The Technology
What Is a Semiconductor?
A number of elements are classified as semiconductors, including silicon, zinc, and germanium. These elements have the ability to conduct electrical current, and their amount of conductivity can be regulated. Silicon is the most widely used semiconductor material because it is easily obtained.
Silicon is basically extracted from sand, which is the starting material. It has been used for centuries to make cast iron, bricks, and pottery. In ultra-pure form, the controlled addition of minute amounts of certain impurities called dopants alters the atomic structure of silicon. The silicon can then be made to act as a conductor or a nonconductor, depending upon the polarity of an electrical charge applied to it ¾ hence, the generic term semiconductor.
Early Developments
As early as 1830, semiconductor materials were studied in laboratories. The first materials studied were a group of elements and compounds that, if heated, were poor conductors. Shining light on some of them would create an electrical current that could pass through them but only in one direction.
By 1874, electricity was being used not only to carry power, but also to carry information. The telegraph, telephone, and later the radio were the earliest devices in an industry that would eventually be called electronics.
Radio receivers required a device called a rectifier to detect signals. Ferdinand Braun discovered one-way conduction in metal sulfide crystals. Radio receivers required a device called a rectifier to detect signals. Braun used the rectifying properties of the galena crystal, a semiconductor material composed of lead sulfide, to create the cat’s whisker diode[1] for this purpose. Thus was born the first semiconductor device.
The Integrated Circuit
Until 1959, all electronic components were discrete: that is, they performed only one function, and many of them had to be wired together to create a functional circuit. Although a great number of identical discrete transistors could be put on a wafer, they then had to be cut up and individually packaged. Packaging each component and hand wiring the components into circuits was extremely inefficient.
New technologies emerged and integrated circuits were soon developed, with various components (transistors, resistors, and capacitors) formed on the same chip, but interconnection of the various components still required tedious hand wiring.
In 1959, Jean Hoerni[2] and Robert Noyce[3] developed a new process called planar technology at Fairchild Semiconductor that enabled them to diffuse various layers onto the surface of a silicon wafer to make a transistor, leaving a layer of protective oxide on the junctions. This process allowed metal interconnections to be evaporated onto the flat transistor surface, and replaced the hand wiring. The new process used silicon instead of germanium, and made commercial production of ICs possible.
The initial resistance to the new IC technology gave way to enormous popularity. By the end of the 1960s, nearly 90% of all the components manufactured were integrated circuits.
How Microprocessors Work
In November 1971, Intel introduced the world's first commercial microprocessor, the 4004, invented by three Intel engineers. Primitive by today’s standards, it contained a mere 2,300 transistors and performed about 60,000 calculations in a second. Twenty-five years later, the microprocessor is the most complex mass-produced product ever, with more than 5.5 million transistors performing hundreds of millions of calculations each second. Today, the microprocessor has roughly 15 million transistors and can perform billions of calculations each second.
Today’s microprocessors are the brains of a PC. On a tiny silicon chip are millions of switches and pathways that help a computer make important decisions and perform helpful tasks. And microprocessors don’t just think for computers — you might find a processor in many other everyday items such as TVs, cars, radios, home appliances, and, of course, computers. Transistors are the main components of microprocessors.
At their most basic level, transistors may seem simple, but their development actually required many years of painstaking research. Before transistors, computers relied on slow, inefficient vacuum tubes and mechanical switches to process information. In 1958, engineers (one of them Intel co-founder Robert Noyce) managed to put two transistors onto a silicon crystal and create the first integrated circuit, which led to the microprocessor. Microprocessors are built in layers on a silicon wafer through various processes using chemicals, gases, and light.
Semiconductor Equipment: The Process
Making the Chips
The semiconductor equipment industry is a broad market with various segments encompassing the design, manufacture, and packaging of semiconductors. Semiconductors are the fundamental electronic components or building blocks used in all modern electronic and telecommunications equipment. Demand for semiconductor and electronic component manufacturing equipment typically mirrors that for semiconductors, which in turn attributes to the continued growth of the PC market and the increasing use of semiconductors in the automotive, telecommunications, and consumer electronics market.
The semiconductor manufacturing process involves two distinct phases — the front-end process and the back-end process. The front-end processing involves thousands of steps applied to a silicon wafer to form millions of circuits on the wafer in an array, each of which will become an individual integrated circuit.
The back-end process begins with the separation of the individual circuits or chips from the wafer, the bonding of each of those chips onto a plated metal leadframe or other carrier, and the connection of the chip on external leads using extremely fine gold, silver, or aluminum wire. After these two processes, the circuits are then encapsulated in an epoxy plastic to protect the chip from any contamination, the devices are separated, and the leads are trimmed and bent to form the final product.
Manufacturing semiconductors can be compared with baking a very thin layer cake — very thin. The process begins with a silicon wafer that is six to 12 inches in diameter. Next, a thin, second layer of material is deposited on the wafer. Then, by etching away the unwanted material, a design is created, as in an electrical circuit. This metallization process is repeated several times, with one or more layers of metal added to act as connections for the various circuits. The wafer is then cut up into many thumbnail-sized pieces, tested, and packaged.
The process of manufacturing a semiconductor, or integrated circuits (ICs or chips), can typically consist of more than 100 steps, during which copies of an integrated circuit are formed on a single wafer. In general, this involves the creation of 20 to 30 patterned layers on and in the substrate, ultimately forming the complete integrated circuit. Layering creates electrically active regions in and on the wafer surface.
Wafer Production
Art, Not Science
Today, most ICs are made of silicon. Turning silicon into chips is an exacting, meticulous process that involves personnel such as engineers, metallurgists, chemists, and physicists.
The material preparation is the very first step in the process of making an IC chip. The process is to grow the wafer. The wafer is a shiny, round silicon disc that will have all of the processing performed on it. First, sand (silicon dioxide) is obtained to be able to grow what will become a semiconductor wafer. The sand used to grow the wafers has to be a very clean and good form of silicon. For this reason, not just any sand scraped off the beach will do. Most of the sand used for these processes is shipped from the beaches of Australia.[4]
The first step from silicon to circuit in semiconductor manufacturing process is the creation of a pure single-crystal cylinder or ingot of silicon. The process is started by inserting or dipping a “seed” silicon crystal into a pot of pure molten liquid silicon[5] that is slowly rotated as it is “pulled” out in order to form an ingot (which resembles a whole delicatessen salami). The crystal ingot is then ground or sanded down to an exact uniform diameter and a diamond saw blade slices the ingot into thin wafers. The wafer is processed through a series of machines, where it is ground smooth and chemically polished to a mirror-like luster — in a series of a combination chemical and mechanical polishing processes called CMP.
Each wafer is given either a notch or a flat edge that will be used subsequently in orienting the wafer into the exact position for later procedures. For some applications, a thin layer of silicon with different electrical characteristics than the underlying wafer is deposited on the wafer’s surface, creating an epitaxial layer. If a layer of silicon is grown onto the top of the wafer using chemical methods, that layer is of a much better quality than the slightly damaged or unclean layer of silicon in the wafer. The epitaxial layer is where the actual processing will be done.
The wafers are then ready to be sent to the wafer fabrication area, where they are used as the building material for manufacturing integrated circuits.
Exhibit 1. Silicon Wafer
Pure single-crystal cylinders of silicon are sliced into thin, highly polished wafers less than one-fortieth of an inch thick. Hundreds of chips are etched onto each wafer.
Most chip designs are developed with the help of computer systems or computer-aided design (CAD) systems. Circuits are developed, tested by simulation, and perfected on computer systems before they are actually built. When the design is complete, photomasks are made — one mask for each layer of the circuit. These photomasks are used in the photolithography process.
Crushed high-purity polycrystalline silicon is melted at 1,400° in a quartz crucible surrounded by an inert gas atmosphere of high-purity argon and then cooled to an exact temperature. A “seed” of single-crystal silicon is put in and slowly rotated as it is “pulled” out, causing a small amount of the liquid to rise with the seed. This results in an ingot with the same orientation as the seed. The Ingot’s thickness is determined by both temperature and the extraction speed.
The heart of IC manufacturing resides in wafer fabrication. This is where ICs are formed in and on the wafer surface. The fabrication process involves a series of operations called oxidation, masking, etching, doping, dielectric deposition, metallization, and passivation.
These steps — primarily due to their complexity — constitute the majority of wafer fabrication costs, more than in any other step in the semiconductor manufacturing process. Typically, it can take up to 30-60 days to complete the fabrication process, which is carried out in the following manner:
Oxidation: Growing a Layer of Silicon Dioxide on the Wafer’s Surface
Oxidation refers to the process of growing a layer of silicon dioxide onto the wafer’s surface. This layer of silicon dioxide acts as an insulator to electricity. The insulator is important for a semiconductor to achieve the correct electrical characteristics. The silicon dioxide layer also serves to protect the inner part of the wafer.
On the wafer, the first layer of silicon dioxide is grown by exposing it to extreme heat and gas. This growth is similar to the way rust grows on metal when exposed to water. However, the silicon dioxide layer grows much faster and is too thin to be seen by the naked eye.
Cleaning: A Rigorous Process
The silicon wafers are pre-cleaned using high-purity, low-particle chemicals (important for high yields). The wafers are rigorously cleaned by a variety of liquid environments and dried in a centrifuge dryer. These wafers are then heated and exposed to ultra-pure oxygen in the diffusion furnaces under carefully controlled conditions, resulting in a uniform layer of silicon dioxide film on the wafer’s surface.
Deposition: A Fundamental Step in IC Manufacturing
Deposition is a fundamental step in IC manufacturing. During deposition, a layer of either electrically insulating (dielectric) or electrically conductive material is deposited or grown on a silicon wafer. Generally, there are several types of deposition:
Physical Vapor Deposition (PVD). This technique is based in the formation of vapor of the material to be deposited as a thin film. The material in solid form is either heated until evaporation (thermal evaporation[6]) or sputtered by ions (sputtering). The most common form of PVD is sputtering, in which a target is exposed to plasma made with an inert gas like argon that is not chemically reactive. It is also possible to bombard the target with an ion beam from an external ion source. This allows the ability to vary the energy and intensity of ions reaching the target surface. The excited gas atoms, accelerated by the electric field, hit the target and physically knock off the metal atoms that deposit onto the wafer below, building up the desired metal film on the wafer’s surface. Alternatively, it is possible to use the ion source to directly bombard the wafer surface thereby imparting a higher energy to the evaporated atoms and result in a film with better properties (adherence, density, etc.).
Chemical Vapor Deposition (CVD). CVD is used to deposit films that function as dielectrics (insulators), metals (conductors), or semiconductors (partial conductors) on a wafer. During the CVD process, gases that contain atoms of the material to be deposited react on the wafer surface, forming a thin film of solid material. The most common films deposited by CVD are silicon dioxide (often called oxide), silicon nitride, polysilicon, and tungsten.
The main goals of CVD can be summarized as follows: 1) to provide films that are uniform in thickness, as well as in chemical, electrical, and mechanical properties; 2) to provide films that are pinhole-free; 3) to provide films that adhere to the host substrate; and 4) to provide films with controlled surface roughness.
There are primarily two types of CVD, thermal CVD and plasma-enhanced CVD or PECVD:
Thermal CVD. Thermal CVD is the most conventional technique, and has been used for more than 50 years. The wafer’s surface to be coated is heated to temperatures typically ranging from 800 to 1,050°C in order to activate a chemical reaction. Lower temperatures, like 500°C, have also been used at lower pressure. Thermal CVD gives a thin, hard, wear-resistant surface layer. Because of the high process temperature, thermal diffusion results in a mixed interface between coating and substrate, giving very good adhesion. If the gases are equally distributed in the process chamber, all surfaces with the same temperature will get an equal coating.
PECVD. PECVD is an important deposition method for semiconductor fabrication. PECVD has two advantages over the more conventional CVD method: low process temperature and flexible film properties. The former satisfies the low thermal budget requirement for most production. The latter makes it possible to tailor film properties for specific device characteristics. However, the plasma nature of the process could damage some films and deteriorate the device.
Photolithography
How Complex Patterns of ICs Are Defined
Photolithography is the process by which complex patterns of ICs are defined. The photolithography process encompasses all the patterning operations needed to transfer an image from one medium to another. The term “photolithography” refers to the use of light as part of the transfer process.
The lithographic process begins by applying a photosensitive material or photoresist evenly over the surface of a wafer. This coating gives the wafer characteristics similar to a piece of photographic paper.
After Coating, Next Steps Are to Align and Expose
At this stage of the process, circuit patterns are imaged from a reticle[7] or mask to the coated wafer surface in a photolithography tool called a stepper.
A stepper is much like a photographic enlarger, taking a negative (the reticle) and projecting an image of it on a target (the silicon wafer), which is coated with photographically sensitive material. While an enlarger expands the negative, a stepper shrinks the image to microscopic dimensions. This reproduction is usually accomplished by transferring light through a photomask onto a photoresist. The areas of the photoresist that have been exposed to light are then dissolved by chemical developers and subjected to further processing, such as etching, ion implantation, and metal deposition.
Masking is used to protect one area of the wafer while working on another. This process is also referred to as photomasking. Photomasks are high-purity quartz or glass plates containing precision microscopic images of integrated circuits and are used as masters (equivalent to “negatives” in a photographic process) to optically transfer images of design patterns onto the surface of a wafer during semiconductor manufacturing.
Photomasks are manufactured from photoblanks, which are highly polished quartz or glass plates coated with ultra-thin layers of chrome and photoresist. The photomask is protected from particle contamination by an ultra-thin, frame-mounted transparent film, called a pellicle. Pellicles are made from nitrocellulose or other polymer[8] solutions that are prepared or purchased. The ultra-thin film is typically precision coated with an anti-reflective layer that improves optical performance. Material properties and manufacturing conditions are carefully tuned to match the pellicle’s light transmission properties with the specific semiconductor lithography application.
The pellicle, when mounted on the photomask, creates a sealed contamination-free environment for the photomask pattern. Photoblanks and pellicles make up about 80% of the materials costs associated with photomask production. The production of photoblanks requires ultra-pure chrome deposition on highly polished and extremely flat quartz or glass substrates. The quality and properties of photoblanks strongly affect the yield and quality of the products.
A photoaligner aligns the wafer to a mask and then projects an intense light through the mask and through a series of reducing lenses, exposing the photoresist with the mask pattern. Precise alignment of the wafer to the mask prior to exposure is critical. The photomask is then inspected for defects, its critical dimensions are confirmed, and any defects are repaired.
There are many patterns to be transferred onto the wafer; the exposure of each wafer layer puts different sets of features onto the device. These layers must all be properly aligned to each other prior to imaging, so that any structures formed on the wafers are correctly placed, exactly stacked on one another, to ensure proper functionality. After the development of the photoresist, the etching process selectively removes material from areas that are not covered by the imprinted pattern.
Successive steps of lithography, deposition, and processing gradually create the multiple layers of conducting, semiconducting, and insulating patterns that make up the millions of transistors found in a modern semiconductor device.
In the semiconductor manufacturing process flow, the etch process follows photolithography, and is used to permanently transfer a mask pattern to a wafer, or to clean the wafer surface. During the etch process, the image in the photoresist produced in photolithography is precisely and permanently transferred into the surface layer of the wafer. The wafer is developed (the exposed photoresist is removed) and baked to harden the remaining photoresist pattern. It is then exposed to a chemical solution or plasma (gas discharge) so that areas not covered by the hardened photoresist are etched away. The photoresist is then removed (cleaned) using additional chemicals or plasma and the wafer is inspected to ensure that the image transfer from the mask to the top layer is correct. After repeated etchings, the resulting patterns of exposed areas interconnect the transistors on the die, and ultimately create an IC.
There are several types of etch:
Dry Etch. Dry etching removes unwanted materials from the surface of a wafer, or from films deposited on the wafer, to make a desired pattern by employing the excited gases, which are called plasma. Plasma is a highly reactive chemical species created in an etch reactor. The reactor not only produces the plasma but also controls the chemical and physical reactions that occur on the wafer surface at an atomic level. Through the etch process, selected materials are removed from the wafer or film, which shapes the profile and critical dimensions of the remaining materials.
The photoresist is used as a mask to protect the unetched area that will be used later on, for example, the poly or metal lines. On the other hand, the etched area will be later filled with the appropriate materials, like the insulator or metal for connection point. After etch, the photoresist needs to be removed cleanly. This can be achieve by employing the asher, which uses the similar principle as the dry etcher to remove only the resist without damaging the underlying layer.
Wet Etch. Wet etch patterning uses wet chemistry in conjunction with a liquid photoresist to define a desired pattern in the polyimide layer. Typically, wet etch processing is used to pattern course features such as bond pads or large vias. The base process involves the spin coating and partial curing of the polyimide layer, using one or more in-line hot plates or sometimes a convection oven. A layer of positive photoresist is coated over the top, baked, and imaged. The photoresist is then developed. The developer will simultaneously wet etch the underlying polyimide in the imaged areas. After develop/etch and a water rinse, the photoresist is stripped using a liquid photoresist stripper. The patterned polyimide wafer is then fully cured to complete the imidization[9] process and remove residual solvent. The process can be repeated up to 18 times to create the various layers necessary for each part’s circuitry.
Atoms with one less electron than silicon, or more than one electron than silicon, are introduced into the area exposed by the etch process to alter the electrical character of the silicon. These steps are repeated several times until all active devices have been formed.
Diffusion. This is the process by which dopants are added to a wafer. By using the appropriate mask, a certain pattern is diffused onto the wafer surface. Doping is done through diffusion by either exposing the wafer to a high temperature of dopant vapor in a carrier gas (gaseous diffusion) or by covering the wafer’s surface which is to be doped with a temporary layer of dopant oxide (non-gaseous diffusion). The most common dopants include arsenic, boron, and phosphorus. Others include aluminum, antimony, beryllium, gallium, germanium, gold, magnesium, tellurium, and tin. Dopant gases such as arsine, silane, phosphine, and diborane are highly toxic. All of these gases, except arsine, are also “pyrophoric” (can ignite on contact with atmospheric oxygen).
Ion Implantation. The diffusion process must be tightly controlled. Ion implantation provides more controlled doping than diffusion. The ion beam implanter is used to alter the near-surface properties of semiconductor materials. Instead of controlling the time and temperature of a diffusion furnace, ion implantation makes use of an extremely high-voltage electron gun (implanter) that accelerates the dopants and “shoots” them or imbeds them into the silicon wafer. By adjusting the high voltage, the implant depth is controlled. It is possible to get very precise profiles by using this method.
Dielectric Deposition and Metallization
The following steps are repeated at this point in the process: dielectric deposition, photolithography, reticle inspection, and etch. The individual devices are interconnected using a series of metal depositions and patterning steps of dielectric files (insulators). Current semi-fabs include as many as three metal layers separated by dielectric layers.
Metal Deposition
Using physical vapor deposition (PVD), agron atoms are shot at a “target” of pure metal. These metal atoms chip off and deposit on the wafer surface.
Passivation
After the last metal layer is patterned, a final dielectric layer is deposited to protect the circuit from damage and contamination. Openings are etched in this film to allow access to the top layer of metal by electrical probes and wire bonds.
Electrical Test
An automatic, computer-driven electrical test system then checks the functionality of each chip on the wafer. Chips that do not pass the test are marked for rejection.
Assembly and Packaging
After a wafer has gone through all the processes and has been tested, it needs to be separated and assembled into different individual chips.
The Assembly Process
IC devices are fabricated on a variety of materials, typically in a wafer form. The wafer itself is divided into individual segments called die. The first step is to slice the wafer into these tiny pieces. This is called die separation.[10] Typically, a diamond saw slices the wafer into single chips. Chips that have been rejected are discarded, and the remaining chips are visually inspected under a microscope before packaging.
The chip is then assembled into a package that provides the contact leads for the chip. These leads extend from the die and connect to the outside of the device. The die is connected electrically to the frame by tiny gold wires. This process is called wire bonding.[11] More than 95% of all packages are assembled with wire-bonding technology, because the high-speed wire bonders meet most of the interconnection needs of semiconductor devices. A wire-bonding machine attaches wires (which are a fraction of the size of a human hair) to the leads of the package. After wire bonding, another sample optical inspection is typically performed, looking for missing or non-sticking wires.
Once the bonding is completed, it is important to protect the IC package by applying encapsulation.[12] The packaging is encapsulated with a plastic coating for protection, and the chip is then tested again prior to customer delivery. The IC is virtually complete at this point. Final quality inspection and packing into shipment tubes, trays or automated tape is all that remains. The completed component has traveled through several hundred processing steps. The assembly and packing segment alone takes up to a week, depending on the materials, design, and purpose of the IC. The semiconductor and semiconductor equipment industries, while basically following the same process steps, have made and continue to make tremendous improvements in the consistency and quality of the products produced.
Each chip is tested at various stages in the manufacturing process to see how fast it can store or retrieve information, including the high-temperature burn-in that tests the circuitry of each chip, ensuring the quality and reliability. This burn-in process provides feedback to the semiconductor manufacturer throughout the process, allowing identification and correction of manufacturing problems.
The completed packages are inspected, sealed, and marked with special ink to indicate product type, date, package code, and speed. The finished goods area ships the chips to computer, peripheral, telecommunications, and transportation customers throughout the world.
[1] A cat’s whiskers are very sensitive sensors of its environment. In the early days of radio, the fine wire that contacted a quartz crystal to form the receiving diode was called a “cat’s whisker.”
[2] Jeam Hoemi was the founder of Fairchild Semiconductor, the basis of today’s microchip technology.
[3] Robert Noyce was the inventor of the silicon microchip and co-founder of Intel.
[4] Western Australia has large deposits of high quality silica. Approximately 60% of the world market for fused silica is for the electronics industry. Fused silica is manufactured by the fusion of high-quality silica sand in an electric arc and/or resistance furnace at temperatures of around 2,000C (electricity typically represents 12.5% of production costs at a scale of 5,000 tons per annum. The fused silica is crushed to market specifications.
[5] The sand is taken and put into a pot where it is heated to about 1,600 degrees Centigrade, where it melts. The molten sand will become the
source of the silicon that will be the wafer.
[6] In thermal evaporation techniques, different methods can be applied to heat the material. The equipment available uses either resistance heating (Joule effect) or bombardment with a high energy electron beam, usually several KeV, from an electron beam gun (electron beam heating).
[7] The reticle is the medium by which a stepper can transfer a chip design from an engineer’s computerized layout to the wafer. In a typical wafer process, as many as 30 reticles may be required to build a chip. These reticles will each vary in design content and specifications.
[8] The word comes from the greek polumeres, meaning “having many parts.” Polymers are large molecules consisting of repeated chemical units (“mers”) joined together, usually in a line, like beads on a string. Each mer is typically made up of more than five and less than 500 atoms; the word “polymer” is applied when you have more than about 50 mers stuck together. Historically, polymers have typically been used to make solid plastics where the chains virtually don’t move. Today, there are applications of polymer liquids, and a key area of research is the modification of the properties of surfaces using thin polymer coatings.
[9Relating to chemical engineering — Polyimides are currently the materials of choice for interlayer dielectrics in microelectronic applications, since polyimides, as a class of materials, best satisfy the requisite properties to survive the thermal, chemical, and mechanical stresses associated with microelectronic fabrication. As more and more functionality is demanded of these polymer dielectrics, e.g., low residual thermal stress, adhesion, photosensitivity, and low dielectric constant, it is increasingly hard to design materials with the desired enhancements without compromising existing properties. The imidization process involves the dissolution of polyamide acid in an amide solvent or a mixture of ether and an amide solvent at a low percentage of solids followed by treatment of this solution with an aprotic organic base, such as triethylamine or pyridine, for a period of time before treatment with an organic dehydrating agent. The latter treatments facilitate the formation of polyimide. This technique improves the melt flow of polyimide. Chemical imidization or cyclodehydration techniques do not cause a significant decrease in molecular weight. This is useful for production of polyimide molding materials, and has such applications as easily processed adhesives, molding powders, and matrix resins.
[10] Die separation is usually done with a diamond saw. The wafer is held in place by adhesive-type tape, while the saw blade precisely makes horizontal and vertical cuts across the wafer. Once the saw is done, the unused sections of the wafer are discarded and the die are ready to be packaged.
[11] Wire-bonding is accomplished by thermosonically bonding a wire to the die using sophisticated equipment and software with a high degree of positional accuracy.
[12] Encapsulation is a commonly used process in which high temperature and pressure liquefy epoxy resin that is forced through a mold over both the die and die frame and into the cavity on the frame, where the die was placed earlier in the manufacturing process.
