I need your help ! This page requires proofreading - If you notice any errors, please notify me
Bulk transport of the reactant molecule AB in a gas/vapor phase (gaseous precursor) to the reactorAB is activated by various energy forms such as chemical, thermal, plasma or photonDiffusion of AB across boundary layerAdsorption and Desorption of reactants onto the heated surfaceAB (g)→AB* (ads)Decomposition of the reactant to metalAB* (ads) →A* + B*Incorporation of A in film A* (ads) →A (s)Desorption of byproduct BB* (ads) →B (g)Transport of B from the surface across boundary layerBulk transport of Bads stand for adsorption1,3 and 8,9  are transport steps4,5,6,7  are reaction stepsAway from the surface of the substrate the velocity is high and pressure is low as the gas get closer to the surface the velocity decrease until It's zero on the surface (no sold-state diffusion at the surface only reaction) and the pressure increase (shorter MFP(λ))Ratio of Adsorption and Desorption depends on sticking coefficient γAB (AB bounces off surface) 0 ≤ γAB ≤ 1 (Good adhesion)Deposition requires deposited material to have good adhesion (The sticking together of particles of different substances)Assuming laminar flow and Knudsen NK < 1 (viscous flow) For transportJd = - D dC/dx Jd = - D/δ(x) (Cg - Cs) Jd = - hg (Cg - Cs) (molecules/cm2.s)where δ(x) is the boundary layer thickness (δ(x)↓ Desired → Jd↑), D is gas diffusivity = λ⟨𝒗x⟩/2 (For solids D = D0 exp -Ea/kT), Cg is gas concentration in bulk and Cs is gas concentration at surface, hg is the gas-phase mass-transport coefficientFor reactionJr = kCs (molecules/cm2.s)where k is the reaction rate constant = k0 e -ΔH/kT, ∆H is the energy change in reactionAt steady stateJr = Jd kCs = hg (Cg - Cs)kCs = hgCg - hgCs kCs + hgCs = hgCg Cs (k + hg) = hgCg Cs = hg/(k + hg) Cg Jr = kCs =  k [ hg/(k + hg) Cg ]Jr = khg/(k + hg) CgJr = Cg / (1/hg + 1/k)Film growth rate (𝒗) = Jr/Nf (cm-2s-1/cm-3)Film growth rate (𝒗) = khg/(k + hg) Cg/Nf (cm/s)Film growth rate (𝒗) = Jr/Nf = (Cg/Nf) / (1/hg + 1/k) (cm/s)where Nf is the number of atoms per cm3 in the film (5 ✕ 1022 cm-3 for Si)In laminar flow conditionsδ(x) = √ηx/ρu0 ⟨δ⟩ = 1/L 0 ∫L δ(x) dx⟨δ⟩ = 2/3 L √η/ρu0Lwhere u0 is gas velocity far from the surface ρ is density, η is viscosity, x is the distance along substrate and L is the substrate lengthReynolds number Re = ρu0L / η, It's a measure to the ease of gas flow⟨δ⟩ = 2/3 L/√ReThe slower process either transport or reaction controls growthFor slow transport, Transport limited growth, hg << k𝒗 = hgCg/Nf𝒗 = hgCg/NfAs hg = D / ⟨δ⟩ so hg = D/L 3/2 √Re𝒗 = Cg/Nf D/L 3/2 √Re𝒗 = 3DCg/2LNf √ReFor D = λ⟨𝒗x⟩/2𝒗 = 3λ⟨𝒗x⟩Cg/4LNf   √Re𝒗 ∝ T½ √u0The rate depend on gas dynamics control and reactor design but has a weak temperature dependanceFor slow layer-by-layer growth of epitaxial SiRequires high T, low pressure (u0↑), low gas viscosity (η↓, Re↑)Most CVD is done using transport limited growthFor slow reaction (constant gas supply), Reaction limited growth, hg >> k 𝒗 = kCg/Nf 𝒗 = Cg/Nf  k0e -ΔH/kT𝒗 ∝ e -ΔH/kT Arrhenius-likeThe rate has a strong dependance on temperature and reactants For Polysilicon at lower temperatureAt low T, reaction is rate-limiting; at high T, diffusion (transport) is rate-limitingΔ𝒗/ΔT > Δ𝒗/Δu0    Reaction limitedΔ𝒗/ΔT < Δ𝒗/Δu0    Transport limitedMass transport limited growth (high T):Should be able to control gas flow and pressure to get uniform filmsReaction rate limited growth (low T, low P): Should be able to control the temperature profile for uniform filmsThe susceptor (a material that absorb EM energy and convert it to heat) holding the substrate is tilted by 3o to 10o for more uniform gas velocity and concentration which in turn in crease the uniformity of film growth rateWhile PVD delivers atoms or small groups of atoms to the substrate surface, CVD carries molecules. These molecules can adsorb/desorb or diffuse on the growing film many times before they decompose Deposition rates are temperature dependent according to the Arrhenius equation, but they are on the order of 0.1–10 nm/s.For materials such as silicon, silica glass, and silicon nitride, and other dielectrics, CVD is the simplest and the most cost effective wayFor conductor materials, physical vapor deposition (PVD) is a more traditional way to deposit; only the cases where PVD cannot or very difficult to achieve, CVD prevail such in step coverage requirements, for example, Tungsten required to have 100% step coverage to fill high aspect ratio via, is dominant by CVDCVD Reactors:Hot wall reactor (The whole reactor walls are heated)It has more uniform temperature distribution which make it easier to control temperature but the surface of reactor gets coated so system must be dedicated to 1 species to avoid contaminationHot-walled LPCVD is mostly used for poly-Si growth also It's used for low pin-hole SiO2 and conformalityCold wall reactor (Only substrate is heated)The reaction rate is reduced which reduce deposition on surfaces so film quality can be better controlled. Better for epitaxial filmsCommon CVD processesSiliconEpitaxial silicon can be deposited on a crystalline silicon wafer using CVDTransport-limited growth are used for slow layer-by-layer growth of epitaxial Si Silane pyrolysisFor crystalline silicon, LPCVD is used because Poly Si is grown at 1 atm and high pressure. Increasing the pressure further etch the filmSiH4 (g) → Si (s) + 2H2 (g) (∼600 ◦C)The deposited film contain hydrogenReactions isn't that simple and form other species also:SiH2 (g) →   Si (s) + 2H2 (g)SiH4 (g) →   SiH2 (g) + H2 (g)     silyleneSiH4 (g) →   Si2H6 (g) + SiH2 (g)disilaneSiH4 (g) →   Si (s) + 2H2 (g)All these different species are present in the chamber and a gas phase equilibrium is set up. The growth rate of the Si is calculated using equations such as (In practical there is even more equations and parameters):1. Total pressure = ∑ partial P for each gasPtot = PSiH4 + PH2 + PSiH2 + PSi2H6 2. The conservation ratio ofSi/H = [ PSiH4 + PSiH2 + 2PSi2H6 ] / [ 4PSiH4 + 2PH2 + 2PSiH2 + 6PSi2H6 ] = constant 3. Each reaction has an equilibrium constantk = [PSiH2 PH2] / PSiH4 = k0e -ΔH/kTFor AB →A + BThe equilibrium constant k = PAPB/PAB The total pressure Ptot = PA + PB + PABSiCl4 can also be used (tetrachloride reduction)SiCl4 (g) + 2H2 (g) → Si (s) + 4HCl (g) (∼1200 ◦C)If run in reverse to clean the substrate, by adding HClLPCVD is also used to deposit poly or amorphous SiDopingPhosphine  2PH3 (g) → 2P (s) + 3H2 (g)Diborane  B2H6 (g) → 2B (s) + 3H2 (g)GaAsTrimethyl Ga (TMG) reduction(CH3)3Ga + H2 →   Ga (s) + 3CH4 Arsine decomposition 2AsH3 → 2As (s) + 3H2Metals and DielectricsSilicon OxideSilicon dioxide deposition (USG - undoped silica glass)SiO2 films can also be formed by low pressure (~100 mTorr) chemical vapor deposition (LPCVD) or plasma-enhanced CVD (PECVD)Thermal oxidation consumes Si from the substrate and is done at very high temperatures. CVD does not consume Si from the substrate and can be done at much lower temperatures (used when there is metallization on the wafer)The CVD process reacts a Si-containing gas such as SiH4 with an oxygen-containing precursor, causing a chemical reaction, leading to the deposition of SiO2 on the substrateHigh-temperature (CVD) Oxide (HTO):SiCl4 (g) + 2H2 (g) + O2 (g) → SiO2 (s) + 4HCl (g) ∼900 ◦CLow-temperature (CVD) Oxide (LTO), better film quality, Silane oxidationSiH4 (g) + O2 (g) → SiO2 (s) + 2H2 (g) ∼425 ◦C - LPCVDTEOS is the name of the precursor molecule tetraethoxysilane Si(OC2H5)4, but it is used as the name for the resulting oxide too (high-quality oxide)Si(OC2H5)4 → SiO2 (s) + gaseous byproducts ∼700 ◦CPOCl3 gas can be added to get phosphorus-doped oxide deposition (PSG - phosphorus-doped silica glass) also borophosphosilicate glass (BPSG), or fluorosilicate glass (FSG) can be depositedSilicon Nitride3SiH2Cl2 (g) + 4NH3 (g) → Si3N4 (s) + 6H2 (g) + 6HCl (g) ∼800 ◦CTungstenCVD tungsten is deposited in two steps. The silane (SiH4) reduction step deposits a thin nucleation layer over every surface in the systemWF6 (g) + SiH4 (g) → W (s) + 2HF + H2 (g) + SiF4 (g)A high-rate blanket deposition with hydrogen reduction is used to achieve the desired total thickness:WF6 (g) + 3H2 (g) → W (s) + 6HF (g) ∼400 ◦C - PECVDThis process is able to fill holes and trenches in multilevel metallizationCVD processes depend on both chemical reactions and flow dynamicsA high flow rate supplies enough reactants but the deposition rate is limited by slow surface chemical reactions (surface reaction limited)A fast surface reaction consumes source gas rapidly and the deposition rate is limited by gas supply (mass transport limited or diffusion limited)Other CVD methods used for epitaxy growthVapor-phase epitaxy (VPE) is the most common method for epitaxy growthMetal Organic CVD (MOCVD) for compound semiconductors are a more challenging example of CVD where not only must a film be deposited, but single-crystal growth must also be maintainedMolecular-beam epitaxy (MBE)Atmospheric Pressure CVD (APCVD)Not used today High Pressure (No vacuum) (equilibrium) (λ is small)Slow mass transport and Large reaction rates (Transport limited) Film growth limited by mass transferQuality of APCVD Si from silane is poorStep coverage is not very goodContaminationStochiometry is hard to maintainLarge number of pinhole defectsQuality of APCVD dielectrics is betterSiH4 (g) + 2O2 (g) → SiO2 (s) + 2H2O (g) (240 - 450 ◦C)Done in diluent gas ambient (usually N2) is used to prevent gas phase nucleation by reducing the reaction rateAdd 4 - 12% PH3 only for doping but ASH3 and PH3 can decrease the deposition rate and create thickness variations in the filmLow Pressure CVD (LPCVD)Low Pressure (Vacuum) (non-equilibrium) (λ is large) which increases the growth rate (P↓ → hg ↑, Dg ↑, δ(x) ↓)Higher hg so higher diffusively of gas to the substrate which improves transport and reduces boundary layerAlso extends reaction-limited regime (the curve moves up). Reducing pressure allows reaction-limited growth at higher T (preferred)LPCVD is reaction-rate limited because in low pressure, non-equilibrium isn't achieved and λ is large, so growth rate is reaction limitedUsed for dielectrics and semiconductorsLow partial pressure is used to reduce nucleation of products in gas phase Requires no carrier gas (Not transept limited)Fewer gas-phase reactionsFewer particulatesFewer defects due to low pressureEliminates boundary layer problemSlow growth rateGood conformal growth (Better step coverage, better film uniformity)Strong temperature dependence to growth rate (Reaction-rate limited)Plasma-Enhanced CVD (PECVD) Plasma (RF or DC) allows lower temperatures and lower pressures to be used and improves film qualityBecause high temperatures cannot be used in many cases such as when oxide needs to be deposited on aluminum (m.p. 650 ◦C) in MOS metallization to avoid Al contact interaction with Si and SiO2, ILD must be deposited at T < 450°C As in sputtering, the energetic ions transfer their energy (enhance surface diffusion at lower temperature) and momentum (enhance step coverage) on the reactant gas molecules and atomsThe energy transfer breaks up the molecules and aids the chemical reactions so higher deposition rates can be achieved at lower temperaturesAt low T, surface diffusion is slow. To supply kinetic energy for surface diffusion, plasma is used in PECVDPlasma is also used to enhance source gas decomposition and reactions. This results in deposition rates similar to thermal CVD, 0.1–10 nm/s, at much lower temperatures, typically around 300 ℃. Lower deposition temperature results in less dense filmsCompared to sputtering, the pressures are higher (50 mtorr – 5 torr) so ions are less energetic and undergo more collisions to lose energy which reduces the sputtering effects on the substrate A simple parallel-plate diode reactor is used for PECVD. Wafers are placed on a heated bottom electrode, the source gases are introduced from the top, and pumped away around the bottom electrodeThe operating frequency is often 400 kHz, which is slow enough for ions to follow the field, which means that heavy ion bombardment is present but if 13.56MHz frequency is used, only the electrons can follow the field, and the ion bombardment effect is reducedThermal CVD is controlled by pressure, temperature and flow rate. In PECVD there is additionally the RF powerIn advanced PECVD reactors, RF power can be applied to both electrodes, and the two power sources can supply different frequencies, duty cycles and power levels. The ratio of 13.56 MHz power to kilohertz power is important for film stress tailoringPECVD shares many beneficial features of both thermal CVD and sputteringThermal oxide or LPCVD nitride is stoichiometric SiO2 and Si3N4 , with ratios 1:2 and 3:4 of atoms, but most PECVD films are non-stoichiometric: for example, plasma nitride is best described as SiNx (x ≈ 0.8)PECVD can be used to deposit mixed oxides, nitrides and carbides, as well as doped oxides just like thermal CVDUse direct and indirect plasma reactorsHigh-density plasma CVD (HDPCVD) has superior gap-fill properties