The preliminary cleaning step is primordial to kickstart the other cleaning processes. The purpose is to remove large impurities that might be present on the pre-processed wafer. 1)
The oxidation reaction removes sulfur residues 2) $$ S_x + 2H_2SO_4→ 3SO_2 + 3H_2O $$
For the HF reaction 3) $$ SiO_2+6HF→H_2SiF_6+2H_2O $$
The purpose of this initial cleaning step is to remove organic contaminants, such as silicon films and trace metals, from the surface of the wafer. 4) This involves submerging the silicon wafer in two consecutive cleaning solutions: SC-1 and SC-2. Complexing refers to the process whereby a positively charged metallic ion shares electrons with the NH₄OH molecule. This is followed by oxidative desorption, a first-order reaction that oxidises the product to be removed, which can then be flushed out by reacting with a strong acid (HF). The result is a water-soluble solution.
The ratio used for the SC-1 solution is as follows: a range of 5:1:1 to 7:2:1 of H₂O, H₂O₂ and NH₄OH, respectively. 5)
Metalic Ion (charged positively) : $$ Cu²⁺ + 4 NH₄OH → [Cu(NH₃)₄]²⁺ + 4 H₂O $$
Desorption reaction: Thermal desorption of gasses 11) Organic components react with the peroxyde in change produces CO2, water and heat.
$$ CₓHᵧ + H₂O₂ → CO₂ + H₂O $$
Overall, the SC-1 step is a slightly exothermic reaction that produces heat. The ideal operating conditions are a temperature range of between 70 and 80 degrees Celsius. As the reaction also presents low thermal stability, temperature control is essential. Immersion time is also an important variable for this step.
The key objective of this cleaning step is to remove alkalyne residues, trace metals and hydroxyde metals. 12)
The composition of the SC-2 solution is as follows: a ratio of H₂O, H₂O₂, HCl by volume ranging from 6:1:1 to 8:2:1. 13)
$$ M_x + 2HCL→ M_xCL_2 + 2H^+ $$
The reactions occurring during these steps are more stable; therefore, the temperature and immersion time in the bath do not need to be controlled as strictly. The solution temperature must be within the range of 70 degrees Celsius +/- 5, with a bath time of approximately 10 minutes.
The aim of this final step is to remove the residues of oxidation that were produced in the previous reactions.
$$ SiO_2 + 4 HF → SiF_4 + H_2O $$
This step involves immersing the wafer in a hydrofluoric acid solution and then rinsing it with deionised water. This makes the oxides soluble. The immersion takes around 15 seconds. It's a cleaning step that can be used multiple times in a cleaning procedure, as it is highly effective at removing silicon oxides formed during etching processes.
This reaction works in the same way as the SC-1 to clean photoresist specifically. Piranha eats away at the photoresist mask and any heavy organic components present on the wafers.
$$ H₂SO₄ + H₂O₂ → H₂SO₅ + H₂O $$
This creates the possibility of producing peroxymonosulfuric acid, which reacts with the organic components of the photoresist film. $$ CₓHᵧ + H₂SO₅ → CO₂ + H₂O + SO₄²⁻ $$
This process seems similar to SC_1; the operating temperatures vary greatly depending on the stability of the reaction. It also appears to be the initial preliminary step applied to wafers, as it uses a mixture of highly reactive acids towards organic components. This means it is used to clean off the larger imperfections attached to the wafers beforehand.18)
This is another step whose goal is to clean the organic components present on the wafer after RCA cleaning. This is done through baths or spray processes, which cover the wafer in highly concentrated ozone solutions.
The solution consists of these components at an O₃ concentration varying between 10 and 15 ppm.
$$ –(CH₂)ₙ– + 3n O₃ → n CO₂ + n H₂O $$
This step illustrates the various ways in which peroxide solutions can be used to clean silicon wafers. APM refers to ammonia peroxide mixtures. BPM stands for basic peroxide mixtures. HPM stands for hydrochloric peroxide mixtures. These steps are mostly related to RCA procedures. The same chemical principles underpin them all, with different mixtures of H₂O₂ being used to remove organic residues. While these steps may resemble the RCA SC procedures, they do not react with trace metals through oxidative desorption.
$$ CₓHᵧ + H₂O₂ → CO₂ + H₂O $$
Some of the reactions associated with different concentrations of H₂O₂ vary in terms of operating conditions. For example, if you look at a solution with the same concentration of H2O2 as the SC-1, the temperature must range from 70 to 75 degrees Celsius. However, with alternative dilution and concentration ratios, the operating temperature can range from 100 to 130 degrees Celsius, with immersion times of 10 to 15 minutes.
Plasma cleaning is one of the dry cleaning methods. When plasma is mentioned in the literature, it can refer to multiple processing steps, such as bulk photoresist removal, etching, predeposition cleaning, and surface conditioning. In processing cleaning, this step only involves removing bulk photoresist material after etching and ion implantation. The intake of 10% H₂ and 90% N₂ (specifically for O₂ plasma) creates a non-oxidising environment when used to feed the devices. They also act as plasma cleaning gases.
$$ CₓHᵧ + O* → CO₂ + H₂O $$
Plasma cleaning is performed at low pressure, depending on the gas used and the specific technique employed. There are multiple techniques, each with a different purpose. Ar and H₂ bombardment, for example, are used to clean native oxides and thin contamination layers.
This cleaning step involves exposing the surface to UV light of a specific wavelength. This light reacts with certain organic molecules and ambient oxygen to create ozone (O₃). The resulting ozone has a strong oxidising potential, enabling it to react with the dissociated molecules that are by-products of the UV reaction. As the reaction continues, the O₃ is also destroyed by absorption.
This is the global formula for ozone cleaning. $$ CxHy+O∙(1D)/O3→CO2+H2O $$
Where $$ O∙(1D)/O3 $$ The generation of ozone is caused by the reaction of O2 with UV.
This is usually done at 120°C and 500 Torr.
It is also a dry-cleaning method.
This is done through the spraying effect of a mixture of cryogenic gases.
It includes the Ohmi Clean process.
Legend: QDR: Quick Dump Rising bath ; FR: Final Rinsing bath ; SD: Spin dryer ; EDR: Dump Rinsing bath
This processing step has the goal of creating a layer of silicon dioxide (SiO2), contrary to the cleaning step, this thin layer of silicon dioxide serves as a growth step, it acts as the dielectric component of the wafer. 19) 20). This process can be achieved with two distinct methods. n
Process in which the wafer is exposed to high temperature water vapor. The oxidation occuring creates a film that will act as the gate dielectric agent. 21) 22)
$$ Si+2H_2O →SiO_2+2H_2 $$
This film is thicker and presents less uniformity than the one created through a dry oxidation process. The offset is that the growth rate present on the silicon wafer is faster when using this method. The temperature of operation is between 800 - 1000° Celsius.
This method exposes the wafer to an environment consisting of only Oxygen. The oxidation occurs more slowly then with the wet processing, but produces a more uniform and denser silicon dioxide layer. 23)
$$ Si+O_2 →SiO_2 $$
Process temperature varies between 900 - 1000 Celsius.
This process consists of adding thin film to the wafer, who's role will either to act as a structural layer or a spacer to be etched. These process are physicochemical processes. 24)
Various precursor gases are sent to the reactor (vacuum chamber) whose's atmosphere consists of pure inert gases. The reaction occurs with the surface of the wafer thanks to the action of the plasma generated with a certain electrical field. The plasma forms radical that in turn react and form a coating on the wafer.
Various chemical reactions occur depending of the precursors. 25) 26)
$$ SiH_4+NH_3→Si_3N_4 + H_2 $$ $$ SiH_4+N_2O→SiO_2+ N_2 + H_2 $$ $$ SiH_4+ CH_4 → SiC + H_2 $$
Pretty low temperature, can vary from room temperature to 400 Celsius.
This deposition method relies on a metallic precursor reacting with the surface thanks to the presence of plasma. Every cycle also includes a purge that removes the excess reactants and by-products. These cycles are composed of an Half cycle A where the precursor is deposited on the surface. The second cycle is where the plasma reacts with the said precursor to form the film. The presence of plasma also eliminates the presence of excess reactants and by products on the surface.
Half cycle A, deposition on the surface $$ –OH* (surface) + Al(CH₃)₃ → –O–Al(CH₃)₂* + CH₄ $$ Half cycle B, reaction of precursor with the plasma $$ –Al(CH₃)ₓ* + O* (plasma) → –Al–OH* + CO₂ + H₂O $$
These reactions can be done at lower operating temperature due to the presence of plasma in the reaction.
Plasma cleaning is often used after the thin film deposition processes to clean out the chamber in which the reactions took place. Plasma reacts with the oxides and film remaining on the walls. 29)
Plasma formation : $$ NF₃ → NFₓ + F* $$ Reaction with the film $$ SiO₂ + 4F* → SiF₄ + O₂ $$ $$ > Si₃N₄ + 12F* → 3SiF₄ + 2N₂ $$
g
The aim of these processes is to create a photoresistive layer. This layer will subsequently be exposed using either deep ultraviolet (DUV) or extreme ultraviolet (EUV) light. The photoresist is formed using a resin.
This process uses two types of laser: either a KrF laser with a wavelength of 248 nm, or an ArF laser with a wavelength of 193 nm.
Hexamethyldisilazane surface functionalization
$$ Si–OH + (CH₃)₃Si–NH–Si(CH₃)₃ → Si–O–Si(CH₃)₃ + NH₃ $$
This reaction makes the surface hydrophobic.
$$ Polymer–O–C(CH₃)₃ + H⁺ → Polymer–OH + C₄H₈ (isobutylene) + H⁺ $$
Photolithography processing is a crucial step in which a pattern from a mask is applied to a substrate surface. In this case, the substrate is a wafer. Photolithography relies on the interaction between a photoresist polymer layered on the substrate and a specific wavelength of light. Patterns are drawn onto the wafer through this interaction, and by repeating this step, wafers with multiple layers can be built. Below is a list of the main stages of the photolithography process, which are repeated in the order presented (note: all these processes are shared by the different photolithography methods). 41)42):
Substrate preparation: The substrate's surface must be cleaned and prepared for good adhesion of the photoresist. Traditionally, this is done by cleaning the wafer surface with the cleaning processes in section 1, performing a dehydration bake to remove the water adsorbed by the substrate, and finally applying an adhesion promoter. The dehydration bake is usually performed at temperatures ranging from 200 to 400°C for 30 to 60 minutes in a dry environment. The adhesion promoter is then immediately applied to the dry surface. These promoters are typically silanes, most notably hexamethyl disilazane (HMDS), which react with the silanol groups on the wafer surface to improve photoresist adhesion.
Photoresist coating (see above): This step involves applying the photoresist to the surface of the wafer, typically using a spin or spray coating method. Before coating, the photoresist material is liquefied by mixing it with a solvent. The difference lies in how the spinning action is performed. Either the wafer is spun with the photoresist solution applied to its surface, or the wafer remains stationary while the solution is coated dynamically. The important parameters of these steps are the uniformity and thickness of the resulting coating.
Post-apply bake (also soft bake): This baking step occurs directly after the photoresist coating. The objective is to remove excess solvent from the photoresist surface. This is traditionally done by baking in a convection oven at 90°C for 30 minutes. Removing the excess solvent increases the stability of the photoresist coating.
Alignement and exposure: In short, this step involves projecting light through a photomask to draw patterns onto a photoresist-coated substrate. The wavelength of the light depends on the technique used: 248 nm or 193 nm for deep ultraviolet, and 13.5 nm for extreme ultraviolet. Between each exposure step, the mask must be aligned with the last pattern drawn on the wafer to maximise the packing capacity of the wafer. An important side effect of this step is the standing wave effect, which occurs when ultraviolet light is reflected off the wafer's surface. This causes ridges to be imprinted between every layer of resist on the wafer.
Post-exposure bake: The objective of this step is to reduce the standing wave effect resulting from the previous processing step. This is achieved by applying a bake to the processed wafer. The heat reacts with the chemically amplified resist, resulting in smoother edges.
Development: In the development step, a developer — typically an aqueous solution containing tetramethylammonium hydroxide (TMAH) — is applied to the surface of the photoresist to control its profile. This is achieved using a variety of application methods, such as spraying and puddling.
Post-bake: This step involves the final heat treatment, which hardens the wafer surface in preparation for the harsh conditions of subsequent processing steps. This baking is carried out at temperatures ranging from 120 to 150 degrees Celsius.
This type of photolithography uses krypton fluoride (KrF) lasers to produce deep ultraviolet light at a wavelength of 248 nm. This light is generated by electrical discharges sent through a chamber containing krypton, fluoride and neon. The optimal pressure conditions for the gas mixture in the chamber are between 2.6 and 3.1 atm. 43) The F2:Kr:Ne gas mixture is 3:70:2600 mbar. This method enables a high power output of 300 W and a reflection coefficient of 99%. Neon acts as the buffer and carrier gas. Between operations, the chamber is purged with pure N₂ to minimise interaction with accumulated air or water. The purge gases are filtered as they are output and recirculated.
The reaction that occurs in the chamber goes as follows 44) : $$ Kr + F₂ + Ne_ → KrF_ → KrF + hν (248 nm) $$ The resulting nodes of lithography vary from 250 nm to 90 nm.
This photolithography process is the next generation of technological improvement to DUV photolithography. It is based on the same principles as KrF photolithography, but uses a laser that operates with a different mixture of gases. A 193 nm wavelength deep ultraviolet light is produced by an argon fluoride laser. 45) The proportions of the gases used are as follows: Helium (94.77%), neon (5%), fluorine (0.23%) and argon (trace amounts). 46) The purge gas is also nitrogen.
The reaction occuring goes as follow : $$ Ar + F₂ + Ne_ → ArF_ → Ar + F + hν (193 nm) $$
The resolution node varies from 65 to 130 nm.
This technique uses the same ArF laser, but a fine layer of ultrapure water is placed between the laser and the wafer surface to improve reflectivity and patterning precision. 47) This technique enables nodes to be made much smaller than before, with a resolution of around 20 nm. 48) The ultrapure water is constantly recirculated in the chamber.
This technique relies on the formation of a laser that generates extreme ultraviolet light with a wavelength of 13.5 nm. This laser is induced by the formation of a tin plasma. Droplets of tin are dropped into a chamber and hit by a first CO₂ laser, which causes them to expand. They are then hit by a more powerful CO₂ laser, which transforms them into plasma. This plasma then emits light at a wavelength of 13.5 nm. 49) For this to work, the environment must be in a near-vacuum state. 50) Hydrogen (H₂) is used for this purpose since it absorbs little EUV light. 600 litres of H₂ are fed into the operating chamber. 51) In the scanner, the atmospheric gases are a mixture of nitrogen and argon. 52)
Capacitor cell
Doping processes are used to introduce conductivity into silicon wafers. These steps are essential for controlling parameters such as junction depth, carrier mobility, leakage current and switch speed. 53)
This method of modifying the conductivity of a patterned wafer involves the physical process of directing ions of a specific element towards the wafer's surface, where they act as either n-type or p-type dopers. These ions are conveyed through doping gases, which are carried by inert gases in a chamber. These ions are produced by bombarding the doping gases with electrons, which are typically generated using a hot tungsten rod. 54)
Here are the electron bombardment formulas showing how ions are generated depending on the desired elements. 55)
$$ BF₃ + e⁻ → B⁺ + 3F + 2e⁻ \; (p-type \;dopant) $$ $$PH₃ + e⁻ → P⁺ + 3H + 2e⁻ \; (n-type \; dopant)$$ $$AsH₃ + e⁻ → As⁺ + 3H + 2e⁻ \; (n-type \;dopant) $$
This is usually done at room temperature in a vacuum environment.
This method uses high-temperature furnaces to drive dopants in gaseous and liquid forms into the surface of the wafer, with inert carrier gas transporting them. These dopants are typically composed of boron and phosphorus. 56)
Gas phase decomposition for phosphorus oxychloride 58):
$$ 4 POCl₃ + 3 O₂ → 2 P₂O₅ + 6 Cl₂ $$ Followed by the doping with P : $$ 2 P₂O₅ + 5 Si → 4 P + 5 SiO₂ $$
Gas phase decomposition for Boron tribromide : $$ 4 BBr₃ + 3 O₂ → 2 B₂O₃ + 6 Br₂ $$ Followed by the doping with B : $$ 2 B₂O₃ + 3 Si → 4 B + 3 SiO₂ $$
The operating temperature of diffusion furnaces hovers around 800–1,200°C.
This process takes place after ion implantation and serves to recondition the semiconductor's crystal lattice. This is achieved by heating the material to a high temperature, which activates the recrystallisation process on the surface. 59)
This method is a type of chemical vapour deposition (CVD), which involves doping the wafer surface using silicon-derived gases to form a crystalline thin film. The formation of the doped film on the wafer surface is referred to as epitaxy. 60) This is achieved by operating at high temperatures, depending on the favoured inputs and reactions.
Silane pyrolysis 63)
Silane pyrolysis $$ SiH₄ → Si (epitaxial) + 2 H₂ $$ Dichlorosilane - H2 reduction $$ SiH₂Cl₂ + H₂ → Si (epitaxial) + 2 HCl + H₂ $$ Trichlorosilane $$ SiHCl₃ + H₂ → Si (epitaxial) + 3 HCl $$ n-type doping : $$ PH₃ → P + 3/2 H₂ $$ p-type doping : $$ B₂H₆ → 2 B + 3 H₂ $$
### Manufacturers
### Equipments
Chemical mechanical polishing (CMP) (also called chemical mechanical planarization) is a process of smoothing surfaces with the combination of chemical and mechanical forces. It can be thought of as a hybrid of chemical etching and free abrasive polishing.[https://en.wikipedia.org/wiki/Chemical-mechanical_polishing] d
Key elements of CMP are:
There are three primary types of CMP: oxide CMP, tungsten CMP, and copper CMP
Oxide CMP is primarily used to flatten dielectric layers. It is used to:
Tungsten CMP is used to remove excess tungsten after it has filled openings. It is used to:
The most widely used approach uses ferric nitrate (Fe(NO₃)₃) or hydrogen peroxide (H₂O₂) as the oxidising agent to convert the tungsten surface to a soft tungsten oxide (WO₃) layer, which is then mechanically removed by the abrasive.
followed by:
Copper CMP is used to form copper interconnections, particularly in Damascene or dual-layer schemes It is used to:
The most specific feature of copper CMP slurry is the use of benzotriazole (BTA) as a corrosion inhibitor.
d
### Equipment
### Misc
### Sources