Planar 1959 Step 12 Oxide Thickness Growth Rate Tradeoff

# Choose Wet Over Dry Oxidation to Hit Masking Thickness Without Blowing the Furnace Schedule: The Tradeoff Step 11 Set Up

## 1. Why the Fast Way to Grow Oxide and the Dense Way to Grow Oxide Are Not the Same Furnace Run

Step 11 derived the growth law; this step applies it to a real scheduling decision this series has to make every time it grows a masking oxide — whether to oxidize in dry oxygen or in steam, because the two oxidants give very different growth rates through the same furnace at the same temperature, and this series cannot afford either extreme on its own. Steam oxidizes silicon roughly an order of magnitude faster than dry oxygen at a shared temperature, because water diffuses through growing oxide far more readily than molecular oxygen does — but the oxide steam produces is measurably less dense and a weaker diffusion mask per unit thickness, while dry oxide is slow enough that reaching Step 1's required masking thickness inside a workday can cost hours this series' furnace throughput cannot spare.

$$B_{\text{wet}} \gg B_{\text{dry}}, \qquad \rho_{\text{wet}} < \rho_{\text{dry}}, \qquad t_{\text{required}} = f\!\left(x_{\text{ox, target}},\ B,\ \rho\right)$$

where $B$ is the parabolic rate constant from Step 11 and $\rho$ the resulting oxide density, which this series' own masking-adequacy requirement from Step 1 depends on along with raw thickness. A furnace schedule that blindly chases speed by running steam the whole cycle risks a mask too porous to trust; one that insists on dry oxygen throughout risks a schedule no production line can sustain — which is why Fairchild's actual furnace recipe for this process runs a dry-wet-dry sequence rather than picking one atmosphere and staying there.

Speed Against Density, Not Speed Against Nothing wet oxidation grows fast and loose; dry oxidation grows slow and tight WET (STEAM) fast growth lower density DRY (OXYGEN) slow higher density, better mask fairchild recipe: dry → wet → dry dense seed layer, fast bulk growth, dense cap layer B₂₅₝ ≫ B⁚ₙ₥, ρ⁚ₙ₥ ≥ ρ₂₅₝ — neither atmosphere alone satisfies both constraints this is the schedule hiding inside every continuous oxide step 1 simply assumed was already there

## 2. Real Diagram: A Three-Layer Oxide That Reads as One Continuous Layer to Every Later Step

Because every later step in this series — the contact window at Step 3, the etch at Step 4, the coverage check at Step 7 — treats the masking oxide as a single uniform layer, the dry-wet-dry sequence has to produce an oxide that behaves as one film electrically and mechanically even though it was grown in three distinct furnace atmospheres with different rates.

Three Furnace Atmospheres, One Oxide Later Steps See dry seed layer, wet bulk fill, dry cap, grown in sequence on the same wafer dry cap, dense, thin wet bulk, fast-grown, carries most of the thickness dry seed, dense, thin silicon, unaware which atmosphere grew the oxide above it step 4's etch rate and step 7's pinhole inspection both assume one oxide, not three recipes stacked

## 3. Why This Tradeoff Was Always Sitting Underneath Step 1's Shortcut

Step 1 made its entire argument rest on "a continuous oxide," as if that oxide arrived for free once the decision not to mesa-etch was made. It did not: growing that oxide thick enough to mask a real diffusion schedule, without stalling the furnace line, required exactly the tradeoff this step names between wet oxidation's speed and dry oxidation's density. Every reliability gain this series has claimed since Step 7's coverage check and Step 8's failure-rate comparison assumes the oxide actually achieved was dense and thick enough to do its job — an assumption that was never automatic, and that Fairchild's furnace engineers had to solve by sequencing atmospheres, not by picking a single one and hoping.

Step 12 does not add a new material or a new mask; it names the scheduling tradeoff that Step 1's "simply don't cut it" idea was always quietly depending on someone having already solved.

Take planar 1959 step 12 oxide thickness growth rate tradeoff further

Ask the copilot about this term, or have our engineers assess it against your process.

Ask CFSGPT about this → Request a technical assessment → Open the CFS app