Selective Emitter

As one of the steps towards high efficient solar cell, Selective Emitter (SE) cells have been gradually implemented into production lines. The concept of this structure is to reduce emitter non-metal area doping concentration while keeping emitter area under metal contact high doping profile to ensure good ohmic contact of metallization. The purpose is to reduce short-wavelength light (e.g. blue light) generated minority charge recombination in emitter layer because the high-energy light has low penetration capability and majority amount gets absorbed in the front layer. Thus, the Internal Quantum Efficiency (IQE) is boosted on the short-wavelength range. SE is commonly used for front side of front-contact cells but similar concept in fact can be used on back side to enhance long-wavelength (e.g. red) IQE. Here are some examples of SE development: approaches and processes.

Fig. 1 Schmid's Etch-back SE Cell Process Flow
(Courtesy of Schmid)

1. Etch Back [1, 2]. This method utilizes traditional doping/diffusion approach with initial high doping level for entire emitter layer to start with. A layer of mask (e.g. wax) is then needed to cover area (metallization region) not to be etched and subsequently followed by an etching step (e.g. HF, HNO3) to remove some of top heavily doped layers to achieve low dopant concentration. In addition, Phosphorous Silica Glass (PSG) removal step is still required. A complete process flow to make such cell from Schmid is pictured in Figure 1. Schmid claimed >0.7% of absolute cell efficiency increase.

Fig. 2 Schematic of Laser Doping System (Courtesy of Synova)

2. Selective Doping. The principle of this idea is to selectively dope different areas to achieve lightly doped emitter area and heavily doped contact area. Based on doping method, the selective doping can be done by laser doping or LDSE [3] and ion implantation. For laser-assisted SE, a lightly doped region is formed by traditional POCl3 diffusion on entire surface and a heavily doped region is selectively obtained by laser doping with phosphoric acid H3PO4 without masking step but with PSG removal step. Brief description of laser doping system is shown in Fig. 2. For ion-implantation, the process requires mask for heavy-doping but no PSG removal step. Ion-implantation process seems not cost effective. But with its developed know-how and extensive application in IC industry, it could turn into less expensive process. Centrotherm is one of the companies currently using LDSE.

Fig. 3 Si Ink Process for SE (Courtesy of Innovalight)

3. Printed Doping. This method utilizes dopant carriers, which can be spin-coated, ink-jetted, or screen-printed, to print and form a layer with dopant source. The dopant layer then annealed to allow dopant atoms drive-in. The dopant precursor concentration, chemical structure, layer thickness, layer structure, and thermal profile can be engineered to achieve desired doping profile and depth. Based on chemistry of printable dopant source, several different inks are being developed or sampled in production lines, for example, Honeywell spin-on glass (SOG) and Innovalight (now DuPont) silicon ink [4-7]. Fig.3 depicts an SE application of silicon ink from Innovalight.  Regardless of ink types, efforts towards reduction of annealing temperature, prevention of dopant evaporation/loss, prevention of dopant cross-contamination, lowering impurity level, and, of course, lowering the cost, have been taken to fasten the commercialization of this technique. In particular, Innovalight's Si ink was claimed being used in the manufacturers' production lines (JA solar, Jinko, and Yingli etc.) although no press release so far has been made to confirm commercialization of its technology in shipped panels.  

Fig.4 Process Step Comparison of Selective Emitter
(Courtesy of Applied Materials)

4. New Approaches. To avoid using masks in Etch Back process, Fraunhofer is developing a new method which combines laser doping with etch back methods.

In summary, selective emitter is a way to further improve cell efficiency. There exist several different approaches to make it happen but cost needs to be further verified. No simple comparison for eff/cost for above processes can be made but AMAT has a picture showing step differences (Fig. 4). Nevertheless, more effort is undergoing for further optimization and cost reduction so that SE structure can be more extensively and effectively adopted into production lines.

References

[1] Thomas Lauermann et. al. "An inline selective emitter concept with high efficiencies at competitive process costs improved with inkject masking technology", 24th EU PVSEC (2009), pp.??

[2] Thomas Lauermann et. al. "The optimal choice of the doping levels in an inline selective emitter design for screen printed muticrystalline silicon solar cells", 24th EU PVSEC (2009), pp. ??

[3]  Finlay Colville "Laser-assisted selective emitter and the role of laser doping", PV International (2009), pp. 1-7.
[4] H. Antoniadis et. al. "All screen printed mass produced silicon ink selective emitter solar cells, IEEE (2010), pp. 001193-001196.
[5] M.L. Terry et. al. "All screen-printed 18% homogeneous emitter solar cells using high volume manufacturing equipment", IEEE (2010), pp. 003618-003622.
[6] D. Poplavskyy et. al. "Silicon ink selective emitter process: optimization of selectively diffused regions for short wavelength response", IEEE (2010), pp. 003615-003569.
[7] H. Antoniadis "High efficiency, low cost solar cells manufactured using 'silicon ink' on thin crystalline silicon wafer", Innovalight's DOE report (2011), 41 pages.