By Richard Hendel

For industrial thin-film cell production, the laser is virtually an indispensable tool. Selective ablation and structuring of thin films cannot be performed with more flexibility and precision in a comparatively clean and simple process. However, in crystalline cell manufacturing, laser technology is only considered if it helps to reduce cost/Wp. This can be due to a simplification of the production process as well as to improved solar cell efficiency. Furthermore, in order to enter industrial mass production, the laser process has to be fast enough to meet cycle times of 1 to 2 seconds.

Laser Process Technology in Photovoltaic Production

Compared to mechanical machining and chemical treatment, the laser offers perfect process control, the process is cost-efficient and gentle to the material. To preserve and develop the strengths of this technology, laser manufacturers like ROFIN are putting great efforts in the further development of laser sources and laser process technology.

The fundamental process of most laser structuring applications on solar cells is the direct laser-induced vaporization and melt ejection by nanosecond laser pulses. The structuring processes, which are already used in many industrial pilot- and volume productions, offer big advantages due to their high speeds (single-pass up to 5,000 mm/s) with smallest ablation diameters (10-100 microns). More and more cutting and drilling applications are using the same principle in a multi-pass process in order to maximize flexibility and processing speed and to minimize the amount of heat load to the wafer.

Diode-Pumped Solid-State Lasers Are the Standard Tool

The dominating sources used for PV-related processes are solid state lasers. This is due to the demand of simultaneously delivering high power, top beam quality, and maximum repetition rates, all of which are needed to realize high processing speeds together with high resolution. Taking a closer look, there are currently four concepts competing with each other. But none of them is capable of occupying the solar market alone. The diversity of various processes leads to the necessity to constantly develop tailored laser solutions.

Up to now, the laser equipment of research institutes as well as solar cell producers mainly consists of diode-sidepumped and diode-endpumped solid state rod lasers, containing a neodymium-doped rod made of YAG or Vanadate. A tried, tested and field-proven technology which is constantly being developed further. Diode-endpumped systems score with high repetition rates, excellent p-to-p stability and small footprints. Whereas side-pumped sources offer high average power and high peak power. But recently, two other geometries of the laser active medium have been challenging the established concepts: fiber and disc.

Disc, Fiber and Femtosecond Laser to Come

The disc laser offers some exceptional properties which are particularly interesting for certain c-Si applications. As a result of the comparatively low volume of the active laser medium, diode pumping takes place via a parabolic mirror, leading to a multi-pass reflection. The laser crystal is water-cooled from the rear side, causing only a one-dimensional temperature gradient. Therefore, compared to a rod, a disc laser offers significantly higher beam quality or, in other words, much higher average power of best beam quality.

Compared to bulk lasers, fiber lasers feature higher wall-plug efficiency, rugged and compact design and nearly maintenance-free operation. Fiber lasers are suitable for generating high average powers with good beam quality. On the other hand, due to nonlinearities there is a lower potential for high pulse energies and peak powers. Nevertheless this young technology did successfully enter the photovoltaics market in replacing diode side pumped lasers in edge isolation applications.

Femtosecond lasers process materials faster than energy can diffuse within the atomic lattice. Therefore, no heat is transferred to surrounding material which eliminates any recast and burr. “old material processing”with ultrafast fs and ps lasers shows enormous potential in lab-scale applications. However, for mass production applications investment costs have to drop substantially.

Laser Applications on Crystalline Solar Cells

Two types of crystalline silicon are used in the photovoltaics industry. The first is monocrystalline, produced by slicing wafers from a high-purity single crystal boule. The second is multicrystalline silicon, made by sawing a cast block of silicon first into bars and then wafers. Laser processing has become a key technology for the industrial production of crystalline silicon solar cells reaching higher conversion efficiencies. Enhancements of the current solar cell technology are achieved by using advanced approaches like selective emitter structures or Emitter Wrap-Through (MWT/EWT), Laser Fired Contacts (LFC) or the Interdigitated Back Contact (IBC). At present, mono- and polycrystalline solar cells are the dominant technology in the commercial production of solar cells, accounting for about 80% of the solar cell market.

Edge Isolation, Grooving

The decisive factor for solar cell performance is the minimization of recombination possibilities. In order to obtain high efficiency, front and rear side must be electrically isolated on the edges. The separation of p-type layers is done by cutting trenches with qs-Nd:YAG or qs-Nd:Vanadate lasers. High power density is necessary in order to effectively eject the melt out of the kerf and to avoid re-deposition of the molten material. Compared to plasma etching, the benefit in production of laser scribing is higher yield due to less breakage in handling processes, and improved material flow by inline processing. Apart from that, there is no need for costly etching gases and their disposal. Compared to wet chemical edge isolation the laser process is much less costly in capital and running costs.

Drilling of Wafers

Rear side contacted solar cells eliminate the otherwise necessary front side strip lines and in this way enhance the solar-active surface and thus cell efficiency. What is more, the entire interconnection of solar cells into modules can be realized without any connections from the front to the rear side: packing density increases whereas costs drop. With the MWT method (Metal Wrap Through), the necessary soldering lines which are required for the interconnection within the module, are positioned on the rear side of the solar cell. For this purpose, 25-50 holes with a diameter of 200-500 μm are drilled in a grid pattern on one solar cell and then filled with conductive material. If, however, the entire contacting of the emitter layer is located on the rear, approx. 15,000 holes with a diameter of 60-70 μm are drilled (Emitter Wrap Through method - EWT). For both techniques, a q-switched disc laser like ROFIN’s StarDisc, operates with up to now unmatched throughput rates: up to 5,000 holes per second with the EWT method and 25-50 holes per second with the MWT method. The q-switched beam source provides both high power in TEM00 fundamental mode and ideal pulse widths for these applications. The combination of high pulse peak power together with material adjusted pulse duration provides for extremely short processing times per solar cell in the range of only few seconds. These are optimal preconditions for inline integration of rear side contacting processes into the whole production process.

Traceability Marking

When it comes to encoding silicon wafers and solar cells with the laser, demands on the marking result are high. Wafer marking in solar cell manufacturing is the key to traceability and improved manufacturing processes. As the wafer backside is entirely covered in a later production step only front-side markings are feasible. DataMatrix codes may be used on solar active areas but they deteriorate in the multitude of coating and etching processes. Thus in small batch R&D applications plain text markings on the wafer edges are preferred. The new ECCI code could offer a potential solution.

Unlike ECC DataMatrix codes, the ECCI code uses differential decoding and majority decisions. Thus the code is highly resistant against changing contrasts, often found on in-homogenous surfaces like polycrystalline solar cells. As long as anything is visible decoding works--even in case of contrast inversions within the marking. Additionally, ECCI code redundancy is adjustable. Markings which might be subject to damage or heavy wear and tear, can use a highly redundant coding.

Cutting and Scribing of Wafers

Fast cutting of mono and polycrystalline silicon wafers can be conducted with very high precision and low heat input by using the same ablation process as for edge isolation and drilling. In the past, flash-lamp pumped Nd:YAG lasers were used to melt cut silicon in a single pass with a coaxial gas jet. Due to rapid cooling of the melt layer at the cut edge, micro cracks were formed. New approaches indicate that a multi-pass cutting process without assist gas gives a better surface quality at the edge. With a qs-disc laser, users can expect typical cutting speeds of up to 200 mm/s for a wafer thickness around 0.2 mm. Specially for concentrator cells and recycling of chipped or broken cells, silicon wafers are not cut completely, but scribed to a depth of 40-60% of the cross section. To separate the wafer, a subsequent breaking, either manual or fully-automated, is required. Typical scribing speeds are in the range of 300-700 mm/s.

Laser Assisted Selective Doping

Selective emitters promise efficiency gains of about one percentage point (up to 17% for multicrystalline and 18.5% for monocrystalline cells). In the past, there has always been a trade-off between the desired heavy doping of the n-type silicon layer underneath the metallized contact regions and light doping between the contact fingers. Heavy doping achieves low contact resistance and good lateral conductivity whereas light doping is necessary for limited recombination and good response to blue light. The solution is selective doping of emitters and most new selective emitter concepts for mass production scale rely on laser material processing.

One approach is selective doping through laser ablation. Previous to dopant diffusion, it generates a dielectric masking layer, which is then selectively opened by lasers in the later contact areas. The laser ablates the dielectric layer in lines slightly wider than the contact fingers which are screen-printed subsequently. Thus the following diffusion step creates different dopant concentrations on masked and non-masked areas. Direct selective laser doping is a promising concept for realizing selective emitters as well. As lasers generate precisely controlled, localized heat input, they offer optimum prerequisites for a selective doping process. The easiest solution, uses the Phosphosilicate Glass (PSG) layer, already grown on top of the emitter during conventional dopant diffusion, as doping source. Other approaches apply a phosphorous doping source on top of the dielectric layer just before front contact forming. A laser melts the silicon lying beneath, incorporates the phosphorus dopants into the molten silicon and removes the dielectric layer thereby exposing the silicon surface for subsequent self-aligned metal contact formation.

Laser Applications on Thin-film Solar Cells

As crystalline silicon wafers make up 40-50% of the cost of a finished module there is substantial interest in finding cheaper materials to make solar cells. The selected materials are strong light absorbers and only need to be about one micron thick, so materials costs are significantly reduced. Most common are amorphous silicon or polycrystalline materials like cadmium telluride and copper indium diselenide. The thin-film semiconductor layers are deposited on to either coated glass, stainless steel sheets or plastic films. A transparent conducting oxide layer forms the front electrical contact of the cell, and a metal layer forms the rear contact. With more than 20 years, thin-film technologies have taken quite a long time to get from the stage of promising research to the first manufacturing plants producing early products. This emerging technology has yet to make significant inroads into the dominant position held by the relatively mature mono- and polycrystalline technology. But unlike c-Si production, there is still some fundamental cost cutting potential, like the continuous roll-to-roll production of flexible solar cells, which is one of the most promising approaches. For this technology ROFIN offers a new, highly customizable roll-to-roll system, which handles step and repeat processes as well as on-the-fly applications.

Thin-film Scribing

Thin-film solar cells are produced through a sequence of vapor deposition and scribing processes. The integrated circuits are generated between the different deposition steps by selective ablation of single layers to achieve electrical isolation. Exceptional beam quality is crucial for laser scribing processes on thin-film cells in order to maintain small scribe widths and constant depths. A large depth of focus, the result of high beam quality, enables the user to work through various material irregularities such as imperfect flatness and thickness of large glass panels. Pulse-to-pulse stability is key in maintaining elevated repetition rates during the high processing speeds. Lasers with best beam quality (TEM00) with very high repetition rates of up to 200 kHz are used for ablating 20 to 50 micron wide paths with scribing speeds of up to 2,000 mm/s without damaging the substrate or the layers underneath. This is the reason why Nd:Vanadat lasers with short ns-pulse widths (max. 100 ns) are the standard laser type for this kind of application. The ideal wavelength for the various processes depends on the composition of the individual layers. Fundamental (1064 nm) and wavelengths with double frequency (532 nm) are commonly used in the production of a-Si, CIS and CdS/CdTe solar cells.

The particularities of solar cell manufacturing ask for laser beam sources which are consistently optimized for this application purpose, like ROFIN’s PowerLine SL. A cost-effective and compact laser source, which was specifically designed for scribing of photovoltaic and other electric thin films. The pump diode and the RF generator are located in the compact laser head. This simplifies integration and makes great distances between the location of use and the supply unit possible as required. The applications include TCO/ITO/AZO layers on glass or flexible substrates, active layers on thin-film modules like a-Si/μ-Si, CdTe and CI(G)S, and the back contact layers like Al, Ag, Mo and combinations of such. The refined temperature management system maintains lasting long-term stability and performance.

Edge Deletion

For electrical isolation and hermetic sealing of the module, the complete removal of all layers from the edges of fully processed thin-film solar cells on glass substrates is required. In order to meet production requirements of a typical 40 MW p.a. manufacturing plant, removal rates have to be in the range of 10-50 cm2/s. Here, the laser challenges conventional techniques, like sand blasting and grinding. Since standard TEM00-lasers (like Nd:Vanadate lasers used for scribing) do not provide sufficient ablation rates for this application, especially developed high-power qs-lasers are applied.

ROFIN’s laser range for edge deletion applications offers laser sources output powers from 200 to 1000 Watt which can be equipped with 400, 600 or 800 μm optical fibers. Typical ablation widths are between 0.7 and 1.5 mm at processing speeds of up to 4,000 mm/s. As the nature of the process is a shot by shot application a square spot geometry has the advantage, that the overlapping of several pulses is constant across the processing direction. Square homogeneous spots are offered by square fibers which have been introduced by ROFIN lately. Processing with square homogeneous spots from square fibers allows the optimization of removal applications. Overlapping of pulses is realized by a displacement of subsequent pulses less then one spot width. Square pulses have the advantage that the overlap is constant from the center of the spot to the edge. Efficient and homogeneous removal results are achieved easily.

For laser manufacturers, photovoltaics is a promising and demanding market at the same time. New high-efficiency concepts as well as optimizing technologies for mass production require tight cooperation between research institutes, production line suppliers and solar cell manufacturers. With selective emitters, a well-known approach, already tested in high-efficiency cells, finds its way to mass production. It won’t be the last. The laser is a key-technology for a magnitude of tried and tested high-efficiency concepts. More than that, selective doping and selective ablation of dielectric layers with the laser generally opens a broad field for new efficiency improvement approaches.  

Richard Hendel is Sales Manager Photovoltaic of ROFIN-BAASEL Lasertech (http://www.rofin.com/).