A solar cell primer for the PCB industry: manufacturing methods, like screen-printed paste, are giving way to photolithography, inkier deposition and plating, as solar cell manufacturing techniques evolve.

Photovoltaics (PV) is the electronic industry's younger brother. Each industry shares some DNA--imaged conductors on dielectric material, layered substrates, severe cost pressures, high dependence on innovation and rapidly evolving electrical designs. But the sibling industries are enticingly different. During 2007 and 2008, I embarked on a fact-finding tour into the world of solar cells, with the charter to learn the players, markets, issues and drivers that are shaping this rapidly growing industry. Just like the older brother PCB industry of 1960 through 1990, PV is on a rapid growth spurt. TABLE 1 gives a snapshot of where we are today. There are major changes in store for PV's adolescence, such as the emergence of thin-film and organic photovoltaics, the efficiency gains of chemical metallization and the soon-to-come huge transition from achievement of grid-parity. Who will win--silicon, thin-film or organic PV? Germany, China or the US? IC, PCB or green-field PV makers? Cast away your preconceived notions; it will require a fresh look to synthesize the wish lists from leading solar cell makers and to uncover the real opportunities for next-generation PV.

Efficiently utilized, my body's mass would provide enough energy to power the country's electricity demand for one month. That is, if you believe Mr. Einstein. We need to take his calculations somewhat on faith. E=[mc.sup.2] predicts that my mortal coil of 95 kg is equivalent to 2.2 trillion kilowatt-hours of power. Per the US Energy Information Agency, that would power us from June to July. We would need trainloads of coal, pipelines of natural gas, billions of barrels of oil and countless splittings of uranium to make up the rest. These alternatives are, by the way, the sources of energy production in America, in order of power supplied. FIGURE 1 shows US energy use. Coal is dirty, producing dust when mined, sulfur when burned and acid rain when washed from the sky. Natural gas is expensive to find, more expensive to transport and produces greenhouse gases when combusted. Crude oil combines the ills of coal and natural gas and adds other problems. Oil is dirty, expensive, in limited supply and necessitates reliance on a supply chain outside US boarders. Nuclear energy is on the rise, now contributing about 7% of domestic power but suffers from "not in my back yard" concerns over new plant locations, safety and spent fuel disposal. Rounding out the list of energy sources are biomass, wind, hydroelectric and solar power. I should clarify that just about all of the energy sources listed above are really just different manifestations of solar energy. Fossil fuels are made from the anaerobic metamorphosis of plants, which stored solar energy as carbohydrates long ago--same story with biomass, just on a more recent timescale. Wind results from sun driven weather patterns. Only hydrothermal and nuclear fission, deriving their power from the radioactive decay of earth itself, are independent of the sun.

[FIGURE 1 OMITTED]

[FIGURE 2 OMITTED]

Catching Photons

The sun emits 370 billion petawatts. Metric prefixes haven't reached this level of quantity yet, so the best I can do is say 370 yottawatts--no kidding. Luckily for us, energy diminishes with the square of distance, and our atmosphere screens the level down to about 1000 watts per meter at Earth's surface. That is a lot of power, enough, if converted efficiently, to power two 70-inch plasma TVs. That's the trick--converting this boundless source of energy in an efficient way. For more than 100 years, scientists have been trying to convert photons to electrons efficiently. The 1950 vintage solar cell converted solar energy to electricity at about 5% efficiency, while today's production model is more like 15%. Why so low? There are a few factors limiting the amount of electrons produced per sunlit area. Most solar cells are made from crystalline silicon, doped with boron and phosphorous to produce a p-type/n-type junction. The theoretical maximum efficiency of such a system is about 25%, but to get captured, the photons need to reach the p-n junction. The photons might bounce off the cell's surface, reflect back into space or pass right through the cell if their wavelength is wrong. Some of their energy will be wasted if they are too strong or never captured if they don't exceed the band-gap energy. When the photons do get captured, again proving Einstein's Nobel-winning photoelectric effect, they will generate an exciton --a separated charge. The electron and hole pair will need to reach their respective destinations to complete the electric circuit without a fatal recombination, or all that energy will be for naught. Successful electrons will suffer resistive losses in the silicon, at the interface with the conductors and within the conductors themselves. Cumulatively, these effects and others bring production cells to about 15% efficiency. This is where innovation promises to save the day; newer designs reduce shadowing, improve conductivity, minimize reflection and place the junctions closer to the conductors. Lab cells with efficiencies above 24% are translating into production cells with efficiencies above 20% at the high end. FIGURE 2 shows some of the record-setting efficiencies.

[FIGURE 3 OMITTED]

Solar Power Players

Crystalline silicon cells dominate the market, accounting for about 90% of the PV energy produced annually. Industry leaders Q-Cells (Germany), Suntech (China), Sharp (Japan) and SolarWorld (Germany) make most of their cells using silicon wafers with a silver conductive front-side grid made from sintered screen-printed silver paste and aluminum and silver paste back side. SunPower, a US company, takes silicon in a new direction by forming all conductors on the rear side, freeing up the front for full light absorption. Japan's Sanyo incorporates layers of photosensitive silicon--crystalline and amorphous--in its HIT design. While First Solar's (US) take is unique among PV leaders with use of non-silicon thin-film technology, utilizing cadmium and telluride. Other silicon innovations include novel wafer growing methods at Evergreen (US) and Schott (Germany) and unique conductor formation with Suntech's Pluto cells--a new take on BP Solar's mothballed laser trench process called Saturn.

Thin-film technologies, as is common with newer industry sectors, exhibit a much more diverse set of materials and manufacturing methods, representing a fragmented set of a dozen or so sizable producers totaling about 10% of PV capacity. Like the PCB industry, PV loosely follows the 80/20 rule, with the top 20 solar cell producers generating close to 80% of the industry's total revenue. Similar to the PCB industry, the usual cast of characters represent the solar energy manufacturing regions--Japan, China, Germany, Taiwan and the US. Similar, too, is the trend toward technology transfer from western R&D labs to production manufacturing in Asia. This trend will not be as dramatic in PV, however, because the large, heavy, fragile PV modules are best built near the location of their intended use. The consumption of solar cells relies on different drivers than our electronics experience. Consumption is based on government subsidy. This explains the quick change from Japan's consumption of 36% of the world's PV power in 2002 to just 9% in 2007, while Europe now purchases about three-quarters of PV power, thanks to attractive government incentives in Germany and Spain. Late in 2008, Congress extended the Investment Tax Credit. Will the US be the world's next leading consumer, thereby benefiting from the inevitable rise in domestic manufacturing? The stimulus package might be the deciding vote.

[FIGURE 4 OMITTED]

America is no slouch in funding alternative energy, even before "Obama bucks" come rolling in. The Department of Energy's Solar America Initiative funds companies and universities in its Technology Pathway Partnership to the tune of about $60 million per year. The DOE programs, as well as the state--and federal-level tax benefits, will grow dramatically once the $787 billion American Recovery and Reinvestment Act (stimulus) funds begin allocation, with $45 billion slated for the Energy and Green Jobs Program. These funds, mirrored in spirit in Japan and Europe, are enough to stoke the jealousy of the under-represented electronics industry around the world. Those interested in tracking the US funding can do so at: wwwl.eere.energy. gov/solar and www.recovery.gov. Overall, the plan is to bring the cost of solar-generated power equal to, or below, the cost of conventional power--the magic threshold known as grid-parity. With this achievement, the manufacture of PV will explode in growth. FIGURE 3 provides an overview of DOE TPP funding.

[FIGURE 5 OMITTED]

[FIGURE 6A OMITTED]

[FIGURE 6B OMITTED]

Guessing on Growth

Everyone understands the appeal of solar power; it's clean, quiet and a renewable energy source. However, it has also been billed as expensive. As solar power efficiency makes its unrelenting climb upward, the cost/watt continues its descent. FIGURE 4 maps the cost of solar electricity. You can argue that the efficiency increases will slow, following the law of diminishing returns, as we approach the theoretical max of solar energy conversion, but the cost declines will accelerate, as the industry matures. More companies will join, manufacturing efficiencies will improve, material costs will drop and equipment will be amortized, but this industry's growth is not governed solely by the laws of Economics 101. Other factors, like we saw with government incentives, are at play. FIGURE 5 lists the market drivers. Forward thinking environmentalists and technogeeks were the early adopters of solar power. Not concerned with ROI, these groups buy PV regardless of the price. Others, with no access to grid electricity, will also pay any price. We see the use of solar modules in remote Africa, in the Rocky Mountains and even in outer space. Don't dismiss those who wish to control their own power source, from the survivalists in New Mexico to the suburbanite in New Jersey with back-up PV. But when we move into the political realm, all bets are off. Global warming and pollution are touchstone issues galvanizing the lobby to expand solar energy. With a government anxious to reduce dependence on Middle East oil, we can begin to understand the success of solar subsidies in the US. Feed-in tariffs in Germany, Spain, Italy and Greece pay solar households 0.30 [euro]-0.55/kWh, (currently about $0.41-$0.75/kWh) and have shown such popularity that we can expect an expansion of pro-solar policy worldwide. Solar power has achieved its tipping point. Recent annual growth rates exceeded 60%, with long-term forecasts at more than 25% per year. Efficiency increases, material cost decreases, production and consumption incentives, as well as economies of scale, are all conspiring to catapult PV into big-time deployment. The current credit crisis is indeed slowing the flood, but only temporarily, as some investors see the investment return from PV generation as more attractive than alternative, low-risk investments. With apologies to Lao Tze, the journey to grid parity begins with a thousand small steps. Bigger steps will come through design and manufacturing innovation. To understand the opportunities for innovation, look at the weaknesses of existing solar cell technology.

The Power Maker

Conventional "first-generation" solar cells are made using crystalline silicon wafers, often of a lower grade than silicon used in semiconductors. The Si (+4) wafer starts as p-type with boron (+3) dopant. To better capture light, the wafer is textured with hydroxide or nitric/hydrofluoric acids to generate mountainous geometry intended to reflect light obliquely into the silicon. The p-n junction is formed with diffusion of phosphorous using vapor deposition, and a surface passivation is applied, again in vacuum equipment, to impart the characteristic blue silicon nitride film. Among these steps, wet chemical processes are interspersed to clean, to neutralize and to prepare the silicon along the way. FIGURE 6A and FIGURE 6B compare PCB and PV manufacturing processes.

Our silicon wafer is now an energy-producing diode, but without the means for extracting those electrons for useful work. The ensuing conductor formation step, creating elemental circuitry on the cell's surface, begins to remind us of the early days of the PCB industry. A conductive silver paste is screen printed on the sunny side of the wafer. The paste is sintered at 900[degrees]C, an obscenely high temperature necessitated not just by the sintering of silver flakes, but in an effort to break through the silicon nitride with tiny glass flits contained in the specialty silver paste. Meanwhile, the circuit is completed on the dark side of the cell with aluminum and silver paste--silver to make contact with silicon and aluminum to form a back surface field. The +3 valence of aluminum helps boost the p-type nature of the junction. For circuit builders, screen-printing is so 1960s, but screening rules the PV world. Rebellions are springing up in various quarters, however. Sunpower and Suntech famously use sputtering and photolithographic techniques, borrowed from a semiconductor lineage, to make better contact. These companies (and others) also apply the cost advantages of the PCB world by building the conductor with wet plating chemicals.

Patterns of Success

Conductor formation techniques, as an alternative to screen-printed paste, can take several incarnations:

* Inkjet resist and etch of anti-reflective coating (ARC)

* Aerosol deposition of silver paste

* Inkjet deposition of metal paste

* Laser ablation of ARC

* Photolithography

In any case, the replacement of screened paste means that some other method for building a conductor is needed. The PCB industry knows all about these techniques: electroless and electrolytic metal plating. Imagine a silicon wafer with a phosphorous-diffused emitter and an aluminum-back surface field. Resist patterning of the wafer's front, followed by fluoride etching of the silicon nitride anti-reflective coating, would result in exposed silicon. A very thin layer of electroless nickel can be used to make electrical contact to the silicon. A low temperature sintering operation bonds the conductor to the silicon by formation of the nickel silicide intermetallic phase. Once the nickel is locked in place, several options are available to build a suitable conductor. Copper is cheap and nicely conductive, but it requires a thin nickel layer to prevent copper from poisoning the silicon; both metals are easily plated on the wafers. There is a trend toward electroless processing, since the connection of very thin wafers in electroplating can cause breakage of these fragile specimens. One elegant idea has been resurrected from semiconductor's past--the use of the current produced by the cell itself to fuel more metal deposition, a process termed "light-induced plating". Whatever the sequence and technique, deposition of metals from wet process chemicals will bring the cost, handling and productivity benefits needed to get solar cells to grid parity.

Non-Silicon

While much of the solar cell industry relies on crystalline silicon wafers (~85% by wattage produced,) there is a growing class of alternatives, looking to evade the cost and processing disadvantages of conventional PV. The most noteworthy segment--thin-film--is represented by a set of techniques where thin coatings of photosensitive materials are deposited on (usually) a rigid glass substrate. The material must meet the requirements of converting energy from ambient light into electron-hole pairs and allow the charge separation through the act of semi-conductivity. Surprisingly, many materials exist to meet these requirements. TABLE 2 shows some of the PV manufacturing options. The US is especially active in the thin-film sector, possessing a very active group of R&D facilities and startup ventures based on thin-film. What is not possible, when discussing thin-film PV, is use of broad generalizations like we used when describing silicon PV. There are dozens of photosensitive materials (CdTe, CIS, CIGS, amorphous Si, etc.), several substrate platforms (glass, steel, organic films, etc.) and even a fascinating set of installation techniques. If this complexity weren't enough, yet another generation of PV methods is upon us, employing superconducting buckyballs, dye-sensitized organics, multiple heterojunction layers for capturing various photon energies and even nano-structured antennae for plucking electromagnetic wavelengths directly into high-frequency electricity (solar energy in the dark?)

What to do?

In our age of limited resources, perhaps the only thing more abundant than the sun's energy is the unbridled enthusiasm over solar energy. Guilty, I am an unabashed enthusiast! A degree of caution is due to counteract the solar lovefest. I cannot recommend a mass exodus of professionals and capital from the strong and mature electronics industry to the more speculative PV industry. The learning curve is a long path, with profound changes happening daily. I hope my comparison of the PCB industry to PV leads to the conclusion that some processing techniques are similar, but there are many other aspects that vastly separate the sibling technologies. While I will warn you that the path to the sun is long and difficult, I will also remind you of Icarus' inspirational quote, "All limits are self imposed." Then again, flying too close to the sun, Icarus found some limits externally imposed.

DON CULLEN is managing director at MacDermid Photovoltaics Solutions and can be reached at [e-mail].

TABLE 1. Electronics industry vs. solar energy industry.

Electronics Industry:        $1,000 billion/year

Solar Energy Industry:       $30 billion/year

Circuit Board Industry:      $50 billion/year

Solar Cell Industry:         $15 billion/year

Circuit Board Production:    90 million [m.sup.2]/year

Solar Cell Production:       35 million [m.sup.2]/year

TABLE 2. PV manufacturing options.

                       Pros                      Cons

Screen Printed Front   Well established and      Shadowing, contact
Contact                optimized                 resistance, conductor
Ag Paste                                         [ohms]

Back Contact Cells     No conductor shadowing,   Mare expensive
                       highest production        structuring
                       efficiency

Emitter wrap-through   Minimal shadowing,        Laser drill
(EWT)                  aesthetics                through-put

Laser patterning

Plated metal           Low series resistance,    Laser groove
conductors             high contact              throughput
                       efficiency

Light-induced silver   Improved trace            Adhesion, chemical
plating                conductivity              stability, contact,
                                                 cyanide

Amorphous Si Thin      Low cost of materials     Low efficiency
Film                                             (5-10%)

CdTe Thin Film         Lowest cost per watt      Low efficiency (10%)
                       in large production       Toxicity

CIS/CIGS Thin Film     Low cost of materials     Complicated
                                                 deposition
                                                 Low efficiency
                                                 (10-15%)

Oxide on Glass         Minimal shadowing,        Sensitive surface
                       full spectrum light
                       absorption

Dye Sensitized         Tolerant of low-level     10-12% efficiency
                       and polarized light       Use of liquids

Organic PV             Flexible, high            Unproven in HVM
                       production capability