A Peek Inside PV

Intermediate

Inside this Article

A single-crystal ingot
A single-crystal ingot, just drawn from the silicon-boron melt.
Monocrystalline boules
Monocrystalline boules created using the Czochralski process.
A poured multicrystalline ingot
A poured multicrystalline ingot, ready to be sawn into cell-sized wafers.
Cut lengths of ribbon cell material
Cut lengths of ribbon cell material, awaiting further manufacturing.
Multicrystalline cells
Multicrystalline cells, with electrical traces added, are ready to be assembled into PV modules.
A robot with pneumatic suction prepares to place a series string of cells.
A robot with pneumatic suction prepares to place a series string of cells.
Delicate silicon wafers are best handled by robotics.
Delicate silicon wafers are best handled by robotics.
A single-crystal ingot
Monocrystalline boules
A poured multicrystalline ingot
Cut lengths of ribbon cell material
Multicrystalline cells
A robot with pneumatic suction prepares to place a series string of cells.
Delicate silicon wafers are best handled by robotics.

Copper, aluminum, and silver are conductors—that is, electricity easily flows through them. Conversely, plastic and porcelain are insulators that resist the flow of electricity. But semiconductors, such as the silicon used in the majority of PV cells, reside in the hazy middle ground between insulators and conductors. To enable semiconductors to easily donate or accept electrons, vital to their function in making electricity, they must be “doped”—adding impurities to the pure semiconductors’ atomic lattices to make them more receptive to electron transfer.

In a PV cell that uses silicon as the semiconductor, the most common dopants are boron and phosphorous. Doping silicon with boron creates a material that can easily accept electrons (positive, p-type; the “absorber” layer) and doping silicon with phosphorous creates a material that can easily donate electrons (negative, n-type; the “emitter” layer). In a PV cell, the junction between p-type and n-type regions results in an electric field known as the cell’s positive/negative (P/N) junction. Photons of sunlight energy give electrons the push they need to hop onto the conductors (traces or grid lines) and into the electrical flow of the circuit. The P/N junction helps keep the electrons from simply recombining with an electron hole within the cell itself. The “pull” of the electrons toward the positive layer is what keeps the flow going: Give them a less resistive pathway to follow and they’ll gladly take it.

Most PV cells fall into one of two basic categories: crystalline silicon or thin-film. Crystalline silicon modules can be fashioned from either monocrystalline, multicrystalline, or ribbon silicon. Thin-film is a term encompassing a range of different technologies, including amorphous silicon, and a host of variations using other semiconductors like cadmium telluride or CIGS (copper indium gallium diselenide). Thin-film technology generates a lot of the current R&D chatter, but crystalline modules currently capture more than 80% of the marketplace.

Crystalline Cell Manufacturing

Monocrystalline (also called monocrystal or single crystal) cells are the most efficient cells available on the market, although they are also the most energy intensive to manufacture. The highest performing, commercially available monocrystalline modules have the ability to convert about 20% of the energy in photons of sunlight into electrical current.

The monocrystalline manufacturing process starts with melting highly purified silicon along with a boron dopant. The purified silicon is the result of a chain of manufacturing that starts with quartz sand, processed into metallurgical-grade silicon, which is further refined into solar-grade silicon. The Czochralski process, one method for growing the crystals, pulls a seed crystal from the top of an approximately 2,500°F silicon/boron melt, and the crystal structure accretes and solidifies as the seed is slowly drawn from the melt. The result is a boule—a long, cylindrical crystal. The boule (usually 5 or 6 inches in diameter) is sawn into thin, round wafers that become the cell’s building block—the p-type layer (as the positive dopant, boron, has been previously added to the purified silicon). The n-type layer is created after wafer sawing, usually by coating the wafer with phosphorous and using heat to allow the phosphorous atoms to partially diffuse into the silicon. Monocrystalline cells all start out life round in shape, but are often squared off to maximize efficient cell coverage in modules. The downsides of monocrystalline cell manufacturing are the very high heat needed, the slow drawing-out process, and the waste from the sawing process, which can account for up to a 50% loss of the original boule.

Multicrystalline Silicon Cells. The process for creating multi- or polycrystalline cells takes a slightly different, less energy intensive (and less expensive) approach, although the cells themselves are about 4% less efficient than monocrystalline cells.

Purified molten silicon and boron are cast in a large block, which cools into an ingot. Instead of creating a monocrystal, the resulting structure has randomly oriented crystalline regions, which causes the lower efficiency and the cell’s random-shard appearance. To create the cells, the ingot is sawn into square wafers, with the n-layer applied the same way as for monocrystals.

Ribbon Silicon Cells. Another type of polycrystalline cell is produced using string-ribbon technology. In this process, a thin strip of p-type crystalline silicon is slowly drawn up out of the silicon melt between parallel strings. The molten silicon is drawn up with surface tension, much like a soap bubble. This thin strip of silicon cools and then solidifies, and a laser cuts the ribbon into individual cell lengths. This technology is less expensive than creating standard polycrystalline cells because it eliminates the sawing process (and related waste) and the PV ribbon is thinner than standard sawn cells, which also saves silicon. 

Crystalline Module Manufacturing

Because the power output of an individual cell is relatively small (typically a few watts), multiple PV cells are electrically connected in series and parallel to make a module that’s more usefully sized. Whether mono-, multicrystalline, or ribbon silicon, the process leading from cell to finished module is similar.

Cells are overlaid with a conductive grid to carry the electrical current. The grid looks much like a transport network, with side roads to carry electrons branching out over the cells, and main highways connecting the cells in series and parallel configurations. Traces are most frequently made of silver that’s screen-printed onto the cell. As the top grid can shadow the cell from some sunlight, efficiency can be gained by keeping the traces thin or laying the grid in laser-etched grooves—or even moving the grid entirely to the back of the cell, the process used in SunPower Corp.’s cell design. 

Next, an antireflective coating is applied to the top of the cells. Reflection is the enemy of the solar cell, as the more light that is reflected (rather than absorbed), the lower energy production will be. Various means are used to conquer reflectivity as the module is fabricated. Silicon starts out as a shiny-gray, highly reflective material. The gray color is modified by changing the thickness and refractive index of the antireflective coating material (typically silicon monoxide), so the cell appears blue or black. Multiple layers of antireflective coating can reduce reflectivity to less than 4%. An additional antireflective technology is acid etching, which creates a textured cell-top surface, like miniature valleys and mountains, which can help capture rays of light.

After the antireflection coatings are dry, the final step in crystalline module production is encapsulation for weatherproofing. Tempered glass is commonly used as a clear top protector. Tedlar, a polyvinyl fluoride film, is frequently used as the module backing, although glass can also serve this purpose. Ethylene vinyl acetate (EVA) laminate is used to seal (or glue) the front and back of the cells to the glass and Tedlar. Modules are enclosed in a mounting-ready frame, usually aluminum, that is riveted or screwed together. Finally, the positive and negative electrical connections are installed on the back of the module.

Thin-Film Cells & Module Manufacturing

“Thin-film” applies to a very broad range of PV module manufacturing techniques. Basically any PV module for which a crystal has not been grown can be classified as thin-film. Instead of a seeded or cast crystalline structure, the semiconductor is deposited (sprayed, through vapor deposition, or even printed) as a film on various substrates. Without a fragile crystalline structure to protect, thin-film applications can be applied to materials like glass, flexible plastic, or stainless steel and other metals.

Materials such as copper indium gallium diselenide (CIGS) or cadmium telluride (CdTe) can be used as the semiconductor material. Silicon-based thin-film products are most commonly called amorphous silicon. Amorphous silicon manufacturing techniques use silane gas, instead of highly processed polysilicon. Plasma-enhanced chemical vapor deposition is a common method used for amorphous silicon manufacture—a process in which silane gas can be deposited on the chosen substrate within a surprisingly compact laboratory machine. 

Thin-film PV cells generally measure only a few micrometers thick, and are comprised of thin layers of semiconductors and dopants, where each layer is subsequently “sprayed” or “printed” on top of the previous layer. Along with the deposition process, a transparent, conductive oxide is overlaid on the entire module surface to serve as the conductive path. In some cases, the first layer applied on the encapsulant will be the conductive path, depending on manufacturing technique. Thin-film modules are made monolithically, where all the layers are deposited in a sheet, creating one large PV cell. Later in the process, the cell is divided up into smaller cells by laser etching.

There are several cost savings realized with thin-film: manufacturing is a faster, lower-temperature process that eliminates the need for growing crystals and does not depend on highly purified silicon; and much less semiconductor material is used. But thin-film currently has much lower energy conversion efficiencies than crystalline technology, ranging from about 4% to 14%. In addition, amorphous silicon thin-film modules undergo significant power decreases after the first few weeks of deployment (between 15% and 35%). This is something that installers need to be aware of so that they can install wiring and components able to handle the initial higher output, yet base energy production predictions on the lower stabilized output.

On the Horizon

PV module manufacturing is following a trajectory similar to the semiconductor device industry, which ramped up in the 1960s and is still viewed as a fleet-footed industry. Much of the equipment used by PV manufacturers is similar to, if not a repurposed version of, existing semiconductor device equipment. The major costs involved in PV manufacturing are much like all manufacturing, including personnel, operations and facilities, equipment, utilities, and material inputs. But what makes the PV industry special is the astonishing growth rate: According to Solarbuzz.com, PV market installations reached 5.95 gigawatts in 2008, which represents a 110% growth over 2007. Whatever shape PV takes in the future, we can rest assured that the quest for higher efficiency modules combined with lower manufacturing costs will continue unabated.

Access

Rebekah Hren is a licensed electrician living in North Carolina. She works with Honey Electric Solar Inc. designing and installing PV systems, is an instructor for Solar Energy International, and co-authored The Carbon-Free Home, 36 Remodeling Projects to Help Kick the Fossil-Fuel Habit.

Comments (0)

Advertisement

X
You may login with either your assigned username or your e-mail address.
The password field is case sensitive.
Loading