Chiggiato 1

C.A.S.
2006
Outgassing
Paolo Chiggiato
TS-MME
Coatings, Chemistry and Surfaces
CERN
CH-1211 Geneva 23
Wednesday, May 17, 2006, CAS, Platja d’Aro, Spain

05 Novembre 2003 1
Preamble
The total outgassing rate Q, together with the applied pumping speed S,
defines the pressure in the vacuum system:
Q
P = + P0
S
P0: ultimate pressure of the pumping system.
In general S varies in a range of three orders of magnitude (≈1 →1000 l.s-1)
while Q can extend over more than 10 order of magnitude (≈10-5 →10-15 Torr
l.s-1.cm-2).
The right choice of materials and treatments is compulsory in the design of
vacuum systems (especially those for accelerators).
In this respect the measurement of outgassing rates is a basic activity for an
ultra-high vacuum expert.
2
Part 1: Syllabus
Thermal outgassing Water outgassing Non-thermal outgassing

P.C. F. Dylla N. Hilleret
Thursday Thursday Friday
Syllabus
1. Definitions, units and methods
2. General features of outgassing for vacuum materials
3. Outgassing of polymers
3
Part 1: Definitions
Outgassing is the spontaneous evolution of gas from solid or liquid.
Degassing is the deliberate removal of gas from a solid or a liquid.
Desorption is the release of adsorbed chemical species from the

surface of a solid or liquid.
4
Part 1: Definitions
¾ The intrinsic outgassing rate is the quantity of gas leaving per unit
time per unit of exposed geometric surface, or per unit of mass, at a
specified time after the start of the evacuation.
¾ The geometric surface is the visible surface without correction for

roughness or open porosity.
¾ The measured outgassing rate is the difference between the intrinsic

outgassing rate and the rate of readsorption in the test chamber.
The readsorption rate depends on test chamber and on method of
test.
5
Part 1: Units of gas quantity
The quantity of gas can be presented in number of molecules (N) or in

pressure-volume (PV) units.
The two values are related by the ideal gas equation of state:
P ⋅V
P ⋅V = N ⋅ K B ⋅ T → N =
K B ⋅T
The pressure-volume units are transformed to number of molecules when

divided by KBT. A given number of molecules is expressed by different
pressure-volume values at different temperatures. In general the pressure-
volume quantities are quoted at room temperature
− 23 ⎡ N ⋅ m Pa ⋅ m3 ⎤
k B = 1.38 ⋅10 ⎢ = ⎥
⎣ K K ⎦
− 23 ⎡ Pa ⋅ m3 ⎤ − 22 ⎡ Torr ⋅ ℓ ⎤ − 22 ⎡ mbar ⋅ ℓ ⎤
k B = 1.38 ⋅10 ⎢ ⎥ = 1.04 ⋅10 ⎢ ⎥ = 1 . 38 ⋅ 10 ⎢⎣ K ⎥⎦
⎣ K ⎦ ⎣ K ⎦
for T = TRT = 296 K
1
k BTRT
= 2.45 ⋅10 20 Pa ⋅ m3[ ]−1
= 3.3 ⋅1019 [Torr ⋅ ℓ ] = 2.5 ⋅1019 [mbar ⋅ ℓ ]
−1 −1
6
Part 1: Units of outgassing rate
The outgassing rate is
presented in:
9. Pa ⋅ m 3 Pa ⋅ m Pa m Torr l mbar l molec mol

=
s ⋅ m2 s s s cm2 s cm2 s cm2 s cm2
9 . Torr ⋅ l Pa m
7.5x10-4 10-3 2.5x1016 4.1x10-8
s ⋅ cm 2 s
Torr l
9 . mbar ⋅ l s cm2
1330 1.33 3.3x1019 5.5x10-5
s ⋅ cm 2
mbar l
10-3 0.75 2.5x1019 4.1x10-5
9. molecules s cm2
s ⋅ cm 2 molec
4x10-17 3x10-20 4x10-20 1.7x10-24
9 . mol 2 s cm2
s ⋅ cm mol
2.4x107 1.8x104 2.4x104 6.02x1023
s cm2
7
Part 1: Units of outgassing rate
Neoprene (10 h pumping):

qH2O ≈ 10-5 Torr l s-1 cm-2 Some examples
qH2O=3.3x1014 molecules cm-2
Bayard-Alpert gauges (W filaments)
Unbaked stainless steel (10 h pumping):
Q≈10-9 Torr l s-1
qH2O =2x10-10 Torr l s-1 cm-2
Q≈3x1010 molecules cm-2
qH2O =6.6x109 molecules cm-2
Bayard-Alpert gauges (Thoria coated W
Baked stainless steel (150º C x 24 h): filaments)
qH2=2x10-12 Torr l s-1 cm-2 Q≈10-10 Torr l s-1

qH2=6.6x107 molecules s-1 cm-2 Q≈3x109 molecules cm-2
Baked OFS Copper (200º C x 24 h): Residual gas analyzer (W filaments)
qH2=2x10-14 Torr l s-1 cm-2 Q≈10-8 Torr l s-1
qH2=6.6x105 molecules s-1 cm-2 Q≈3x1011 molecules cm-2
8
Part 1: Methods for measurement
Pressure-rise (accumulation) method
Apparatus
Sample No external pumping during the measurement
dN dN V dP dP k BT
P, T, V = Q and = → = Q
dt dt k BT dt dt V
P qs
Q = Aa qa + As qs
qa
⎡ molecule ⎤
( Aa qa + As qs ) ⎢ ⎥⎦
⎣ s
ΔP = k BT ⋅t
V
P
⎡ Torr ⋅ l ⎤
( Aa qa + As qs ) ⎢
⎣ s ⎥⎦
ΔP[Torr ] = ⋅ t[ s ]
V [l ]
If repumping is negligible, the pressure in the system

increases linearly and the total outgassing rate is obtained
t from the slope of the curve.
The sensitivity of the method is limited by the outgassing
of the apparatus (walls, valves and gauges) and by the
sensitivity of the pressure gauge.
9
Example: for a baked spherical system made of regular stainless steel containing
the sample to be measured:
V = 10 ℓ
Torr ⋅ l
qa = 2 ⋅10 −12 [ ]
s ⋅ cm 2
Torr ⋅ l
System contribution 2245[cm 2 ] ⋅ 2 ⋅10 −12 [ ]
ΔPa [Torr ] = s ⋅ cm 2
⋅ t[ s ] = 4.5 ⋅10 −9 t
10[l ]
The signal of the sample has to be at least 25% of that of the apparatus to be
significant.
Torr ⋅ l
qs ⋅ As > 1.1 ⋅10 −9 [ ]
s
Optimistic
For a sample area of 100 cm2:
Torr ⋅ l
qs > 1.1 ⋅10 −11[ ]
s ⋅ cm 2
This excludes most of the metals applied in the UHV. This methods is used for metals
when: upper limits of outgassing rate are needed, the sample is the system itself, or
the sample surface is very large
10
Measuring Throughput method

apparatus
Continuous pumping is applied during the
⎡l ⎤ ⎡l ⎤ measurement
qs S eff ⎢ ⎥ = σ ⋅ Ca ⎢ ⎥
⎣s⎦ ⎣s⎦
P dN ⎛1 ⎞
= Q − σ ⋅ AP ⋅ ⎜ n ⋅ v ⎟ and
dN
=
V dP
→
dP k B T
=
⎛1 ⎞
Q − σ ⋅ AP ⋅ ⎜ v ⎟ P
dt ⎝4 ⎠ dt k B T dt dt V ⎝4 ⎠
qa Q = Aa q a + As q s
Ca is the conductance
of the pump opening dP[Torr ] ⎡ Torr ⋅ l ⎤ ⎡l ⎤
in the system V = Q⎢ ⎥ − σ ⋅ C a ⎢ ⎥ P[Torr ]
dt ⎣ s ⎦ ⎣s⎦
σ is the pump capture
S probability in the
system Q ⎛ −
t
⎞ V
ΔP = ⎜1 − e pτ ⎟ τp =
S eff ⎜ ⎟ S eff
⎝ ⎠
P
In general τp is very small compared to the time scale of
τp the measurement, so:
⎡ Torr ⋅ l ⎤
Q⎢
⎣ s ⎥⎦ Aa qa + Aa qa
ΔP[Torr ] = =
⎡l ⎤ S eff
S eff ⎢ ⎥
t ⎣s⎦
11
Throughput method
τp is the characteristic time of pumping.
Typical values for τp for test systems are less than one minute.
Example: V=10 l S=10 l/sÆ τ =1 s
This implies that the transient of P is very short

Cand can be neglected Æ the quantity
of gas to built up the gas phase in the volume V is very smallsample
in the UHV range.
S Q
Here again, the sensitivity of the method is limited by the outgassing of the apparatus
P2 P1
(walls, valves and gauges) and by the sensitivity of the pressure reading.
Example 1: for a baked spherical apparatus made of regular stainless steel (V=10l)
evacuated by an effective pumping speed of 10 l/s.
The contribution of the system is ΔP=4.5x10-10 Torr. The detection limit of the
outgassing rate is 1.1x10-9 Torr l s-1 (25% of the system background). If the
outgassing of a standard BA gauge is also taken into account the detection limit is
1.4x10-9 Torr l s-1.
12
Throughput method
Example 2: Extreme conditions.
Same apparatus as before, but made of high temperature treated stainless steel:
q≈10-14 Torr l cm-2 s-1 Æ Qa=2.2x10-11 Torr l s-1
and equipped with a very low outgassing rate BA gauge
QBA= 1x10-10 Torr l s-1
The system background is Q=1.2x10-10 Torr l s-1 , dominated by the effect of the gauge.
The detection limit (25% of the system background) for the outgassing rate is 3x10-11
Torr l s-1. This limit can be attained by increasing the surface area of the sample
(whenever possible) or by coupling the accumulation and the throughput methods.
13
Measuring Accumulation Coupled method

system system The sample is isolated by a valve in a separated
vessel where the gas is accumulated for a time ta.
qa Then the valve is opened and the quantity of
P accumulated gas is measured in the test dome.
qs t a + Δt
Q = S eff ⋅ ∫ P(t )dt

ta
This method can be applied only if the outgassing

S rate is constant in the time of accumulation and
repumping is negligible.
P The detection limit is defined by the outgassing of

the accumulation system and by the sensitivity of
the pressure reading.
The advantages of this method are two:
ta •no effects of gauge on the quantity of gas

accumulated
t •ta can be long enough to attain gauge
valve open sensitivity
14
Effect of readsorption
P The gas released by the sample, instead of being evacuated by

the pumping system, can be adsorbed by the system wall,
pumped by the gauge, or eventually reabsorbed by the sample
itself.
Let’s call Sadd the resulting additional pumping speed:
S
⎛ S add ⎞ Qint
Qint = P (S add + S eff ) = PS eff ⎜1 +
Ion gauges remove
⎟⇒ S
= 1 + add
molecules from the gas ⎜ S eff ⎟ Qm S eff
phase through two ⎝ ⎠
mechanisms:
-Cracking on the hot

filaments (does not affect Example: for a 10 l spherical chamber, supposing a sticking
monoatomic molecules like probability of 10-3 for water on the system wall and a sample
rare gases). surface area of 100 cm2:
-Ionization by electrons
Qint
(affects all gases).
≅3
For ordinary B-A gauges Qm
typical values are ≈1 l/s
and ≈10-2 l/s respectively.
15
Weight loss
This method is used for high outgassing rate materials
Sample holder only, namely organics.
25 oC The test consists in measuring the weight loss of a
sample following a defined thermal cycle in vacuum
(TML total mass loss) and the weight gain of a
collector cooled at room temperature and placed in
front of the sample (CVCM Collected Volatile
Condensable Material) .
Collector The set-up and procedure are described in the ASTM
E595-93 standard.
If the throughput method is applied in the same
Sample 125 oC apparatus, data are obtained about the evolution of
the outgassing rate at room temperature and at the
temperature of the treatment.
This method is used largely in space applications and a wide database is available through
the NASA’s web site. http://outgassing.nasa.gov/
16
Key points
1. The quantity of gas are presented in number of molecules or in pressure-

volume units. KBT converts the former to the latter.
2. Pressure-volume values are always referred to the temperature of

measurement.
3. Outgassing rates are measured by the accumulation or by the throughput

method. For both methods gas repumping has to be taken into account.
4. Pressure gauges interfere with the outgassing measurements by pumping

and releasing gas molecules.
17
Part 2: General features of outgassing
The surface and the bulk of materials are both source of gas
molecules:
• Those from the surface have to overcome an energy barrier before

being released.
• Those from the bulk in addition have to diffuse along the material
lattice before encountering the surface.
18
Single desorption energy: mean stay time
Ed [Kcal/mole] Cases τd [s]

0.1 Helium 1.2x10-13
1.5 H2 physisorption 1.3x10-13
Ed
3-4 Ar, CO, N2, CO 1x10-11
physisorption
10-15 Weak chemisorption 3x10-6
The mean stay time
(sojourn time) is 20 H2 chemisorption 100
given by the
Frenckel’s law: 25 6x105
one week
Ed
30 CO/Ni chemisorption 4x109
τ d = τ oe k BT 100 years
40 1x1017
where the value of τo is age of Earth
usually assumed to be 150 O/W chemisorption > age of universe
about 10-13 s
(≈h/KBT).
19
Single desorption energy: pressure evolution without repumping
dP NΘ quantity of
V = − SP + s gas leaving
BA P
variation of
the quantity of dt τd the surface
gas in the gas

phase dΘ Θ
=− quantity of
dt τd gas removed
by the pump
S
t
Ns −
τd
P(t ) ≅ e for t > τ d
S ⋅τ d
The solution is plotted for:
Θ= fraction of sites occupied V=10 l, S=10 l /s, Ns=2245x3x10-5 Torr l

The total number of sites is assumed to
be ≈1015 cm-2Æ 3x10-5 Torr l s-1cm-2 and different energies
20
T=296K
P[Torr]
P(0)=0 T=373K
P[Torr]
0.75 eV
0.001 0.8 eV 0.001
0.85 eV
0.0001 0.0001 1 eV
0.9 eV
0.00001 0.95 eV
0.00001
1. ´ 10- 6 1 eV 1. ´ 10- 6
1. ´ 10- 7 1. ´ 10- 7
10 100 1000 10000 10 100 1000 10000

Pumping Time [s] Pumping Time [s]
21
T=296K
Po=1 Torr
P[Torr]
1
0.75 eV
0.01
0.8 eV
0.85 eV
Effect of the 0.0001
molecules 0.9 eV
already in the
gas phase at 0.95 eV
t=0
1 eV
1. ´ 10- 6
10 100 1000 10000

Pumping Time [s]
22
Single desorption energy: pressure evolution with repumping
dP NΘ
V = − SP + s − SW P
dt τd
quantity of
dΘ Θ gas
= − + SW P reabsorbed by
dt τd the walls
The general solution exists and for the following hypothesis:

τp SW
< 1 and >> 1
τd S
the pressure can be approximated to:
t
−
⎛ S +S ⎞ It is again an exponential
Ns τ d ⋅⎜ W ⎟
P(t ) ≅ e ⎝ S ⎠ function
τ d ⋅ (S + SW )
23
The same result can be obtained by simple arguments:

S
• the probability that a molecule leaving the surface is pumped outside
the system is: S + SW
1
•the probability for a molecule to leave the surface in one second is:
τd
•therefore, the probability for a molecule to be definitively removed in S 1
one second is:
S + SW τ d
S + SW
• the average time the molecule stays in the system is: τ d
S
The increased residence time of the molecule in the system is due to the
surface readsorption and reemission.
Ns
For t << τd Æ P ( 0) ≅
τ d ⋅ (S + SW )
t
−
⎛ S +S ⎞
Ns τ d ⋅⎜ W ⎟
P (t ) ≅ e ⎝ S ⎠
τ d ⋅ (S + SW )
24
Which is the τd that, at time “t”, gives the maximum value of P(t)?
t ⎛ ⎞
−
⎛ SW + S ⎞ ⎜ ⎟
dP(t ) Ns τ d ⋅⎜ ⎟ t
≅ 2 e ⎝ S ⎠
⋅ ⎜1 − ⎟=0
dτ d τ d ⋅ (S + SW ) ⎜ ⎛ SW ⎞⎟
⎜ τ ⋅
d ⎜ + 1⎟⎟
⎝ ⎝ S ⎠⎠
t
τ d ,max =
⎛ SW ⎞
⎜ + 1⎟
⎝ S ⎠
Hence, the maximum pressure that can be obtained in a desorption process (at
constant temperature) is: N s −1
PMAX (t ) ≅ Edwards’
⋅t upper
S ⋅e limit
In the case of one monolayer coverage Ns=1015 /3.3x1019:
N s −1 1 .1 ⋅ 10 −5 3 ⋅ 10 −9 Torr l
q MAX (t ) ≅ ⋅t = = [ ]
e t[ s ] t[ h ] s cm 2
This behavior could be that of a real system only if adsorption sites of different energy
exist and they are in the worst possible combination at each pumping time.
25
D. Edwards Jr. Journal of Vacuum Science and Tech.,14(1977)606 and 14(1977)1030
q(10h)=2.8x10-10 Torr l s-1 cm-2
Experimental value valid

for most of the metals:
2 ⋅10 −9 Torr l
q (t ) ≅
t[ h ] s cm 2
A=2067 cm2 The outgassing rate of an unbaked

S=19.6 l/s material depends on pumping
St. steel
time, it is not an intrinsic value!
26
D. Edwards Jr. Journal of Vacuum Science and Tech.,14(1977)606 and 14(1977)1030
The Edwards’ upper limit is very close to the

experimental values. In the context of the
analysis reported in page 25 it could mean
that:
1. water should be adsorbed on many

energy states
2. the distribution of the water molecules in

the different sites is, at any time, the
worst.
3. this is true for all the metals!
Other models explain the 1/t variation; they

are based on diffusion from the bulk of ad
hoc initial distribution, adsorption in small
porosity in the oxide layer, and quasi-
equilibrium approximations allowing the
application of standard isotherms.
27
The simple models presented suppose the existence of a single desorption site or the
coexistence of many but completely independent desorption sites.
In addition, when readsorption is considered, the pumping speed of the wall is
assumed constant.
Much more complicate models are needed to take into account variable pumping
speed of the wall and the interdependence of the population of the different
adsorption sites.
However, the essential feature of desorption are already underlined by the simplest
models:
1. The outgassing rate of a single desorption site decreases as an exponential

function of time, whose decay time is the desorption time.
2. Higher temperatures allow a faster degassing.
3. The decay time increases in case of repumping by the system walls.
4. The outgassing rate has a t-1 evolution only if desorption from sites of different
energy is considered.
28
Outgassing of molecules dissolved in the bulk of materials: solubility

Exothermic absorption
Molecules of A in gas phase and in solid are in
thermodynamic equilibrium when:
A (gas) = A (solid)
ΔG 0
Endothermic absorption
− aA
The equilibrium constant K is: K =e k BT
=
PA
ΔΗs
where aA is the activity of A in the solid. For diluted
solution : x
K= A
→ x A = K ⋅ PA
PA
where xA is the molar fraction and K the solubility of A in the solid. The temperature
dependence is obtained: ΔH s 0
−
xA = B ⋅ e k BT
⋅ PA
x2 A
For diatomic molecules adsorption: A2 (gas)= 2 A (solid): K=
PA
ΔH s 0
−
xA = B ⋅ e 2 k BT
⋅ PA Sievert’s law
29
P [Torr]
10000
Solubility of H2 in stainless steel:
1 wt. ppm
1000 0.114 [eV ]

−
k B ⋅T
xH [at.ppm] = 71.8 ⋅ P[Torr ] ⋅ e
100
The H content in standard austenitic
stainless steels is about 1 ppm in weight
(≈56 ppm atomic H).
10 After evacuation, in an isolated stainless
steel vessel the hydrogen equilibrium
pressure is about 7 bar at RT !
But no humans will measure it. Why?
20 50 100 200 500 1000
T [°C]
30
Outgassing of molecules dissolved in the bulk of metals
Gas molecules are dissolved into the bulk of materials

during the production processing and during their
permanence in air.
In vacuum, the lighter molecules diffuse and, after
reaching the surface, they are released.
•Only H atoms have enough mobility in metals to attain the surface where they
recombine to form H2.
On the other hand, in organics most of the lighter molecules quickly diffuse
toward the surface where they are released.
•The models that take into account all the steps in the outgassing process are
quite complicate and, in general, they give only asymptotic solution for limit
conditions.
Two mechanisms are considered:
∂c Concentration gradient
1. diffusion limited outgassingÆ q (t ) ∝ −
∂x
2 Square of the
2. recombination limited outgassingÆ q(t ) ∝ cs concentration on the
surface
31
B.M. Shipilevsky and V.G. Glebovsky have shown that, for each metal, a characteristic
critical number of dissolved monolayers defines the limit between the two stages; below
it, recombination becomes the controlling mechanism.
exothermic metals endothermic metals
Fe, Ni,
Nb,Ti,
Cu, Mo,
Pd,Zr,
st. steel
…
The recombination and desorption energy is The recombination and desorption energy is
higher than the dissolution energy. lower than the dissolution energy.
Recombination has to be taken into account The desorption barrier is transparent.
in most of the problems. Kinetics is controlled by diffusion, except
when concentration is very low (much less
than 1 ML dissolved; do not forget that the
recombination rate is ≈ (cs)2).
32
Diffusion limited outgassing
Diffusion is a random process and in most of the cases

of interest is described by the Fick’s equations:
Eb
∂c( x, t )
−D = Γ ( x, t )
−
Eb ∂x where c(x,t) is the
D(T ) = D0 ⋅ e k BT
concentration in the solid and Γ
∂ 2 c( x, t ) ∂c( x, t ) is the flux of molecules per cm2
D =
∂x 2
∂t
In the limit of this model, the outgassing rate is equal to the flux of molecules arriving
at the surface by diffusion:
∂c( x, t )
q (t ) = − D
∂x x = SURF .
33
Semi-infinite solid approximation:

c(x,t)
co
∂c( x, t ) ∂c( x, t ) 2
z
= ∫ e dη
2
−η
D erf ( z ) =
∂x 2 ∂t x π
c( x, t ) = co ⋅ erf 0
I.C. c( x,0) = co 2 Dt
B.C. c(0, t ) = 0 the gaussian character of the
solution reflects the brownian
diffusion length
nature of diffusion
x
∂c( x, t ) D ⋅ co
q (t ) = − D = ∝ t −0.5
∂x 0 π ⋅ D ⋅t
The t-0.5 evolution holds also for solid of finite dimension L (slab) when L>>(Dt)0.5
The total amount of substance M(t) which has left the solid is:
2co
M (t ) = ⋅ D ⋅t
π
34
Slab approximation
c(x,t)
∂c( x, t ) ∂c( x, t )
D =
∂x 2 ∂t − D⋅( 2 n +1) 2 π 2t
I.C. c( x,0) = co c( x, t ) − co 4 ∞ (−1) n (2n + 1)π ⋅ x
= 1− ∑ cos e L2
cw − co π n = 0 2n − 1 L
L
B.C. c(± , t ) = cw
2
L x
4 ⋅ (c0 − cw ) ⋅ D ∞ −( 2 n +1) 2 π 2 ⋅ L2
D⋅t
q(t ) =
L
∑n =0
e
For Dt>0.05 L2 only the first term of the series is relevant : For diatomic molecules
4 ⋅ (c0 − cw ) ⋅ D −π 2 ⋅ L2
D⋅t
q (t ) ≈ e
L
When equilibrium is assumed between gas and surface, the Sievert’s law defines the
surface concentration cw:
cw = K P
35
Slab approximation: thermal history

If the slab is heated to a temperature TH for a time tH, the outgassing rate, when back at
room temperature is:
T
TH - tH
4 ⋅ (c0 − cw ) ⋅ D ⎡ D(TH ) ⋅ t H ⎤
q≈ exp ⎢− π 2 ⋅ ⎥⎦
TRT
L ⎣ L2
⎡ tH
⎤
t
4 ⋅ (c0 − cw ) ⋅ D
⎢ ∫ D (T ) ⋅ dt ⎥
For an arbitrary thermal cycle it can be show that: q≈ exp ⎢− π 2 ⋅ 0 ⎥
L ⎢ L 2 ⎥
tH ⎢ ⎥
⎢⎣ ⎥⎦
∫ D(T ) ⋅ dt
The dimensionless number: Fo =
0
2 is called the Fourier number
L
Fo records the thermal history of the material and determines how much of the initial
concentration is depleted.
When thermal treatment of Fo≈ 3 are applied, the solid is actually emptied or in
equilibrium with the surrounding gas phase.
36
Recombination limited outgassing

In a pure recombination limited process, the concentration in the bulk is flat and the
outgassing rate is given by:
q (t ) = K R (T ) ⋅ c(t ) 2
where KR is called recombination coefficient. Recombination is an activated process and
so: E R
−
K R (T ) = K R , 0 ⋅ e k BT
In this model, the hydrogen concentration in the solid can be obtained in the following
way:
dc(t ) dc K 1 1 K
L = − K R c 2 → 2 = − R dt → − = Rt
dt c L c(t ) c0 L
c0
c(t ) =
⎛K c ⎞
1+ ⎜ R 0 ⎟ ⋅t
⎝ L ⎠
37
Trapping sites are present in metals: they can block the hydrogen
migration to the surface, hence providing a sort of internal pumping.
pumping
The trapping effect can be taken into

account by introducing an effective diffusion
coefficient Deff
Assuming local equilibrium between the

traps and the interstitial sites:
cL
Deff = D
cL + cTrap (1 − θTrap )
For θ<<1: cL
Deff = D
cTOT
cL, cTrap: H atoms in the regular interstitial site, trapping sites
38
Carbides and nitride are deep trapping sites in austenitic stainless steels
39
Key points
1. Both surface and bulk of materials are source of gas.
2. The desorption of gas molecules from the surface defines an exponential
decay of the pressure. The decay time is related to the desorption energy
by the Frenckel’s law.
3. When repumping is considered, the molecules remain in the test chamber
for a longer time.
4. An upper limit to the outgassing rate of adsorbed molecules can be
obtained. It is very similar to that of water in metallic systems!
5. The outgassing of molecules dissolved in the bulk is described by
diffusion limited or recombination limited models. Endothermic metals
should be well depicted by the former model, except for very low
concentration
6. In the diffusion limited model, the Fourier number records the thermal
history of the samples, and the diffusion length gives an indication of the
penetration in the solid of the diffusion process.
40
Part 3: Outgassing of polymers
Outgassing rate of polymers is known to be much higher than that of metals. Two
reasons explain this phenomenon: a polymer contains much more gases than a metal,
and the gas mobility in polymer is orders of magnitude larger than in metals.
Example
PEEK
O
Water solubility:
[O O C ] 0.1 to 0.5 wt.% (4.4 to
-6
n
10 22x1019 molecules/cm3)
H O diffusivity in PEEK *
2
10 to 50 times larger than
the H total content in as
Diffusivity [cm s ]
-1
-7
10 produced austenitic stainless
2
steel
RT
-8 Water diffusivity at RT:
10
5 x 10-9 cm2 s-1
2000 times larger than that reported
-9
10 for H in austenitic stainless steel.
2.6 2.8 3.0 3.2 3.4 3.6 3.8
1000/T [K]
* After G.Mensitieri et al., J.Appl.Polym.Sci., 37, 381, (1989)
41
Transport of gas in polymers
In the rubbery state the gas molecules are dragged by the thermal movement of the
chain.
In the glassy state the gas molecules diffuse through the volume of the polymer and also
through the excess volume.
The diffusion process in amorphous or semicrystalline polymers is not always well

described by Fick’s law. The diffusion process can be distinguished by evaluating the
Deborah number
τ
De =
ν
where τ is the polymer-penetrating molecule relaxation time, and ν=L2/D is the
characteristic time for diffusion.
For De>>1 the time for diffusion is shorter than the time for relaxation, the diffusing
molecules does not record change in the polymer structure Æ Fick’s law is valid. (He)
For De<<1 the penetrating molecules swell the polymer and hence allow the adsorption
of additional moleculesÆ non fickian process.(H2O)
42
Outgassing of Elastomers
In elastomers the molecular chains are reticulated by vulcanization.

The process of vulcanization was developed by Goodyear and Hancock (1844) by heating
natural rubber with sulfur. Chemical bonds are formed by chains of sulfur atoms which
react with the unsaturated bonds of the primary macromolecules. Modern elastomers
utilize a wide variety of chemicals to produce the permanent network.
Fluorocarbon rubbers are prepared by cross linking a copolymer of tetrafluoro-ethylene

and hexafluoro-propylene. These rubbers are thermally stable and have very low
coefficient of friction.
Viton®
Viton is a fluorocarbon rubber typically employed for O-ring in vacuum technology.
Typical elements of the Viton composition are:
• Viton resin 100 parts by wt.

• Carbon black 25 //
• MgO 15 //
• curing agent 1.5 //
43
Outgassing of Elastomers
The MgO is added as an acid acceptor to remove small amount of HF which results
from the curing of the resin. The reaction is:
MgO + HF Æ MgF2 + H2O
Therefore, Viton is created with a built-in source of water.
Three simple warnings:
1. Viton should never be cleaned with solvents, because the solvent is dissolved in
the material and its outgassing could remain for long time.
2. Ozone can cause cracking of O-rings.
3. Pre-baking of O-ring gaskets at 200°C for some hours in air or vacuum is a

necessary operation when possible contaminants need to be removed.
44
The outgassing rates for polymers reported in the literature have a large spread. This
could be due to:
1. The large spreading in the composition and the source of the resin
2. Different history of the samples
3. The relative humidity of the laboratory
In case of seals, the benefit of the baking is hindered by atmospheric gas permeation.
3x10-10 for 100°Cx16 h
cm-2
R. N. Peacock, J. Vac. Sci. Technol., 17(1), p.330, 1980
45
R. N. Peacock, J. Vac. Sci.

Technol., 17(1), p.330,
1980
10-6
unbaked Viton
baked Viton
10-10
46
Permeability of elastomers
The permeation flux of atmospheric water through a Viton O-ring, 5 mm cross section
diameter, 6 cm torus diameter is 10-7 Torr l s-1. The stationary condition (ultimate
permeation) will be attained after about two months.
months
R. N. Peacock, J. Vac. Sci. Technol., 17(1), p.330, 1980
47
L. Laureson and N.T.M. Dennis, J. Vac. Sci. Technol. A3(3), p.1707, 1985
In the Anderson-Stuart model, the activation energy for molecular diffusion

in glassy materials is estimated to be equivalent to the strain energy
generated in the glassy network when an orifice of radius rd is distended
so as to accommodate a diffusing atom of radius r.
In the Frenkel view, the elastic energy of this dilatation is:
E ÷ ( r – rd )2
W. G. Perkins, J. Vac. Sci. Technol., 10(4), p. 543, 1973
48
Case study 1: Outgassing of 4 different glues SAMPLES

Araldite 103 with hardener HY 991
ASTM set-up
Araldite 106 with hardener HV 953 U
Dow Corning R4-3117 RTV
Penning
Nuvovern LW polyurethane
Permeation Outgassing
Pirani
Penning
Sample on the
Liquid nitrogen QMA flange
Ionization
trap
N2 H2 Ar CO Gauge
SYSTEM 1 SYSTEM 2
Figure 1: Experimental set-up 49

-4
10 Dow Corning
Dow Corning Araldite 106
Polyurethane
Araldite 103 10
-10
Outgassing [Torr l/s cm ]
Araldite 106
2
-5
10
Ion Current (A)

-11
10
-6
10
-12
10
-7
10
0.1 1 10 100
Time [h]
0 20 40 60 80 100
Mass (a.m.u)
Figure 2: Outgassing of the 4 glues

Figure 4.c: Comparison between Araldite 1
50
-8
10
-3
10 Polyurethane (new components)
Polyurethane (expired components)
Polyurethane (expired components)
Polyurethane (new components)
-9
10
-4
Outgassing [Torr l/s cm ]
10
2
Ion Current [A]

-10
10
-5
10
-11
10
-6
10 -12
10
-7 -13
10 10
0.01 0.1 1 10 100 0 20 40 60 80 100
Time [h] Mass (a.m.u)
Figure 5: Effect of the polyurethane components

aging on the outgassing Figure 6: Effect of the polyu
on the degassing
51
Case study 2: Outgassing of silicone rubber
The CERN weight loss system:

ASTM E595-93
52
10
H H H
-1712.9/T
y = 133.87 * e
-3675.3/T
y = 7272.2 * e
H C H H C H H C H
H
1
TML & CVCM [%]
Si O Si O Si C H
H H
C
H H
C C
H H H
0.1 H H H
TML
CVCM
0.01
0.0025 0.003 The feature of the spectra does not depend
1/T[K] on the pumping time, except for a general
reduction of the peak intensities. The main
TML and CVCM as a function of the temperature of
peak corresponds to 73 amu suggesting the
the heating treatment applied to the sample. The
presence of methyl-silane with a missing
heating duration is 24 hours. The experimental data
methyl group
are fitted by Arrhenius like exponential curves
(termination of silicone chain).
53
Application of polymers in UHV technology is hindered by the huge

outgassing and permeability.
This drawback has been partially overcome in other technological domains

(i.e. packaging) by coating polymers with metals or metal oxides
(more recently also with a:C-H).
How effective are coatings in reducing outgassing and permeation rate?
54
In materials made of several layers the total permeability depends on the permeability
and thickness of each layer through the equation:
LTOT L L L
= 1 + 2 + 3 + ...
Π TOT Π1 Π 2 Π 3
¾Since the permeability of metals is negligible for all gases, metallic coatings should
entirely block the polymer outgassing and permeation.
¾However, experimental results show that only a partial reduction of the flux is
attained.
¾This is attributed to defects on the coating (pinholes or scratches) that cause
discontinuity on the surface coverage.
¾Pinholes are produced during the deposition process and they are presumably due to
atmospheric dust particles.
55
Examples of pinholes on Al coating deposited on PEEK
2.00 μm 20.0 μm
56
Transmitted light optical microscopy has been used to detect uncoated

surfaces on metal films, showing that they mainly consist of pinholes of
different diameters.
25
Scanned area = 40 mm
2 Al/PEEK
Counts in the scanned area
20
15
10
5 Normalised uncoated surface

areas of the order of 10-4 are
measured
0
0 5 10 15 20 25 30
Diameter of the pinholes [μm] C. Bellachioma, PhD Thesis,
University of Perugia
57
However, experiments have shown that the permeability is reduced much less than
expected from the uncoated fraction
Permeability of uncoated PEEK 0

l s-1 cm-2 x cm Torr-1 ]
10
-11
10 Example:
Al/PEEK(0.125 mm
Normalised permeability
thick)
The uncoated fraction
10
-1 is 10-4, but the
[Torr
-12
permeability of He is
10
10-2 of that of the
Permeability
uncoated polymer.
-2
10
-13
10
0 1 2 3 4 5 6
C. Bellachioma, PhD Thesis,
Al film thickness [µm]
58
This apparent inconsistency can be justified considering that the pinhole

gas throughput is enhanced by lateral diffusion.
ρ =Normalised permeability Æ order of 10-2

Θ =Normalised uncoated area Æ order of 10-4
qH
For L/R0 >0.3 :
*
⎛ LL ⎞⎟
coating
ρ = Θ × ⎜⎜ 1 +1.18
1.18
R00 ⎟
R
polymer
R0
⎝ ⎠
L
Amplification factor
qC q⊥ qC
* After W.Prins and J.J.Hermans, J.Phys.Chem., 63 (1959) 716.
59
In the literature the barrier efficiency is called “barrier improvement factor” or BIF. It is
the inverse of normalized permeability.
Normalized permeability: Barrier improvement factor:
⎛ L⎞ 1 1
ρ = Θ × ⎜⎜1 + 1.18 ⎟⎟ BIF = =
R0 ⎠ ρ ⎛ L ⎞
⎝ ⎜
Θ × ⎜ 1 + 1 .18 ⎟⎟
⎝ R0 ⎠
The fact that ρ depends on L means that for L>10R the flux value does not depend
on the polymer thickness. In other terms, all the concentration gradient is localized
near the pin-hole.
It follows also that, for similar coating, the improvement due to the coating is less
significant for thicker polymer substrates.
60
The BIF for Al coated PEEK.

For thickness larger than 5 mm, the coating
is useless.
4
10
-4
for Θ = 10
3
R =5 μm
0
10
2
10
BIF 1
10
0
10
-3 -2 -1 0
10 10 10 10
PEEK wall thickness [cm]
C. Bellachioma, PhD Thesis, University of Perugia

61
Key points
1. Polymers have a much higher gas content and gas mobility than metals.
As a consequence their outgassing is order of magnitude higher.
2. Permeation of atmospheric gases is not anymore negligible as it is in
metals. Helium is the gas with the higher diffusivity and water is the
molecule with the higher solubility.
3. The gas transport mechanism in polymers depend on the structure. In
rubbery polymers gas molecules are drained by the macromolecule
chains. In glassy polymers the gas molecules jump from void to void, and
excess volume has an important role. The Deborah number defines the
limit of validity of the Fick’s laws in polymers.
4. Elastomers should be heat treated in air or in vacuum before any
application in high vacuum.
5. Metal coatings act as gas barrier, but their efficiency is strongly reduced
by pinholes and scratches.
6. The effect of metal coating is negligible for thick polymer substrates.
62
Part 3: Complementary techniques
∂ 2 c( x, t ) ∂c( x, t ) Permeation
D =
∂x 2
∂t If the whole permeation process is considered to be
I.C. c( x,0) = co dominated by diffusion through a uniform lattice, the
following equation can be applied:
B.C. c(0, t ) = cw c( L, t ) = 0
∞ n 2π 2 D⋅t
q −
= 1 + 2∑ (−1) n e
1 L2
c/cw q∞ n =1
0.8
Where q∞ is the steady-state flux through the
membrane.
0.6 This function, its integral and its derivative give a
number of characteristic points and limiting values from
which the diffusion coefficient can be evaluated.
0.4
The steady-state flux is:
cw D K ⋅ Pa ⋅ D Pa ⋅ Π q ⋅L
0.2 q∞ = = = where Π = K ⋅ D = ∞ a
L L L P
is the permeability. a=0.5 for dissociating diatomic

0.2 0.4 0.6 0.8 1
molecules. a=1 for non dissociating molecules.
x/L
64
Part 8: Work in progress
Work in progress
• Calculation of pressure profiles with localised and distribited

outgassing and pumping
• Discharge cleaning (Calder, Mathewson, Dylla)
• Ougassing of graphite, ceramics.
• Glasses: outgassing and permeation
• Special polymers outgassing
65
Fig. 1
SCHEMATIC VIEW OF THE PERMEATION SYSTEM
Penning Capacitance
BA gauge RGA gauge gauge
Permeation cell
Low pressure side High pressure side
Penning
gauge
H2
TMP station TMP station
260 l/s 70 l/s
Exhaust
66
Typical permeation curve: H2 through stainless steel
0.8
0.6
inf
q/q
0.4
0.2
0
0 4 4 4 4 4
0 10 1 10 2 10 3 10 4 10 5 10
Time [s]
67
Evaluation of the diffusion coefficient: time lag
4
3 10 The following relation gives the asymptotic
behavior of the quantity of gas permeated
4
2.5 10
t
q (t ) ⎛ l2 ⎞
t
q(t ) 2 10
4
Limt →∞ ∫ dt = Limt →∞ ⎜⎜ t − ⎟⎟
∫ dt q 6 D
Integral[(q/q )dt
q∞ 0 ∞ ⎝ ⎠
inf
0
4
1.5 10
This is a straight line that intersects the
4
time axis at:
1 10
3 time lag L2
5 10 τL =
6D
0
0 10
0 10
0
1 10
4
2 10
4
3 10
4
4 10
4
5 10
4 This quantity is known as the time lag,
and gives a simple method to calculate the
Time [s]
diffusion coefficient. It is quoted in most of
the literature on permeation
measurements.
68
⎛ J⎞ ⎛ l2 ⎞
Limt →∞ ⎜ ⎟ = Limt →∞ ⎜ t − ⎟
⎝ J∞ ⎠ ⎝ 6D ⎠
Evaluation of the diffusion coefficient: inflexion and breakthrough points
1
The inflexion time is the time at which
the derivative of q/q∞ is a maximum.
0.8
inflexion point L2
τ i = Ln(16)
3π 2 D
0.6
inf
Similarly the breakthrough time is

q/q
0.4
defined as the time at which the
tangent to the normalized permeation
curve at the inflexion point intersects
the time axis
0.2
breakthrough point
L2
τ b = 0 .5 2
0 π D
0 4 4 4 4 4
0 10 1 10 2 10 3 10 4 10 5 10
Time [s]
69
Other methods are available for the evaluation of the diffusion coefficient D from the
permeation curve.
The D values obtained with the different methods should be equal. If it is not the
case, it means that the permeation process does not follow exactly Fick’s law, and so
diffusion model is not enough to characterize the gas transport process.
Surface recombination or internal trapping could also play an important role in metals.
Strong deviation from the Fick’s law has been observed in amorphous vitreous
polymers. For example linear time dependence of the flux, instead of square root, was
measured; in other cases oscillations in the permeating flux were recorded.
70
TDS
TDS is a very useful technique to provide an insight on both the thermodynamics and the
kinetics of gas-metal systems.
It consists in measuring the desorbing gas spectrum obtained while heating a sample at a
constant rate of temperature rise.
In particular, the diffusion coefficient may be quantified by varying the heating rate of
the sample. In the assumption of an initial uniform concentration, the diffusion model
gives: t −
Eb
D(t )dt D0 ∫0 e
k BT ( t )
4 ⋅ c0 ⋅ D(t ) ∞ −( 2 n +1) 2 π 2 ⋅Fo (t ) dt
∑
t
q(t ) = e where F0 (t ) = ∫ =
L n =0
0 L2 L2
The degassing rate has a maximum when (considering the first term of the series only):
Eb
dq(t ) b k B ⋅ D0 ⋅ π 2 − k BTMAX
=0→ 2 = e "b" is the heating rate
dt TMAX Eb ⋅ L2
⎡ b ⎤ ⎛ 1 ⎞
In the plot Ln ⎢ 2 ⎥= f ⎜⎜ ⎟⎟ the slope of the linear function is the diffusion energy.
⎣ TMAX ⎦ ⎝ k B ⋅ TMAX ⎠
71
Test Chamber
TDS system Test Chamber
Feedthrough
RGA
Gate valve
BA Gauge
Conductance
Water cooled Variable leak valve
double wall Sample Penning & Pirani
TMP Gauges
Sample: 60 cm long, 1 cm wide, 1 mm thick

Thermocouple: 0.1 mm diameter S type
72
TDS peaks calculated for H2

in stainless steel Ln(b/T
MAX
)
1
-1.4 10
10 K/min
l s-1 cm-2]
-6
1 10 E =0.5 eV
1 b
-1.45 10
-7
8 10
H outgassing [Torr
5 K/min
-7 1
6 10 -1.5 10
-7
4 10
1
-1.55 10
-7
2 10 1 K/min
2
1
0 -1.6 10
0 200 400 600 800 1000 14 15 16 17 18 19
Temperature [C] 1/KT
MAX
73
TDS spectra could indicates the presence of multiple interactions between the gas and
the solid; some of them are surface blocking and bulk trapping
5 K/min
J-P. Bacher et al., J. Vac. Sci. Technol. A 21, 167 (2003)
74
Key points
1. Permeation and thermal desorption experiments can clarify the
mechanism of gas transport in solids. Very important intrinsic properties
of materials can be obtained, namely solubility, permeability and
diffusivity.
2. Permeation and thermal desorption measurements can identify deviation

from the classical Fick’s laws and underline the role of trapping centers
and surface oxide.
75
C. Bellachioma, PhD Thesis,

76
Polymers
The term polymer was introduced by Berzelius in 1830. Its etymology is from Greek
πολι−many μεροσ −parts. Actually polymers are made of basic unit (monomer) repeated
to form chain. The degree of polymerization is the number of monomers in the
polymeric chain.
DP DP DP=
=1 =1 2
n- m- (n+m)-
mer mer mer
DP= DP=(N+
DP= M M)
N
77
Polymer classification
Line Linear chain According to topology:
are -Linear
-Branched
-Cross-linked
Catene
branched
polimeriche
Ramif lineari
icato According to morphology:
-Thermoplastics: soften when heated and return to
original condition when cooled
-Thermosets: solidify or "sets" irreversibly when heated
Lunghe catene con “braccia” che -Elastomers: can be stretched to several times their
partonocross-linked
dai punti di ramificazione
Reti length and rapidly return to their original dimension
colat when the applied stress is released.
o
According to polymerization reactions:

-Chain reaction (polyaddition)
-Step reaction (polycondensation)
Catene legate insieme con
crosslinking
78
The polymer structure can be:
Crystalline: macromolecules form three-dimensionally

ordered arrays:lamellar (plate-like) crystals
with a thickness of 10 to 20 nm in which the
parallel chains are perpendicular to the face
of the crystals.
Amorphous:no long-range order (spaghetti like material)

amorphous polymers are softer, have lower
melting points, and are penetrated by
solvents more than are their crystalline
counterparts.
Semicrystalline: mixing of crystalline and amorphous

regions; crystals are small and connected to
the amorphous regions by polymer chains so
there may be no sharp well-defined
boundaries between the two types of regions
Only amorphous and semicrystalline polymers will be considered.
79
Tg and excess volume
-Semi-crystalline solids have both amorphous and crystalline regions. According to the
temperature, the amorphous regions can be either in the glassy or rubbery state. The
temperature at which the transition, in the amorphous regions, between the glassy and
rubbery state occurs is called the glass transition temperature.
-Below Tg, amorphous polymers are stiff, hard and often brittle; in
this state the molecules are frozen on place. This generates
stable empty space called excess volume.
-Above Tg, portions of molecules can start to wiggle

around: the polymer is in the rubbery state, which lends softness
and flexibility .
80
Transport of gas in polymers
In the rubbery state the gas molecules are dragged by the thermal movement of the
chain.
In the glassy state the gas molecules diffuse through the volume of the polymer and also
through the excess volume.
The diffusion process in amorphous or semicrystalline polymers is not always well

described by Fick’s law. The diffusion process can be distinguished by evaluating the
Deborah number
τ
De =
ν
where τ is the polymer-penetrating molecule relaxation time, and ν=L2/D is the
characteristic time for diffusion.
For De>>1 the time for diffusion is shorter than the time for relaxation, the diffusing
molecules does not record change in the polymer structure Æ Fick’s law is valid.
For De<<1 the penetrating molecules swell the polymer and hence allow the adsorption
of additional moleculesÆ non fickian process.
81

Chiggiato 1

Uploaded by

Copyright:

Available Formats

Chiggiato 1

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Chiggiato 1

Uploaded by

Copyright:

Available Formats

C.A.S.

Wednesday, May 17, 2006, CAS, Platja d’Aro, Spain

Thermal outgassing Water outgassing Non-thermal outgassing

1. Definitions, units and methods

2. General features of outgassing for vacuum materials

Outgassing is the spontaneous evolution of gas from solid or liquid.

Degassing is the deliberate removal of gas from a solid or a liquid.

Desorption is the release of adsorbed chemical species from the

¾ The geometric surface is the visible surface without correction for

¾ The measured outgassing rate is the difference between the intrinsic

The quantity of gas can be presented in number of molecules (N) or in

The pressure-volume units are transformed to number of molecules when

9. Pa ⋅ m 3 Pa ⋅ m Pa m Torr l mbar l molec mol

Neoprene (10 h pumping):

qH2=2x10-12 Torr l s-1 cm-2 Q≈10-10 Torr l s-1

qH2=2x10-14 Torr l s-1 cm-2 Q≈10-8 Torr l s-1

qH2=6.6x105 molecules s-1 cm-2 Q≈3x1011 molecules cm-2

If repumping is negligible, the pressure in the system

Measuring Throughput method

Example: V=10 l S=10 l/sÆ τ =1 s

This implies that the transient of P is very short

Example 2: Extreme conditions.

q≈10-14 Torr l cm-2 s-1 Æ Qa=2.2x10-11 Torr l s-1

and equipped with a very low outgassing rate BA gauge

QBA= 1x10-10 Torr l s-1

Measuring Accumulation Coupled method

Q = S eff ⋅ ∫ P(t )dt

This method can be applied only if the outgassing

P The detection limit is defined by the outgassing of

ta •no effects of gauge on the quantity of gas

P The gas released by the sample, instead of being evacuated by

-Cracking on the hot

1. The quantity of gas are presented in number of molecules or in pressure-

2. Pressure-volume values are always referred to the temperature of

3. Outgassing rates are measured by the accumulation or by the throughput

4. Pressure gauges interfere with the outgassing measurements by pumping

• Those from the surface have to overcome an energy barrier before

Single desorption energy: mean stay time

Ed [Kcal/mole] Cases τd [s]

Single desorption energy: pressure evolution without repumping

gas in the gas

The solution is plotted for:

Θ= fraction of sites occupied V=10 l, S=10 l /s, Ns=2245x3x10-5 Torr l

0.001 0.8 eV 0.001

10 100 1000 10000 10 100 1000 10000

10 100 1000 10000

Single desorption energy: pressure evolution with repumping

The general solution exists and for the following hypothesis:

The same result can be obtained by simple arguments:

q(10h)=2.8x10-10 Torr l s-1 cm-2

Experimental value valid

A=2067 cm2 The outgassing rate of an unbaked

The Edwards’ upper limit is very close to the

1. water should be adsorbed on many

2. the distribution of the water molecules in

3. this is true for all the metals!

Other models explain the 1/t variation; they

1. The outgassing rate of a single desorption site decreases as an exponential

2. Higher temperatures allow a faster degassing.