printability of the paste are the keys to optimising the cell efficiency and cost. With SOL9600, minimum paste consumption savings is 10% relative to reference and can save even further with reducing finger opening and EOM.
90
Z. Fortu and L.X. Song / Energy Procedia 33 (2013) 84 – 90
Between meshes, 400/18 mesh will always have advantage over 290/20 mesh in terms of cost with minimum changes in efficiencies. Trade off in FF and Isc happens in choosing between 290/20 and 400/18 meshes. Figure 5 above, can be used as a guide in choosing the correct EOM for a specific finger opening. To balance resistance losses, additional finger lines may need to be added. It is not always true that higher EOM will give higher finger or better aspect ratio. At finer finger, 40 um for example, the trend in aspect ratio is inversely proportional to EOM thickness. If the EOM is too high and the finger opening is narrow, the paste transfer will be affected.
Acknowledgements Thanks to the entire Heraeus Team, for giving their expertise and support for the experiment. Thanks to Brave C&H for providing the screens required for this experiment. References [1]
Top Contact Design; http://www.pveducation.org/pvcdrom/design