Gundyabhaoo: How A Microprocessor Is Made

How A Microprocessor Is Made

Silicon: The foundation for all Intel microprocessors
A microprocessor is an integrated circuit built on a tiny piece of silicon. Silicon is used because it is a semiconductor. Semiconductors are a class of materials whose electrical conductivity is between that of a conductor and that of an insulator. Silicon can be altered to be either an insulator, which blocks an electrical charge, or a conductor, which lets the electrical charge pass through.

A microprocessor contains millions of transistors, which are interconnected through extremely fine wires made of aluminum or copper. The transistors work together to store and manipulate data so that the microprocessor can perform a wide variety of functions.

Microprocessor production: Approximately 300 steps from start to finish
Making microprocessors is a complex, demanding process involving more than 300 steps. Microprocessors are built by layering materials on top of thin rounds of silicon, called wafers, through various processes using chemicals, gases and light.

Wafers are usually 200 mm, or 8 inches, in diameter. However, starting in 2001, Intel will also use wafers that measure 300 mm, or 12 inches, in diameter. The wafers are made from silicon, the principal ingredient of common beach sand, that has been purified and liquefied and grown into long, cylindrical tubes called "ingots." The ingots are sliced into thin wafers, which are polished until they have flawless, mirror-smooth surfaces.

In chip making, very thin layers of material, in carefully designed patterns, are put on the blank silicon wafers. The patterns are computerized designs that are miniaturized so that up to several hundred microprocessors can be put on a single wafer.

Because the patterns are so small, it is impossible to deposit material exactly where it needs to be on the wafer. Instead, a layer of material is deposited or grown across the entire wafer surface. Then, the material that is not needed is removed and only the desired pattern remains.

While there are more than 300 steps required to make a working microprocessor, the chip making process can be summarized in a few steps that involve growing silicon dioxide and creating conductive properties, testing, packaging, and shipping.

Growing silicon dioxide and creating conductive properties
The microprocessor manufacturing process begins with "growing" an insulating layer of silicon dioxide on top of a polished wafer. This oxide layer also acts as an electrical "gate" that either enables or prevents the flow of electrical current within the microchip. The silicon dioxide is grown on the surface of the wafer in a furnace at very high temperature. The thickness of the oxide layer depends on the temperature and the amount of time the wafers are in the furnace.

Photolithography, the process in which circuit patterns are printed on the wafer surface, is next.. First, a temporary layer of a light-sensitive material called a "photoresist" is applied to the wafer. Ultraviolet light shines through the clear spaces of a stencil called a "photomask" or "mask" to expose selected areas of the photoresist. Masks are created during the design phase and are used to define the circuit pattern on each layer of a chip. Exposure to light chemically changes the uncovered portions of the photoresist.

The exposed areas of photoresist are removed, revealing a portion of the silicon dioxide underneath. This revealed silicon dioxide is removed through a process called "etching." Then, the remaining photoresist is removed, leaving a pattern of silicon dioxide on the silicon wafer. Additional materials, such as polysilicon, which conducts electricity, are deposited on the wafer through additional lithography and etching steps. Each layer of material has a unique pattern. Together, they will form the chip's circuitry in a three-dimensional structure.

In an operation called "doping," the exposed areas of the silicon wafer are bombarded with various chemical impurities called "ions," which provide positive and negative charges, thereby altering the way the silicon in these areas conducts electricity. The electrical charges help the transistor to turn on and off, thereby passing electrical current through the transistor's gate.

To provide a link to the additional layers put on the wafer, "windows" are formed by repeating the masking and etching steps.

This layering is repeated 20 to 25 times over a period of several weeks. This process creates a skyscraper effect of layers on top of the wafer.

Metal is applied to fill in the "windows," thereby forming electrical connections between the chip's layers. Intel introduced copper metal on its 0.13-micron process technology, the most advanced microprocessor process generation in production today. Previously, Intel used aluminum metal in its 0.18-micron and older process technology. Copper and aluminum are excellent electrical conductors.

Testing
Once the layering is complete, the wafers are prepared for testing. In order to withstand the processes and equipment used in the layering process, wafers must be relatively thick. This thickness must be reduced by 33 percent before the wafers can be cut into individual microprocessors. Thus, the wafer goes through a series of steps to reduce its thickness and to remove impurities from its backside. Once the wafer's thickness is reduced, a layer of another material is deposited on the backside of the wafer to provide a good surface for die to be attached at assembly. This also provides an electrical contact from the back of the integrated circuit to the external package during the assembly process.

The wafers are then tested to determine the quality of each processing step. Separate components, such as transistors, resistors and capacitors are tested to determine whether or not the chips function properly. If a processing problem exists, this data can be analyzed to determine what processing step caused the problem.

Electrical probes are then placed on each die on the wafer and each die is energized. A computer completes a series of tests to determine if the circuit meets specifications.

Packaging
After wafers are tested, they are sent to Intel's assembly facilities where each wafer is cut into tiny rectangles, called "die," each containing a complete integrated circuit. A precision saw separates the die from one another. The non-functional dies are then discarded.

The individual dies are then assembled into external packages. These packages protect the die from the environment and provide the electrical connections for the die to communicate with the circuit board onto which it will later be mounted. Tiny balls of solder are bonded from designated areas on the die to the electrical leads on the package. Now, electricity can travel from the printed circuit to the die and back again.

After package assembly, a test is run to determine if the die are still functional. Non-functional units are rejected. The functional units are then put through stress tests. During these tests, each individual unit is put through different humidity and temperature levels and checked for electro static discharge levels. After each stress test, the unit is tested to determine the functional level. The units are then placed in specific bins, depending on the speed and power performance of the unit.

Shipping
Qualified integrated circuits are then given an outgoing inspection. This is to verify that all previous tests were correct and the integrated circuit meets or exceeds customer standards. All integrated circuits that pass inspection are marked and packed in boxes to be sent to customers.

About Intel
Intel (NASDAQ: INTC), the world leader in silicon innovation, develops technologies, products and initiatives to continually advance how people work and live. Additional information about Intel is available at www.intel.com/ pressroom and blogs.intel. com.