| Manufacturing the Microchip | ||
|
|
Silicon is a nonmetallic chemical element that has a crystalline structure much like a diamond and is able to form compounds with metals and nonmetals, which is why it is the main component to manufacturing microchips (Alacritude LLC, 2003). In the instance of manufacturing microchips, silicon is combined with wood and coke to form the silicon that is used in the production of microchips (Texas Instruments, 1996). The silicon used to manufacture microchips comes mainly from the sandy beaches that many of us do not view as being useful in the industrial world.
In 1947, the transistor was invented and in its second generation gave birth to the microchip (Schramm, 1988). This invention has been a major contributing factor to the last half century of technological inventions. Nowadays, microchips are used in a variety of ways, including CMOS chips, computer memories, audio amplifiers, and other linear devices (Hallmark, 1991). The information age continues to grow and the microchip remains the core of so many technology related products, making microchip manufacturing a seriously needed industry (Texas Instruments, 1996).
The manufacture of microchips can be broken down into seven common steps, related to all microchip production (Texas Instruments, 1996). These include silicon conversion, wafer fabrication, final cleaning, etching, layering, ion implantation, and all final steps leading up to the packaging of the microchip. Before these steps occur, though, the raw silicon has to be converted from sand to the uncontaminated silicon used in the process of manufacturing.
Raw silicon dioxide is heated with wood and coke to produce MGS silicon which contains about 1 impurity per 100 particles (Texas Instruments, 1996). The MGS silicon is then transformed into EGS silicon through many complicated gaseous phases to reduce the impurity count to one impurity per billion particles. These silicon particles are constantly tested, though impurity readings can be quite misleading because of the chemical makeup of EGS silicon. Once the silicon has reached an acceptable purity level, it is grown into a silicon ingot, or large cylinder, which will later be sliced into the silicon wafer boards. The silicon ingot is grown by exposing a small seed of purified silicon to a molten bath of silicon, reaching 1500-2000 degrees Celsius, rotating in opposite directions. The molten silicon must be very near the freezing point, though, as the molten silicon must attach to the seed to form the silicon ingot. Once the ingot has reached the desired size, it is removed and processed into wafers.
The large ingot of silicon is placed on a precision lathe to ensure that the silicon chunk is precisely the same size (Texas Instruments, 1996). The large ingot is then placed under X-ray observation to check for any impurities created in the silicon conversion process. A flat surface is then ground on one side of the silicon ingot, with water acting as the main lubricant. The ingot is then sliced into wafers as thin as 20 thousandths of an inch with a diamond wire saw. The wafers are tested, constantly checking for uniform flatness, resistance, thickness, and width. The wafers are lapped with aluminum slurry to improve surface flatness to +/- 3 microns. All damaged silicon is removed by chemical etching, using potassium hydroxide. The wafers are then polished and shipped to the wafer fabrication plant, or "wafer fab."
The wafer fabrication plant is where the wafers are converted from the raw silicon wafer to a computerized microchip (Texas Instruments, 1996). The fabrication plant is a very sterile environment, containing on average one particle of contaminant in the air per square foot. To put this into perspective, the average hospital operating room as an average of 10,000 times more contaminants in the air. People who work in these plants wear lint-free clothing, as well as special eye wear, shoes, and gloves to prevent contamination of the silicon wafers. If only a few particles get on the silicon wafers in the manufacturing process, the wafer is ruined and unusable.
An electronic film is placed on the wafer and then it is bathed in a combination of complex chemical acids to remove any minute contaminants (Texas Instruments, 1996). This is called deposition. Then, in a process called metallization, an insulating layer is placed on the silicon wafer, and microscopic pieces of metal are sputtered over the insulating layer. Next, gas flows over the insulating layer, diffusing any impurities and annealing the impurities back to a monocrystalline form.
The surface of the wafer is coated with photoresist, a carbon that is soluble when exposed to ultraviolet light (Intel Corporation, 2003). The wafer then undergoes a process called photolithography. This is where a mask is placed over the wafer, allowing ultraviolet light to reach the wafer where it is exposed. This creates a gooey surface of photoresist where the wafer was exposed to the UV light. Each layer of the microprocessor will have its own mask with a different pattern on it that is covered in photoresist and then exposed to UV light. The wafer of silicon now has a gooey surface of photoresist on it. The next step is to dip the wafer into a bath of developer to remove the photoresist (Texas Instruments, 1996). After the wafer is dipped into the developer, a pattern emerges, showing where the mask was on the silicon wafer (Intel Corporation, 2003). The wafer is then dipped into an acid bath, which triggers the chemical etching of the exposed silicon dioxide. After the exposed silicon dioxide is etched away, creating an intricate pattern on the wafer, the rest of the photoresist is removed, leaving ridges of silicon dioxide on the base of the wafer.
Since microprocessors require many intricate
designs, more layers have to be etche
The next step, called ion implantation, is a process in which the wafer is bombarded by a stream of ions through an ion implanter with precise targets (Texas Instruments, 1996). The ions are implanted into the silicon wafer to alter the way the silicon in the implanted areas conduct electricity, which is crucial in the functionality of the microchip (Intel Corporation, 2003). The ions actually alter the surface properties of the semiconductor surface, silicon dioxide (Case Technology, Inc., 1999). After the implantation of the ions occur on this layer, the next layer is created by repeating the steps described in the preceding two paragraphs. Then, atoms of metal are deposited in between the layers, creating electrical connections between the layers. This entire process is then replicated, often twenty times or more, depending on the complexity of the microchip, creating a three dimensional circuitry flowing throughout the silicon wafer.
Ion Implanter by Case Technology, Inc. Click on the picture to see more information on how the implanter works.
Because the majority of these processes are automated and controlled tightly by computers, many microchips are built on a single wafer (Intel Corporation, 2003). Each microchip needs to be tested and then cut apart from the other microchips on the same wafer.
The purpose of testing is obvious, in most regards. Everyone wants a microchip that functions correctly, and that includes the manufacturers. The purpose of testing by the manufacturers is to maintain the reliability and confidence that comes along with the name of that particular microchip manufacturer.
While still on the large wafer that contains many microchips, each one is tested to ensure the functionality of the microcircuitry that was created by the complex steps above. The microchips are tested by using a probe card, which has many contacts for testing all kinds of microchips built for various devices (Texas Instruments, 2003). The probe card tests every microchip on the wafer, and then after passing this test, the microchips are cut off of the wafer using an intricate diamond saw. The microchips are then examined by humans using an atomic microscope to ensure that there are no visible problems that could affect the performance or packaging of the microchip. The separated microchips are then placed onto a lead frame where they are bonded to miniscule gold wires, finallizing the circuitry and functionality of the microchip. Next, the microchip is covered in a plastic epoxy, which is the final step in the actual production of the product. Next, the finished product is inspected for visible defects and compliance with codes relevant to the product the microchip will be used in. Finally, the microchip is packaged and shipped to its buyer.
The manufacture of the microchip entails many complex steps, which are very hard to describe to the every-day information seeker. Hopefully by viewing this research site, one will have a better understanding of how these miniscule products go from being sand on a beach to the backbone of many computers, cell phones, televisions, calculators, and countless other household products.
Key terms and definitions: coke (noun) - The solid residue of impure carbon obtained from bituminous coal and other carbonaceous materials after removal of volatile material by destructive distillation. It is used as a fuel and in making steel (Lexico Publishing Group, LLC, 2003). MGS silicon (noun) - Metallurgical grade silicon, which is about 98% pure (Wilson, 2003). EGS silicon (noun) - A pure form of silicon which is converted into a single silicon crystal (Wilson, 2003). polysilicon (noun) - Silicon with a crystalline structure, which acts as a conductor of electricity. It is used as the gate in MOS transistors as well as an interconnect between them (The Computer Language Company, Inc., 2003).
Alacritude LLC. (2003). silicon. Retrieved December 1, 2003, from Encyclopedia.com Web Site: http://www.encyclopedia.com/html/s1/silicon.asp Texas Instruments Incorporated. (1996). Making of a microchip: the process of manufacturing an integrated circuit [Videorecording]. Dallas, TX: Texas Instruments Incorporated. Schramm, W.L. (1988). The story of human communication: cave painting to microchip. New York: Harper & Row. Hallmark, C.L. (1991). The master IC cookbook. Blue Ridge Summit, PA: Tab Books. Intel Corporation. (2003). Intel Education: Learning About Technology: How Chips are Made. Retrieved November 30, 2003, from the Intel Corporation Website: http://www.intel.com/education/makingchips/index.htm Case Technology, Inc. (2003). The Ion Implanter. Retrieved November 30, 2003, from the Case Technology, Inc. Website: http://www.casetechnology.com/implanter/intro.html Lexico Publishing Group, LLC. (2003) coke. Retrieved December 2, 2003, from the Dictionary.com Website: http://dictionary.reference.com/search?q=coke Wilson, B. (2003). Silicon Growth. Retrieved December 2, 2003, from Rice University, Connexions Web Site: http://cnx.rice.edu/content/m1033/latest/ The Computer Language Company. (2003). TechEncyclopedia. Retrieved December 2, 2003, from TechWeb Web Site: http://www.techweb.com/encyclopedia/defineterm?term=polysilicon
|
|
| Created by Troy Stackhouse Last Modified: 12/04/2003 11:39 AM -0500 |
