Venting of Vacuum Systems

R Gordon Livesey
1st Vacuum Symposium UK
11/02/2010

Lynthorpe Consultancy
 Outline

Bringing a vacuum chamber back up to atmosphere

(or some intermediate pressure) appears to be a
simple process. However, in rapid venting, there is
more to consider than might be thought.
The talk will cover:

• Dust movement

• Temperature effects

• Relations for Vent time

 Particulates

• Dust particles are a serious problem in many vacuum

processes e.g. semiconductor, telescope mirror coaters
• Venting may introduce dust or cause resuspension - since
it is difficult to avoid turbulent gas eddies
• Historically, the guideline adopted has been to ensure that
Reynolds numbers are kept below 2300
• However, experimental evidence indicates that Reynolds
numbers ≤ 5000 or gas flow velocities ≤ ½ maximum do
not cause resuspension of particles in vent lines (and, by
implication, in the vacuum chamber)
 Resuspension of particles in vent line

600
From: H. P. Bader and M. Bader
For particles > 0.19 µ Vacuum, vol 41. pp 1846-1848, 1990

500

Vent via capillary: L=140mm, d=1.4mm

400
Particles/cu. ft

Laminar Fully turbulent

300

200

100
Choked
flow
0
0 2000 4000 6000 8000 10000 12000
Reynolds Number
 Resuspension of particles in vent line

10000
Vent geometry
Manifold
1000 Capillary
Orifice
Diff. chamber
Particles/cu. ft

For particles > 0.19 µ

100

10

From: H. P. Bader and M. Bader

Vacuum, vol 41. pp 1846-1848, 1990

0.1
0 0.2 0.4 0.6 0.8 1 1.2
Flow/Max flow
 Particles subject to molecular drag
 8 + π  2πM 2
Drag force on a spherical particle*: F =   r Pu
 3  R0T
M, T, P = molecular mass, temperature and pressure of gas through
which particle is travelling, r = radius and u = velocity of particle

2 2πR0T ρr  X 
Range: X max = U0 Velocity: u = U0 1 − 
8+ π M P  X max 
U0 = initial velocity and ρ = density of particle

• Particles injected into a vacuum chamber at high velocity

can travel large distances and may damage delicate
equipment
• In the next slide the air inlet flow is taken to be choked, so
its exit temperature will be 244K and velocity 313 m/s.
Chamber air is assumed to be at 293K
* Epstein P. S. Phys. Rev. 23, 710-733 (1924)
 Range of silica particles injected at 313 m/s

100

Particle diameters
1 micron
0.5 micron
10
0.2 micron
0.1 micron
Range: m

0.1

0.01
0.01 0.1 1
Pressure: mbar
Suggested vent arrangement

vent
valve
vent Re < 5000
gas

Vent flow

‘Showerhead’

baffle
 Thermodynamics of venting

P2 , T, n2

P0 , V0 , T0 , n0

P1, T0 , n1

n0 moles of gas enter the chamber, and the pressure and temperature rise
from P1 , T0 to P2 , T

From the first law: ∆Q = ∆U + ∆W = 0 for an adiabatic process

∆U = change in internal energy

∆W = Work done on the gas = - P0V0

Adiabatic temperature on venting

U final − Ustart = PV
0 0

n2CvT − (n1CvT0 + n0CvT0 ) = P0V0 = n0 R0T0

n2CvT − n1CvT0 = n0T0 (Cv + R0 ) = n0CpT0

( n2 − n1 )CpT0 = n2CvT − n1CvT0

 PV PV 
 2
− 1
 γT0 = PV
2 − PV
1
 T T0 

∴ γP2
T = T0 If P1 << P2 then T → γT0
P2 + ( γ − 1)P1
 Venting temperatures

• This implies an adiabatic bulk temperature of ~ 410K,

137oC (for source air at 293K) - a temperature rise of
almost 120K
• But this is only the bulk mean temperature of the gas
in the chamber, local temperatures in parts of the
incoming jet can significantly exceed this temperature
• The actual bulk mean temperature will be < 410K due
to heat transfer to the chamber walls and this
temperature will generally be greatest in larger
chambers
• Note that, once the vent valve is closed, the pressure
in the chamber will fall as the gas cools
 Venting of a 10 litre chamber (Air)
900 90

800 80

700 70

Temperature: Celsius
600 60
Pressure: mbar

500 Temperature 50

400 Pressure 40

300 30

200 20

Valve open
100 10

0 0
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5
Time: sec
 Vent time formulas - assumptions

• Derivation of vent time expressions make the

following assumptions:
– Source gas pressure and temperature are
constant
– Inflow is through a short duct or orifice
– Formulas for gas flow rate through vent tubulation
are given by thermodynamic analysis
– Entrance to vent tubulation is smooth (orifice
coefficient = 1)
– Isothermal case assumes gas accommodates to
chamber temperature after inrush
– Adiabatic case assumes no heat transfer in
chamber
Vent time - vacuum to source pressure

Isothermal: tvi = τ 2(γ + 1) V

where τ=
Adiabatic: tva = tvi / γ Ac

V = chamber volume
A = vent port cross-section area
c = velocity of sound (at source condition)
= 343 m/s for air at 20oC
Isothermal: = 2.19τ
for air at 20oC
Adiabatic: = 1.56τ
 Vent time - published results

4.9 litre 29.1 litre

τ = 0.45s P1+ = Pi /Ps ≡ r0 τ = 40.5s tchar ≡ τ

tv = 1.47τ Calculated times allowing

tv = 1.51τ
for initial finite pressure
Even for rapid filling (small tank < 1 sec) the results are closer to isothermal

Dutton J.C. and Coverdill R.E., ‘Experiments to Study the Gaseous Discharge and Filling of Vessels’
Int. J. Engng Ed. Vol. 13, No. 2, p. 123-134, 1997
 Inaccuracies in vent time formulas

• The time formulas assume isothermal or adiabatic

conditions in the volume
• Actual time will lie between tvi and tva
• Again note that, once the vent valve is closed, the
pressure in the chamber will fall as the gas cools
• The vent valve may have some length, which will
tend to reduce vent flow rates. However, (continuum)
flow rates through short ducts are very insensitive to
length
 Flow rate vs Tube Length
1

0.9

0.8

0.7 Choked

0.6 10% pressure difference

Q/Qmax

Molecular flow
0.5

0.4

0.3

0.2

0.1

0
1 10 100
Length/diameter
 The End

Thank you

Thanks to for some of the material

used in this presentation