14 March 2016
MOSFET Transistors- Materials
Transistors are an important part of today's world. They are in most of the electronic devices people use on a daily basis. Without them, different advancements in technology would not be possible. However, because of transistors, large amounts of materials are being thrown into dumps and polluting the Earth. This is further explored in the next two papers of the trilogy. To create a tiny transistor takes a massive amount of materials, and as time goes on, the processes become more involved and wasteful. Trying to salvage resources by recycling them is difficult and expensive because the transistor is so small. Analyzing all of the materials used in the creation of a MOSFET transistor shows that this seemingly simple switch is actually very complicated and industry is constantly finding better ways to make it smaller and more efficient.
At its most basic level, a transistor is a switch. It was thought of by Julius Edgar Lilienfield in 1925, but it was not until 1960 that the first MOSFET, by Dawon Kahng and Martin Atalla, became available (Emmerson). It took so long to create it because scientists needed the elements with the right qualities to create the desired effect. Also, these materials needed to very pure. They found their solution in silicon. Mechanically, a transistor is a switch that allows positive or negative current to go through. There are two fundamental types, an n-channel and a p-channel. They are structurally the same, except the location of the p-doped material and n-doped material in an n-channel is opposite to that of a p-channel transistor. In an n-channel, there is are two n-terminals surrounded by p-doped substrate. A gate is placed on top of the gap between the two terminals. When positive voltage is applied to the gate, a path is created to connect them. This is because the positive voltage attracts negative electrons, creating a mostly negative charge between the terminals. If negative voltage is applied to the gate, nothing happens as the path between the negative terminals is obscured by positive charges. This is analogous to switching a kitchen light on, except instead of moving the switch up and down to turn it on and off, the voltage is changed from positive to negative. Unlike the kitchen switch, the amount of current that is allowed through is dictated by the amount of voltage that is applied. The more positive voltage applied to the gate in an n-channel transistor will attract more electrons. This makes the path wider and allows more current to go through (What is a MOSFET).
Fortunately, the electronic industry is using a fairly common resource to make most of the transistor. Silicon dioxide or silica makes up 27% of the earth’s crust (Silica). An integral part of human culture is created from something people unknowingly walk on every day. The silica used in semiconductors must be extremely pure, so industrial sand is used. It is mined from hard rock quartzite formations and untouched beds of silica sand. The mining site must be at least 95% silicon dioxide (Industrial Sand). These mining sites can be found all around the world including Australia, Wisconsin, Minnesota and Europe. The basic process includes pulling the silica out of the mining site and feeding it into a processing plant to remove any impurities. Then the quartz is separated out into different sizes based on where the silica is being sent to (Introduction of Silica Sand Production). The purest silica is sent to polysilicon plants to become even purer and eventually become part of a transistor. This process creates a lot of air pollution as tiny dust particles are being flung into the air during the mining process (Silica a Sand Mining). To limit the health risks, the mining site is constantly dampened with water. More materials used in mining silicon include gas and machinery needed to mine and transport the raw material to polysilicon plants where it will be processed.
As new technology presents itself, the materials that are used in the transistor are replaced by more efficient ones. Engineers are constantly finding new ways to make it smaller as this equates to most other electronic devices becoming smaller as well. This makes finding exactly what is inside a transistor difficult. The three main components are the terminal, surrounding material and gate. Originally, the substrate material is made with silicon crystal doped with some element. Depending on the element the silicon is doped with, the transistor is n-channel or p-channel. In a p-channel transistor, the terminals are silicon crystal doped with boron, and the surrounding material is doped with phosphorus. However, the terminals can be doped with anything that has three valence electrons and the surrounding material can be doped with anything that has five valence electrons as long as it is small enough to fit into the silicon structure. Different elements have been favored over the century based on performance and cost. Engineers make the terminals and surrounding material mainly from silicon crystal because of its semiconductor qualities. There are other materials that do similar things and are even better in some respects but silicon provides the correct combination to make more circuitry work. For example, gallium arsenide is a better conduction than silicon, but it does not have the same insulator qualities (Transistors). However, there are a lot of resources devoted to find some chemical that is cheaper and better than silicon. As of now, Intel and IBM have started using silicon germanium in MOSFET channels, driving the industry to yet another technological advancement. Finally, the gate has two components, an insulator and a conductor. In the beginning the conductor was a metal gate made out of aluminum, but as time went on, researchers found that polysilicon was much better. In both cases, the insulator was silicon dioxide. As transistors become smaller, the silicon dioxide layer shows to be an inadequate insulator, so industry has shifted to using high-k dielectrics. There are many materials that could be researched to analyze a transistor, however, because companies are not willing to divulge their secrets, it is difficult to find information on recent developments. Therefore, older technology will be discussed. This assumes the terminals and surrounding material are silicon crystal doped with boron and phosphorus and the gate is silicon dioxide and polysilicon. (Sah, Chih-Tang)
Polysilicon is a large portion of the transistor. It is used as the gate and, once properly crystalize, in the n and p doped portions. Because the polysilicon must be 99.9999999 % pure, the silicon must be extracted and refined. There are several processes to do this, but the following is a description of the most common one. First the mined silicon dioxide is combined with twice as much carbon and heated in an electric furnace to 1500-2000 degrees Celsius. If the sand is pure enough, this creates a brown metallurgical-grade silicon which is 97-98.5% silicon. Then powder mixture is dissolved in hydrochloric acid to create the gas trichlorosilane to remove impurities such as iron, aluminum and boron. The trichlorsiliane is heated and diluted with hydrogen gas so that it will turn back into silicon (Manufacturing Silicon). Now it has been purified enough to be used as a gate to the transistor, though it needs to be reformed to become part of the substrate and terminal (Calvert).
The body of the chip is made out of silicon crystal. To do this industry takes the very pure polysilicon and changes the molecular structure samples that have a singular crystal orientation. To do this, the polysilicon is broken into pieces and melted down in quartz lined furnace at 1420 degrees Celsius. If the transistor is an n-channel, the silicon is exposed to boron in this step. Otherwise, it is exposed to phosphorous. Then a monocrystalline silicon seed is inserted into the furnace. As the molten silicon is rotated one way, the seed is rotated the other way. This creates a round cylinder. The silicon sticks to the seed as it cools and when the diameter becomes large enough, the seed is very slowly drawn out of the furnace (Calvert). Once the silicon has been pulled out, it is placed in another furnace to slowly cool for seven hours. This allows the structure to become more stable (Manufacturing Silicon). The very brittle silicon crystal is sanded down so the diameter of the cylinder complies with the standard. Then the crystal is cut into very thin slices using spools of wire so that a fresh piece of wire is always in contact with the block. This creates an imperfect surface, so the wafers have to go through surface polishing. The silicon is now ready for etching and photolithography. (How Do They Make Silicon Wafers and Computer Chips?)
After the silicon crystal is sliced into wafers and the insulator is added to the chip. Ironically, the insulator is often silicon dioxide. This is the same chemical formula as the raw material, but due to the transistor’s size engineers cannot put sand on the transistor. To create a very thin layer, manufacturers used thermal oxidation where the heated silicon crystal comes in contact with oxygen. This can be done two ways. The first is with pure oxygen. The second is with steam that produces hydrogen gas as a byproduct. Sometimes hydrogen chloride is also used in conjunction to the other two methods to clean any metal ions that may be in the silicon. This creates a coating of a couple nanometers thin. These reactions take place in furnaces at 800-1200 degrees Celsius. Because of the high temperatures this reaction requires, creating the silica dioxide always takes place before the doping of the source and drain terminals (Vanheusden). After this step, the process of photolithography begins (Calvert).
Engineers use photolithography and etching to imprint the materials onto all silicon based chips. This makes the different components. The etching uses bases and acids to eat through different layers and the photolithography uses polymers. This process has changed over time, just like the processes of the other components, but it is analogous to an artist using spray paint. They would use tape to cover up areas that the artist does not want to spray paint over. First the insulator is applied to the chip. Then polymers are placed on the chip to allow etching through the silicon dioxide to the silicon crystal (Photolithography). In the analogy, this is like a painter covering the area with tape. Only the area not covered with chemicals from the photolithography are affected when the area is exposed to phosphorous to create an n-doped region. More layers are added and taken off to allow the conductor to be deposited precisely between the terminals. Finally, all layers and chemicals from the etching and photolithography are removed. Depending on the process, there can be more than twenty different layers applied to the transistor (Calvert).
Assuming the transistor is an n-channel, the terminals are exposed to phosphorous at high temperatures to diffuse the element into the silicon. Depending on the temperature, the element will bond with the silicon at different rates. Phosphorous is chosen since it has five valence electrons. Because silicon has one less valence electron than phosphorous, one electron from the silicon will be free to more around. This creates extra electrons in the lattice structure allowing negative charges to travel through this section of the silicon. The p-type part of the transistor is created with a similar, yet different idea in mind. The charges of the n and p-doped areas must have opposite charges to develop the channel between the terminals. Therefore, a lack of an electron instead of an extra one is needed. This means the element in the p doped are must have three valence electrons so that there are three bonds with the silicon crystal and a hole. This creates a positive overall charge. Typically, the element used is boron. This is why silicon is so important to modern electronics. With only a small change in the crystal structure, silicon can go from being a very good insulator to a conductor of current. It allows these elements to change small parts of the crystalline structure (Calvert). Aluminum and gold wiring is placed over the terminals. The tiny transistor is complete.
Because transistors are such a necessary part of most electronics, they are transported all over the world to go into computers everywhere. Those computers are then shipped everywhere else. It takes a lot of fossil fuels. Because they are so small and sensitive, the individual transistors are shipped in anti-static or static shielding bags. Anti-static bags are typically made out of plastic and aluminum. The Aluminum is embedded in the plastic and is hardly noticeable. Its purpose is to create a cage so that any charge that the bag comes in contact with is quickly dissipated through the aluminum. Static shielding bags are made from polyethylene and are designed to hold the charges in the bag instead of dissipating them. Both protect the transistors form electric shock and moisture (How to Compare Anti-Static Bags to Static Shielding Bags).
Since a transistor is so small and permanently installed into a computer; once the computer stops working, the transistor is no longer useable either. In this way, it is not feasible to re-use the transistor once the part the transistor has broken. Similarly, one cannot perform maintenance on a transistor either. Transistor components can be reused in breadboards as different components are placed for testing and they can be taken out. There is protocol for recycling the product. This entails removing heavy metals from the product as stated in the waste paper of this product. Unfortunately, many people do not follow these procedures and the metals go into the ground where they litter the environment. This is because it is expensive to remove the part as people mostly do it by hand and only have a small amount of material taken out of them. As for making usable products out of the bi-products of the process, there are none to date. Though very useful to expanding human capabilities, a transistor is not the best thing for the environment (Kavitha).
To conclude, transistors, though used as a simple switch, are very complicated to make. Even the older technologies need very pure materials to be successfully created. This requires a lot of materials that do not go into the final product and produces a slew of waste. Yet, it is a necessary part of people’s lives and productivity would drastically decrease without transistors. Technology is stripping the earth of its precious natural resources, but as it progresses, maybe it can help spark human ingenuity and find solutions to environmental issues it is creating.
Calvert. "Silicon." Physics. N.p., 15 Sept. 2007. Web. 13 Mar. 2016. <http://mysite.du.edu/~jcalvert/phys/silicon.htm>.
Emmerson, Andrew. "Who Really Invented the Transistor." The Porticus Centre. Beatrice, n.d. Web. 13 Mar. 2016. <http://www.beatriceco.com/bti/porticus/bell/belllabs_transistor1.html>.
How Do They make Silicon Wafers and Computer Chips. YouTube. N.p., 5 Mar. 2008. Web. 13 Mar. 2016 <http://www.youtube.com/watch?v=aWVywhzuHnQ>.
"How to Compare Anti Static Bags to Static Shielding Bags." Bennett & Bennett. Bennett & Bennett, n.d. Web. 13 Mar. 2016. <http://www.bennettnbennett.com/showResource.cfm?id=15>.
Introduction of Silica Sand Production. Tochu. Tochu Corporation, 2010. Web. 3 Mar. 2016. <http://www.tochu.com/english/sand/>.
Kavitha, A. V. "EXTRACTION OF PRECIOUS METALS FROM E-WASTE."Journal of Chemical and Pharmaceutical Sciences ISSN 974: 2115.
"Manufacturing Silicon." Green Rhino Energy. Green Rhino Energy, n.d. Web. 13 Mar. 2016. <http://www.greenrhinoenergy.com/solar/technologies/pv_manufacturing.php>.
“Photolithography” Wikipedia: The Free Encyclopedia. Wikipedia Foundation, Inc. 11 March 2016. Web. 13 March 2016.
Sah, Chih-Tang. "Evolution of the MOS transistor-from conception to VLSI."Proceedings of the IEEE 76.10 (1988): 1280- 1326.
"Silica." Mineral Company. Farzin Rock Stone Company, n.d. Web. 3 Mar. 2016. <http://www.mineralco.net/silica/index.php>.
Silica Sand Mining. Minnesota Polution Control Agency. N.p., n.d. Web. 3 Mar. 2016. <https://www.pca.state.mn.us/air/silica-sand-mining>.
“Transistor” Wikipedia: The Free Encyclopedia. Wikipedia Foundation, Inc. 13 March 2016. Web. 13 March 2016.
Vanheusden, K., et al. "Non-volatile memory device based on mobile protons in SiO2 thin films." (1997): 587-589.
"What is a MOSFET, What Does it Look Like, and How Does it Work?" Techpowerup. Techpowerup, n.d. Web. 13 Mar. 2016. <https://www.techpowerup.com/articles/overclocking/voltmods/21>.
"What Is Industrial Sand." National Industrial Sand Association. National Industrial Sand Association, 2011. Web. 3 Mar. 2016. <http://www.sand.org/what-is-industrial-sand>.
14 March 2016
The Energy Story
The global landscape has been dramatically altered by the explosive growth of digital technology, arguably one of the greatest innovations in recent human history. At the heart of this amazing technology lies the humble transistor, which, when working together by the millions, gives us everything from kitchen timers to supercomputers. Computers operate and communicate using binary numbers, and the transistor, by its ability to be either on or off, provides the needed binary states. To cram more and more transistors into ever more complex devices, materials engineers are faced with the challenge of shrinking these switches in a continual battle with the limits imposed by processing technology and the laws of physics. These are closely guarded secrets – the materials used, the processing technologies, the energy consumed in their fabrication. However, recent studies do offer a glimpse into the energy required to fabricate and operate Field Effect Transistors (FET’s), which are made on silicon wafers, highlighting the incredible consumption footprint left by the lifecycle of this technology.
Before beginning a rigorous investigation into the actual energy calculations, it is helpful to first look at the lifecycle as a whole. The story of the transistor begins in the earth, where the main material silicon, is extracted in the form of silica, more commonly known as quartz. Through chemical processing and massive energy input, the silicon is purified and grown into a large crystal called an ingot, which is then sliced into thin wafers. Once sliced, the wafer is treated appropriately with impurities to adjust the conductivity properties in specific locations, and layers of other materials are applied. A special coating, known as a photoresist, is then applied, and the wafer undergoes the process of photolithography where the schematic of the transistor is “drawn” on. Once the drawing is complete, it is permanently etched in, and finally all that artwork is packaged into a protective housing, giving birth to the transistor. During operation, the transistor will consume electricity, until its eventual death, upon which it will be returned to the earth (Ho; Decker; Iles). All these stages contribute to the energy footprint of a transistor, adding layers to an energy iceberg that serves as a great model when looking at the lifecycle of this device. The process of building these transistors is complex, and calculations can be very cumbersome, but understanding the energy consumption of this technology is important. Using simple models, and focusing solely on the steps, the iceberg’s size can be roughly estimated. The units used in this paper are kilowatt-Hours, one of the most typical units when describing large values.
Like many products, semiconductors trace their origin back to minerals in the earth. Silicon is extracted from one of the most abundant materials, silica, a process that requires 0.2 kWh/kg of energy (“Silicon Wafer Fabrication”; Stappato 226). Even though silica is so plentiful, the mineral is usually found in an amalgamation of other particles, making purification next to impossible (Calvert). Thus, ore with high concentrations of crystalline silica, or quartz, is required, which forms in areas of the Earth where molten rock low in aluminum content can slowly solidify, most commonly Africa’s Great Rift Valley. (Calvert). One of the largest ports in this particular region is Mombasa, Kenya (“Africa’s Ports”), and assuming the silica is transported 10,900 miles via container ship running on diesel to the port of New Orleans, the total energy required for shipping per year is 176,000,000 kWh (“Distance Calculator”; Elert; “Fuel Efficiency”). In order to produce metallurgic grade silicon, silica is reacted with carbon to produce silicon and carbon dioxide, utilizing the chemical energy found in carbon (Honsberg). Using standard enthalpy of formation calculations, the chemical energy involved is negligible, the support processes and environmental control consuming the bulk of the energy (Honsberg). Most of this energy is consumed by the transportation of the carbon sources required for the process, as well as the thermal energy required to drive the reaction. Considering that 234,000 tonnes of silicon was extracted for use in semiconductor production in the year 2013, energy totals now rise to 3.92 billion kWh/year, enough to power 343,486 American homes, which consume 10,900 kWh, for a year (Honsberg; “FAQ”). However this is only the energy required to produce metallurgic grade silicon which is only 98% pure; further energy is required to purify silicon until it is 99.999999999% pure, or 9N. On the iceberg of energy, only the tip has been revealed, the rest lurking in the depths below.
As stated before, metallurgic grade silicon does not possess the purity required for transistor fabrication and must undergo further chemical processing (“Silicon Wafer Fabrication”). The silicon is treated with hydrochloric acid to remove impurities of boron, aluminum, and iron which form their respective halides when exposed to the chlorine anion (Honsberg). An estimation of the chemical energy used can be calculated using the methods of enthalpy formation for the halides mentioned earlier. Assuming the silicon is 98% pure, and an even distribution of impurities, the calculations for the chemical energy yields 0.136 kWh/kg, which pales in comparison to the electrical input. This example demonstrates an important pattern, that the chemical energy consumed during semiconductor production is insignificant in comparison with other forms. Henceforth, chemical energy will not be considered in calculations, but it is an important form of energy. Once the silicon is purified, it must be grown into a monocrystalline structure, or ingot (“Silicon Wafer Fabrication”). During this crucial step, thermal and mechanical energy are used to melt the polycrystalline silicon and to spin the molten silicon around to produce the ingot as the molten silicon cools. Once these crystals have been synthesized, they must be sliced into wafers, requiring the use of electric saws (“Silicon Wafer Fabrication”). The energy values associated with purification and wafer creation are substantially higher than the energy to convert silica to metallurgic grade silicon, totaling 498 billion kWh/year or the power consumed by 68.5 million homes for a year (Williams 5507). Often times the purification of materials is excluded from a lifecycle analysis, but as shown, the purification process consumes nearly 160 times the amount of energy used to simply obtain silicon, making it an important factor in the lifecycle. The next phase of wafer processing, which involves building the transistor, requires even more energy, diving deep into the water, exposing more of the iceberg.
The chip fabrication process, photolithography, etching and packaging, consumes the most energy after wafer fabrication in the lifecycle of any semiconductor transistor product (Ciceri, Garetti 2). During this phase, primarily electricity is consumed, split between environmental control, lithography processes, etching, and packaging. Environmental controls are responsible for the majority of the energy used, for they run continuously, independent of the quantity of wafers being processed (Branham 4298). Examples of such systems are water distillation, vacuum chambers, and air conditioning to cool down the assembly process (Branham 4298). Another process that consumes large quantities of energy is the commonly used photolithography process, which requires the use of focused high energy ultraviolet light to pattern the surface of the photoresist to make the traces and transistor mounts. The photoresist is vapor deposited onto the wafer surface, usually in multiple layers as the wafer is spun (Malloy 1). The form of energy associated with this step is electromagnetic energy, as the photons from the UV light carve the photoresist (Malloy 1). This process also requires very pure water, leading to more energy consumption (Murphy). Distillation of water accounts for 9% of the energy consumed by the entire process itself, making it a major consumer of energy (Branham 4298). The most common practice in use is reverse osmosis, and often water produced from this process is further conditioned with electrodeionization, requiring electric energy to remove any traces salts (Lui 5). During reverse osmosis water is pushed through a semipermeable membrane, requiring massive quantities of mechanical energy to generate substantial pressure, using electric pumps (Liu 5). Once the photolithography process is complete, the next phase of etching can commence. There are two primary methods of etching, wet and dry, both of which remove material based on the pattern outlined during the lithography process (Jansen et al. 14). Wet etching heavily relies on chemical removal of the silicon oxide layers using the chemical energy found in extremely corrosive acids and caustic bases (Kulkami 176). This particular process uses the diffusion of the chemical through the silicon as the main driving force for removal, yet another form of chemical energy. Dry etching requires a combination of radiofrequency energy as well as ion energy (Jansen et al. 15). This method has been growing more popular with industry due to its increased versatility and efficiency compared to wet etching. Once etching is complete, the vapor deposition of the traces can be accomplished, completing the fabrication of the silicon part of the chip. The final step involves packaging of the silicon chips into the protective cases to form the integrated circuit. Unfortunately, the efficiency of all these processes is shockingly low, ranging between 10-3 and 10-6 thermodynamically speaking (Branham 4299). The total energy consumed during this phase comes to 1242 billion kWh/year, or enough to power 114 million homes for a year. A majority of the entire embodied energy of the product is revealed, leaving only the usage and disposal energy left to explore on the energy iceberg.
One area where the embodied energy analysis has a major gap is the distribution phase. Most silicon devices are sold by the companies that manufacture them, however, there are other retailers that sell devices in bulk, such as Digi Key, but from personal experience the distribution means of both modes is essentially the same. Due to the sheer number of devices that are transported around the world on a daily basis, it is arduous to attempt an energy calculation without incredible simplifications of the model. It is reasonable to assume the main mode of transport is truck or aircraft, considering that most silicon devices are ordered through parcel services, thus the primary energy source is fossil fuel. However it is also reasonable to assume the energy contribution of this step is insignificant, as the energy required to haul the silica from Kenya to the United States was 176 million kWh to go 10,900 miles, a number that is 3 orders of magnitude lower than the entire acquisition step (“Distance Calculator”; Elert; “Fuel Efficiency”). The iceberg has a fingernail’s thickness of ice at this level.
Once these silicon chips are made and delivered, they are placed on circuit boards and used in every computing device produced for consumer use. Every day, millions worldwide use these chips in their phones, computers, and even vehicles, consuming energy nearly continuously. One prime example is the internet, which lives in a series of severs that function on silicon transistors and run continuously, consuming electricity as their main source of energy (Walsh). As the internet becomes more and more prevalent the energy consumption of this technology will only increase. One website alone had a carbon footprint of 10,000 tonnes of carbon dioxide, equaling 42.8 million kWh (“Digital Carbon Footprint”). Most of the energy comes from the personal devices subscribers use to access the website, thus one can conclude that consumer devices are responsible for the majority of the energy footprint of the internet. Many computers do not actually consume massive quantities of energy when in use, and during the lifetime of a typical device, will use approximately a twelfth of the energy of fabrication (Decker). Following this model, assuming energy use is distributed evenly across all components and summing the energy values discussed in the previous paragraphs, transistors will use 145 billion kWh/year, enough to power 13.3 million homes. This is a unique property to semiconductor devices, for most products will consume more energy during their use (Decker). This can create a misconception about the amount of energy actually being used by electronic devices. Many modern devices contain CPU’s that run at approximately fifteen watts, which seems like a rather small number compared to many other appliances like microwaves and automobiles. This is why, when it comes to energy, the usage phase of this product doesn’t communicate the sheer amount of energy actually used. In fact, the energy that goes into manufacturing the memory of a laptop is sufficient to operate the machine for 500 to 1000 days continuously (Decker). However, digital technology has allowed for many other processes to become much more efficient, leading to less energy and material consumption. This presents an energy tug of war, where the energy of manufacturing digital technology is weighed against the energy saved by implementing these technologies into other products like cars. At this point in the life cycle, the iceberg of energy has nearly been completely exposed, and only the waste and recycling energy remain.
Once silicon devices approach the end of their lives, they are thrown away or recycled and through waste management, return to their starting place, the Earth. When semiconductor devices like computers are recycled, the transistors themselves are often not the parts recycled, thus the lump sum of energy is lost (Iles 85). However, some of the precious heavy metals are in fact extracted and re-used in other product lifecycles, transferring some of the energy elsewhere (Kavitha 147). Most electronic waste is shipped to Los Angeles, where it is redirected halfway around the world to Dubai in the United Arab Emirates, where it is once again shipped to Asia, traveling a total average distance of 19,254 miles equaling 310 million kWh/year (Iles 78; “Distance Calculator”; Elert; “Fuel Efficiency”). When these components are recycled they are mainly processed through human labor, and according to a study done by the Silicon Valley Toxics Coalition, approximately one million rural people were employed in the industry in both India and China bringing the workforce up to two million (Iles 84). Assuming a 70 hour work week, a 52 week year, and an average power output of 100W per person, the total energy associated with this human labor comes to 728 million kWh, bringing the total energy calculation for this phase of the lifecycle to 1.083 billion kWh. It is important to keep in mind that this is an extremely conservative estimate, as hauling on land was not considered in the energy calculation. With this final calculation, the massive iceberg of energy has been completely revealed.
Silicon technology has revolutionized the world, increasing access to information and boosting productivity, but at a hefty energy price. When all stages of the lifecycle of these wafer based transistors are accounted for, the iceberg of energy weighs in at 1890 billion kWh per year, enough power to keep 173 million US residential household running. This enormous amount of energy comes mainly from electricity, which is a secondary source derived from the burning of fossil fuels. With the amount of semiconductor devices predicted to double by the year 2030, society is faced with a daunting task of being able to continue to supply the massive energy required to produce one of the physically smallest innovations in human history. The iceberg will only continue to grow, and society must find a way to support such a large burden to ensure future generations’ prosperity.
"Africa’s Biggest Shipping Ports." BusinessTech. BusinessTech, 8 Mar. 2015. Web. 28 Feb. 2016. <http://businesstech.co.za/news/general/81995/africas-biggest-shipping-ports/>.
Branham, Matthew S., and Timothy G. Gutowski. "Deconstructing energy use in microelectronics manufacturing: an experimental case study of a mems fabrication facility." Environmental science & technology 44.11 (2010): 4295-4301.
Calvert, J. B. "Silicon." Silicon. N.p., n.d. Web. 24 Feb. 2016. <http://mysite.du.edu/~jcalvert/phys/silicon.htm>.
Ciceri, N. Duque, T. G. Gutowski, and M. Garetti. "A tool to estimate materials and manufacturing energy for a product." Sustainable Systems and Technology (ISSST), 2010 IEEE International Symposium on. IEEE, 2010.
Decker, Kris D. "The Monster Footprint of Digital Technology." Low-Tech Magazine. Low-Tech Magazine, 16 June 2009. Web. 25 Feb. 2016. <http://www.lowtechmagazine.com/2009/06/embodied-energy-of-digital-technology.html>.
"Digital Carbon Footprint: Steps in the Right Direction." The Guardian. Guardian News and Media, 03 Oct. 2012. Web. 18 Feb. 2016. <http://www.theguardian.com/sustainability/sustainability-report-2012-digital-carbon-footprint>.
Elert, Glenn. "Energy Density of Diesel Fuel." Energy Density of Diesel Fuel. The Physics Factbook, 2006. Web. 03 Mar. 2016. <http://hypertextbook.com/facts/2006/TatyanaNektalova.shtml>.
"Fuel Efficiency: Modes of Transportation Ranked By MPG." True Cost Analyzing Our Economy Government Policy and Society through the Lens of Costbenefit. N.p., 27 May 2010. Web. 02 Mar. 2016. <http://truecostblog.com/2010/05/27/fuel-efficiency-modes-of-transportation-ranked-by-mpg/>.
"Frequently Asked Questions." How Much Electricity Does an American Home Use? US Energy Information Administration, 21 Oct. 2015. Web. 23 Feb. 2016. <http://www.eia.gov/tools/faqs/faq.cfm?id=97&t=3>.
Ho, Joshua. "An Introduction to Semiconductor Physics, Technology, and Industry." AnandTech. Purch, 9 Oct. 2014. Web. 10 Mar. 2016. <http://www.anandtech.com/show/8223/an-introduction-to-semiconductor-physics-technology-and-industry>.
Honsberg, Chistiana. "Refining Silicon." Refining Silicon. PV Education, n.d. Web. 03 Feb. 2016. <http://www.pveducation.org/pvcdrom/manufacturing/refining-silicon>
Iles, Alastair. "Mapping environmental justice in technology flows: Computer waste impacts in Asia." Global Environmental Politics 4.4 (2004): 76-107.
Jansen, Henri, et al. "A survey on the reactive ion etching of silicon in microtechnology." Journal of micromechanics and microengineering 6.1 (1996): 14.
Kavitha, A. V. "EXTRACTION OF PRECIOUS METALS FROM E-WASTE."Journal of Chemical and Pharmaceutical Sciences ISSN 974: 2115.
Kulkarni, Milind S., and Henry F. Erk. "Acid‐Based Etching of Silicon Wafers: Mass‐Transfer and Kinetic Effects." Journal of the electrochemical Society 147.1 (2000): 176-188.
Liu, Chuanfeng. "Polygeneration of Electricity, Heat and Ultrapure Water for the Semiconductor Industry." Thesis. Department of Energy Technology at the Royal Institute of Technology, 2004. Print.
Malloy, Matt, and Lloyd C. Litt. "Technology review and assessment of nanoimprint lithography for semiconductor and patterned media manufacturing." Journal of Micro/Nanolithography, MEMS, and MOEMS 10.3 (2011): 032001-1 - 032001-13.
"Metallurgical-grade Silicon Making Inroads in PV." Metallurgical-grade Silicon Making Inroads in PV. Semi, n.d. Web. 03 Feb. 2016. <http://www.semi.org/en/IndustrySegments/Materials/p044005>.
Murphy, Cynthia F., et al. "Development of parametric material, energy, and emission inventories for wafer fabrication in the semiconductor industry."Environmental Science & Technology 37.23 (2003): 5373-5382.
"Silicon Wafer Fabrication Process." Silicon Wafer Fabrication Process. Brigham Young University, Department of Electrical & Computer Engineering, n.d. Web. 22 Feb. 2016. <http://www.cleanroom.byu.edu/EW_formation.phtml>.
"Standard Enthalpy of Change of Formation." Wikipedia. Wikimedia Foundation, 7 Feb. 2016. Web. 23 Feb. 2016. <https://en.wikipedia.org/wiki/Standard_enthalpy_change_of_formation_(data_table)>.
Stoppato, A. "Life cycle assessment of photovoltaic electricity generation."Energy 33.2 (2008): 224-232.
"The Monster Footprint of Digital Technology." 'LOW-TECH MAGAZINE' Low-Tech Magazine, 16 June 2009. Web. 03 Feb. 2016. <http://www.lowtechmagazine.com/2009/06/embodied-energy-of-digital-technology.html>.
"Transit Time, Distance Calculator & Port to Port Distances." SeaRates. N.p., n.d. Web. 28 Feb. 2016. <https://www.searates.com/reference/portdistance/?A=ChIJfSzAjucSQBgRWtvQNbyLYcs&D=10327&G=16569&>.
Williams, Eric D., Robert U. Ayres, and Miriam Heller. "The 1.7 kilogram microchip: energy and material use in the production of semiconductor devices." Environmental science & technology 36.24 (2002): 5504-5510.
Walsh, Bryan. "The Surprisingly Large Energy Footprint of the Digital Economy [UPDATE] | TIME.com." Science Space The Surprisingly Large Energy Footprint of the Digital Economy. Time Magazine, 14 Aug. 2013. Web. 18 Feb. 2016. <http://science.time.com/2013/08/14/power-drain-the-digital-cloud-is-using-more-energy-than-you-think/>.
14 March 2016
Waste Products of MOSFETs: There’s A Lot
Building a MOSFET, also known as a Metal Oxide Semiconductor Field Effect Transistor, requires an extremely high amount of purity and precision during every aspect of the process. Each step of this process produces high amounts of waste products and emissions. Although transistors are a necessity in today’s society, the amount of waste produced in their life cycle, throughout its entire lifetime, is immense. The negative effect of the wastes generated by their production is enough to bring to attention whether or not their advantages outweigh their impact to the environment.
The first stage of building a MOSFET is obtaining the ultra-pure electronic grade silicon wafer that is used as the base material for the MOSFET. Pure silicon is not found naturally in nature; it has to be obtained from mining quartz (High Purity Quartz 1). Quartz is found in rocks like granite, gneiss, and sandstone. It is acquired by open pit mining using tractors and other big excavation machines. Explosives are used sparingly due to quartz being damaged by extensive temperature change. Due to there being low concentrations of needed minerals in the ore, high amounts of solid waste is generated. Mining generates an incredible amount of solid waste products such as waste rocks and dust. Moreover, due to there being so much waste, it is often disposed of at very low costs in order to save money. This causes further damage to the environment. For example, waste rocks may be displaced to nearby canyons or just left in the mines after mining is done. This leads to extensive erosion of the soil, drying of the soil, loss of nutrients in the soil, and can contribute to desertification of the environment. Physical changes in the environment, coupled with chemical spillage and displacement of surrounding sediment and minerals in the water, can poison and destroy aquatic life. An increase in particles in the water due to spillage may lower the water quality or even poison it for local users. Moreover, the dust particles can have a particularly adverse effect on surrounding humans, crops, and animals. It can “contribute to chronic respiratory illnesses such as emphysema and bronchitis and have been associated with increased mortality rates from some diseases” (Environmental Contamination 203).
However, there are ways to lower the negative effects of mining such as recycling and mine closure plans that are required by most regulatory agencies. Some of these solid wastes can be used in construction such as in the production of cement, pavements, and other structures that require rocks (Terec). It has been shown that waste management strategies focused on “source reduction and resource recovery, reuse, and recycling have proven to be more cost effective over the long run, and they are less damaging” (Environmental Contamination 202). However, despite efforts towards the proper disposal and treatment of waste, there are accidents or intentional neglect that occurs during quartz mining. Recycling waste products conserves limited resources as well as preventing or lowering damage to the environment. Moreover, it can save energy, reduce harmful gas emissions, and stop toxins from leaking out into the surrounding environment from landfills (Environmental Contamination 202). Efforts towards the rehabilitation of mining sites have also made significant contributions to the environment. Rehabilitation of mining sites is dependent on the individual characteristics—ecological, social, and geological—of the mining sites. The general process involves extensive landscaping of the area by filling in pits and burying large rocks. The next goal is to restore life and fertility to the land. This is done by returning the rocks, soil, and in particular the fertile topsoil that was removed during the initial stages of mining. Due to important nutrients and seeds being present in the first layer of dirt, mining operations that aim to rehabilitate the land store them until mining is done. After rehabilitation has been done, the area is monitored to ensure that the land is recovering.
After the quartz is mined it needs to undergo many purification processes that often require many chemicals and results in many chemical by-products being produced. First, the oxygen needs to be extracted from the silica to leave only silicon behind. This is done by heating it to over 2000 degrees Celsius in the presence of carbon, causing carbon dioxide emissions to be formed from the reaction. Then it is treated with oxygen to rid it of calcium and aluminum impurities, resulting in metallurgic-grade silicon (~99% pure). However, it needs to be 99.999999% pure silicon, so it is further treated with gaseous hydrogen chloride, a corrosive (techradar 1). They react to form trichlorosilane gas, which contains many chlorinated metal impurities such as iron, aluminum, or carbon (Mulvaney 2). These impurities are removed by distillation and treated as waste products. The next step is to add heat and hydrogen to transform it into silane gas. Accidental releases of silane are especially hazardous due to its properties as an explosive. Afterwards, it is deposited on an ultra-pure silicon rod to result in electronic-grade silicon.
However, this silicon’s polycrystalline structure is unsuitable for semiconductor products because it has too many defects in them. It has to be rotated and held above 1414 degrees Celsius, right above its melting point, to form into an ingot. It is then sliced into wafers with a wire saw, generating kerf (aka silicon dust). This dust is hazardous if it is breathed in, so the Occupational Safety and Health Administration (OSHA) greatly recommend respiratory masks. However, these precautions do not appear to be entirely effective in preventing over exposure to silicon dust. Due to how many wafers are cut from a silicon rod, lots of the silicon is wasted or lost in the form of silicon dust: as much as 50%. Finally, it is etched with a variety of dangerous and corrosive chemicals like nitric, hydrofluoric, and acetic acid. Hydrofluoric acid is much more dangerous than normal acids because it can “penetrate the skin and bind with bone calcium” (Vigil 3). Interfering with the calcium in the body can poison the body and lead to cardiac arrest and or death. These acids and chemical by products can’t just be tossed away or flushed into a sewage system without treatment. It needs to be treated with even more chemicals before it is safe to be disposed of in the environment. In order to maintain high degrees of precision for etching with acids, photolithography is used extensively: up to fifty times. The photoresist that is used in photolithography is dispensed on the wafers from bottles. A considerable amount of photoresist is wasted due to bottles being tossed before they had been completely drained of photoresist. One method that can be used to reduce this waste is to install automated supply systems to dispense the photoresist. This reduced wasted photoresist by 50% as well as lowering photoresist expenses by 18% (Gemar 211).
In order to prevent harm to the general public, the semiconductor industry has to adopt various strategies to deal with waste products. These strategies fall into three categories: waste elimination, reduction, and treatment (Vigil 5). Waste elimination involves attempting to prevent the production of waste by-products in the first place. This is done by trying new chemical processes or alternative materials that cause non-toxic by-products. An example would be switching from strong acids or organic solvents to citric acid. This can reduce negative effects to the environment and reduce costs (Vigil 5). Or they can find new materials or processes to replace ones that result in high amounts of waste. One example is switching from strong acid washing of reactors to dry plasma cleaning instead. Waste reduction aims to reduce the amounts of waste created in the process of semiconductor grade silicon. Increasing efficiency and using alternative processes can help conserve precious resources as well as lower negative by-products. Due to there being limits to how much waste can be eliminated or reduced, there will always be remaining waste products. As a result, industries must resort to waste treatment; it is viewed as least desirable due to the considerable expenses that accompany waste treatment. Harmful byproducts need to be properly treated in order for them to be disposed without harmful effects to the environment. Gaseous waste can be burned in the case of the explosive silane gas or passed through a wet scrubber in order to get rid of pollutants. For example, due to the Clean Air Act requiring less than 0.5 parts per million of SO2 (sulfur dioxide) in the air, it is treated with limestone to convert it into the safer CaS (calcium sulfide) (Industrial Grade Silicon 3). Liquid effluents, waste that is disposed of in a river or the sea, are divided into three categories: acid, base, and organics. Acids and bases need to be neutralized before they can be sent to local wastewater treatment systems. Organic waste may also need additional pretreatment before it is sent to wastewater treatment systems.
After the MOSFETs are finished, they need to be prepared for consumer use. They are tested to see if they are defective before they are packaged and shipped off. However, due to the precision required in making MOSFETs, a high amount of them are defective, as much as 40%, and are tossed before they even make it to the market (techradar 2). If they pass inspection, they are encased in a processor like package to protect the fragile chip and packed in anti-static packaging made from polyethylene terephthalate plastic (PET). This packaging is not recyclable so it is sent to landfills afterwards. MOSFETS are typically not bought by consumers directly, but are instead in electronics that are sold such as computers, phones, and radios. As a result, they are usually not individually disposed or recycled either. When electronics that contain MOSFETs in them are disposed of, assuming they were not incorrectly thrown in the garbage, are recycled or reused as much as possible. Old computers are often updated with new software and used again, but if these computers are no longer able to keep up with the times then they are recycled. They are taken apart and scrapped for their raw materials like metals and resold.
Overall, despite the fact that MOSFETs are incredibly important to our way of life, it contributes to considerable negative effects on the environment. Due to how many of them are used on a daily basis, it comes as no surprise that major effort has been put towards reducing their impact. However, despite efforts to prevent and reduce these harmful effects, MOSFETs are still contributing to considerable harm to the environment. This brings to question whether or not enough is being done to reduce the impact of a MOSFET’s lifecycle on the environment.
Baumgärtel, Gerd. "The Siemens Thermal Waste Recycling Process." Journal of Analytical and Applied Pyrolysis 27.1 (n.d.): n. pag. The Siemens Thermal Waste Recycling Process. Elsevier Science Publishers B.V., Oct. 1993. Web. 03 Feb. 2016.
Esfahani, Shaghayegh, and Mansoor Barati. "Purification of metallurgical silicon using iron as
an impurity getter part I: growth and separation of Si."Metals and Materials
International 17.5 (2011): 823-829.
Gemar, Claudia. "A REVIEW OF SOLVENT WMTE MINIMIZQTION OPPORTUNITIES IN THE ELECTRONICS INDUSTRY." (n.d.): n. pag. Web. 12 Mar. 2016.
"High Purity Quartz: A Cut Above." Dictionary of Gems and Gemology (n.d.): 862. Web. 12 Mar. 2016.
"How Sand Is Transformed into Silicon Chips." Techradar. Future US, 24 May 2009. Web. 3
Feb. 2016. <http://www.techradar.com/us/news/computing-components/processors/how-sand-is-transformed-into-silicon-chips-599785>.
Kang, Hee Nam, Jin-Young Lee, and Jong-Young Kim. "Recovery of Indium from Etching Waste by Solvent Extraction and Electrolytic Refining." Hydrometallurgy110.1-4 (2011): 120-27. Web. 03 Feb. 2016.
Mei, et al. "Purification of Metallurgical Silicon by Horizontal Zone Melting." Solar Energy Materials and Solar Cells 98 (2012): 233-39. Environmental Sciences and Pollution Management. Web. 3 Feb. 2016.
Menezes, Romualdo R., Lisiane Navarro L. Santana, Gelmires A. Neves, and Heber C. Ferreira. "Recycling of Mine Wastes as Ceramic Raw Materials: An Alternative to Avoid Environmental Contamination." (n.d.): 200-20. Www.intechopen.com. Department of Materials Engineering, Federal University of Paraiba. Web. 12 Mar. 2016.
Mulvaney, Dustin. "Sidebar_headline." Hazardous Materials Used In Silicon PV Cell Production: A Primer. N.p., n.d. Web. 12 Mar. 2016.
Murphy, Cynthia F., et al. "Development of parametric material, energy, and emission inventories for wafer fabrication in the semiconductor industry."Environmental Science & Technology 37.23 (2003): 5373-5382.
Nomura, Kenji, et al. "Thin-Film Transistor Fabricated in Single-Crystalline Transparent Oxide
Semionductor." Science 300.5623 (2003): 1269-72. Print.
"Production of Industrial Grade Silicon." (n.d.): n. pag. Web. 12 Mar. 2016.
"PV Manufacturing." Photovoltaics Manufacturing, Polysilicon. Green Rhino Energy, n.d. Web. 03 Feb. 2016.
"REC Silicon's Fluidized Bed Reactor (FBR) Process." REC Silicon. REC Silicon, n.d. Web. 03 Feb. 2016.
Tangstad, M., and B. S. Xakalashe. "Solar Cells. Datasheet Information and Product Data for Crystalline Silicon Solar Cells." Southern African Pyrometallurgy 2011(n.d.): n. pag. Web. 11 Mar. 2016.
Terec, Liana-Rodica. "RECYCLING OF QUARTZ AND FELDSPAR MINE WASTES IN THE BUILDING AND CONSTRUCTION INDUSTRIES." (2014): n. pag. “Gheorghe Asachi” Technical University of Iasi, Romania RECYCLING OF QUARTZ AND FELDSPAR MINE WASTES IN THE BUILDING AND CONSTRUCTION INDUSTRIES. Web. 12 Mar. 2016.
"Transistor." Infoplease. 6th ed. Columbia UP, n.d. Web. 2 Feb. 2016.
Vigil, Samuel A. "Environmental Management of Wastes in the Semiconductor and Electronics Industries." Proceedings of International Symposium on Advanced Waste and Emission Management (2000): n. pag. Web. 11 Mar. 2016.
"What Happens to Mine Sites after a Mine Is Closed? - See More At: Http://www.miningfacts.org/environment/what-happens-to-mine-sites-after-a-mine-is-closed/#sthash.HseOOTix.dpuf." Fraser Institute, 2012. Web. <http://www.miningfacts.org/environment/what-happens-to-mine-sites-after-a-mine-is-closed/>.