A semiconductor, as you’ve learned before, is a device that can shift between a conductor and a nonconductor, hence its name. In other words, it can allow electricity to flow, or keep it from flowing, whenever necessary.

In order for a TV to turn on, what must be done? It needs to be plugged into a power outlet, of course. The same goes for semiconductors, which is where the metal interconnect process comes into play.

Laying down metal highways that bring semiconductors to life

Utilizing metal’s conductive properties, the metal interconnect process creates metal circuits along the pre-designed patterns. The metal used for semiconductor manufacturing must meet the following requirements:

1. Adhesiveness to the semiconductor substrate (wafer):
The metal needs to easily and strongly adhere to the semiconductor substrate in thin film form.

2. Low electric resistance:
Since the metal circuits deliver the electric current, the substance must have low electric resistance.

3. Thermal and chemical stability:
It is important that the attributes of the metal do not change during the metal interconnect process.

4. Easy formation of patterns:
Regardless of the quality of the metal, it is essential that the material can easily form patterns, especially during the etching process.

5. High reliability:
With the advancement of integrated circuit technology, the metal interconnect material needs to be durable even in minute scale.

6. Manufacturing cost:
Even if the above conditions are met, the cost also has to be suitable for the mass production of semiconductors.

The metals typically used in semiconductor manufacturing that meet the above criteria are aluminum (Al), titanium (Ti) and tungsten (W).

Now, let’s find out how the actual metal interconnect process is carried out.

Popular metals for the interconnect process

Aluminum, one of the main substances used in metal interconnect for semiconductor manufacturing is known to have two merits: great adhesiveness to silicon dioxide and high processability.

Because aluminum and silicon tend to react with each other, aluminum circuits on semiconductors that are made of silicon may get damaged. To prevent this, a barrier metal is deposited.

Aluminum circuits are created through deposition. When a mass of aluminum is boiled in a decompressed vacuum chamber, the chamber gets filled with aluminum particles. A wafer is then inserted into the vacuum chamber, where the aluminum particles adhere to the wafer and form a thin film. Because aluminum is vaporized and deposited in a high-vacuum environment, this is called the evaporator process. Physical vapor deposition (also known as sputtering) using plasma is also a method widely used today.

A contact is a point where a basic element and a metal interconnect meet. Should the contact be in the form of a narrow tunnel and is difficult to fill, tungsten then comes into play. In such cases, the metal interconnect process is carried out using chemical vapor deposition (CVD) instead of the evaporator process, so as to uniformly deposit the metal as a thin film.

As continued advancements are made in semiconductor technology, semiconductor fabrication processes are also experiencing changes. The metal interconnect process that we discussed today is undergoing a transition from evaporator to chemical vapor deposition so as to better meet the demands of finer design rules. Replacing traditional metals, copper is becoming the material of choice for semiconductor fabrication, thanks to its cost-effectiveness and better conductivity properties.

So there you have it. We have now covered all of the essential processes to design and build the semiconductor circuits on a silicon wafer.

In the next part of our series, we will explore the final step of semiconductor manufacturing, the testing and packaging process, in which silicon wafers are transformed into the individual chips that we see in electronic devices. Stay tuned!

In Korean, http://samsungsemiconstory.com/183.

If you get a chance to see a cross section of a semiconductor chip under a high-resolution electron microscope, you will see layer after layer of materials piled up like a skyscraper.

To fabricate such structures, the photolithography and etching processes (which we discussed previously), along with a few others, are repeated a few hundred times until layers of thin film material result in a semiconductor chip.

Today, we will be going over this “thin-film process,” in which a semiconductor chip gets its electrical properties.

Too thin for seein’

The dictionary defines a film thinner than 1 micrometer (μm, one millionth of a meter) as “thin film.” This thickness cannot be manufactured mechanically.

In order for a semiconductor chip to get the desired electrical properties, materials at atomic or molecular levels are densely piled up in the thickness of a thin film. This film is so thin that very elaborate and precise technology is required for it to be deposited evenly on the wafer.

Let’s say we are forming a 1 micrometer-thick film on an 8-inch (200mm) wafer. This would be the equivalent to evenly piling up sand 1mm thick on a schoolyard that is 200m large in diameter. Pretty high-tech, huh?

Deposition: Air-brushing with chemicals

Semiconductor deposition structure

Deposition refers to a series of processes where materials at atomic or molecular levels are deposited on the wafer surface as a thin layer to contain electrical properties.

The deposition methods can be largely divided into physical vapor deposition (PVD) and chemical vapor deposition (CVD).

Physical vapor deposition (PVD) is mainly used for depositing thin metal films and does not involve chemical reactions.

Chemical vapor deposition (CVD) occurs as particles from the chemical reaction of gas are deposited in the form of vapor activated by an external energy source. CVD can be used on conductors and nonconductors, as well as semiconductors.

For this reason, CVD is more commonly used in today’s semiconductor manufacturing processes. The CVD method is further broken down into thermal, plasma-enhanced and optical CVD depending on the source of external energy used. Plasma-enhanced CVD, in particular, yields many benefits as it can be processed at lower temperatures in large volumes while offering greater control over thickness uniformity, making it a preferred method of choice these days.

The thin film fabricated through the deposition process can be categorized into metal (conducting) layers for electrical connections between circuits, and dielectric (insulating) layers that electrically isolate the internal layers or protect them from contaminants.

For semiconductors to develop electric properties, a process of implanting ions on the deposited layer must follow. Ion implantation is the process of implanting electrically charged particles onto a semiconductor surface with circuit patterns. These ions are referred to as “impurities” and include boron (B), phosphorus (P) and arsenic (As). By inserting impurities into the wafer surface in the form of fine gas particles to a desired depth, the silicon wafer acquires its electrical conductivity. (Review the diffusion process to learn more about implanting impurities.)

Today, we looked at how the initially pure, nonconductive silicon wafer is transformed into a semiconducting wafer through the deposition and ion implantation processes. Just how thin and evenly the layer is formed during this deposition process can determine the final chip’s quality.

Stay tuned for the next part of our series, coming next week!

In Korean, http://samsungsemiconstory.tistory.com/172.

An art in itself

Etching is a method widely used in print-making for illustrations that require elaborate and detailed lines. Rembrandt, known as “the master of light and shadow,” and Goya, a leading artist of the late 18^th century, both used etching techniques in their artwork.

In etching, an anti-corrosive material is coated on a metal plate, then carved out with sharp tools like scribes and needles to draw the desired design. The plate is then dipped in a corrosive material such as nitric acid, and the level of corrosion is controlled. The exposed area is etched away, while the remaining area remains intact.

The Etching Process

Similarly, the etching process in semiconductor fabrication uses a liquid or gas etchant to selectively remove unnecessary parts until the desired circuit patterns are left on the wafer surface. By repeating this process on multiple layers, a semiconductor chip is eventually born.

The two fundamental types of etchants are liquid-phase (“wet”) and plasma-phase (“dry”), and the method used depends on which material is to be etched. Whereas engraving utilizes sharp tools to physically carve out patterns, etchants are used in semiconductor manufacturing to remove the undesired portion of the wafer while the circuit patterns are protected with an anti-corrosive layer formed during the photo process.

Compared to the wet etching technique, dry etching is more costly and complicated. However, with continued innovation in semiconductor technology and the circuitry now in nanoscale, dry etching is the more widely used technique, as it produces a higher yield.

Dry etching to rid of unnecessary parts

So how does dry etching remove the unnecessary substrate materials?

Dry etching, also called plasma etching, starts with the generation of plasma. Plasma is a state of matter —along with solid, liquid and gas— that consists of a large number of free electrons, ions and neutrons, or molecules in the form of ionized gas. When something is ionized, it means a neutron or molecule has changed its state of electrical charge by either losing or gaining electrons.

The Generation of Plasma

When a magnetic field is applied to gas, its free electrons become energized and start bumping into neighboring neutrons or molecules. This collision, or ionization, of free electrons and neutrons produces even more free electrons.

This chain reaction of ionization, called the avalanche effect, causes the number of ions to increase exponentially, resulting in a plasma state. A radical atom dissociated from this plasma state becomes volatile and moves itself away from the wafer surface, consequently peeling off surface material that was not previously coated and protected with photoresist.

There are a couple of things that we need to pay special attention to during the dry etching process.

The first is to maintain uniformity of the etching speed throughout the wafer’s surface. If the time it took to etch varies in different areas of the wafer, then there will be inconsistencies in the etch depths. This could lead to malfunctioning chips or chips containing differing properties in certain areas.

The second is etch rate, which refers to the amount of surface material removed in a given amount of time. Etch rate can differ depending on the number of reactive atoms and ions, as well as the amount of energy the ions carry which causes the surface reaction to occur.

In addition, selectivity and profile are also considered to be important elements of dry etching. All these factors must be closely controlled so as to improve the overall yield.

Now you know a bit more about the etching process that constructs the circuit patterns of semiconductors. In the next part of our series, we’ll take a look at how a semiconductor wafer gets its electrical properties.

In Korean, http://samsungsemiconstory.com/152.

Before people began to take pictures with digital cameras or smartphones, there were analogue film cameras.

Drawing a circuit on a wafer via the photolithography—or ‘photo’ for short—process is quite similar to taking a picture and having a film developed. Let’s take a closer look.

Designing circuits and creating a mask

The first step is to use computer-aided design (CAD) software and devise the circuits to be drawn onto the wafer. The size of this electronic circuit pattern can measure anywhere between 10 to 50 meters wide. This vast canvas bears a precisely designed, complex pattern that will end up on the semiconductor chips, which are about the size of a fingernail. Before this is carried out, an engineer actually steps onto the blown-up version of drawing to examine whether each part of the circuit is properly designed. The scale and the level of meticulousness required to accomplish this are quite amazing.

After examination, the pattern design is duplicated onto a glass substrate made of ultra-pure quartz with a beam of electrons, also known as an e-beam. A patterned substrate called a photoreticle, more commonly known as a photomask, which works like a negative film, is born.

Throughout the manufacturing process, the microscopic circuit has zero tolerance for any particles. As the patterned mask, which is larger than the actual chip size, passes through a reducer lens to transfer the design onto the chip, any particles that may be present on the mask could be shrunk as well. As such, particle contamination can be a significant problem during semiconductor manufacturing.

Picture-perfect technology

The photolithography process got its name from its role to transfer the circuit design onto a wafer by exposing the patterned mask to light. Making a replica on a wafer is like printing a black-and-white negative on light-sensitive paper.

Since the main focus of semiconductor technology is to scale the circuitry as small as possible, the first step to success is determined by the edge in photo process technologies. This is why continuous research in the photolithography field is a must to stay ahead.

Let’s take a closer look.

1. Making the wafer surface into a photographic print

First, photoresist (PR; highly sensitive to light) is applied evenly over the wafer surface. The additive can either be positive or negative depending on its reaction to light. Areas with positive photoresist are removed during the developing process when exposed to light, while those with negative photoresist remain. In other words, they create the image in the opposite way.

The PR layer needs to be thin, even and highly sensitive to ultraviolet rays to get the desired results.

2. From film to print

The next steps are comparable to the process for developing photographs. After a wafer is prepared with the PR layer, it then goes through the stepper where the circuit design on the patterned mask is projected and transferred onto it with ultraviolet light. Due to the scale in semiconductor manufacturing, the area exposed to light is highly controlled and selective.

As developing solutions come in contact with the photoresist, certain areas are selectively removed to create the final pattern.

As such, our circuit designs are nicely traced onto the wafer.

Stay tuned for the next part of the series, where we will see how currents run through the microscopic circuits.

In Korean, http://samsungsemiconstory.tistory.com/136.

A reliable oxide layer that shields the wafer’s surface

Before it can be used as a raw material for the integrated circuit, silicon extracted from sand goes through a purification process and is shaped into an ingot. This conic object is then cut to a uniform diameter, polished and eventually becomes a wafer.

The polished wafers start out pure in a non-conductive state. To make them semi-conductive, various substances are transferred onto the wafer, and then the circuit pattern is etched onto the surface.

Oxidation, the groundwork for the sequential procedures mentioned above, is a process in which a thin layer composed of various materials is deposited. The technique forces oxygen, or vapor, to diffuse into the wafer surface at high temperatures between 800 and 1200°C so that a thin, smooth layer of silicon dioxide can be created.

This layer protects the surface from chemical impurities and pollutants that permeate during the processes. Even tiny contaminants invisible to the naked eye can alter resistivity or conductivity and consequently damage the circuit’s electrical properties. Therefore, shielding the surface from these substances with a protective layer is crucial.

Oxides that protect the silicon surface

The silicon dioxide layer doubles as a trustworthy guardian against unintended adulteration during the ion implementation stage, and as an insulator that separates each part of the electrical circuit on the wafer to prevent a short circuit.

So what kind of chemical reaction creates this dependable oxide layer?

When exposed to oxygen in the atmosphere or within chemicals, an oxide layer begins to build on the wafer’s surface, just as iron (Fe) rusts when it becomes oxidized in the air.

There are a variety of oxidation methods, such as thermal oxidation, electrochemical anodic oxidation and plasma-enhanced chemical vapour deposition (PECVD). Among them, the thermal oxidation procedure performed at a high temperature is most widely used.

Thermal oxidation can be either wet or dry. Dry oxidation only uses oxygen to forge a thinner layer, whereas wet oxidation uses both oxygen and vapour to fashion a thicker layer.

Although oxides created by the dry method have excellent electronic properties, they grow much slower when compared to the wet method. Under identical time and temperature conditions, the wet method can present an outcome that is five or ten times thicker than that of the dry method.

This brings us to the end of the second part of the series. Stay tuned for Part 3, as Samsung Tomorrow will explain next week how the circuit pattern is imprinted on the oxide-clad wafer.

In Korean, http://samsungsemiconstory.com/110.

Although an integrated circuit (IC), also known as a semiconductor chip, may deceive you with its fingernail-sized form factor, it is actually packed with billions of electronic components—transistors, diodes, resistors, and capacitors—which all work together to perform logic operations and store data.

So, what does it take to manufacture this kind of circuit, you ask?

The technology behind engineering an IC goes far beyond the simple assembling of individual components. In fact, microscopic circuit patterns are built on multiple layers of various materials, and only after these steps have been repeated a few hundred times is the chip finally complete.

Today, we are introducing a new series that will walk you through the entire manufacturing process of this advanced device, from the raw material stage to the final testing of the semiconductor chip. The series will consist of eight parts and will be published weekly.

Read on for the first part of the series, which introduces the “canvas” for integrated circuits, otherwise known as the silicon wafer.

What’s a wafer?

A wafer, also called a disc, is a thin, glossy slice of a silicon rod that is cut using specific diameters. Most wafers are made of silicon extracted from sand. The main advantage of using silicon is that it is rich in supply, being the most abundant element in nature, just after oxygen. Its environmentally friendly properties are an added bonus.

Building an ingot, the foundation for wafers

Once silicon is extracted from sand, it needs to be purified before it can be put to use. First, it is heated until it melts into a high-purity liquid then solidified into a silicon rod, or ingot, using common growing methods like the Czochralski (chokh-RAL-skee) process or the Floating Zone process.

Ends cut off from silicon rods, or ingots

The popular Czochralski method uses a small piece of solid silicon (seed) which is placed in a bath of molten silicon, or polycrystalline silicon, and then slowly pulled in rotation as the liquid grows into a cylindrical ingot. This is why the finished wafers are all round discs.

Giving new meaning to the term “wafer-thin”

Before it is completely cooled, the cone-shaped ends of the ingot are cut off while the body is sliced into thin wafers of uniform thickness with sharp diamond saw blades. This explains why an ingot’s diameter would ultimately determine the size of a wafer. In the early days of the semiconductor industry, wafers were only three inches in diameter. Since then, wafers have been growing in size, as larger wafers result in more chips and higher productivity. The largest wafer diameter used in semiconductor fabrication today is 12 inches, or 300mm.

Smoothing things out – the lapping and polishing process

Sliced wafers need to be prepped before they are production-ready. Abrasive chemicals and machines polish the uneven surface of the wafer for a mirror-smooth finish. The flawless surface allows the circuit patterns to print better on the wafer surface during the lithography process, which we will cover in a later posting.

Know your wafer

Each part of a finished wafer has a different name and function. Let’s go over them one by one.

1. Chip: a tiny piece of silicon with electronic circuit patterns

2. Scribe Lines: thin, non-functional spaces between the functional pieces, where a saw can safely cut the wafer without damaging the circuits

3. TEG (Test Element Group): a prototype pattern that reveals the actual physical characteristics of a chip (transistors, capacitors, resistors, diodes and circuits) so that it can be tested to see whether it works properly

4. Edge Die: dies (chips) around the edge of a wafer considered production loss; larger wafers would relatively have less chip loss

5. Flat Zone: one edge of a wafer that is cut off flat to help identify the wafer’s orientation and type

This brings us to the end of the first part of the series. Want to know what happens next? Then, stay tuned for Part 2, as Samsung Tomorrow will take you through the disc production stage by discussing the oxidation process of the wafer next week.

In Korean, http://samsungsemiconstory.com/95.