Optical and thermal modeling

In laser processing of materials, when the laser light is incident on the material surface (as shown in Fig. 3), a portion of the laser light is reflected, a portion is absorbed, and in some cases it may be partially transmitted. The amount of light reflection can be calculated using the optical constants n, k of the material, where n is the refractive index and k is related to absorption. For normal incidence, the surface reflection ( R ) and transmission ( T ) are given by ³ :

Figure 3.

Light reflection and refraction at the interface.

When light is incident at an angle, the polarization state of the incident light must be considered. For S-polarized light:

For P-polarized light:

For electrically conducting media the index will be complex, $\tilde n$ = n + ik, where k ₁ = 0 for the incident media and k ₂ = k and reflectance for normal incidence will be given by

From tabulated n and k values, surface reflectance can be calculated. For normal incidence on metals with conductivity σ and incident light frequency ω it can be approximated as ^{Reference Born and Wolf4} :

For Cu metal, σ = 6.7 × 10⁷ (Ω m)⁻¹ and ω = 7 × 10¹⁴ so R = 0.97. The light absorption is given by Beer's law: I(z) = I ₀e^−αz where absorption coefficient α = 4πk/λ, I ₀ is the incident light intensity and z is the material thickness. For laser processing, the calculation of surface reflectivity is important as it determines the amount of light coupled into the material.

The following equation describes the thermal behavior of the material under laser irradiation:

where T is the temperature, H is the enthalpy, k is the thermal conductivity of the material, and S is the energy deposited by the laser heating source. The source term can be written as ^{Reference Kannatey-Asibu2} :

where α is the light absorption coefficient, R is the surface reflectance, I ₀ is the laser intensity, r is the radial distance, a is the laser beam radius, and z is the light penetration depth. The solution of the nonlinear thermal equation requires numerical methods. There are commercial codes like COMSOL, ANSYS, Fluent, etc. that can provide the rise in temperature and its distribution. From the laser wave length, power, pulse duration and repetition rate, beam size, it is possible to calculate various parameters of interest.

Diode lasers

Diode lasers are based on direct band gap semiconductors like GaAs with a p–n junction, and output wavelengths are in the range of 785–1060 nm. A single junction can generate watts of optical power and by arranging them to form bars or arrays it is possible to achieve kilowatts of optical power. The electrical to optical power conversion efficiency is very high (>30%) and they are highly reliable with lifetimes of over 50,000 h. Though the beam quality of high-power diode lasers is relatively poor, they are available with fiber delivery. Diode lasers can be CW or pulsed with modulation rates of up to 50 kHz. ⁵

Fiber lasers

Fiber lasers are produced by coupling optical fiber and semiconductor lasers. ⁶ The beam from a diode laser is coupled into optical fibers containing rare earth doped ions to generate the light which remains inside the fiber. The fiber laser provides high beam quality (M ² < 1.5, where M ² is the beam quality factor, which compares the laser produced beam to an ideal Gaussian laser beam with M ² = 1), and high optical powers (over 20 kW) at relatively low cost. Fiber lasers are compact, need little maintenance, have high wall plug efficiency, and exhibit long operational life. They are extensively used in various applications such as material processing, scientific research, and medical applications, and can be easily integrated for robotics applications. With Q-switch technology, the fiber lasers can provide CW as well as pulsed operation with pulse width in the range of 1–400 ns. The peak power can be in megawatts (MW) and pulse energy over 50 mJ. Figure 4 shows the evolution of fiber laser technology. ^{Reference Jauregui, Limpert and Tunnermann7}

Figure 4.

Evolution of the average output power of nearly diffraction limited fiber lasers (emitting either a CW or ultrashort pulses) over the past 25 years. ^{Reference Jauregui, Limpert and Tunnermann7}

Optical beam delivery systems

For material processing, the laser beam needs to be focused on the object. The beam can be delivered in two ways: one by free space optics using mirrors and lenses and second by fiber optics. A fiber optic beam delivery is commonly used to provide a collimated beam which can be focused with an external lens on the object as shown in Fig. 5. Fiber beam delivery systems use flexible cables. Laser beams can be focused to 10–20 µm in size, allowing the generation of very high power densities. A laser beam is coupled into the single mode or multimode fiber and beam can be delivered away from the laser source. The output of the laser beam is collimated by a lens and then focused by another lens on the materials to be laser processed.

Figure 5.

(a) Free space delivery system and (b) fiber optic beam delivery system.

Introduction

While lasers are currently used by some manufacturers during the processing of crystalline silicon solar cells, the most widespread application of laser technology in the field of photovoltaics (PVs) is the scribing of thin-film solar cells to make large-area, series-connected PV modules (Section “Laser processing for thin film solar cells”). Most thin-film solar cell manufacturers utilize laser scribing for at least part of the interconnection process as the thin-film cells are patterned and series-connected into large-area monolithic modules.

Some laser processes for crystalline silicon solar cells, such as laser edge isolation (Section “Laser edge isolation”) and laser grooving for selective emitter formation (Section “Laser marking, grooving, and cutting”), were technically successful, but have generally been replaced by less expensive alternatives. A laser marking process is currently being utilized by some crystalline silicon solar cell manufacturers for single wafer identification to assist in quality control and process improvement (Section “Laser marking, grooving, and cutting”).

However, new laser technology is actively being investigated by a number of prominent research organizations for the processing of crystalline silicon solar cells. ^{Reference Bogue8–Reference Booth17} Lasers are being used to form both emitter and ohmic contacts for crystalline silicon solar cells using processes such as laser doping (Section “Laser doping”), laser-fired contacts (LFCs) (Section “Laser-fired metal contacts”), laser transfer processes (Section “Laser transfer”), and laser ablation of dielectric passivation layers (Section “Laser ablation of dielectric passivation layers”). In addition, several organizations are utilizing laser drilling of vias (holes) in silicon wafers for metal-wrap-through (MWT) and emitter-wrap-through (EWT) back-contact solar cells (Section “Laser drilling”). Also, lasers are being used to form low-reflectivity microtextured surfaces (Section “Light trapping by laser microtexture”) and to form thin crystalline silicon films by laser crystallization of amorphous silicon films (Section “Other laser processing: annealing, sintering, crystallization, defect repair”).

Laser edge isolation

The p–n junction in most crystalline silicon solar cells is formed by diffusing phosphorus into a p-type wafer at temperature of ∼850 °C using POCl₃ gas as a dopant source. This process usually results in the creation of a leakage path (or shunt) from the front surface emitter to the rear surface base contact due to diffusion of phosphorus into the edges of the wafer and some wrap-around of diffused phosphorus onto the rear surface. A pulsed laser can be used to scribe an isolation groove around the outer periphery of the top or rear surface (see Fig. 6), effectively removing the shunt path. ^{Reference Hauser, Hahn, Spiegel, Feist, Breitenstein, Rakotoniaina, Fath and Bucher18} This is usually accomplished using a Q-switched Nd:YAG or Nd:YVO₄ laser operating in the infrared wave length to scribe a narrow isolation groove or trench near the edge of the emitter on the top surface of the wafer as shown in Fig. 6(b). The isolation groove is about 10–30 μm deep and 40–80 μm wide and is placed as close to the edge of the wafer as possible to maximize the solar collection area. ^{Reference Dunsky19}

Figure 6.

Schematic of silicon solar cells with laser isolation grooves on (a) the rear surface and (b) on the front surface. ^{Reference Carlson10}

However, in recent years, many manufacturers are replacing laser edge isolation with a single-sided wet etching process since although the etching process requires the handling of hazardous materials, it appears to be more cost effective. ^{Reference Hofmann, Schmidt, Kohn, Rentsch, Glunz and Preu20}

Laser doping

The depth of p–n junctions in silicon solar cells is typically less than a micrometer, and laser-induced diffusion is ideally suited for forming shallow p–n junctions in silicon. Fairfield and Schwuttke ^{Reference Fairfield and Schwuttke21} were the first to demonstrate laser doping of silicon when they used a laser to create a p–n junction in silicon in 1968. In their work, a polished silicon wafer was coated with phosphorus and scanned with a pulsed laser beam.

In an example of more recent laser-doping work, ^{Reference Ventura, Schunck, Muller, Barthe, Vetrella, Pirozzi and Salza22} a spin-on-dopant solution is first spin coated on a silicon surface. The coated silicon surface is then exposed to laser illumination, and above a certain laser fluence or energy density, the silicon surface can be melted. The emitter dopant from the spin-on-dopant film will diffuse into the molten silicon to form the junction. The schematic of the process is shown in Fig. 7. The advantage of this approach is that the selective doping can be achieved with a direct laser write process without requiring photolithography. Also, in the laser doping process, only the silicon surface is heated, so it minimizes the incorporation of impurities and reduces the thermal budget. However, for large area applications of laser doping, the throughput needs to be optimized with larger beam size using high-power lasers or multiple laser beams.

Figure 7.

Laser doping process: (a) Step 1: laser pulse irradiation; (b) step 2: molten pool formation and dopant diffusion; (c) junction formation.

Laser-fired metal contacts

High-performance solar cells with efficiencies as high as 21.7% ^{Reference Schneiderlochner, Preu, Ludemann and Glunz23} have been fabricated by scientists at Fraunhofer ISE using a laser firing process, which is a variant of laser doping. ^{Reference Hofmann, Saint-Cast, Suwito, Seiffe, Schmidt, Kambor, Gautero, Kohn, Nekarda, Leimenstoll, Wagenmann, Erath, Catoir, Wolke, Janz, Biro, Grohe, Rentsch, Glunz and Preu24} The laser firing process involved forming localized p⁺ contacts by first depositing a passivation layer on the rear surface of a silicon wafer followed by the deposition of a layer of aluminum. The Al layer was then fired through the passivation layer using a Q-switched Nd:YAG laser with a wave length of 1064 nm. The laser pulse locally disrupts the dielectric passivation layer, melts the underlying Si, and allows Al to diffuse into the molten Si forming a localized p⁺ contact. The rear surface passivation layers in these high-performance cells consisted of ~70 nm of amorphous silicon (a-Si:H) covered with ~100 nm of SiO _x . Both layers in the stack were deposited by plasma-enhanced chemical vapor deposition (PECVD). ^{Reference Schneiderlochner, Preu, Ludemann and Glunz23}

LFCs have also been made using other dielectric passivation layers such as thermally-grown silicon dioxide ^{Reference Schneiderlochner, Preu, Ludemann and Glunz23} ; PECVD silicon nitride (SiN _x :H) ^{Reference Schneiderlochner, Preu, Ludemann and Glunz23} ; PECVD silicon carbide (SiC _x :H) ^{Reference Hofmann, Kambor, Schmidt, Grambole, Rentsch, Glunz and Preu25} ; PECVD silicon oxide (SiO _x :H) ^{Reference Hofmann, Kambor, Schmidt, Grambole, Rentsch, Glunz and Preu25} ; stacks of PECVD SiO _x :H, SiN _x :H, and SiO _x :H ^{Reference Hofmann, Kambor, Schmidt, Grambole, Rentsch, Glunz and Preu25} ; PECVD aluminum oxide (AlO _x ) ^{Reference Saint-Cast, Kania, Billot, Weiss, Hofmann, Gautero, Kohn, Biro, Rentsch and Preu26} ; and atomic layer deposited (ALD) AlO _x capped with PECVD AlO _x . ^{Reference Saint-Cast, Kania, Billot, Weiss, Hofmann, Gautero, Kohn, Biro, Rentsch and Preu26} LFCs have also been made with different forms of Al. Most LFCs have been made using evaporated Al ^{Reference Schneiderlochner, Preu, Ludemann and Glunz23} or sputtered Al, ^{Reference Nekarda, Reinwand, Hartmann and Preu27} but laser firing has also been performed using screen-printed Al paste or Al foil, ^{Reference Nekarda, Hörteis, Lottspeich, Wolf and Preu28} and high efficiency (>20%) solar cells have been made using all these forms of Al.

While most of the LFC research has been performed using pulsed nanosecond Nd:YAG lasers operating at a wave length of 1064 nm, other pulsed nanosecond lasers such as ytterbium fiber lasers ^{Reference Panek, Drabczyk, Kulesza and Zięba29} and Nd:YVO₄ (Ref. Reference Fischer30) have been used. LFCs have also been formed using millisecond single-mode fiber lasers. ^{Reference Raghavan, Blecher, Palmer, Reutzel and DebRoy31}

In the vast majority of the published research, laser firing has been used to make ohmic base contacts for solar cells by firing Al into p-type Si wafers. However, solar cells have also been made by laser firing Al into n-type Si wafers and forming localized emitter contacts. ^{Reference Glunz, Schneiderlöchner, Kray, Grohe, Hermle, Kampwerth, Preu and Willeke32,Reference Granek, Hermle, Fleischhauer, Grohe, Schultz, Glunz and Willeke33} Solar cell conversion efficiencies as high as 19.4% were obtained by laser firing Al through a thermal oxide on the rear surface of 100 Ω cm n-type Si wafers. However, the efficiency fell significantly to ~13.5% when 1 Ω cm n-type Si wafers were used. The difference in performance is due to injection-dependent recombination in a laser-induced damaged region surrounding the localized laser-fired emitters. ^{Reference Granek, Hermle, Fleischhauer, Grohe, Schultz, Glunz and Willeke33} Simulations indicated that the damaged region extended ~5 μm beyond the laser-fired emitter region and that the minority carrier lifetime was reduced to ~0.3 μs within the damaged region. The effective lifetime is almost an order of magnitude greater in the 100 Ω cm silicon wafer since the device operates in the high-injection regime, and the recombination rate within the laser-induced damaged region is reduced by almost one order of magnitude. ^{Reference Granek, Hermle, Fleischhauer, Grohe, Schultz, Glunz and Willeke33}

LFCs have recently been made using materials other than aluminum. Antimony (Sb) was laser doped into n-type crystalline silicon by Takahashi et al. ^{Reference Takahashi, Matsumoto, Murakami and Hayashi34} in 1985 to form a low resistance ohmic contact, but unlike the laser-firing process developed by Fraunhofer, there was no dielectric passivation layer on the surface of the silicon wafer. More recently, Song et al. ^{Reference Song, Balaji, Kim, Choi, Ryu, Park, Ju, Lee, Lee, Lee, Lee and Yi35} laser fired Sb through a thermal oxide into n-type Si to form localized back-surface field contacts, and they fabricated solar cells with efficiencies as high as 17.2%. In even more recent work, He et al. ^{Reference He, Hegedus, Das, Shu, Bennett, Zhang and Birkmire36} laser fired Al/Sb stacked layers through either silicon nitride or p-type amorphous silicon to make ohmic contacts to n-type Si, and they demonstrated efficiencies as high as 16.9% in interdigitated back-contact silicon heterojunction solar cells.

Interdigitated back-contact solar cells with efficiencies >18% were fabricated by laser firing a silver paste on the top of a dopant paste through a dielectric into n-type Si. ^{Reference Sidhu, Bennett, Hmung, Ren, Zou, Carlson, Dong, Li, Lu, Song, Tao, Xia, Yang and Xing37} These back-contact solar cells exhibited good values of the open-circuit voltage (632 mV) and the short-circuit current density (39.7 mA/cm²), but the fill factors were limited to values ≤0.72 due to high series resistance. Imaging with a high-resolution, mid-IR thermal camera showed that resistive losses were occurring due to current crowding at both the laser-fired n⁺ and p⁺ contacts. Figure 8 shows both a thermal image and an optical microscope image of a laser-fired p⁺ contact. The thermal image shows nonuniform resistive heating under bias. The nonuniformities appear to be associated with variations in the topography and reflectivity of the Ag paste. In addition, most of the Ag paste has been removed from the central region of the laser-fired emitter due to laser ablation.

Figure 8.

A thermal image (on the left) and an optical microscope image (on the right) of a localized laser-fired emitter made by laser firing a Ag paste over a dopant paste through an ALD AlO _x /PECVD SiN _x passivation stack.

The LFC process developed by Fraunhofer ISE is being commercialized by Hanwha Q-Cells for multicrystalline Si solar cells with passivated rear point contacts, and conversion efficiencies well above 18% have been demonstrated in a continuously running pilot line. ^{Reference Kloter, Buss, Hussack, Kowalzik, Peters, Schutze, Stegemann and Muller38}

Laser transfer

In the laser transfer process, a thin film on a substrate is transferred to the silicon surface. The laser-transferred film can be used to form a p–n junction or highly doped regions in silicon for metal contacts with low contact resistance. The laser-transferred film can also be used for patterning metal contacts on silicon solar cells. Figure 9 shows the laser transfer process using pulsed lasers. ^{Reference Wang, Carlson and Gupta39} One of the advantages of the laser transfer process is that the transfer can be achieved selectively so that a photolithographic process is not required. The film transfer process can be achieved through a passivation dielectric layer, where the transferred film can make electrical contact to a silicon surface or can dope the silicon. Hoffmann et al. ^{Reference Hoffmann, Röder and Köhler40} demonstrated that an antimony film could be laser transferred through an 80 nm silicon nitride layer and could form a selectively doped n⁺ layer in silicon.

Figure 9.

Schematic of laser transfer process for contact formation on both the front and rear sides of a silicon solar cell. ^{Reference Wang, Carlson and Gupta39}

When a laser beam is incident on a thin film on a substrate, part of the light is absorbed by the film. The absorbed laser light energy raises the temperature of the film at the interface with the substrate and can cause melting and vaporization. The vapor pressure keeps rising since the vapor is contained in a small volume until a threshold value is reached where the surface tension cannot hold the molten region on the substrate. Then the metal film is transferred to the silicon surface due to the forward pressure generated at the metal film substrate interface. The vaporized metal in conjunction with the laser beam can locally melt silicon and can form an alloy with the molten silicon to generate ohmic contacts. Typical morphology of the transferred metal contacts is shown in Figs. 10 and 11. ^{Reference Wang, Carlson and Gupta39} When high levels of dopants such as phosphorus are transferred, the dopant diffuses into p-type silicon to form a p–n junction or highly doped regions are formed for ohmic metal contacts in n-type silicon. For example, a thin film of a-Si doped with phosphorus deposited by CVD on a glass substrate can be laser transferred to silicon. Due to melting of silicon by the laser transfer process, a junction will be formed in silicon in a very short interaction time (typically on the order of a few tens to a few hundreds of nanoseconds) due to the fast diffusion of phosphorus in molten silicon.

Figure 10.

Microscope images of transferred metal spots created by single laser pulses with different energy densities: (a) 1.1 J/cm², (b) 0.90 J/cm², and (c) 0.69 J/cm². ^{Reference Wang, Carlson and Gupta39}

Figure 11.

Scanning electron microscope (SEM) images of laser-transferred metal contacts on Si solar cell: (a) Ti contacts and (b) Al contacts. The insets are optical images of the transferred metal contacts. ^{Reference Wang, Carlson and Gupta39}

Laser ablation of dielectric passivation layers

Solar cells with efficiencies as high as 21.4% have been fabricated by using laser ablation of the rear dielectric passivation layer to form openings for localized contacts. A Nd:YVO₄ laser operating at a wave length of 355 nm (pulse duration = 30 ns) was used to locally ablate a rear surface thermal SiO₂ passivation layer over the boron- and phosphorus-diffused regions before applying the interdigitated finger metallization in an EWT solar cell. ^{Reference Hermann, Engelhart, Merkle, Neubert, Brendemühl, Meyer, Harder and Brendel41}

Das et al. ^{Reference Das, Tous, Hernández, Jaffrennou, Ngamo, John, Posthuma, Baert and Poortmans42} also used a UV (355 nm) nanosecond Nd:YVO₄ laser to ablate 80 nm of silicon nitride on the front surface of a textured Si surface with a 60 Ω/sq emitter to form openings for both point and line electroplated Ti/Cu contacts. They demonstrated conversion efficiencies as high as 18.55% using line contact ablation on full-size solar cells. Laser ablation can cause laser-induced damage at high-energy densities, but Correia et al. ^{Reference Correia, Lossen, Wald, Neckermann and Bähr43} showed that nanosecond lasers could remove dielectrics such as SiO₂ and silicon nitride from the surface of silicon wafers with phosphorus-diffused emitters with no significant laser-induced damage and minimal redistribution of the emitter dopant.

Hernández et al. ^{Reference Hernández, Tous, Allebé, Philipsen, Schlenker, John, Baert and Poortmans44} used a UV (355 nm) nanosecond Nd:YVO₄ laser to open contact vias in the silicon nitride antireflection coating on the front surface of cells made on p-type CZ-Si wafers. They also laser ablated openings in a rear surface oxide/nitride dielectric stack and then formed localized base contacts by depositing aluminum and firing. They obtained conversion efficiencies as high as 19.4% by contacting the emitter with a thin layer of Ti.

Laser drilling

Lasers have also been used to drill arrays of small holes through wafers in the fabrication of MWT and EWT back-contact solar cells. In both types of cells, laser-drilled holes are used as vias to carry the current collected by the emitter on the front surface to busbars on the rear surface of the cell. In some cases, an emitter surface may also be formed over a portion of the rear surface of the Si wafer. In the MWT solar cell, a relatively small number of laser-drilled holes (~16 per wafer) are filled with metal and used to carry the photocurrent from a fine grid pattern on the front surface to busbars on the rear surface (see Fig. 12). In the EWT solar cell, a relatively large number of laser-drilled holes (∼15,000–25,000 holes per wafer) are formed and an emitter surface is formed on both the front surface and on the walls of the vias. In this case, there is no front surface grid and the conductivity of the emitter surface on the walls of the vias is sufficient to carry the collected front surface current to the busbars on the rear surface.

Figure 12.

Schematic of a MWT solar cell.

Lee et al. ^{Reference Lee, Lim, Kim, Moon, Seo, Lee and Cho45} used a single-mode Q-switched UV (355 nm) laser with a pulse duration of 40 ns for drilling holes for MWT silicon solar cells. Mingirulli et al. ^{Reference Mingirulli, Grohe, Dohrn, Hofmann, Schubert, Roth, Biro and Preu46} used both a Nd:YAG laser (1064 nm wave length) with pulse durations in the range of 60–90 μs and a Yb:YAG laser (1030 nm wave length) with a pulse duration of 1 μs to drill holes in Si wafers for EWT silicon solar cells.

Lamers et al. ^{Reference Lamers, Tjengdrawira, Koppes, Bennett, Bende, Visser, Kossen, Brockholz, Mewe, Romijn, Sauar, Carnel, Julsrud, Naas, de Jong and Weeber47} fabricated MWT solar cells on p-type multicrystalline Si wafers and demonstrated efficiencies as high as 17.9%. They laser drilled 16 holes in each wafer for the MWT connections that carried photocurrent from the fine grid pattern on the front surface to the busbars on the rear surface. Their final processing step involved using a laser to form an isolation trench around each hole and also around the edge of the cell. Drews et al. ^{Reference Drews, Clement, Spribille, Thaidigsmann, Linse, Gutscher, Werner, Reitenbach, Ould Chighali, Wolf, Zimmer, Nekarda, Haedrich, Tranitz, Eitner, Wirth, Wilson, Biro and Preu48} fabricated MWT solar cells in a pilot line with efficiencies as high as 19.3% using a cell structure similar to that shown in Fig. 12.

Hermann et al. ^{Reference Hermann, Engelhart, Merkle, Neubert, Brendemühl, Meyer, Harder and Brendel41} fabricated EWT cells (92 cm²) on float-zone Si with efficiencies as high as 21.4%. In addition to using lasers to drill the vias, they also used a picosecond Nd:YVO₄ laser (wave length 532 nm) to ablate the rear silicon dioxide layer and form localized selective emitter contacts as well as localized contacts to the back-surface field regions.

Laser marking, grooving, and cutting

Laser ablation processes have been used by solar cell manufacturers to mark solar cells for identification, ^{Reference Wanka, Rychtarik, Muller, Geissler, Kappe, Spallek, Vom Bauer, Ludwig and Wawer49} to form grooves in solar cells for selective emitters ^{Reference Wenham and Green50} and to cut silicon ribbons into wafers. ^{Reference Hanoka51}

Solar cell manufacturers such as Q-Cells have used laser marking for single wafer identification, which improves quality control and assists in process improvement. ^{Reference Wanka, Rychtarik, Muller, Geissler, Kappe, Spallek, Vom Bauer, Ludwig and Wawer49} Both Nd:YAG lasers (1064 nm) and Nd:YVO₄ lasers (1064 nm) have been used to create dots ~100 μm in diameter and 10–30 μm deep (using 20–50 ns laser pulses) for marking silicon wafers.

The laser-grooved buried contact solar cell was invented at the University of New South Wales ^{Reference Wenham and Green50} and was commercialized by BP Solar in 1995. ^{Reference Bruton, Heasman, Nagle, Cunningham, Mason, Russell and Balbuena52} Either a high-power Nd:YAG laser or a Nd:YVO₄ laser operating in the infrared laser scribed grooves through both the antireflection layer and the shallow emitter of a textured p-type Si wafer (see Fig. 13). The grooves were typically about 30 μm wide and about 35 μm deep. After laser grooving, an additional phosphorus diffusion was used to form a heavily-doped selective emitter in the grooves. The rear surface and the grooves were then plated with Ni, Cu, and Ag. Efficiency as high as 18.3% was achieved in full-size Saturn cells (147.5 cm²) and efficiency over 20% was obtained in small area cells. ^{Reference Bruton, Mason, Roberts, Hartley, Gledhill, Fernandez, Russell, Warta, Glunz, Schultz, Hermle and Willeke53} Manufacturing of the laser-grooved buried contact solar cell ceased in 2008 due to relatively high manufacturing costs compared to conventional screen-printed silicon solar cells.

Figure 13.

A laser-grooved buried contact solar cell. ^{Reference Bruton, Heasman, Nagle, Cunningham, Mason, Russell and Balbuena52}

Lasers have also been used in production to cut silicon ribbons into wafers and to also mark the wafers for identification. Evergreen Solar developed a process to grow silicon ribbons using strings of refractory material, and they also developed an automated laser-cutting system to cut the silicon ribbon into wafers for cell processing. ^{Reference Hanoka51} However, Evergreen Solar ceased commercial production of solar cells in 2011.

Light trapping by laser microtexture

The silicon surface of a typical smooth wafer reflects about 30% of sunlight due its high refractive index in the visible spectral region. For solar cells, surface light reflection needs to be minimized to achieve higher current density and hence higher solar cell efficiency. A quarter wave thick layer of a material such as Si₃N₄ or TiO₂ on the top of a silicon surface can reduce the surface reflectance through its antireflection properties. These antireflection coatings are only effective at one given wave length. The solar spectrum is broad from the UV region to the near-infrared region, so antireflection coatings are not very effective for broad wave length sources.

Anisotropic chemical etching of silicon surfaces to generate surface microtexture has been used for light trapping applications. The chemically etched surface in combination with an antireflection layer has been shown to reduce the surface reflection to below 5%.

The silicon surface can also be microtextured to reduce surface reflection by pulsed laser processing. Under certain laser processing conditions, self-organized cone type microtexture can be formed on a silicon surface as shown in Fig. 14. The microtextured silicon (also called black silicon) shows very low reflectance (~1%) over broad spectral range as shown in Fig. 15. ^{Reference Nayak, Iyengar and Gupta54} Solar cells with efficiency of over 16% have been demonstrated using laser microtextured silicon surfaces. ^{Reference Nayak, Iyengar and Gupta55} The throughput of laser microtexturing process still requires improvement in order for this process to be cost effective.

Figure 14.

SEM image of laser-microtextured Si surface. ^{Reference Nayak, Iyengar and Gupta54}

Figure 15.

Light reflection from polished bare silicon and laser-microtextured silicon surface. ^{Reference Nayak, Iyengar and Gupta54}

Other laser processing: annealing, sintering, crystallization, defect repair

Alemán et al. ^{Reference Alemán, Streek, Regenfuss, Glunz and Willeke56} used lasers to produce electrical contacts to the emitter of silicon solar cells by sintering metal powders. They spread metal powders such as silver, molybdenum, and tungsten on the top of the silicon wafer, and then used a laser beam to sinter the powder forming an electrical finger pattern contact to the emitter. After removing the nonsintered powder, they electroplated silver to increase the conductivity of the contacts. The electroplated fingers were 40 μm wide and 18 μm thick. They demonstrated conversion efficiencies of ~14.5% in small (1 cm²) p-type Si solar cells by laser sintering tungsten particles to a 50 Ω/sq emitter.

Lasers have also been used to repair defects present in the surface region of silicon solar cells. Abbott et al. ^{Reference Abbott, Trupke, Hartmann, Gupta and Breitenstein57} used a Nd-YAG laser (1064 nm, 200 ns pulses) to ablate grooves to locally isolate shunted regions in industrial silicon solar cells. They were able to improve the performance of a shunted industrial screen-printed solar cell from a conversion efficiency of 9.6 to 13.3% by laser isolating shunts.

Lasers have been used to crystallize thin amorphous silicon films on glass to form a seed layer for the epitaxial growth of thicker Si films (>10 μm) for thin crystalline silicon solar cells. Nd:YAG lasers have been used in the CW mode to grow relatively large crystals (tens of micrometers wide and >100 μm long), ^{Reference Andrä, Bergmann and Falk58} but the scanning speeds were not adequate for commercialization. In other works, ^{Reference Nayak, Eaton, Selvan, McLeskey, Gupta, Romero and Ganguly59,Reference Andrä, Bochmann, Falk, Gawlik, Ose and Plentz60} the laser scanning speed was increased by first heating the a-Si coated borosilicate glass to 600 °C, and then scanning with a CW diode laser operating at a wave length of 806 nm. This approach allowed the scanning rate of a line-focused (30 mm × 0.1 mm) laser beam to be increased to several cm/s. ^{Reference Andrä, Bochmann, Falk, Gawlik, Ose and Plentz60}

Andrä et al. ^{Reference Andrä, Höger, Bergmann, Gawlik, Plentz, Ose, Falk, Burghardt and Schippel61} used both a CW diode laser and a diode-pumped disk laser (512 nm, pulse duration 250–600 ns) to crystallize thin boron-doped a-Si films (60–400 nm thick) at 600 °C on borosilicate glass coated with silicon nitride. After increasing the thickness of the thin silicon films using epitaxy, they fabricated silicon solar cells with open-circuit voltages over 500 mV and short-circuit current densities as high as 20 mA/cm².

Solar thermal

The solar thermal systems generate electric power by absorbing the sunlight to generate heat and then heat is converted into electric power. The key element in solar thermal systems is the solar receiver, which captures the sunlight and transfers the heat energy to flowing fluid, as shown in Fig. 18. ⁷⁹ The solar receivers are fabricated with a multilayer approach such that solar light is efficiently absorbed but infrared radiation from the heated receiver is minimized to minimize the radiated thermal loss. The authors in Ref. Reference Ungaro, Gray and Gupta81 have demonstrated a high efficiency solar receiver by laser sintering of tungsten nanoparticles on a stainless steel substrate. Through the control of laser processing parameters, the morphology of the sintered nanoparticle layer is governed such that high solar absorptance can be achieved with low thermal emittance. Figure 19 shows the infrared reflectance spectra of laser-sintered layer on stainless steel substrate. The efficiency of solar thermal systems increases with higher operating temperatures as governed by the Carnot cycle limit. So, solar thermal receiver must operate at temperatures higher than 500 °C and preferable greater than 650 °C.

Figure 18.

Schematic of solar receiver in a solar thermal system. ⁷⁹

Figure 19.

Measured reflectance for gold, stainless steel, tungsten, and laser-sintered tungsten. ^{Reference Shah and Gupta80}

Thermophotovoltaics

The efficiency of semiconductor solar cells is lowered by the thermalization of high-energy photons (with energies greater than the band gap) and the transmission of low-energy photons (with energies less than the band gap). The solar cell efficiency can be enhanced by absorbing the broad solar spectrum and emitting the radiation into a narrow spectral range to match the band gap of semiconductor solar cell. Thermophotovoltaics are based on this approach and Fig. 20 shows the operation of a thermophotovoltaic system. The concentrated sunlight is absorbed by the receiver and the emitter generates the infrared radiation in a narrow spectral range to match the solar cell band gap. High-power lasers can be used to fabricate solar receivers and emitters. A laser microtextured surface described previously under light trapping structures section can be an efficient solar receiver. The morphology of the emitted surface can be controlled by laser processing to cause the IR emission in a narrow spectral range.

Figure 20.

Schematic of thermophotovoltaic system. ^{Reference Ungaro, Gray and Gupta81}

Energy sources

Electric power is generated in thermoelectrics due to charge carrier flow under the generated temperature gradients. The figure of merit is given by the dimensionless value ^{Reference MacDonald82} :

where σ is the electrical conductivity; S is the Seebeck coefficient; k = k _e + k _L is the thermal conductivity, which contains contributions from electron (k _e) and lattice (k _L) thermal conduction ^{Reference Nolas, Sharp and Goldsmid83,Reference Kittel84} ; and T is the absolute temperature. The Seebeck coefficient characterizes the thermopower of the material, which describes the voltage developed when a temperature gradient is present. Optimizing ZT is a significant challenge as improving one parameter often adversely affects another. For instance, according to the Wiedemann-Franz law, ^{Reference Franz and Wiedemann85} the electrical and thermal conductivity of metals are proportional to each other, making it difficult to reduce k without also reducing σ.

Recently, it was shown that the thermal conductivity could be strongly reduced through nanostructuring of the material. ^{Reference Bux, Blair, Gogna, Lee, Chen, Dresselhaus, Kaner and Fleurial86} A promising approach involves nanocrystalline materials with grain sizes of a few tens of nanometers, in which enhanced phonon scattering at grain interfaces leads to a strong reduction of the thermal conductivity without negatively affecting the electrical conductivity. ^{Reference Wang, Lee, Lan, Zhu, Joshi, Wang, Yang, Muto, Tang, Klatsky, Song, Dresselhaus, Chen and Ren87–Reference Ma, Hao, Poudel, Lan, Yu, Wang, Chen and Ren91} These “nanograined” materials behave like simultaneous “electron crystals” and “phonon glasses”. ^{Reference Snyder and Toberer92}

The nanograined material based films have been prepared by DC hot pressing where grain growth occurs and quantum confinement does not play a role. The pulsed laser sintering method can avoid the grain growth and hence quantum effects could be utilized for the enhancement of thermoelectric power. ^{Reference Stoib, Langmann, Matich, Antesberger, Stein, Angst, Petermann, Schmechel, Schierning, Wolf, Wiggers, Stutzmann and Brandt93}

Laser processing can provide manufacturing advantage for fuel cells. Laser drilling and laser sintering are being evaluated for alkaline fuel cell (AFC) applications. ^{94–Reference Scotti, Trusheim, Kanninen, Naumenko, Shulz-Ruhtenberg, Snitka, Kallio and Franssila96} The AFC converts hydrogen into electric power. Laser-drilled holes perforate the metal substrate for gases to permeate. Currently, over 4000 holes per second can be drilled in thin metal substrates. The laser sintering of powder on a substrate is an alternative method to achieve porous substrate for gas permeability.