Advanced Pump Down Time Predictor

Predict chamber pumpdown curves with pump speed curve, conductance, and gas load (leak, permeation, outgassing). Understand why base pressure won't drop or why pumpdown tails occur.

Calculator

Basic Conditions

Pump Model

S_pump is the pump curve (not spec); C limits effective speed at chamber.

Line/valve/throttle combined. Use large value (e.g. 1e6) for no limit.

Gas Load

Outgassing: Q_out = Q0_fast·e^(-t/τf) + Q0_slow·e^(-t/τs). Fast: adsorbed gas; Slow: pumpdown tail.

Results

Target reached: Not reached in simulation time
P(t_end): 760.0000 Torr

Pumpdown Curve P(t)

x: time (min), y: Pressure (Torr, log). Target line shown.

Effective Pumping Speed S_eff(P)

Throughput Balance

Pump term (S_eff·P) vs Q_total. Where they cross → tail starts.

Technical Explanation

Why This Tool?

Semiconductor equipment (PVD, Etch, ALD, etc.) chamber pumpdown curves are often assumed to follow simple exponential decay. In reality, pump speed varies with pressure, conductance limits effective speed, and gas loads (leak, permeation, outgassing) create tails and floor pressure. This predictor uses a physics-based ODE model to simulate pumpdown with pump curve, conductance, and time-varying gas load, helping engineers understand why base pressure won't drop or why tails appear.

Physics Model

The governing equation is:

V · dP/dt = −S_eff(P) · P + Q_total(t)

where V is chamber volume, P is pressure (Torr), S_eff is effective pumping speed (L/s), and Q_total is total gas load (Torr·L/s). Effective speed combines pump and conductance in series:

1/S_eff = 1/S_pump(P) + 1/C

S_pump(P) is interpolated from the pump curve in log(P) space. Q_total = Q_leak + Q_perm + Q_out(t), with outgassing modeled as a two-component exponential decay (fast + slow).

Gas Load Components

  • Leak (Q_leak) — Constant gas inleak. Dominates at base pressure; limits how low P can go.
  • Permeation / Virtual leak — Trapped gas or permeation through seals. Often modeled as constant.
  • Outgassing — Q_out(t) = Q0_fast·e^(-t/τf) + Q0_slow·e^(-t/τs). Fast: adsorbed gas (e.g. water). Slow: polymers, residual films, recontamination; creates pumpdown tail.

Throughput Balance

At steady state, S_eff·P = Q_total. The throughput balance chart plots S_eff·P (pumping term) vs Q_total (gas load). When these curves approach each other, pressure stops dropping quickly and the tail begins. The intersection indicates when gas load matches pumping capacity.

Numerical Method

The ODE has no closed-form solution for arbitrary S_pump(P) and Q_out(t). We use the RK4 (4th-order Runge–Kutta) method for numerical integration. Pressure is clamped to a minimum (e.g. 1e-12 Torr) to avoid negative values.

References & Disclaimer

Pump presets are illustrative. Use actual pump curves for accurate predictions. For critical applications, validate against real pumpdown data.

Engineering Guide

Why this calculation matters

Post-vent pumpdown time affects PM validation and production restart. Simple exponential models ignore pressure-dependent pump speed, conductance bottlenecks, and outgassing tails.

In semiconductor equipment work: PM qualification forecasts, comparing turbo vs cryo strategies, estimating time to base pressure, and explaining long low-pressure tails after chamber exposure.

Input parameters

Chamber volume V
Process volume in L or m³.
Initial / target pressure
Start and goal pressure with unit selection.
Pump curve
Preset turbo/cryo or custom CSV of pressure vs speed.
Conductance C
Total molecular-flow conductance of foreline (L/s).
Gas load Q
Leak, permeation, outgassing (constant or modeled).

Output interpretation

p(t) curve
Simulated pressure vs time.
Time to target
Estimated seconds/minutes to reach p_target.

If simulation reaches target quickly but real tool does not, increase Q or reduce S_eff assumptions. Long tails usually mean outgassing or virtual leaks, not insufficient catalog pump speed.

Assumptions

  • Well-mixed chamber volume; uniform pressure.
  • 1/S_eff = 1/S_pump + 1/C in series.
  • Pump speed interpolated in log(p) space from curve.
  • Gas load model as configured (constant or decaying outgassing).

Limitations

  • Not a CFD or multi-volume model; transfer lines simplified.
  • Transition/viscous roughing regime approximated by pump curve only.
  • Outgassing is order-of-magnitude — measure and refine with log analyzer.

Worked example

Post-PM — 120 L chamber to 1×10⁻⁶ Torr

  1. V = 120 L, p₀ = 760 Torr, target = 1e-6 Torr, Generic Turbo preset.
  2. C = 45 L/s, Q = 5e-6 Torr·L/s constant outgassing estimate.
  3. Note predicted knee and tail — compare with pumpdown log analyzer on actual run.