Polysilicon to Wafers and Cells

Silicon High purity silicon ("Polysilicon") is the key feedstock for almost all solar cells and modules produced today. Silicon-based PV cells and modules accounted for 91% of all PV production in 2005, according to Photon Consulting. In unpurified form, silicon is widely abundant, making up more than 25% of the Earth's crust. While there is no fundamental limitation on the availability of unpurified silicon material, the standard "Siemens" purification process requires several costly steps, including the processing of volatile chemicals (Trichlorosilane) at 1,100 degrees Celsius. The solar PV industry and semiconductor manufacturers are the two main consumers of polysilicon. In 2000 the solar industry consumed only 10% of the world's silicon supply. In 2006 the PV industry consumed more than half of the world's available supply of polysilicon for the first time ever, according to Photon Consulting. This historic shift illustrates the growing size and importance of the solar PV industry, and polysilicon manufacturers are expanding production capacity dramatically to meet the growth in the PV industry. However, the time lag between planning new polysilicon capacity and the actual production of polysilicon is typically 1.5 to 2.5 years, thereby contributing to the current shortage of silicon supply. The impact of this shortfall is evident in the sharp rise in polysilicon prices since 2004.

Ingot Silicon consumers in the solar PV industry must convert silicon feedstock into silicon ingots to enable further processing into wafers, cells and modules.   Silicon-based solar modules fall into two categories: monocrystalline and multicrystalline. In each category, the polysilicon must be converted into a crystalline structure. A monocrystalline ingot is comprised of one large crystal structure, which yields a uniform color and texture throughout the ingot. A multicrystalline ingot contains numerous smaller silicon crystals and often has a mottled or flecked appearance. Trina Solar uses monocrystalline technology because this solution produces solar cells and modules with a higher efficiency conversion rate of sunlight to electricity. The most common technology used in the production of ingots for monocrystalline solar cells is based on a technique called the Czochralski Process. In this process a silicon seed crystal at the end of a metal rod is lowered into a quartz crucible of molten silicon liquid. As the rod and seed crystal are slowly pulled out of the crucible, a single cylindrical silicon crystal forms on the seed crystal. The production of monocrystalline ingot requires precise specifications and careful monitoring to ensure uniform crystal growth and contaminant-free ingots. Completing a single cylindrical silicon crystal ingot takes between 36 and 40 hours and yields an ingot of approximately 2 meters long and 6 to 8 inches in diameter.   Single Crystal Growing for Wafer Production (Page 1 of 2)              Integrated circuits are built on single-crystal silicon substrates that possess a high level of purity and perfection. Single-crystal silicon is used in VLSI fabrication instead of polycrystalline silicon since the former does not have defects associated with grain boundaries found in polysilicon. Such defects have been known to limit the lifetimes of minority carriers.    Aside from the need to be single-crystalline in nature, silicon substrates must also have a high degree of chemical purity, a high degree of crystalline perfection, and high structure uniformity. The acquisition of such high-grade starting silicon material involves two major steps: 1) refinement of raw material (such as quartzite, a type of sand) into electronic grade polycrystalline silicon (EGS) using a complex multi-stage process; and 2) growing of single-crystal silicon from this EGS either by Czochralski or Float Zone process.                  Czochralski Crystal Growth      Czochralski (CZ) crystal growth, so named in honor of its inventor, involves the crystalline solidification of atoms from a liquid phase at an interface.  The basic CZ crystal growing process is more or less still the same as what has been developed in the 1950's.      CZ crystal growing consists of the following steps. 1) A fused silica crucible is loaded with a charge of undoped EGS together with a precise amount of diluted silicon alloy. 2)  The gases inside the growth chamber are then evacuated. 3) The growth chamber is then back-filled with an inert gas to inhibit the entrance of atmospheric gases into the melt during crystal growing. 4)  The silicon charge inside the chamber is then melted (Si melting point = 1421 deg C). 5)  A slim seed of crystal silicon (5 mm dia. and 100-300 mm long) with precise orientation tolerances is introduced into the molten silicon. 6)  The seed crystal is then withdrawn at a very controlled rate. The seed crystal and the crucible are rotated in opposite directions while this withdrawal process occurs.              Continue to next page...                     Fig. 1.  Examples of Czochralski Pullers   Float Zone Crystal Growth    The float zone (FZ) process is another method for growing single-crystal silicon. It involves the passing of a molten zone through a polysilicon rod that approximately has the same dimensions as the final ingot.  The purity of an ingot produced by the FZ process is higher than that of an ingot produced by the CZ process. As such, devices that require ultrapure starting silicon substrates should use wafers produced using the FZ method.     The FZ process consists of the following steps. 1.  A polysilicon rod is mounted vertically inside a chamber, which may be under vacuum or filled with an inert gas. 2.  A needle-eye coil that can run through the rod is activated to provide RF power to the rod, melting a 2-cm long zone in the rod. This molten zone can be maintained in stable liquid form by the coil. 3. The coil is then moved through the rod, and the molten zone moves along with it. 4. The movement of the molten zone through the entire length of the rod purifies the rod and forms the near-perfect single crystal.           FZ growing equipment can also use a stationary coil, coupled with a mechanism that can move the silicon rod through it.                Fig. 2.  Examples of Float Zone Crystal Growing Equipment                                  After the single-crystal silicon ingot has been manufactured, it undergoes a routine evaluation of its resistivity, impurity content, crystal perfection, size and weight.  It is then ground using diamond wheels to make it a perfect cylinder that has the right diameter. It then undergoes an etching process to remove the mechanical imperfections left by the grinding process.        Fig. 3.  A Single-Crystal Silicon Rod                            The cylindrical ingot is then given one or more 'flats' by another round of grinding. The largest flat, called the primary flat, is used by automated wafer handling systems for alignment.  Flats (primary and secondary) are also used to identify the crystallographic orientation and conductivity of the wafer.     The ingot is then sawn into thin wafer slices, each of which will be subjected to further etching and polishing until it is ready for use as substrates for VLSI fabrication.  The above process of silicon growing, grinding, shaping, sawing, etching, and polishing to produce input wafers is known as wafering.   Fig. 4.  An ingot slicer (left) and a wafer grinder/polisher (right) Wafer Wafer sawing is the process of cutting the monocrystalline ingot into thin slices to enable the processing of silicon into solar cells. Producing thinner wafers and reducing silicon waste is a major area of focus in the solar industry's campaign to lower the cost of module production and ensure more efficient use of silicon. A large source of lost silicon is "kerf", the silicon dust produced during the sawing process. "Kerf loss" refers to the silicon removed from the ingot in the sawing process used to produce the wafers. Because the sawed grooves are approximately the same width as the produced wafers, kerf loss can approach 50% of the total silicon in the ingot. Wire sawing is the standard technique used to slice ingots into wafers. The raw ingot is first cooled and then the top and tail of the ingot are cut off and can later be reused in the ingot production process as reclaimable silicon. Next, the ingots are cut into 400-500 cm long sections and the cylindrical shape is 'squared' into four equal sides, so as to be mounted safely in the wire saw machine. In the sawing process a single strand of stainless steel wire hundreds of kilometers in length and 160 to 200 microns thick is pulled over the ingot by grooved rollers. To complete this process commonly, a mixture made up of oil and normally an abrasive material known as slurry is pumped over the wires to provide the friction needed for the cutting action. CROPPER Features – Designed to slice tops and tails of mono- and multi-crystalline silicon ingots – Cropper process uses a 250 µm-structured wire leading to a 350 µm kerf loss – Same machine base as wire saw. Key Benefits – Minimal silicon loss – High throughput – Possibility to crop mono- and multi-crystalline silicon SQUARER Features – Used to slice mono- and multi-crystalline ingots into pseudo-square ingots using a cross multi-wire web for most ingot sizes – Same machine base as wire saw Key Benefits – High silicon savings with a minimum kerf loss – Low consumables cost – High throughput – High work load WIRE SAW Features – E500SD-B5: slices mono and polycrystalline bricks into wafers Key Benefits – Production of thin & ultra thin wafers – High mechanical accuracy – High throughput – High load capacity