Screen Print Diffusions for Buried Contact Solar Cells

Neil Bennett, Matt Edwards, Jeff Cotter
Research Theme: Energy, Resources & The Environment

Introduction

Background

• Solar Cells produce clean, renewable energy but are expensive and further research is required to increase efficiency and reduce costs to make them competitive.
• Screen printing of dopant sources offers a number of possible benefits including the ability to perform selective diffusions cheaply and easily and simultaneously diffuse boron and phosphorus.

Objective

• To devise a method to dry and fire phosphorous dopant paste on silicon wafers while maintaining high electrical quality. Leading to a phosphorus/boron co-firing regime that will simultaneously diffuse localised regions of the wafer. (See Figure 1)

P Diffusion B Diffusion

Figure 1: Left, the Interdigitated Backside Buried Contact cell (UNSW, 2004) could benefit from a simplified production process that utilises simultaneous local diffusions. Right, an example diffusion pattern is shown.

Experiments and Results

1. Establishing Drying and Firing Routine

• Phosphorus paste was printed and dried on both sides of the wafer.
• The effect on the sheet resistance from firing the wafers at varying times and temperatures was observed. (Figure 2)

Figure 2: Scatter Plot of Sheet Resistance Results from Diffused Wafers

• Position in the furnace affected the sheet resistance, wafers on the outside of the tray had higher sheet resistances then the rest of the batch. Possibly due to the interior wafers receiving extra diffusion through the gas-phase.
• Manual printing process may deposit unequal amounts of film causing inconsistencies. Using a fully automatic screen printer should alleviate this problem.

2. Observations of Printed Film

• Printing accurately in a defined area is necessary for selective diffusions.
• The paste was printed in a finger pattern and observed with a microscope at various stages of the process. (Figures 3 – 6)

Figure 3: Screen Finger (5x)

Figure 4: Printed Finger (5x)

Figure 5: Dried Finger (5x)

Figure 6: Diffused, Oxidised Finger (5x)

• Significant changes in finger width were observed during the processes (See Table 1)

Table 1: Width of Screen Printed Fingers

• Observations of oxidised diffused wafer showed regions of heavy and light diffusion as well as cracking of the film (Figures 7-9).

Figure 7: “Creek Bed” effect caused by cracking of paste (10x)

Figure 8: Uneven diffusion due to impurities (5x)

Figure 9: No oxidation due to impurities (20x)

• Defects in printed film could be due to paste drying dynamics, printing defects or impurities.
• Reduce printed area with metal stencil printing instead of wire-mesh screens.
• Overlapping areas of P and B in a co-fired cell could cancel expansion effects.

3. Optimisation to Preserve Electrical Quality of Wafers

• The electrical quality of the cells was assessed by measuring the bulk lifetime of both p-type and n-type diffused wafers.
• Low bulk lifetimes were observed, possibly caused by contamination from the process (especially from contact with metal) and carbon entrapment occurring during the firing of the paste.
• New procedures were developed including a clean metal free process and a longer dry designed to prevent carbon entrapment. (See Table 2 and 3)

Table 2: Max Bulk Lifetime (p-type wafers)

Table 3: Max Bulk Lifetime (n-type wafers)

Conclusions

• Screen printing dopant pastes has potential to be a quick and cheap method for diffusions on silicon
• Bulk lifetimes are still inconsistent, but potential for improvement has been shown and further research is required.
• Phosphorus / Boron co-firing should be possible with current firing regime.

[Top of Page]