Connections: The Veins of the Semiconductor

After the processes explored in the previous episodes including oxidation, photolithography, etching, and deposition are completed, semiconductor devices are finally formed on the wafer’s surface. For memory chip makers like SK hynix, they line up transistors and capacitors¹ on the wafer’s surface, while foundries or CPU manufacturers line up three-dimensional transistors such as FinFETs² on the wafer.

¹Capacitor: A storage battery that can store electricity and is often used for electronics. In this article, it refers to a device used to store data in a semiconductor.

²FinFET (Fin field-effect transistor): A type of three-dimensional MOSFET that has an electric current with a path that resembles a fish’s dorsal fin.

▲ Figure 1. Layers of a semiconductor device and areas of metal wiring (Image Source)

However, these devices prove to be useless if they are isolated. Just as individual devices on top of an electronic substrate cannot work unless they are soldered, transistors on a wafer cannot function without being interconnected with each other. These transistors operate by receiving external power so they can perform various tasks such as moving processed data to the next area. Accordingly, a process is required to connect devices to one another, or to a power source. As for these devices, they must interact with one another using electricity as semiconductors are essentially electronic circuits. This is where the metallization process comes in to enable the operation of a semiconductor. As for different types of semiconductors like CPUs or GPUs, they are created using the same devices that are interconnected in different ways.

▲ Figure 2: Metal wiring (yellow) connecting the device’s layers (red). *Some structures omitted. (Image Source)

Metallization is not a single process such as photolithography, etching, and deposition—processes that make it possible to apply metal wiring on a semiconductor. Additionally, the metallization process is distinguishable for using materials, including metals, that have different characteristics from materials that were used in the previous process of forming device layers.

Thus, there is no “metallization equipment” like there is etching equipment. Instead, equipment that is used in other processes is also used to make the metal wiring in the metallization process. When a material needs to be carved out, etching equipment is used, while deposition equipment is generally used to fill empty spaces. All the while, photolithography is involved between these processes.

Contacts: Linking Wires and a Device

When connecting devices on the substrate, electric wires are first connected and then soldered on. As layers are stacked starting from the bottom on a semiconductor, a junction called a contact—which connects metal wiring with a device—is created after the formation of the bottom device layer. The wiring is then connected on top of it.

▲ Figure 3. The use of tungsten in the formation of contacts, and an example of barrier metal (Image Source)

It might seem like metal can be connected directly to the device without a contact, but an issue arises as semiconductors go through miniaturization. This problem is related to the gaps that inevitably form on the semiconductors. Although the deposition process has a “gap-fill” property that fills in these voids to prevent devices from disintegrating, metals such as aluminum are not able to fill in deep holes even with deposition. This results in faulty wires which are desolate in the middle. Due to this problem, the gap must be filled beforehand by depositing a metal like tungsten (W) that has a superior gap-fill property so the deep metal wiring can be formed. This applies when there is significant distance between the device layer and the metal layer. Otherwise, high-temperature heat treatment is required following the formation of metal contacts. Furthermore, even when a metal or a material that is not heat-resistant like aluminum is used, contacts need to be formed with tungsten while aluminum wires are placed on top.

Barrier Metals: Reducing Metal-to-Metal Resistance

Meanwhile, a metal—or metal compound—called a barrier metal is needed between the device and the contact. In semiconductor processing, it is very difficult to precisely connect non-metallic and metallic materials. When two materials with different characteristics are directly connected, a high resistance caused by the difference in conduction bands³ between the metal and silicon occurs at the boundary. This leads to an increase in the power consumption of the semiconductor. To avoid this, it is necessary to add a barrier metal. To make a barrier metal, materials such as titanium (Ti) or cobalt (Co) are applied on top of the silicon layer of a semiconductor device, and the metal reacts with the silicon atoms. This process is called ‘silicidation’ while the corresponding area is known as a contact silicide.

³Conduction band: Among the two bands that are separated by an energy gap in the energy band structure of a solid, the band on top contributes to the conduction of the solid.

Barrier metals are also used to prevent unwanted damage to the device during processes. Aluminum, for example, tends to react with silicon (Si) which is the main material in wafers. When the aluminum metal wiring closely passes by the silicon in the device layer, a barrier metal such as a titanium compound needs to be placed between the two as a barricade.

▲ Figure 4. The purpose of the barrier metal when using aluminum wiring

In addition, the increased use of copper conductors due to the limitations of aluminum has resulted in further application of barrier metals. Unlike aluminum, copper has the tendency of diffusing into silicon dioxide (SiO₂) which is more stable than silicon. If left unchecked, the oxide film that is supposed to block the current will contain copper atoms and cause current leakage. To prevent this, a metal called tantalum (Ta) is used to create a boundary between the copper conductor and the device layer.

Conductors: Wires Connecting Semiconductor Devices

After soldering is completed, the wires need to be connected. The process of connecting wires in a semiconductor has similarities to the process of connecting wires in a regular circuit. For example, both processes create a part that is the equivalent to a sheathed cable. However, while completed wires are connected with each other in a normal electronic circuit, the wires are instead ‘created’ on top of the circuit for a semiconductor.

▲ Figure 5. Comparison of reactive ion etching (RIE) and the damascene process (Source: Hanol Publishing [Understanding Semiconductor Manufacturing Technology, p. 293])

The process of creating and connecting the wires unfolds very differently depending on the type of metal that’s intended to be deposited. If aluminum is going to be deposited, the wires can be made using the etching and deposition techniques that were explained in previous episodes. After applying a metal film over the entire surface of the wafer, photoresist needs to be applied on top before progressing to exposure. Then, after shaving off unwanted aluminum, various dielectrics—or insulating materials—are filled around the remaining aluminum.

In the case of copper (Cu), the order of metal and dielectric deposition is reversed. When copper is used to make wiring, the dielectric is deposited first and then etched afterwards using photolithography. A copper seed layer is then created, and the copper is filled in between the dielectric before the remaining copper is grinded away.

As a matter of fact, the order in which metal and dielectric deposition is carried out is significant. As explained earlier, copper requires the application of a seed layer which, in turn, requires a new deposition process. Moreover, a technique not used for aluminum called electroplating⁴ also needs to be applied. In order to overcome the limitations of aluminum, it was necessary to not only apply copper as a new material in the semiconductor process but also develop various processes to enable its use. For a long time, it has been widely known that copper has better electrical properties than aluminum. However, for semiconductor companies, the manufacturing process also needed to evolve to allow copper to be applied to transistors that are produced on a larger scale at lower costs. Looking at metal wiring, it is thicker towards the top layers of the device. Although it is advantageous to place semiconductor devices that need to send and receive a lot of data frequently close to each other, devices that do not send and receive data back-and-forth as much can be connected over longer distances. These devices are connected by thicker metal wiring at the top.

⁴Electroplating: The process of coating metal from one side to another side using the principle of electrolysis.

This thick metal wiring on the upper parts does not require complex techniques to be created. Just as aluminum was used to create a bottom layer wiring of a certain thickness in the past, an aluminum wiring with the same thickness can be made on the upper part. So, the application of the top wiring layer does not have to involve advanced technology but can use processes that have been frequently used before. This not only cuts costs for the semiconductor company but also saves time as there is no need to learn and use a new process.

Combining Technologies to Achieve Progress

These technologies mentioned above do not exist in isolation but are combined in various ways by semiconductor manufacturing companies that produce semiconductors based on their strategies. Unlike memory chip makers like SK hynix, companies that manufacture logic semiconductors⁵ need to be able to fully control the current in their transistors. To achieve this, they need to widen the area by introducing a three-dimensional structure such as a FinFET and make the channel where the current passes into a three-dimensional shape. It can be presumed that it is much more difficult to form contacts on a transistor with this structure compared to flat transistors used in DRAMs and other products. Looking at the left side of Figure 6, it is easy to bring the contact and the channel where the current flows close together. In contrast, the right side shows it is difficult to make a contact that is close to the channel.

⁵Logic semiconductor: A semiconductor with operational purposes such as a CPU or GPU.

▲ Figure 6. Diagrams showing it is more difficult to form contacts in FinFETs of logic semiconductors (right) than planar transistors in DRAM (left)

As miniaturization progresses, these challenges are bound to continue. Memory companies like SK hynix do not face issues concerning the operation speed of the circuit but, instead, struggle maintaining charge capacity when stacking capacitors up high. The important thing to note is that the challenges of miniaturization, as well as metallization, will vary depending on the business environment of each company.

Semiconductor Manufacturing as a Collective Effort

Semiconductor manufacturing involves tens of thousands of workers while a single wafer must go through hundreds of steps to become a finished product. Therefore, those involved in each process contribute to less than 1% of the final product, but the semiconductor will not function if any one of these processes fail. Accordingly, the work that goes on in a semiconductor company is not completely independent; the work of one person influences the work of another, creating a big synergy effect.

One of the goals of this article is to help readers understand the technologies behind the processes. Nevertheless, a more important goal is to understand the relationship between the technologies. The material used in the deposition process affects subsequent processes that use heat and, also, the etching process. When etching is proceeded extensively, problems can occur for the entire product if the following deposition process does not have good gap-fill properties. If there are not enough steppers to draw fine patterns, a process involving additional hard mask deposition and etching, which is called multi-patterning⁶, is required.

⁶Multi-patterning: A technique that uses steppers multiple times to draw very fine patterns on a wafer.

As evident from the explanations above, the semiconductor industry does not only rely on high-tech processes but also trust. It is a sector that requires good communication skills, honest work, and creativity in coming up with new ideas. The development of semiconductor technology is a continuation of identifying problems by honestly sharing various challenges that arise when developing a new micro-process. Afterwards, the problem needs to be solved with the creativity of the organization which holds the solution. As seen in the photolithography process, a new problem caused by an immersion stepper was solved by using a photoresist instead of a stepper.

▲ Figure 7. Example of a photoresist solving an issue with a stepper

It is the hope that readers, especially those who desire to join the semiconductor sector, understand these aspects of the industry. Understanding these technologies can go a long way in assisting the development of careers while helping to create a virtuous cycle of relationships with many related organizations, which can eventually lead to creating the world’s best semiconductors.

Currently, semiconductor technology is facing major challenges in miniaturization. The voices of the people who use semiconductors need to be heard even more going forward. Likewise, the abilities of those who will communicate and develop technology within semiconductor companies will prove to be even more valuable in the future.

Read articles from the Front-End Process series

Read articles from the Back-End Process series

The post Semiconductor Front-End Process Episode 6: Metallization Provides the Connections That Bring Semiconductors to Life first appeared on SK hynix Newsroom.

▲ Figure 1. Adding chocolate syrup and another layer of the cookie on top

We can first get a better idea of the deposition process by returning to the cookie-making analogy used in previous episodes. As shown in Figure 1, in order to make a chocolate-filled cookie, chocolate syrup is first added to the etched surface of the cookie and then another cookie is placed on top to create another layer. This thin layering process is akin to deposition.

Deposition: Adding Materials to the Wafer

The deposition process is very intuitive. Once the wafer is prepared for processing, it is inserted into the deposition device. As time passes, a sufficiently thick film will be formed on the surface of the wafer before unnecessary parts are removed in order to move on to the next process.

Just as etching is one of the many processes that removes materials from a wafer’s surface, deposition is also part of a group of processes that adds materials to a wafer’s surface. For example, the photoresist coating process involves applying various films to the wafer’s surface, while the process of oxidating the wafer—or silicon—also adds materials to the wafer’s surface. So, what makes the deposition process stand out among these various processes?

It has to do with the miniaturization of semiconductors that became increasingly necessary for high-performance and low-power electronics. With such miniaturization, it was required to add thin films composed of various materials such as metal that can handle different roles. In the past, semiconductor companies used aluminum for the metal wiring inside the chip due to its high-conductivity¹. However, as the miniaturization of aluminum reached its limit, manufacturers switched to using copper for the wiring as it has a higher conductivity than aluminum. But the problem with copper atoms is that, unlike aluminum, they have the tendency to even spread into areas that manufacturers don’t want to be interfered with—such as those containing silicon dioxide (SiO₂). To prevent this, a high-quality thin film is applied to the area where the copper wiring is to be coated. It acts as a protective film by restricting the passing of copper.

¹Conductivity: The measure of a material’s ability to pass an electric current. Materials such as metals have high conductivity.

To make the layers of the semiconductor’s core device and wiring that is only one-thousandth the thickness of a human hair, it is necessary to apply these materials thinly and evenly. This is why the deposition process—which is referred to as the thin film deposition process within the semiconductor industry—is crucial in the semiconductor manufacturing process.

Types of Thin Films and Their Roles

Since semiconductors cannot operate with pure silicon alone, the process of adding materials is very important in semiconductor manufacturing. It separates the two areas that must not be interfered with while using wires to connect areas together. Adding materials is also necessary in other cases such as when using a specific film to strengthen or weaken an electrical field, or to facilitate the next process in the semiconductor manufacturing procedure by producing a thin film in advance.

Among the various roles of thin films in a semiconductor, the most crucial function is that of a protective barrier. These films increase the reliability of operation by creating boundaries among circuits to prevent interference of core semiconductor devices and to block leakage in currents. If needed, the films can be applied at the end of the manufacturing process to protect the chip from external shock. In addition, when etching is used following the stacking of semiconductors, the films can be used to prevent etching in unwanted areas. Some examples of such protective films are STI² and IMD³, while the materials that are used for the films include silicon dioxide (SiO₂), silicon carbide (SiC), and silicon nitride (SiN).

² STI (Shallow Trench Isolation): A trench-shaped protective film that prevents leakage current in the device’s boundary.
³ IMD (Intermetal Dielectric): A protective film that prevents unwanted current flow between the layers of metal wiring.

▲ Figure 2. STI preventing leakage current at the device’s boundary

Another material that is used is metal. It makes sure the semiconductor’s bottom device, or the transistor, fulfills its role by connecting the transistor with other devices and power sources. As such, the transistor is useless by itself without such connections. To make these connections, metal wiring made from materials such as titanium, copper, and aluminum is needed while a contact that connects the metal wiring and components needs to be created. This process is just like soldering wires to connect the components on an electronic circuit board inside home appliances. The wires connected on top of the circuit board serve the same purpose as the metal wiring inside a semiconductor, and soldering has the same function as contacts inside a chip.

Moreover, deposition is used for many other purposes, such as the production of transistors to form the gate insulating film and applying hard masks that are used in multiple patterning⁴. As mentioned earlier, deposition is used in almost every step of semiconductor manufacturing and it, at times, replaces other processes. As an example, gate insulating films were made through the oxidation process in the past. But, nowadays, deposition is the primary method as the emphasis on precision and quality increased due to the miniaturization of semiconductors.

⁴ Multiple Patterning: A technology that makes semiconductors even finer. It repeats the processes of exposure and etching several times.

Key Aspects of Deposition: Uniformity and Step Coverage

▲ Figure 3. Examples of high and low uniformity

It is also useful to know the terminology that relates to the quality of the deposition process. Some of them will sound similar to those that were introduced in the etching episode. The first term is “uniformity” which is the measure of how evenly the materials are formed during the deposition process. As the entire wafer is placed inside a machine during deposition just as it is during etching, the thickness may vary in different parts of the wafer. A higher uniformity signifies that the material was evenly applied to the entire wafer.

The next term is “step coverage.” As we saw in the etching and oxidation processes, the thickness of the film may not be evenly formed if there are sharp or uneven edges on the wafer’s surface. Step coverage refers to the difference in thickness between the top and bottom films—or top and side wall films—that are on the rough surface where deposition is performed. If the step coverage is close to a value of 1, it means that there is minimal difference between the top and bottom or side film thickness. If the step coverage is significantly less than 1, it indicates that the bottom or side wall film is very thin compared to the top.

▲ Figure 4. Examples of step coverage

Types of Deposition: CVD, PVD and ALD

Similar to other processes, deposition methods can also be divided into chemical and physical types: chemical vapor deposition (CVD) and physical vapor deposition (PVD). CVD is a method that deposits materials onto the wafer’s surface using a chemical reaction. The most common method is to utilize catalytic activity to provide energy into a mixture of gases. If material A needs to be deposited on the surface, two gases—B and C—which can produce A are injected with the addition of energy or something similar that can trigger a reaction. The below equation shows how the material is made:

B + C + (energy, etc.) → A + byproducts

The chemical method is optimal due to a high deposition rate and excellent step coverage. However, there is the downside that various impurities can contaminate the materials as it is impossible to completely remove byproduct gases that can be constantly generated during the reaction process. This method, therefore, is used to create various thick shields or dispensable films like hard masks rather than being used in areas where property control must be very precise.

▲ Figure 5. CVD and PVD deposition methods

Meanwhile, PVD is a method that deposits materials onto the wafer’s surface by gasification. As shown in Figure 5, material A is vaporized into atoms which will then be deposited onto the wafer. Just as in etching, the common method of PVD is sputtering⁵ which uses plasma ions—typically inert gases—moving at high speeds to release atoms from the target, material A. The separated atoms travel in the opposite direction until they are deposited onto the wafer.

⁵ Sputtering: A physical method that causes the surface of a material to break apart by impinging high energy on it.

Given that there are no byproduct gases, high purity is one advantage of this method. Moreover, it is possible to deposit nonreactive, pure materials such as pure tungsten and cobalt. Due to these characteristics, PVD is commonly used in metal wire manufacturing where pure materials are heavily used.

Meanwhile, there is also a unique process called atomic layer deposition (ALD). While the processes we have discussed so far involved chemical bonding of activated gas to a wafer’s surface or depositing materials via sputtering, ALD uses a slightly different method. To thinly deposit material A onto a wafer, two reactant materials in the form of B and C that are used to make A need to be prepared. Material B is a precursor that can be easily adhered to the wafer’s surface while material C is highly reactive. To begin with, the atoms of material B stick onto the wafer’s surface. If these atoms have the characteristic of not sticking well to each other, only a single atomic layer of material B will remain on the wafer’s surface. Next, the remnants of material B are removed and material C is injected. Material B and C react to form material A and, also, create a byproduct gas, which should be removed afterwards. Repeating this process can control the thickness of the film at the atomic level.

▲ Figure 6. Concepts of CVD and ALD (Source: The Understanding of the Semiconductor Manufacturing Technology, p. 293)

This method is ideal for its excellent uniformity and step coverage. The precursor material not only can stick to various surfaces—whether vertical or horizontal—but it also allows only one atomic layer to be produced per ALD cycle. However, given that the method operates at the atomic-layer level, the downside is that the process is slow. Due to these characteristics, ALD is commonly used in components such as DRAM capacitors that have a high aspect ratio⁶ but require a high-quality film.

⁶ Aspect Ratio: The ratio of depth to width. A high aspect ratio means that the structure is narrow but tall.

Having reviewed the types of deposition, it is clear that there is a trade-off between precision and processing speed in deposition processes, just as in other processes. In other words, when improving properties like uniformity to enhance precision, the processing speed is going to be inevitably slower. This balance between precision and processing speed is a constant dilemma for semiconductor manufacturers, and the deposition process is no exception.

Difficulty in Choosing Materials

At times, there are reports over discoveries of new materials that are expected to greatly improve specifications. However, when it comes to the semiconductor sector, there are not that many instances where the new material featured on the news is actually used in the industry. This is because better qualities of materials do not necessarily guarantee better performance, while the properties required for deposition materials are as diverse as those required for deposition equipment. The next section will look at some of the effects that material properties have on manufacturing.

▲ Figure 7. Pattern damage due to thermal expansion

Any change in size when a material is heated is called thermal expansion. If we take train tracks as an example, there are gaps between the tracks to prevent them from bending under the summer heat due to thermal expansion. While thermal expansion also occurs in semiconductor manufacturing, this can be an issue as each material has a different degree of expansion. For example, the coefficient of thermal expansion in aluminum is over 40 times that of silicon oxide. Consequently, if a high-temperature process is applied to an aluminum thin film made on silicon oxide, the internal structure may bend and get damaged. If the material previously used for a specific thin film is replaced with a material with a significantly different expansion coefficient, the manufacturing yield may change significantly at high temperatures.

▲ Figure 8. The concept of electromigration

There is also a phenomenon called electromigration (EM) where moving electrons hit metal wiring atoms and change positions when the electricity flows through the metal wiring. This phenomenon mainly occurs in light metal wires made of aluminum. To avoid this, copper wiring was introduced and, consequently, many additional processes were discovered including the need to introduce a diffusion barrier. As miniaturization progressed further, EM also appeared in copper wiring, and to solve this problem, a major tech company introduced cobalt wiring in the metal wiring layer. Since the material of the core wiring layer changed, tremendous process changes occurred in the layers above and below as the attempts to improve the EM properties required a major change in the process.

It is worth noting that semiconductor manufacturing is a very tightly intertwined operation of hundreds of processes. In other words, when assessing the quality of a material, not only the material’s traits but its relationship with other processes should be taken into consideration as deposited materials do not exist in a vacuum.

Deposition’s Relationship with Other Processes

Deposition is a vital aspect of semiconductor manufacturing which has been shown to have close relationships with other processes, even replacing them on some occasions. As mentioned earlier, it is possible to produce the same materials through deposition and other processes with differing results. As an example, silicon dioxide (SiO₂) can be made through oxidation as well as deposition, but the properties of the material can change depending on the process that has been taken.

In some cases, the same methods can be used in various processes for different purposes. For example, sputtering is used in physical etching and deposition. The only distinguishing factor is whether sputtering is used to cut the wafer itself or to attach materials that have already been cut. Chemical etching and CVD also share similarities. Most notably, one of the important factors in chemical etching is whether the byproducts generated by the reaction between the etching gas and the reactant are vaporized. This also applies to CVD as the byproducts generated from this deposition process should be vaporized well and discharged easily so that the process proceeds smoothly.

Deposition’s importance and consideration of other processes can also be seen through the choice of materials used. We have seen that semiconductor manufacturers do not merely select materials that have a few good physical properties, as thermal expansion must be considered for deposition materials. If an excessively high temperature is required in the material deposition process, the previously deposited material may change. Whereas, if a material that is excessively sensitive to temperature is used, it becomes difficult to use heat in subsequent processes. Moreover, having more control over the material’s deposition rate and purity level provides more options for the process.

Opening the Door for Next-Generation Technologies

Semiconductor manufacturing involves making a single product by combining hundreds of manufacturing processes. Among these, deposition is essential in the age of miniaturization to keep apart areas that should not be interfered with and to connect vital components. Thus, deposition can be seen to support miniaturization, allowing more functions to be added to devices and paving the way for more advanced, energy-efficient products.

Read articles from the Front-End Process series

Read articles from the Back-End Process series

The post Semiconductor Front-End Process Episode 5: Supporting Wafer Miniaturization Through Deposition first appeared on SK hynix Newsroom.

▲ Figure 1. Steps to carving out the center of a cookie to fill it with chocolate syrup

Returning to the cookie analogy in the previous episode, how can chocolate syrup be added to the middle layer of the cookies shaped like the SK hynix logo, the Wings of Happiness? The easiest way that comes to mind is to remove the middle part of the cookie to pour the chocolate syrup into it. The process of carving out the part where the chocolate syrup needs to go is equivalent to “etching” in semiconductor manufacturing. When making cookies, it involves placing a cover (acting as a photomask) with holes on top of the cookie and applying a solvent that removes the uncovered areas. Afterwards, the cover is removed from the cookie so the chocolate can be poured in. Removing the excess chocolate syrup and making another layer of the cookie on top of it will allow chocolate syrup to fill the crack in the cookie.

Going back to semiconductors, there are various types of processes used to remove materials on the wafer including rinsing and etching. Rinsing refers to washing the entire wafer to remove unwanted impurities, while etching is a process that uses a photomask to carve out the desired fine pattern.

Etching Characteristics: From Selectivity to Uniformity

As etching has many important properties, the figure below will help explain many of the terms that are related to the process.

▲ Figure 2: Characteristics of isotropic and anisotropic etching

The first term to know is “selectivity,” which is the measure of how well an etching process removes only the targeted materials. During etching, some materials which are supposed to remain on the wafer such as the photoresist can also be slowly dissolved. Therefore, high selectivity means efficiently removing only the targeted materials and minimizing the removal of the areas which are to remain.

“Directionality” refers to the direction of the etching, and this can be divided into isotropic etching and anisotropic etching. Isotropic etching occurs in all directions with the exposed part of the photoresist as the starting point, while anisotropic etching only reacts well in specific directions.

The “etch rate” is the measure of how fast etching occurs. While it is generally preferable to opt for a fast etch rate, accuracy also needs to be considered. Therefore, there needs to be a balancing act between accuracy and speed during the etching process. For example, the pressure of the gas must be lowered to increase the anisotropy of etching, but lowering the pressure leads to reducing the amount of gas that is reacted and, eventually, to the slowing down of the etching.

“Uniformity” measures how evenly etching occurs on the entire surface of the wafer. Unlike photolithography, etching exposes an entire wafer to gas. For etching to proceed, substances must be circulated by injecting reactant gas and removing by-products. However, applying this evenly to the entire wafer is a difficult task, which is why different areas on the wafer have varying etch rates.

Dry and Wet Etching

Like oxidation, etching is also divided into wet and dry. While wet oxidation involves using steam as the reactant gas, wet etching dips the wafer in a reactant liquid. This method of etching has the advantages of being fast and having a high selectivity as a chemical process. However, the nature of this method leads to the etching being strongly isotropic. The liquid moves freely and reacts with the substances when the wafer is dipped in it. This leads to low precision as parts on the back of the photoresist that are not meant to be removed are taken off quickly. Moreover, due to surface tension, the etching liquid can’t pass through the gap between the photoresist and the wafer if the gap is too small. Even if a stepper draws a fine pattern, it proves to be useless if the circuit cannot be made according to the blueprint. As a result, wet etching can’t be used in core layers of modern semiconductors.

▲ Figure 3: Liquid moving freely inside a gap during wet etching

Dry etching is a process in which a wafer applied with a photomask is exposed to gas. Some examples of this type of etching include plasma etching, sputtering, and reactive ion etching (RIE). Unlike wet etching, these processes remove materials in a variety of ways, so the anisotropic and isotropic properties cannot be clearly explained. For instance, dry etching that’s performed chemically will be isotropic and physical etching will be anisotropic. Nevertheless, as RIE became a prominent dry etching method, its properties of being highly anisotropic and reasonably quick resulted in dry etching being considered as anisotropic. The exact parameters of RIE’s mechanism to remove materials will be further explained in the next section.

Chemical and Physical Etching

In addition to dry and wet etching, the etching process can also be categorized into chemical and physical etching. The chemical method uses a substance that reacts well with the material to be removed. There are various materials on the bottom of the photoresist that need to be removed such as an oxide film formed in the oxidation process or other materials applied in the deposition process. These materials are successfully taken off with the spraying of a substance that only reacts well with these parts and not with the photoresist.

The type of etching liquid and gas differs according to the material that needs to be removed, but fluorine or chlorine-based compounds are frequently used in chemical etching. As a chemical reaction is the main mechanism of this process, it has high selectivity.

The other method is physical etching, also referred to as sputtering. This process involves high-energy particles colliding with the wafer’s surface and removing the surface of materials. When the pressure of gas–usually inert gas–is lowered and high energy is applied to it, the gas separates into positive atoms and negative electrons. After the electric field is applied towards the direction of the wafer, the atoms accelerate and collide with the wafer.

Although physical etching is quite a simple process, it does have its limitations. As low pressure equates to low amounts of gas, the etch rate is slow. Additionally, physical etching removes a large proportion of materials that should remain on the wafer as it relies on force–which doesn’t distinguish between materials. The most important method that’s used in practice is RIE which combines the two methods mentioned above. As a type of dry etching, RIE converts the gas used in etching into plasma. When strong energy is applied after injecting mixed gas–formed with substances such as reactant gas and inert gas–into the equipment, the etching gas separates into electrons, positive ions, and radicals¹⁾. Electrons are lightweight and don’t have a big impact, but positive ions can perform physical etching if they accelerate toward the direction of the wafer’s surface with an electric field. Since positive ions carry an electric charge, they are highly directional when accelerated in an electric field. Up to this point, there aren’t too many differences between RIE and physical etching.

¹⁾Radical: A highly reactive atom, molecule or ion that has at least one unpaired electron.

▲ Figure 4: Overview of the RIE process

However, positive ions produce an additional effect. They weaken the bond between the collided materials as they are highly directional due to the electric field. Consequently, they tend to collide in the red area shown in Figure 4. This causes the side area to remain strongly bonded, while the bond at the front becomes weaker. When the radicals that are highly reactive come into contact later on, the front part of the surface is etched even quicker. In the end, this increases the anisotropy of the etching.

Plasma etching technology achieves three different feats. In addition to physically etching by producing positive ions, the method weakens the material to be etched while increasing the reactivity of the gas being used in the etching. Therefore, it possesses both the advantages of high selectivity and anisotropy found in chemical and physical etching, respectively. Nevertheless, even if RIE is used, etching alone cannot create all of the intended patterns.

Etchants and Etching Gases

It becomes clear by now that the gases used in etching are crucial and that the key to etching is chemical reactions. Therefore, etchants must be picked according to the material that is meant to be removed. Major factors to consider when choosing the gas include seeing whether the resulting by-products are easily removable and knowing how fine the selectivity and reaction rates are. Compounds of the halogen family that have high reactivity rates–including fluorine, chlorine, and bromine–are commonly used.

▲ Figure 5: Types of gases for plasma etching (Source: The Understanding of the Semiconductor Manufacturing Technology, p. 443)

Large numbers of materials can be applied to the top of a wafer and etched. This next section will look at several of these important materials. Generally, silicon-based materials can be easily removed with fluorine gas. When silicon comes into contact with fluorine, it has the tendency to form silicon fluoride, which can be removed quickly as it vaporizes well.

Silicon dioxide, which is commonly used as an insulating or protective material, can also be easily removed by gas containing fluorine. But unlike pure silicon, silicon dioxide is in a stable state as it is bonded with oxygen and, thus, needs to be used with gas that produces heat. So, gases that have carbon atoms bonded to fluorine are usually used for etching. The silicon atom is taken from oxygen by the thermogenic action of the gas.

In the HKMG²⁾ and BEOL³⁾ processes, it’s necessary to etch metallic materials. While metals generally react with halogen-based gases like chlorine and fluorine, it’s notable that metals generally have by-products with high evaporation points. Therefore, it’s more difficult to remove them. For copper, the evaporation point of its by-products from reacting with gas is over 1,000 degrees Celsius. This means that copper adheres like rust. However, if the temperature of the wafer is raised to 1,000 degrees Celsius to remove these by-products, the heat can damage important devices. As a result, copper–regardless of its exceptional electrical properties–could be introduced with a new method of construction called Damascene⁴⁾ only after the electrical properties of aluminum have reached their limit. It’s important to bear in mind that a new material is not valuable in and of itself, but that it’s only valuable when a new process capable of mass production is introduced and harmonizes with existing processes.

²⁾HKMG (High-K Metal Gate): A new MOSFET gate developed to effectively reduce leakage current. A transistor where metal replaces the gate that was polysilicon and high-K replaces the insulation film that was silicon dioxide.
³⁾BEOL (Back End of Line): A process of creating extremely fine wiring to connect billions of unit devices together.
⁴⁾Damascene: A process used to create copper wiring. After etching the metal space, the metal goes under deposition and its excesses are physically removed.

Note that the above reactions are not perfectly controlled according to the type of substance. For example, gases that efficiently remove silicon can also tend to remove silicon dioxide, and vice versa. Therefore, there needs to be special attention to the mixture of gases when silicon and silicon dioxide are exposed together and it’s necessary to remove more than one specific material.

Additive gases are also critical, as adding various gases like oxygen, nitrogen, and hydrogen to the etching gas can bring about desired properties. In the case of hydrogen, it generates a lining that increases anisotropy if added during the process of removing silicon. Inert gases are also added partially. Neon gas is a prime example as it can control the concentration of the etching gas or provide the effects of physical etching.

Another Factor to Raising Density

Etching is a key step in the semiconductor manufacturing process that combines physical and chemical methods to create desired fine patterns on a wafer. Although it does not directly draw a precise pattern like a stepper, it is a very important task that helps hundreds of billions of transistors across the wafer to have nearly identical shapes. It adjusts various factors such as the gas ratio, temperature, the intensity of the electric field and pressure.

The importance of etching has grown even greater recently as the increase of density through the development of steppers has reached its limit. FinFET⁵⁾ from products such as CPUs and APs is an example of this occurrence.

DRAM and NAND, the two core products of SK hynix, rely heavily on etching. For DRAM, it has the problem of having to gradually make the capacitor higher to store more data, while more than 100 layers need to be etched to upgrade to 3D NAND memory.

These products require a very high aspect ratio⁶⁾ and, in order to ensure high reliability, there are innumerable factors that etching must solve such as having the starting point of the etching be almost identical to the diameter of the floor.

⁵⁾FinFET: A type of 3D MOSFET that has a passage of current that looks similar to the shape of a fish’s fin.
⁶⁾Aspect ratio: The value from dividing the etch depth by the etch base. The bigger the aspect ratio, the deeper the removal.

▲ Figure 6: The internal structure of DRAM. Numerous thin and deep structures in the cell area are capacitors.

As explained in this episode, silicon and silicon dioxide are very easy to remove because they vaporize and disappear immediately when they come into contact with fluorine. With this in mind, it is possible to anticipate what happens after a silicon wafer changes to germanium or another material. Germanium is of no use if it can’t be manufactured through processes like etching or deposition, regardless of how strong its properties are.

As we can see, the main etching processes are effective in removing unwanted surface materials and carving the desired pattern on the wafer’s surface. The etching process is one of the most crucial steps in the fabrication of semiconductors as a precise pattern needs to be etched to ensure the chip functions correctly.

Read articles from the Front-End Process series

Read articles from the Back-End Process series

The post Semiconductor Front-End Process Episode 4: Etching Fine and Identical Wafer Patterns first appeared on SK hynix Newsroom.

Just Like Baking Cookies

The emergence of the MOSFET (Metal Oxide Semiconductor Field Effect Transistor) allowed more transistors to be packed into a fixed space. As MOSFETs became smaller, they were able to use less power while obtaining more functions as the number of transistors increased. Consequently, making MOSFETs smaller is essentially killing two birds with one stone and proves to be critical to the development of semiconductors.

The process for manufacturing semiconductors shares similarities with baking cookies. Imagine making a batch of cookies in the shape of SK hynix’s “Wings of Happiness” logo. It would take a very long time to make hundreds of these cookies by hand. So, what are the alternative options?

▲ Figure 1: How to swiftly make cookies in the same shape

The best solution would be to use a cookie cutter on the dough. This makes baking 100 cookies relatively simple. But what if the cookies needed to be made smaller? Then, a cookie cutter with a smaller shape could be used to create the cookies. The role that cookie cutters play is similar to the role that the machine called a “stepper” plays in semiconductor manufacturing. The difference between MOSFETs and cookies, nonetheless, is that people prefer MOSFETs that are small and tightly packed, as having two small MOSFETs is much more useful than having one large one.

The process of producing semiconductors is similar to the steps described above. However, the process becomes more complex when the wing-shaped cookies need to be colored.

▲ Figure 2: The order of coloring the wing-shaped cookies

▲ Figure 3: Cookies can be quickly sprayed with color when they are bunched together

By following the steps shown in Figures 2 and 3, color can be added to the cookies. After cutting the cookies into the desired shape, the areas that don’t need to be colored are covered before the whole surface is sprayed with pigment. This method allows one to bake cookies with a specific shape or color in a short amount of time. But this still doesn’t explain how the black cover in the above image is made. This is, in fact, the main function of the process called “exposure.” While the cookies in the images above only have two layers of color, semiconductors require dozens of layers that begin from the device layer to multiple metal line layers. This is why exposure becomes such an essential process.

Photolithography: The Process of Making Semiconductor Patterns

In the semiconductor industry, the process of applying the pattern of a cookie to a wafer is known as photolithography. One of the first steps in this process is to apply a light-sensitive solution called a “photoresist” which changes properties when exposed to light. After applying the photoresist, light (laser) is shot at the wafer so the areas covered in the photoresist can be printed with the required pattern.

▲ Figure 4: The general sequence of photolithography

To ensure that only the desired areas of the wafer are exposed to the light, a disc with the required pattern is placed in front of the light. This disc is called a “photomask.” The intended pattern is printed on top of the wafer by shining light through the photomask and onto the wafer.

Once the pattern is created, it goes through the “develop” process which will later be explained in more detail. As areas covered with the photoresist are removed with exposure to light, a stamp with the desired pattern is created.

Selecting Positive or Negative Photoresists

There are two types of photoresist used in the photolithography process: positive or negative photoresist. A positive photoresist softens when it is exposed to light making it more soluble when the solvent is applied, while a negative photoresist will harden when exposed to light and remain on the wafer. With a positive photoresist, the areas exposed to light are removed during the develop process. Conversely, the areas not exposed to light are protected by the photoresist and are not removed during the develop process or the subsequent processes of etching and deposition.

Semiconductor manufacturers choose the type of photoresist according to the purpose of the process. For example, a negative photoresist is not suitable for making fine patterns because the hardened areas that were exposed to light will absorb some of the solution during the develop process and start to swell up. That’s why a positive photoresist is usually used when making fine patterns. On the other hand, a negative photoresist holds the advantage of being cheaper and yielding a higher resistance to processes like etching.

▲ Figure 5: Positive and negative photoresists

After choosing which photoresist method to use, a device called a “coater” is used. When drops of photoresist fall on the wafer, the coater spins rapidly and spreads the photoresist evenly. After applying the photoresist, manufacturers remove excess photoresist found on the back or edges of the wafer. They also evaporate excessive soluble components by heating them inside an oven, all the while preparing for the next stage of the process.

Over time, the structure of photoresists has become more complex. Photoresists usually consist of multiple layers, and one of these layers is called the bottom anti-reflective coating (BARC). It was developed as the need for miniaturization grew and as light from the stepper started to reflect off the wafer and affected the formation of patterns. The BARC is a substance that is applied to the wafer’s surface before applying the photoresist to prevent the reflection of light (hence the name “bottom,” as it’s located under the photoresist). Additionally, with the development of lithography systems that use water (ArF immersion¹), a waterproof coating called the top anti-reflective coating (TARC) was developed to repel water droplets and prevent damage.

Rather than memorizing every structural detail of a photoresist, it’s more helpful to examine how the industry has addressed new challenges it has faced when new technologies were introduced. Problems that surfaced after using an extreme ultraviolet (EUV) stepper² are a good example. After an EUV stepper was used, high-energy EUV rays hit the photoresist and rebounded back, resulting in the contamination of the photomask. The problem was only solved with research on photoresist materials and by introducing a protective film on photomasks called a pellicle.

¹ArF immersion: An argon fluoride (ArF) immersion stepper that uses water instead of air as the medium for light, improving its performance.
²EUV stepper: A machine that makes ultra-fine patterns using extreme ultraviolet rays.

Creating the Pattern Outline with Photomasks

▲ Figure 6: Operation of a stepper

After applying the photoresist, patterns are drawn on top of it. In order to create the desired pattern on the photoresist, a transparent plate containing the pattern called a photomask is needed. The photomask is largely opaque with some transparent areas to allow light to pass through. As the name suggests, the photomask is placed between the light source and the wafer to create the required pattern. The pattern on the photomask is designed taking into account possible issues such as light interference, so it might look different from the pattern that the manufacturer originally intended to make.

The pattern on the photomask is essentially the design of the semiconductor and, consequently, determines its usage. Photomasks used in semiconductor memories such as DRAM and NAND flash have very regular and repetitive patterns that are hard to see with the naked eye. On the other hand, photomasks used for making logic semiconductors such as CPUs and GPUs have very complex patterns.

Additionally, semiconductor manufacturing requires multiple photomasks. After using a photomask for exposure, various processes like etching, deposition, and oxidation are performed. Then, the above processes are repeated to build up the next layers. Ultimately, the design process involves creating photomasks for each layer of a semiconductor in order to provide the semiconductor chip with the desired function of the manufacturer.

Since the photomasks are prepared in advance, the next step is to accurately find the starting position for exposure. This operation is called alignment. As for exposure, it can be performed multiple times in semiconductor manufacturing. And since the patterns on a semiconductor are only tens of nanometers apart, small errors can accumulate and, thus, cause large defects. Such problems can be prevented before carrying out exposure by precisely adjusting the array according to the alignment mark produced in the previous process.

Exposure and Refining Patterns with Steppers

The stage where light is actually shone onto the wafer is called exposure. Light, or a laser, is shone onto a small area of the wafer that’s the size of a chip. After a certain amount of time, the stepper moves slightly to the side of the wafer and repeats the process.

The capability of the stepper to distinguish between two objects and analyze them is called “resolution.” Resolution is conveyed by the following formula: d=λ/(2NA) (where λ is the wavelength of light and NA is the numerical aperture). If the resolution is high, two nearby objects may seem as one. So, no matter how fine the pattern on a photomask is, the pattern is not etched accurately on the wafer’s surface.

Therefore, the key is to reduce resolution, and the light’s wavelength is the most important factor when reducing resolution. As the energy of the laser increases, the wavelength of the light decreases. The EUV stepper is a machine that can draw finer patterns than a deep ultraviolet (DUV) laser by reducing the wavelength by 1/14, or increasing the energy of the light. Another way to improve resolution is to increase the numerical aperture (NA). The NA can be increased by making the lens of the light source larger or by using a medium with a high refractive index. An example of the former is High NA EUV while DUV (ArF immersion), which is still commonly used, is an example of the latter.

The NA is a measure that may be difficult to understand initially. This concept can be clarified by referring to Figure 7 below, which shows that resolution improves, or decreases, when the lens of the light source becomes larger.

▲ Figure 7: Numerical aperture and resolution

Finding a suitable light source for the stepper is very difficult. In the early 2000s, researchers were able to discover a better light source, but it took over 10 years to commercialize a 13.5nm light source for EUV after developing the argon fluoride (ArF) 193nm laser. This is due to shorter wavelengths of light being less refracted and more likely to be absorbed when they hit a material. This is why the development of EUV steppers has been the subject of intense debate among semiconductor manufacturers.

Exposure is also critical for achieving high yields in semiconductor manufacturing. As explained earlier, exposure cannot be performed on multiple wafers at once—unlike oxidation. Thus, it’s impossible to create a uniform light source that can process a wafer with a 300mm-wide diameter. The latest steppers are very expensive, costing over 76.8 million US dollars per unit. Nevertheless, they can only process about 100 wafers per hour. The money invested in exposure itself is 12 times as much as the money spent on oxidation. For EUV, it was much harder to achieve commercially meaningful throughput than it was to create a light source. To solve this problem, the materials sector also had to work hard to find a photoresist material that would still react strongly to less light.

Once exposure is complete, overlay may be performed. Overlays are small marks placed on the wafer during the exposure process. If an overlay mark is engraved in a shape that shares the center and varies in size at every exposure, it becomes possible to measure how misaligned the exposure is or whether the wafer is slightly rotated. However, unlike the alignment process, overlay measurements are not performed on all wafers.

The Develop Process and Finalizing the Pattern

After light is shone on the photoresist, the areas that were exposed to the light either soften or harden depending on if a positive or negative photoresist was used, respectively. This process is called develop, and it involves removing the parts of the wafer that were exposed to the light.

Before this process, the wafer is heated again in an oven during a process called post-exposure bake (PEB). This helps to further intensify the changes in the photoresist that was exposed to the light.

Once the develop process is complete, a chemical agent known as a developer is applied to remove the areas of the photoresist that were affected by the light. Depending on the material used for the photoresist, the area may be rinsed. The solution used for rinsing differs according to the substance used for the photoresist. The rinsing equipment can also vary, and the choice of equipment often has a trade-off relationship between the processing speed and the rate of defects.

After the photoresist has been properly processed, the semiconductor frame is finally ready to be used. Transistors and wires can then be made by applying photomask on the photoresist that’s on top of the wafer or by carving out the desired parts.

The Development of Photolithography and Holistic Thinking

The photolithography process illustrates how important it is to have a deep understanding of the underlying technology rather than merely memorizing specific details. When the ArF laser’s light source reached its limit at 193nm, EUV technology wasn’t ready yet. But since microfabrication had to continue, industry professionals developed ArF immersion equipment to reduce the wavelength of the same light source. This allowed the semiconductor industry to approach processes below 100nm. Achieving this feat required contributions from other sectors.

▲ Figure 8: New technology added for the ArF immersion equipment

To use the immersion equipment, the wafer must be covered with a liquid that has a high refractive index—or water. However, this poses challenges as the semiconductor process is so advanced that even small errors can cause problems. Impurities in the liquid can cause defects in the semiconductor or photoresist and can even dissolve the photoresist over time. To address these issues, it’s necessary to develop technologies that produce high-purity water and to apply a protective coating that can be easily removed from the photoresist. Since the coating changes the properties of the photoresist, changes are also required in the develop process.

These challenges are something that the semiconductor industry and its professionals will have to work together to solve. Recently, SK hynix collaborated with a materials company to develop a dry photolithography process.

As explained in the previous article that focused on the process of oxidation, a dry process is a process that doesn’t involve water. Unlike the immersion method described above, the dry process involves attaching the photoresist directly to the wafer and not rinsing it during the develop process. Among the numerous reasons for developing these new technologies, the most important is that miniaturization has advanced to the extent that fine patterns are damaged in the process of applying and cleaning the photoresist even when the stepper shines light with a small photomask. Such challenges will continue to appear down the road, but the semiconductor industry will keep coming up with innovative solutions to combat them.

Photolithography is Just One Step

After a pattern has been successfully created using photolithography, the next step is to fill or carve off the insides of the pattern. While photolithography is one of the crucial processes of semiconductor manufacturing, it’s not the only important one. Creating a fine pattern through photolithography is only one step, as utilizing it for a specific purpose is a completely different story.

Read articles from the Front-End Process series

Read articles from the Back-End Process series

The post Semiconductor Front-End Process Episode 3: Forming Patterns on Wafers Through Photolithography first appeared on SK hynix Newsroom.

Producing and Connecting the Semiconductor Components

Before explaining the process of producing semiconductors, it’s easier to think about the makeup of ordinary electronic devices. It’s clear from looking at disassembled electronic equipment that a number of parts such as transistors, batteries, storage batteries, and coils are soldered on a PCB¹ board. When manufacturing this equipment, the discrete devices are made before being connected together.

¹PCB (Printed Circuit Board): A semiconductor board that’s made up of electronic circuits and has components soldered on its surface. These boards are found in most electronic devices.

▲ Figure 1. Just as computer CPUs were made in the past, various devices are soldered onto the PCB. (Image Source)

This order of procedure is also applied to the MOSFET (Metal Oxide Semiconductor Field Effect Transistor) that is placed on top of a silicon wafer. The first step in wafer manufacturing is to make various types of discrete devices. Although “made,” it’s more accurate to say that they are engraved on the wafer’s surface. This process is called FEOL (Front End Of Line). Afterwards, a process tantamount to soldering follows, but such small devices cannot be directly soldered on. So, techniques that are similar to FEOL are used to create fine wires that interconnect billions of semiconductor devices together. This process is called BEOL (Back End Of Line). The combination of these two processes is a wafer manufacturing procedure called the front-end process.

▲ Figure 2. The actual flow of the manufacturing process. A MOSFET is formed in the FEOL area, and on top of this area are metal wires which connect the components of FEOL instead of soldering them together. (Image Source)

Various processes including oxidation, photolithography, and etching are methods used in FEOL and BEOL. The frequency of using certain equipment varies according to the purpose of each respective step in the process, but the main purpose remains the same: creating many fine patterns.

▲ Figure 3. A simplified version of the semiconductor manufacturing process and the companies involved.

The eight main semiconductor processes are wafer manufacturing, oxidation, photolithography, etching, deposition, metallization, testing, and packaging, but such a categorization of the processes is not absolute. As evident from Figure 3, wafer manufacturing does not strictly occur within a semiconductor factory. Moreover, in the case of metallization, testing, and packaging, they are not single processes such as photolithography, etching, and deposition, but rather are broad categories of processing that are executed for a certain purpose.

Oxidation: Protecting Silicon Wafers Through Glass Coating

As shown in Figure 2, the semiconductor manufacturing process starts at the bottom and flows upward. To evenly create various shapes inside the semiconductor, numerous operations are required such as removing unnecessary parts and covering essential components with materials. As these steps involve a variety of highly reactive chemicals, semiconductor fabrication can be compromised if the chemicals come into contact with unwanted areas. In addition, there are parts in the semiconductor that may short-circuit if they are in direct contact with each other. Thus, there needs to be a way to create some type of a barrier between them, and this process is called oxidation.

The oxidation process involves covering a protective film on a silicon wafer. When silicon (Si) reacts with oxygen, it becomes silicon dioxide (SiO₂), otherwise known as glass. Glass is not only strong but also has low reactivity, so it is used to store various beverages as well as chemicals including hydrochloric acid and sulfuric acid. As a protective material that possesses these characteristics, oxide film blocks the entry of other substances and proves to be useful in the ion implantation process².

Additionally, oxide films are also used to intentionally block current flow. The core part of the MOSFET structure is the gate. Unlike previous transistors such as BJT³, the gate of the MOSFET does not directly touch the path of the current (the area between S and D in Figure 4) but only exerts an indirect influence. An oxide film, which is also called a gate oxide, is often used as the material to block the gate and the current’s path. In the case of the latest advanced semiconductors, the reduction in semiconductor sizes leads to various alternative materials such as HKMG⁴ being used as gate insulating films.

²Ion Implantation: The process of implanting group 3 or group 5 ions to convert pure silicon wafers into semiconductors in the semiconductor manufacturing process.
³BJT (Bipolar Junction Transistors): A transistor made using a PN junction, which refers to the boundary between two regions of a P-type semiconductor and an N-type semiconductor inside a semiconductor.
⁴HKMG (High-K Metal Gate): A newly developed MOSFET gate that effectively reduces leakage current. A transistor that replaces the polysilicon gate with metal and substitutes silicon oxide insulating film for High-k.

▲ Figure 4. A gate insulating film (shown in the red box) blocking the gate labeled G and the current path between S and D. Silicon dioxide (SiO₂) was previously widely used as such a film but now other materials are taking its place. (Image Source)

The 3 Types of Oxidation Processes

The oxidation process can be divided into three types: wet, dry, and radical.

▲ Figure 5. Types of Oxidations (Source: The Understanding of the Semiconductor Manufacturing Technology, p. 143)

Wet oxidation is a method in which a silicon wafer reacts with high-temperature vapor (water). In other words, it’s similar to rusting the surface of silicon using high-temperature water. A silicon oxide film can grow very fast through a wet process, but characteristics such as overall uniformity and density of the oxide film are deteriorated. As it’s not easy to adjust the characteristics of the film, it becomes difficult to use it in key instances where the product’s performance is significantly affected. In this process, hydrogen (H₂) is produced as a by-product.

In the case of dry oxidation, high-temperature oxygen gas is directly sent to the silicon wafer. Oxygen molecules are heavier than water molecules⁵, so they penetrate the silicon wafer relatively slower. Unlike in wet oxidation, hydrogen (H₂) is not generated as a by-product and, instead, a dense and uniform oxide film is produced. Due to these characteristics, the oxide film in the semiconductor gate—which has a significant effect on the performance of the final product—is made by dry oxidation.

⁵When the atomic mass of hydrogen (H) is 1, the atomic mass of oxygen (O) is 16. Therefore, the molecular weight of oxygen (O2) is 32 and the molecular weight of water (H2O) is 18, so the oxygen molecule is heavier.

Radical oxidation proceeds in a somewhat different process from the two oxidations discussed above. During wet and dry oxidation, natural gases are raised to a high temperature to emit energy and react with the wafer’s surface but radical oxidation takes this a step further. When oxygen atoms are mixed with hydrogen molecules at high temperatures, they are converted into highly reactive gases called “radicals” which react with silicon wafers. Radicals can produce oxide films of much higher quality compared to the dry oxidation process as radicals are highly reactive.

▲ Figure 6. Characteristics of Radical Oxidation (Source: The Understanding of the Semiconductor Manufacturing Technology, p. 149)

In addition, the thickness of the oxide film can be uniform even in a three-dimensional structure by using radical oxidation. Silicon wafers used by semiconductor companies are single-crystal, and silicon atoms on the entire surface of the wafer have the same crystal orientation.

The images at the top of Figure 7 show the structure of the silicon atoms, while the numbers at the bottom (100, 110) indicate the direction of the silicon crystals. In the case of wet and dry oxidation, the formation rate of oxide film in the direction above the wafer (100) is slow and the oxidation rate is fast in the side (110) direction. The reason for these differences lies in the fact that the silicon atoms arranged in the (100) direction are denser. If the atoms are closely arranged, it is difficult for dry or wet oxidizing gas to penetrate the crystal and react. However, radical oxidation is significantly less impacted by this issue.

▲ Figure 7. The appearance of silicon atoms according to the Miller indices.

In addition, radical oxidation creates a uniform oxide film even on the corners where it was previously difficult to do so. Radical oxidation induces silicon nitride (Si₃N₄)⁶ to cause an oxidation reaction as it has a relatively low reactive rate.

⁶Silicon nitride (Si₃N₄): One of the materials that acts as a protective film and is covered by deposition in the semiconductor device manufacturing process.

As the miniaturization of semiconductors seems to be getting closer to its technological limits, semiconductor companies are increasingly introducing three-dimensional structures into semiconductors. For this reason, the technology of making a uniform high-quality film is becoming more critical in semiconductors. This also applies to the oxidation process.

Oxidation Equipment: Gas Reacts With the Wafers

Although oxidation equipment is complex, the simplified image below shows the key components.

▲ Figure 8. Structure of wafer oxidation equipment

Gas is injected into the inlet of the equipment and heated inside to react with the wafer. Additionally, some dummy wafers are inserted alongside the loaded wafers. The dummy wafers are used to fill in the difference in reaction speed between the wafers at both ends of the device, which inevitably occurs due to the structure of the device. As dozens of wafers are simultaneously injected into the oxidation equipment, oxidation is considered to be a relatively quick process.

Read articles from the Front-End Process series

Read articles from the Back-End Process series

The post Semiconductor Front-End Process Episode 2: Protecting Key Components Through Oxidation first appeared on SK hynix Newsroom.

But how familiar people are about the history and rise of semiconductors is a different matter. To provide more context on these indispensable materials that power everything from household appliances to mobile phones, this series will trace back the origins of semiconductors and explain why they became such a significant part of everyday life as we know it.

Starting with ‘Computers and Transistors,’ a total of six chapters including ‘Process and Oxidation,’ ‘Photolithography,’ ’Etching,’ ‘Deposition,’ and ‘Metal Wiring’ will help explain the nature and processes of semiconductors. This series will put a special focus on the correlations between all of these technologies.

The Advent of Computers

As people continuously look for ways to simplify their day-to-day activities at home, work, and wherever they need to go, the need for technological devices has always been on the minds of innovative thinkers. Starting with simple machines that only knew how to make basic calculations, people eventually progressed to developing more advanced and accurate machines that would become of more practical use.

Inventing such a machine required tremendous contributions from various people. A major experiment that came out from one of these individuals was Charles Babbage’s Analytical Engine in 1871. Users could insert a thin plate called a punched card into the machine and perform a numerical calculation. When inserted into the machine, the inner analysis engine repeats various arithmetic operations according to specific commands and a result value is printed out from another part of the machine.

Although the Analytical Engine was never manufactured, it stands as an interesting case study. First, the Analytical Engine has all the elements of a computer. The punched card and the part of the engine where the result value is printed out is the same concept as the computer memory. So, the Analytical Engine is essentially the primitive CPU^* (Central Processing Unit).

^*CPU: Abbreviation for Central Processing Unit. A device that acts as the computer’s brain.

Simply put, the Analytical Engine was a computer that operated on steam and consisted of a memory and CPU unit made from pieces of metal and wood. Consequently, we can assume that people in the past knew how computers structurally functioned. It’s also clear that computers and ‘electronic circuits’ are completely different concepts. Therefore, there’s a need to know why electronic circuits became the heart of the modern computer.

Electrically Controlled Computers

Electronic circuits have advantages over other devices based on steam, manpower, or hydraulic power, because the control of signals is fast and efficient. Looking at steam, reaction rates are slow, as steam needs to physically reach a specific location. Furthermore, since steam is transmitted at a high pressure, the pipes need to be thick and, overall, it lacks efficiency. Now, let’s suppose there’s a device that opens and closes its door automatically when a rope is pulled. If steam is used as the energy source here, the operator will have to open the boiler valve and wait for the high-pressure steam to push the door in order for it to be closed. But when electricity is the energy source, a button and a motor is all that is required. The size of the entire device becomes smaller, while energy efficiency and reaction speed both increase.

▲ Figure 1. A steam-based automatic door (left) and an electric automatic door (right)

With the invention of electricity, controlling computers with it became the general trend. After numerous attempts to create an electricity-based computer, the ENIAC (Electronic Numerical Integrator And Computer) was eventually made. Unlike the Analytical Engine that used gears and steam, the ENIAC operated through the combination of a type of light bulb called the vacuum tube and various electronic circuits. By looking at the components of the ENIAC that resemble light bulbs, it’s quite clear that its energy source was electricity.

▲ Figure 2. ENIAC (Source: View Original Document)

The ENIAC was a huge computer that almost took up a whole room and used up to 170 kW of electricity, equivalent to operating 170 microwave ovens. Nevertheless, it was able to accomplish numerous tasks that were needed back then. Its operation speed was still a lot faster as it used more than 170,000 vacuum tubes instead of gears that squeaked and moved slowly. Since its development, the ENIAC contributed towards achieving many milestones, including formulation of the simulation methodology.

However, we know that the ENIAC’s performance was not even a match for the portable calculators of the 1990s. Efficiency was a major issue, and it wasn’t possible to supply these commodities on a large scale due to their size. This is why the world needed another innovation called the transistor.

The Emergence of Transistors

As aforementioned, the ENIAC was built using a vacuum tube similar to a light bulb. But it’s important to know why these devices were needed in the first place. People knew that the ability to control signals would lead to the creation of some type of computing device, such as the automatic steam door that we looked at above.

Thus, a computer is basically a device that adds a lot of inputs and outputs to an automatic steam door along with various logical structures that are added by connecting thousands of pipes in the interior. Automatic steam doors can only do simple tasks such as opening and closing the door. But a computer can be built when it’s possible to carry out more complex tasks such as simultaneously opening two doors with a single rope or making a safety door that does not close when a person is under the door. Ropes and steam pipes basically play the role of basic devices corresponding to a vacuum tube.

▲ Figure 3. An automatic steam door that opens multiple doors with one operation (left),
and an automatic door that opens only when two operators agree to open the door (right)

So, how can the performance of a steam computer be enhanced while adding extra functions to it? The number of steam tubes could be increased to add more functions, or a boiler with a higher pressure and temperature could be installed to increase the speed at which the steam rises. The problem with these solutions is that they are not easy.

Steam engines are very large by themselves, so adding a tube from the boiler to another area creates an even bigger burden on space. It requires too much energy, and the risks of explosion or other malfunctions also increase when trying to enhance the performance of the boiler. Vacuum tubes were merely the best devices available to engineers at the time. As they operate on electricity, there are no risks of explosions like with a high-pressure boiler. And the operating speed was clearly faster than the steam engine. Of course, there were frequent accidents such as individual vacuum tubes malfunctioning due to too much power being used. To make a better computer, it was necessary to find more advanced components.

In 1947, the transistor was invented. Transistors were innovative devices that could regulate the flow of large currents with very small currents. Scientists found that by using two types of semiconductor materials, as shown below, it was quite easy to disconnect and connect signals. Although the structure may look complex, the nature of its operation is essentially the same as controlling the movement of steam by pulling a rope. In the same year that the first transistor was invented, the BJT^* (Bipolar Junction Transistor), which is widely used to this day, was also invented. At this juncture, semiconductors also began to be known by the public.

^*BJT: The Bipolar Junction Transistor, within a semiconductor is a transistor made by using a PN junction, or the boundary between two domains of a P-type semiconductor and an N-type semiconductor.

▲ Figure 4. The structure of the transistor. Both N-type and P-type semiconductors are used.
(Source of the right image: The Understanding of the Semiconductor Manufacturing Technology, p. 143, Table 4-6)

Semiconductors for everyone: MOSFET’s Revolution and Its Manufacturing Technology

While working at Bell Labs in 1959, Dr. Martin Mohammed John Atalla and Dr. Dawon David Kahng developed a new type of transistor called the MOSFET (Metal Oxide Semiconductor Field Effect Transistor). The two scientists formed two types of semiconductor layers on a silicon disk and then placed metal on top of it to create a flat transistor. Although the operating principle of this transistor was slightly different, its usage was not too distinct from the transistors introduced above. But what made this transistor stand out was its productivity.

▲ Figure 5. Dr. Dawon Kahng’s MOSFET model structure (Source: Hanol Publishing Co., Ltd.)

Due to their flatness, numerous MOSFETs could be made on a silicon wafer simultaneously. If the outline could be made smaller, it was possible to make ten times more MOSFETs on wafers of the same size. Additionally, a set of already-connected MOSFETs could be manufactured simultaneously as well. Let’s say a CPU needs to be built using a BJT. No matter how efficient the BJT’s production process is, it’s necessary to solder hundreds of millions of BJTs together and attach them to the circuit board, as CPUs were made by connecting BJTs. As for MOSFETs, hundreds of millions of transistors are already soldered to the circuit board when they are produced.

Eventually, the whole point of semiconductor factories is to make MOSFETs more affordable. The succeeding chapters will explain how terms like exposure, etching, deposition, and other processes of making semiconductors contributed to producing affordable MOSFETs.

Read articles from the Front-End Process series

Read articles from the Back-End Process series

The post Semiconductor Front-End Process Episode 1: The Birth of Computers, Transistors, and Semiconductors first appeared on SK hynix Newsroom.