A flat wafer sits inside a hot chamber. Gas flows left-to-right above it. Right against the wafer surface the gas is nearly still — that stuck-still layer is important later. Somewhere in that gas are the useful molecules (the "reactant"). They must travel down to the surface, react into solid film, and the leftover gas must travel back up and leave.
Everything below is just naming the pieces of this picture.
Why the topic needs it: the whole reaction is fed by reactant molecules. More crowding → more collisions with the surface → faster growth. So C is the fuel gauge.
Two special values appear in the parent note:
Cg — concentration up in the bulk gas (the "g" is for gas/bulk), where flow keeps it well-stocked.
Cs — concentration right at the surface (the "s" is for surface), where reaction is eating molecules up, so it is usually lower.
Why the topic needs it: growth speed is flux. If more molecules land and react per second, the film thickens faster. Everything in the derivation is an equation for F.
Units picture: molecules per (area × time), e.g. molecules/(cm2⋅s).
Because the boundary layer is nearly still, reactant cannot be carried through it by flow — it must diffuse (wander) across on its own. That crossing is slow, and it is one of the two bottlenecks.
Why the topic needs it: it is the "distance" molecules must struggle across to reach the surface. A thicker or slower-crossing boundary layer means fewer arrivals per second.
We need a number that says how quickly the concentration difference Cg−Cs turns into arriving flux. That number is hg.
Why the tool: we chose a linear law (flux ∝ difference) because for small differences diffusion genuinely is proportional to the concentration gap — this is the simplest law that matches the physics, so we use no fancier tool than it needs.
hg shrinks at high pressure and grows at low pressure — the reason LPCVD works (diffusivity ∝1/P).
Arriving is only half the job. Once at the surface, a molecule must actually react into solid. That eagerness is ks.
Notice it uses Cs (surface value), not Cg — the surface can only react with what has reached it.
This is called "first-order": flux is proportional to the first power of Cs (not Cs2, etc.). We pick first-order because for many CVD reactions doubling the surface reactant doubles the rate — again the simplest matching law.
ks is not a fixed number; heat makes molecules react much more eagerly.
Let us earn every symbol in that formula:
T = absolute temperature in kelvin (K). Picture: a thermometer starting from the coldest possible point, 0 K. Hotter = molecules jiggle harder.
Ea = activation energy = the energy "hill" a molecule must climb to react. Picture: a hump between "reactant" valley and "film" valley. High hill = hard to react.
k = Boltzmann's constant = the exchange rate between temperature and energy, k=8.617×10−5eV/K. It converts "how hot" into "how much jiggle-energy."
k0 = a top-speed constant = how fast reaction would go if the hill were free (no barrier).
e−Ea/(kT) = the fraction of molecules with enough energy to clear the hill. It is a number between 0 and 1.
The parent note sets F1=F2. Here is why that is allowed.
Why the topic needs it: it lets us set arrivals = reactions (F1=F2=F), which is the one equation that pins down the unknown Cs and gives the boxed flux formula. Without it we would have two unknowns and no way to solve.
Solving hg(Cg−Cs)=ksCs gives the parent note's result
F=ks+hgkshgCg,
which you now have every symbol to read.
The parent note keeps saying "conformal." Define it against its opposite.
This is exactly why CVD (gas everywhere) beats PVD (line-of-sight) for filling deep holes, and it links to Epitaxy and Thermal Oxidation as film-forming cousins. Measuring the result is the job of Thin Film Metrology.