Case study

Vacuum Gauge Placement and Pumpdown Misdiagnosis Case Study

Vacuum chamber case study for pump-side gauge misdiagnosis, local pressure, conductance, outgassing, rate-of-rise testing, mean free path, validation, and release criteria.

Branch: Engineering Physics
Content: Case study
Updated: Jun 22, 2026
Revision: v1.0.0 · reviewed

This case study follows a realistic vacuum test failure. A chamber was accepted because the pump-side pressure gauge reached the specified pressure. The optical instrument inside the chamber still failed contamination and stability checks. The root cause was not a weak pump. The root cause was an incomplete measurement boundary: the accepted pressure was measured near the pump, not near the instrument region where outgassing, conductance limits, and thermal surfaces controlled the environment.

The case teaches a common engineering physics lesson: a vacuum number is useful only when it describes the region, gas load, species, and operating state that matter for the device.

Case Summary

Item	Engineering relevance
System	Small optical instrument in a thermal-vacuum test chamber.
Acceptance requirement	Local pressure near the optical assembly below $2.0\times 10^{-3}\ \text{Pa}$ for 30 minutes before exposure.
Reported passing evidence	Pump-side ion gauge showed $1.5\times 10^{-3}\ \text{Pa}$ .
Failure observed	Optical throughput drifted and witness samples showed film contamination.
Hidden weakness	No gauge or residual-gas evidence near the optical assembly.
Dominant mechanisms	Outgassing, conductance restriction, thermal gradients, and gauge placement.
Corrective action	Local gauge, rate-of-rise test, residual-gas trend, bakeout control, and revised release criteria.

The central engineering question was:

Did the pressure evidence describe the optical assembly environment, or only the pump inlet environment?

The answer was the latter. The original acceptance test measured the easiest place to pump, not the place where the instrument was sensitive.

Initial Test Configuration

The chamber contained:

an optical assembly with coated mirrors, detector window, wiring harness, and adhesive-bonded brackets;
a heater plate used to bring the assembly to operating temperature;
a turbomolecular pump connected through a short duct and isolation valve;
a pump-side ion gauge mounted near the pump inlet;
a chamber wall thermocouple;
no local gauge near the optical assembly;
no residual-gas analyzer during the original acceptance run.

The acceptance procedure said “chamber pressure below $2.0\times 10^{-3}\ \text{Pa}$ .” It did not specify the measurement location, gas species, thermal state, or whether pressure stability had to be verified after heater operation.

That ambiguity mattered. During early pumpdown, pressure differences were modest. During warm operation, surfaces inside the optical assembly released water vapor and hydrocarbons. Conductance between the optical volume and pump-side gauge was limited by the instrument enclosure, baffles, cable routing, and apertures.

Event Timeline

The empty chamber reached its target pressure during commissioning.
The assembled optical instrument was installed after ordinary clean handling.
The pump-side gauge reached $1.5\times 10^{-3}\ \text{Pa}$ after overnight pumpdown.
The optical heater was enabled.
Optical throughput drifted during exposure.
A witness sample near the optical path showed visible film after the run.
The team initially suspected the pump, then the optical coating.
A temporary gauge mounted near the optical assembly showed local pressure several times higher than the pump-side reading.
Rate-of-rise and residual-gas evidence indicated outgassing and a virtual-leak-like trapped volume, not a large external air leak.

Evidence After the Failure

A repeat test used two gauges:

Measurement point	Pressure after warm stabilization
Pump-side gauge	$1.5\times 10^{-3}\ \text{Pa}$
Local optical assembly gauge	$6.0\times 10^{-3}\ \text{Pa}$

The local pressure was:

\displaystyle \frac{6.0\times 10^{-3}}{1.5\times 10^{-3}}=4.0

So the optical assembly region was operating at about four times the pressure implied by the accepted pump-side reading.

The local gauge also showed a slow decay rather than a stable plateau. When the chamber was isolated from the pump for a rate-of-rise test, pressure near the optical assembly increased from:

p_1=1.8\times 10^{-3}\ \text{Pa}

to:

p_2=7.8\times 10^{-3}\ \text{Pa}

over:

\Delta t=20\ \text{min}=1200\ \text{s}

The estimated gas load for a chamber volume:

V=0.12\ \text{m}^3

is:

\displaystyle Q=V\frac{\Delta p}{\Delta t}

Pressure rise:

\Delta p=7.8\times 10^{-3}-1.8\times 10^{-3}=6.0\times 10^{-3}\ \text{Pa}

Gas load:

\displaystyle Q=0.12\frac{6.0\times 10^{-3}}{1200}=6.0\times 10^{-7}\ \text{Pa}\,\text{m}^3/\text{s}

This gas load was too high for the local contamination requirement. It also did not behave like a large open leak. The pressure rise slowed after bakeout and dry nitrogen purge changes, which pointed toward water, trapped gas, and material outgassing.

Mean Free Path Check

At the accepted pump-side pressure, the gas is rarefied. For air at:

T=293\ \text{K}

with molecular diameter:

d=0.37\ \text{nm}

mean free path can be estimated by:

\displaystyle \lambda=\frac{k_BT}{\sqrt{2}\pi d^2p}

Using local optical-region pressure:

p=6.0\times 10^{-3}\ \text{Pa}

gives:

\displaystyle \lambda=\frac{(1.381\times 10^{-23})(293)}{\sqrt{2}\pi(0.37\times 10^{-9})^2(6.0\times 10^{-3})}

\lambda\approx 1.1\ \text{m}

For a local aperture of characteristic length:

L=30\ \text{mm}=0.030\ \text{m}

the Knudsen number is:

\displaystyle Kn=\frac{1.1}{0.030}\approx 37

This is a free-molecular regime. Gas movement, contamination transport, and gauge interpretation depend strongly on surfaces, geometry, and line of sight. A continuum pressure-drop intuition is not enough.

Failure Modes

The failure was not a single component fault. It was a chain of engineering assumptions:

the word “chamber pressure” did not define the measurement location;
the pump-side gauge was treated as representative of the instrument region;
no local pressure evidence was collected during warm operation;
the pumpdown curve was not compared with an installed-instrument baseline;
outgassing from adhesives, cables, or trapped volumes was not separated from pump performance;
no residual-gas trend was recorded before optical exposure;
contamination acceptance was treated as pressure-only rather than pressure, species, surface, time, and temperature;
the acceptance report preserved a passing number without enough context to interpret it.

These are system-level measurement failures. The pump and gauge both worked; the test definition did not.

Corrective Engineering Decision

The team changed the acceptance boundary from “pump-side pressure below threshold” to “local optical assembly environment below threshold under operating thermal state.”

The revised release criteria required:

local gauge pressure below $2.0\times 10^{-3}\ \text{Pa}$ near the optical assembly;
pump-side and local gauge readings recorded together;
rate-of-rise gas load below the process limit after isolation;
residual-gas trend showing water and hydrocarbon reduction after bakeout;
heater-on pressure stability for 30 minutes before optical exposure;
witness sample or optical throughput stability within acceptance limits;
controlled vent and dry purge procedure after maintenance;
preserved pumpdown curve for comparison after future service.

The original unit would not have passed this revised criterion. After cleaning, vent-procedure revision, harness bakeout, and local gauge installation, the local pressure stabilized below the requirement and the witness sample remained acceptable.

Validation Evidence

The corrected validation package included:

Claim	Evidence
The gauge represents the sensitive region	Local gauge mounted near the optical assembly and compared with pump-side gauge.
The chamber is not dominated by a large external leak	Isolation rate-of-rise data and helium leak test.
Outgassing is controlled	Pumpdown curves before and after bakeout plus residual-gas trend.
Thermal operation is represented	Heater-on pressure and temperature stability during acceptance.
Optical performance is protected	Witness sample and optical throughput trend during exposure.
Future maintenance can detect regression	Baseline pumpdown curve, gauge locations, vent procedure, and service record.

Transferable Lessons

Vacuum acceptance is not a single pressure number. A defensible vacuum test states:

where pressure is measured;
which gas species or contamination limit matters;
whether the system is cold, warm, empty, or fully assembled;
which conductance restrictions separate the gauge from the sensitive region;
whether outgassing, virtual leaks, or trapped volumes dominate the gas load;
how pressure uncertainty affects the pass/fail decision;
what evidence proves recovery after maintenance.

This case is transferable to thermal-vacuum testing, semiconductor tools, electron microscopes, coating chambers, x-ray systems, optical payloads, and laboratory instruments. In each case, the measurement boundary must match the engineering decision. A good gauge in the wrong location can produce a precise but misleading answer.

REF

Disciplines