Industrially feasible multi-crystalline metal wrap through (MWT) silicon solar cells exceeding 16% efficiency

ARTICLE IN PRESS Solar Energy Materials & Solar Cells 93 (2009) 1051–1055

Contents lists available at ScienceDirect

Solar Energy Materials & Solar Cells journal homepage: www.elsevier.com/locate/solmat

Industrially feasible multi-crystalline metal wrap through (MWT) silicon solar cells exceeding 16% efficiency Florian Clement , Michael Menkoe, Tim Kubera, Christian Harmel, Rene Hoenig, Winfried Wolke, Harry Wirth, Daniel Biro, Ralf Preu Fraunhofer Institute for Solar Energy Systems (ISE), Heidenhofstr. 2, D-79110 Freiburg, Germany

a r t i c l e in fo

abstract

Article history: Received 3 January 2008 Received in revised form 17 September 2008 Accepted 27 November 2008 Available online 20 January 2009

On the way to higher efficiencies, back contact solar cells seem to be a promising alternative to the conventional screen-printed solar cells. Especially, the metal wrap through (MWT) solar cell concept with only two additional process steps is appropriate for a fast transfer to industry. Hence, an industrially feasible process based on a new contact design was developed and tested at the pilot-line of the Photovoltaic Technology Evaluations Center (PV-TEC). A maximum cell efficiency of 16% is achieved. Compared with conventionally processed cells made of the same mc Si-block, an efficiency gain of 0.5% absolute is observed. Due to a cell interconnection on the back the serial resistance losses in the tabs decrease. Therefore, a fill factor of almost 77% and an efficiency of 15% for a MWT module prototype (16 MWT cells) is reached. & 2008 Elsevier B.V. All rights reserved.

Keywords: Solar cell Back contact Multi-crystalline MWT Module technology

1. Introduction The metal wrap through (MWT) solar cell [2] is a back contact solar cell based only on industrially feasible technologies. Due to a design and process sequence, which is very similar to that of conventional screen-printed solar cells, the MWT solar cell is a promising alternative for industrial mass production. Only two additional process steps are required: via drilling for interconnection between front and back side and metallisation of the holes. The contact isolation on the back can be done in the same step as the edge isolation. Furthermore, the MWT cell concept offers two main advantages. First, a gain in active cell area and thus an increase in short circuit current is achieved due to less shadowing caused by the absence of the bus bars on the front. Second, having the n- and pcontact on the back simplifies the interconnection of the cells and allows the use of new tabbing technologies with the objective of series resistance reduction [3]. Moreover, the optical appearance of the module is more uniform due to the absence of tabbing material on the front. Particularly recent R&D results in cell and module production show a very high potential for the MWT cell concept [4,5]. Hence, the transfer of the MWT solar cell process to industry is still

 Corresponding author. Tel.: +49 761 4588 5479; fax: +49 761 4588 9250.

E-mail addresses: fl[email protected], fl[email protected] (F. Clement). 0927-0248/$ - see front matter & 2008 Elsevier B.V. All rights reserved. doi:10.1016/j.solmat.2008.11.059

ongoing. First cell and module producers successfully implemented the MWT solar cell concept in their production lines [6].

2. Experimental procedure 2.1. Cell production and design In this work an industrial feasible process flow for MWT solar cells was developed and realised. Therefore, several batches of multi-crystalline silicon (mc-Si) MWT solar cells (125  125 mm2) were produced at the pilot-line of the Photovoltaic Technology Evaluation Center (PV-TEC) [1]. For comparison conventionally processed reference cells of the same multicrystalline silicon block (neighbouring wafers) were fabricated in each batch. In Fig. 1 the process flow of the MWT solar cell process is shown. First, a certain amount of vias was drilled by a laser system. Usually, about 50 vias with a diameter of approx. 100 mm are necessary for a 125  125 mm2 wafer. After the via drilling the saw and laser damage was removed during the acidic texturing process. The following emitter diffusion (sheet resistance: approx. 50 O/sq) was carried out in a POCl3 tube furnace. After the phosphorus silicate glass removal a SiNx layer (anti-reflection coating) was deposited on the front by the use of plasma enhanced chemical vapour deposition or Sputter technology. The second specific MWT process step is the metallisation of the holes and the n-contacts (bus bars) on the rear. Therefore, a modified

ARTICLE IN PRESS 1052

F. Clement et al. / Solar Energy Materials & Solar Cells 93 (2009) 1051–1055

Fig. 3. Picture of the newly developed MWT measurement chuck for 125  125 mm2 cells. All contact pins are located beneath the cell and the glass fixes the cell during the measurement.

Fig. 1. Process flow for MWT solar cells. All additional or modified process steps compared with the conventional solar cell process are in italics.

MWT prototype module consisting of 16 MWT solar cells was fabricated. The MWT cells were arranged in four strings, so that each string consists of four cells. To optimise the module fill factor, a tab design with the objective of minimisation of the serial resistance losses was chosen. The cells were soldered with standard equipment, but only from the back. Afterwards a conventional lamination step was carried out. For comparison two mini-modules (one cell per module) were fabricated: One MWT mini-module and one reference minimodule produced with standard module technology. 2.3. I–V measurement set up

Fig. 2. Picture of a MWT solar cell (125  125  0.24 mm3) with three rows of holes and 48 fingers. The positions of the holes are marked.

screen-printing step with a special silver paste was introduced in the process. The other screen-printing steps (back and front) were done afterwards. Next, a contact firing process was performed in a fast firing furnace. Last, the contact and edge isolation is required. In principle, both isolation steps can be done at the same time, for example by a laser system working simultaneously from both sides. However, in this work, two separate laser process steps were used. In Fig. 2 the front of a MWT solar cell processed at the PV-TEC is shown. A design with three rows of vias is chosen because the serial resistance losses in the fingers are smaller and the shading is not significantly higher than for two rows. The positions of the holes are marked. Altogether, a mechanical yield 496% was achieved in the MWT cell process, a very good value for pilot-line production.

For the I–V measurement of MWT solar cells a new measurement chuck was developed in co-operation with the company AESCUSOFT [7]. The measurement chuck is shown in Fig. 3. It allows measurements for 125  125 mm2 MWT cells with two and three n-contact bus bars on the back. Due to the use of pairs of current and voltage contact pins a good determination of the fill factor is guaranteed. The adjustment of the cell during the measurement is arranged by a special glass which fixes the cell. Therefore, a mismatch factor due to transmission and reflection effects of the glass was determined by transmission measurements. The temperature during the measurement is logged and controlled by external cooling. Nevertheless, the measurement of the open circuit voltage is done immediately after starting the illumination, so that a cell temperature of 2570.5 1C is assured. The reference cells and the mini-modules were measured with standard equipment. All cell and mini-module I–V measurements were performed under steady-state one-sun illumination (AM 1.5 g). The MWT module prototype was measured at the Calibration Laboratory of the Fraunhofer Institute for Solar Energy Systems (ISE) by flash light one-sun illumination (AM 1.5 g). The measurement uncertainty is 73% relative for all measurements.

3. Optimisation steps First MWT cells processed in 2006 (see Fig. 7) were limited by losses due to shunting and high values of the series resistance. Hence, a few optimisation steps were carried out. 3.1. Via contact optimisation

2.2. Module production and design To show the capability of the MWT cell concept for module assembly, a new tabbing technology was developed. Hence, a

For the optimisation of the via contact, a reliable screenprinting step of the back side silver paste which enables low values for the via resistance is necessary. In Fig. 4 two crosssections of vias are shown. In case of the high via resistance (Fig. 4,

ARTICLE IN PRESS F. Clement et al. / Solar Energy Materials & Solar Cells 93 (2009) 1051–1055

Fig. 4. Cross-sections of two vias of MWT cells: Left: Cell with a high via resistance. A cavity appears and front side paste covers a part of the via walls. Right: Cell with a low via resistance. The via walls are completely covered with back side silver paste.

left side) the back side silver paste does not completely cover the via walls and thus front side silver paste covers a part of the via walls. Therefore, a cavity appears between the different pastes and the via resistance increases. Hence, the screen-printing process was modified. The focus of this modification was a complete covering of the via walls with back side silver paste after the printing step (Fig. 4, right side). Due to the modified printing process the value of the via resistance was reduced from over 70 mO to below 7 mO. Furthermore, the back side silver paste has to meet two conditions: a satisfying via (wall) metallisation and the paste should have a low shunting behaviour, especially in the regions (via and back n-contact region) where no anti-reflection coating is realised. The fact that almost no ohmic shunting appears if the via walls are completely covered with back side silver paste, confirms that the used back side silver paste meets both conditions pretty well. Nevertheless, the MWT cells show a low non-linear shunting behaviour, which could be caused by the back side silver paste [8]. 3.2. Optimisation of the back contact structure For the optimisation of the back contact structure two problems were focussed. First, the influence of the width w of the back n-contact region on the series resistance was calculated. Therefore, the series resistance RS,base,lat due to the lateral conductivity of the base was determined: RS;base;lat ¼ XðwÞrbase

0:5w2 3d

d is here the wafer thickness and rbase the base resistance. The factor X(w) (in%) is necessary because only a certain amount of the back is covered with n-contact regions and thus causes such series resistance losses. So, X(w) is the covered fraction of the back and strongly depends on w. The results of the calculation are presented in Fig. 5. The series resistance decreases for smaller n-contact regions and for a lower base resistance. But, soldering for interconnection of the cells and a reliable contact isolation must be possible. So, the optimum width of the back n-contact region w is about 3 mm. Second, the influence of the solder pad regions on the cell performance was examined. Therefore, spectrally resolved light beam induced current (SR-LBIC) measurements were performed. Thereby the distribution of the effective diffusion length Leff is

1053

Fig. 5. Calculation of the influence on the series resistance due to the lateral conductivity of the base in the back n-contact regions. The calculation was performed for different width w of the back n-contact regions and is based on a design with three back n-contact regions.

determined. In Fig. 6 two measurements are compared. The area of the solder pad is varied (the width w of the back n-contact region and the finger distance is also varied). A decrease of Leff due to the p-contact solder pad is clearly shown in both cases. An explanation for the decrease of Leff is the small amount of aluminium in the solder pad paste, which causes almost no back surface field in this region and thus a minor passivation. However, in case of the small solder pads (Fig. 6, right side) the mean diffusion length over the whole cell area is higher. Furthermore, the high value of Leff in the back n-contact regions is probably caused by a double-side collection of charge carriers in this region. 3.3. Optimisation of the isolation process For the edge and contact isolation a UV laser system was used. Different process parameters were tested and the ohmic shunt resistance RP was determined with dark I–V and Suns-VOC measurements [9]. With optimised process parameters RP values over 5 kO cm2 were realised. Nevertheless, the MWT cells show a low non-linear shunting behaviour, which could be caused by the laser trenches [8].

4. Results 4.1. Cell results The mean and maximum efficiencies as well as the standard deviation of different batches of MWT cells are shown in Fig. 7. The efficiency gain compared with reference cells is also shown for every batch. If the first (may06 and sep06) and latest (jun07 and nov07) results are compared, a clear efficiency improvement due to the optimisation steps is observed. The decrease of the efficiency between the batch jun07 and nov07 could be explained by the use of different mc material and a different front side paste. However, the increase of the efficiency gain shows that a further improvement of the MWT cells is achieved in the batch nov07. Hence, a maximum efficiency gain of 0.5% absolute is reached on cell level (batch nov07). In Table 1 the best MWT cell results are presented (batch jun07). In this batch about 50 MWT were processed. A very good homogeneity over the whole batch is reached. Hence, the applied MWT process seems to be very stable and reliable. The maximum

ARTICLE IN PRESS 1054

F. Clement et al. / Solar Energy Materials & Solar Cells 93 (2009) 1051–1055

Fig. 6. Spectrally resolved light beam induced current (SR-LBIC) measurements of two MWT cells. The distribution of the effective diffusion length Leff is presented. The bad influence on Leff due to the solder pad regions is shown. Left: MWT cell with large solder pads on the back. Centre: Line scan in the region of a large solder pad. Right: MWT cell with small solder pads on the back.

Table 2 Results of I–V measurement of the MWT and the reference mini-module (1 cell) as well as of the MWT module prototype (16 MWT cells). Mini-module

A (cm2) VOC (V) ISC (A) FF (%) Z (%)

Prototype

MWT

Ref.

MWT

156.3 0.62 5.19 74.8 15.4

156.3 0.62 5.01 74.6 14.8

2627 9.84 5.19 76.6 14.9

All cells were fabricated within the batch jun07.

Fig. 7. Development of the maximum and mean efficiency of MWT cells during the last 2 years. The gain compared with reference cells made out of the same mc Siblock is also presented in the upper part of the figure.

Table 1 Results of I–V measurement of the best mc-Si MWT solar cells (batch jun07).

Mean Best

VOC (V)

JSC (mA/cm2)

FF (%)

Z (%)

61172 611

33.470.1 33.5

77.570.4 78.0

15.870.1 16.0

efficiency of 16% is also a very good result for a conventionally screen-printed mc-Si solar cell. To reach efficiencies over 16% further cell improvement is necessary. Therefore, two industrially feasible additional process steps were tested. First, an annealing step (350 1C, 10 min, forming gas) after the contact firing step was carried out and thus an efficiency increase up to 0.3% absolute was reached. Second, a light induced silver plating step as second front side metallisation step was tested, which leads to a fill factor gain of about 1% absolute.

(16 MWT cells of batch jun07) were characterised (see Table 2). The measurements of the mini-modules were done with a mask, so that only the active cell area is illuminated and thus used for the calculation of the module efficiency. For the calculation of the module efficiency of the prototype the illuminated area A (without frame) was used. The current gain for the MWT mini-module of about 4% relative verifies the cell measurements. It can be explained by less shading on the front and additionally by double-side collection of charge carriers in the n-contact regions on the back. The low fill factors of both mini-modules are caused by large serial resistance losses in the tabs, which connect the external contacts with the tabbed solar cell in the middle of the mini-module. However, these losses in both cases are the same. Hence, an efficiency gain of 0.6% absolute is achieved for the MWT mini-module. Compared with cell measurements of the same cells (gain of 0.4%) the gain is increased. This can be explained by less series resistance losses in the tabs in case of the new MWT tabbing technique. The achieved efficiency of the module prototype (about 15%) is also a very good value compared with currently available multicrystalline modules [10]. Furthermore, the fill factor loss between cell and module is only about 1% absolute, a clear evidence that the use of the new tabbing technology was very successful.

5. Conclusion 4.2. Module results To verify the good results two mini-modules (one MWT and one reference mini-module) and the MWT module prototype

A reliable pilot-line process for MWT solar cells based on a new contact design was successfully developed at the PV-TEC. Therefore, a few optimisation steps were carried out. Especially the via contact, the back side structure and the isolation process were

ARTICLE IN PRESS F. Clement et al. / Solar Energy Materials & Solar Cells 93 (2009) 1051–1055

optimised. Furthermore, a new measurement chuck for MWT cells was developed. Moreover, first module prototypes were produced with a new tabbing technology for MWT cells. Cell efficiencies up to 16% and a module efficiency of about 15% were reached for MWT cells. An efficiency gain up to 0.6% absolute compared with reference cells (modules) of the same mc Si-block was achieved. The gain can be explained mainly by less shading on the front and less series resistance losses in the module due to the new tabbing technology. Therefore, a very small fill factor loss between cell and module of only about 1% absolute was realised.

Acknowledgements The authors would like to thank the following persons for technical support: E. Scha¨ffer, A. Leimenstoll, A. Herbolzheimer, H. Reitenbach, J. Geilker, J. Hohl-Ebeniger, M. Kasemann, A.M. Hassan, A. Drews, W. Kwapil, N. Mingirulli, K. Kordelos and the whole PV-TEC team. The German Ministry for the Environment, Nature Conservation and Nuclear Safety (BMU) is gratefully acknowledged for financially supporting this work within the project ‘‘PV-TEC’’ (0329984). Applied Materials is gratefully acknowledged for financially supporting this work within the project ‘‘WEST’’. Du Pont is gratefully acknowledged for the supply of the silver pastes.

1055

References [1] D. Biro, et al., PV-Tec: photovoltaic technology evaluation center—design and implementation of a production research unit, in: Proceedings of the 21st European Photovoltaic Solar Energy Conference, Dresden, Germany, 2006, p. 621. [2] E. van Kerschaver, et al., A novel silicon solar cell structure with both external polarity contacts on the back surface, in: Proceedings of the Second World Conference on Photovoltaic Energy Conversion, Vienna, Austria, 1998, p. 1479. [3] J.H. Bultman, et al., Fast and easy single step module assembly for backcontacted C-Si solar cells with conductive adhesives, Osaka, Japan, Osaka, Japan, 2003, p. 978. [4] H. Knauss, et al., Industrially applicable metallisation wrap through solar cell process resulting in efficiencies exceeding 16 %, in: Proceedings of the 21st European Photovoltaic Solar Energy Conference, Dresden, Germany, 2006, p. 1192. [5] A.W. Weeber, et al., 17% cell efficiencies on large back-contacted multicrystalline silicon solar cells, in: Proceedings of the 21st European Photovoltaic Solar Energy Conference, Dresden, Germany, 2006, p. 605. [6] E. Van Kerschaver, et al., Back-contact solar cells: a review, Progress in Photovoltaics: Research and Applications 14 (2006) 107–123. [7] Aescusoft GmbH, Orschweiertstrasse 39a, 77955 Ettenheim, Germany, email: [email protected]. [8] F. Clement, et al., Processing and comprehensive characterisation of screenprinted mc-si metal wrap through (mwt) solar cells, in: Proceedings of the 22nd European Photovoltaic Solar Energy Conference Milan, Italy, 2007, p. 1399. [9] R.A. Sinton, et al., A quasi-steady-state open-circuit voltage method for solar cell characterization, in: Proceedings of the 16th European Photovoltaic Solar Energy Conference, Glasgow, UK, 2000. p. 1152. [10] Photon 02/2007, Solar Verlag GmbH, Germany.