UDK 533.5:669.14.018.8:669.788	ISSN 1580-2949
Original scientific article/Izvirni znanstveni članek	MTAEC9, 46(2)161(2012)
OUTGASSING OF HYDROGEN FROM A STAINLESS STEEL VACUUM CHAMBER
RAZPLINJEVANJE VODIKA IZ NERJAVNEGA JEKLA
Sefer Avdiaj1,3, Bojan Erjavec2
1University of Prishtina, Faculty of Natural Sciences and Mathematics, Mother Teresa Av. 3, Prishtina 10000, Kosova 2Institute of Metals and Technology, Lepi pot 11, 1000 Ljubljana, Slovenia 3Lotrič, d.o.o., Selca 163, 4227 Selca, Slovenia sefer.avdiaj@uni-pr.edu
Prejem rokopisa - received: 2012-02-07; sprejem za objavo - accepted for publication: 2012-02-22
Due to outgassing from the walls of vacuum calibration chambers, the generation of the calibration pressure in primary vacuum calibration systems operating below 10-6 Pa becomes un-accurate. Austenitic stainless steel (SS) is the most common construction material for ultrahigh vacuum (UHV) and extremely high vacuum (XHV) chambers. Hydrogen is the predominant residual gas at very low pressures in SS vacuum systems, i.e., in the UHV and XHV range. Therefore, the reduction of the hydrogen outgassing rate is the most challenging problem in achieving pressures in those ranges. In this paper, a vacuum chamber with a wall thickness of 2.62 mm and made from AISI type 304 L SS was examined with the aim of mitigating the outgassing rate. The heat treatments were carried out at 250 °C for 380 h and at 350 °C for 140 h. After baking at 250 °C for 380 h (corresponding to a dimensionless time Fo = 3.09), an outgassing rate q = 2.86 X 10-13 mbar L s-1 cm-2 was achieved at room temperature (RT). This RT outgassing rate was further reduced to q = 5.7 X 10-14 mbar L s-1 cm-2 after baking for another 140 h at 350 °C (Fo = 8.66, resulting in a total dimensionless time I.Fo = 11.75).
Keywords: hydrogen outgassing, stainless steel, dimensionless time, pressure-rise method, throughput method, gas-flow calibration coefficient
Zaradi razplinjevanja sten vakuumskih kalibracijskih posod postane vzpostavitev kalibracijskega tlaka v primarnih kalibracijskih vakuumskih sistemih, delujočih v tlačnem območju manjšem od 10-6 Pa, nenatančna. Avstenitno nerjavno jeklo je najprimernejši material za gradnjo ultra visokovakuumskih (UVV) in ekstremno visokovakuumskih (EVV) posod. Pri zelo nizkih tlakih v UVV- in EVV-področju je vodik prevladujoč rezidualni plin v vakuumskih sistemih, narejenih iz nerjavnega jekla. Tako je zmanjšanje razplinjevanja vodika najpomembnejše za doseganje nizkih tlakov v naštetih področjih. V tem članku je predmet raziskav vakuumska posoda z debelino stene 2,62 mm, ki je narejena iz nerjavnega jekla AISI type 304 L. Postopki pregrevanja vakuumske posode so se izvajali pri 250 °C 380 h in pri 350 °C 140 h. Po pregrevanju pri 250 °C 380 h (kar ustreza brezdimenzijskemu času Fo = 3,09) je bila pri sobni temperaturi dosežena gostota pretoka razplinjevanja q = 2,86 X 10-13 mbar L s-1 cm-2. Z nadaljnjim pregrevanjem pri 350 °C 140 h (Fo = 8,66, kar rezultira v celotni brezdimenzijski čas "LFo = 11,75) se je gostota pretoka razplinjevanja pri sobni temperaturi zmanjšala na q = 5,7 X 10-14 mbar L s-1 cm-2.
Ključne besede: razplinjevanje vodika, nerjavno jeklo, brezdimenzijski čas, metoda naraščanja tlaka, pretočna metoda, kalibracijski koeficient za plinski pretok
1 INTRODUCTION	gases may have adsorbed onto the chamber inner
surface while it was exposed to the environment and
The development of the science and technology of	then slowly released as the pump removed the gas
ultrahigh vacuum (UHV) and extremely high vacuum	from the vacuum chamber.
(XHV) devices has been strongly coupled to the	The performance of a vacuum system is limited
development of increasingly larger and more sophisti-	because it is impossible to eliminate all of the gas
cated devices for physics research, such as particle	sources. Outgassing occurs even in the best-designed
accelerators, magnetic fusion devices, gravity wave	vacuum systems for UHV/XHV. The ultimate pressure in
observatories and many surface analysis techniques.	the system is related to the magnitude of the gas load;
Scientific advancements in the understanding of the	therefore, the measures taken to reduce outgassing are
outgassing limits in UHV/XHV conditions are associated	critical for the production of UHV/XHV. These effects
with these technological developments.	extend the time required to reach UHV/XHV.
Outgassing refers to the spontaneous liberation of	Due to the vaporisation or sublimation of atoms and
gases from the walls of a vacuum chamber or other	molecules from some materials with a vapour pressure
components placed inside the vacuum chamber. This gas	higher than or comparable to the residual pressure within
release results from two processes:1	a vacuum chamber, these materials will increase out-
•	Gas diffusion from the interior of vacuum chamber	gassing. Thus, not all types of materials can be used to walls to their inner surface. This is followed by gas	build vacuum systems, and sometimes it is difficult to desorption into the chamber volume that contributes	select appropriate materials that will fulfil the require-to the vacuum system outgassing.	ments of different processes. Among technical alloys,
•	Release of gases or vapours previously adsorbed onto	austenitic stainless steel (SS) offers several advantages the inner surface of the vacuum chamber walls. These	compared to other alloys, and it is the most widely used
construction material for building vacuum chambers, instruments and components. SS is the most important construction material for UHV/XHV systems because of its good vacuum and mechanical properties.2 It is nonmagnetic, corrosion resistant and chemically inert. At room temperature (RT), SS exhibits negligible vapour pressure and negligible permeation of atmospheric gasses.3 SS is relatively cheap, and it can be machined and welded by standard procedures.
In the high-vacuum range, the predominant gas is usually water vapour resulting from the exposure of the system to humid air.4 The outgassing of hydrogen from SS is the main gas load in UHV/XHV systems,5 particularly in large systems (i.e., accelerators and storage rings). Thus, hydrogen outgassing is the most significant limiting factor in reaching outgassing rates below 10-12 mbar L s1 cm-2 in SS vacuum systems.6 By reducing the hydrogen content in the bulk, the hydrogen outgassing rate can be reduced.
Many techniques have been reported that mitigate outgassing in vacuum systems, but reducing the outgass-ing rates of SS remains a challenge. These outgassing techniques include the following:
a)	Surface treatments to reduce surface roughness, such as electro-polishing and surface machining under special conditions 3
b)	Surface treatments to create oxide films to act as a barrier to the diffusion of hydrogen from the bulk3
c)	High-temperature bake-out (vacuum firing) to reduce the amount of dissolved H (for SS as high as 1 000
0C)4,7
d)	Baking the vacuum system to remove water vapour, which has to be performed at 150-450 0C4
e)	Degreasing and chemical cleaning4
f)	Deposition of thin films to serve as a barrier layer on inner surfaces8
g)	Choosing metals with a low solubility for hydrogen (e.g., copper)4
h)	Reduction of hydrogen surface mobility by introducing surface trapping to reduce recombination8
i)	Thin-film getter coatings to serve as a pump for hydrogen diffusing from the bulk9
The typical outgassing rate of vacuum components made from SS is on the order of 10-11 mbar L s-1 cm-2 without additional processing. An outgassing rate of 10-13 mbar L s-1 cm-2 is "routinely" achievable with a sufficient bake-out treatment. Outgassing rates lower than 10-14 mbar L s-1 cm-2 are achievable only with exceptional care, particularly in thin-wall vessels.8 Until now, several reports have been published describing the reduction of SS outgassing rates by conventional bake-out or vacuum firing, but the resulting outgassing rates are not consistent: similar bake-out temperatures and times gave significantly different numerical values for the outgassing rates.
In 1967, Calder and Levin10 published a paper on the reduction of outgassing of hydrogen from metals. They
proposed the diffusion-limited model (DLM), which means that the atomic hydrogen concentration at the surface is zero and the hydrogen atoms desorb and recombine as molecules as soon as hydrogen reaches the surface. However, their assumption was not correct for lower outgassing rates, whereas their assumption was reasonable for higher outgassing rates. The experimental values obtained by Calder and Levin were two orders of magnitude larger than expected according to the DLM.2 Therefore, their assumption that the surface density of hydrogen when solving the diffusion equations is always zero is questionable. To overcome this problem, Moore11 calculated the outgassing rate using a recombination-limited model (RLM). Moore assumed that the surface density can never be zero and that the boundary conditions must be included when solving the diffusion equation.
Nemanic, Setina, Bogataj and Zajec12-14 used very thin-wall vessels instead of using thicker materials to reduce the outgassing rate. They also introduced a Fourier number (or dimensionless time) Fo to describe the heat treatment intensity. They reported extremely low outgassing rates of approximately 10-15 mbar L s-1 cm-2, which were achieved using thinner materials.12 Therefore, the use of thin materials is more efficient and economically suitable because shorter baking times can be used. However, the insufficient mechanical strength of very thin walls may make them unsuitable for the practical realisation of a vacuum chamber.
1.1 Measuring of outassing flux and outgassing rate
The outgassing flux of a SS sample can be measured by two general methods4. In the first method, the pressure increase is measured as a function of time in a closed vacuum chamber after its evacuation to a low pressure. This method is called the rate-of-rise (RoR) method or the gas-accumulation method. The outgassing flux Q is obtained as
dP
Q = ^dlT
where V is the chamber volume and dP/dt is the measured rate of the pressure change in a closed chamber at a constant temperature.
In the second method, the vacuum test chamber is pumped through an orifice of known conductance, and the pressure drop across the orifice is measured. This method is known as the throughput method or the orifice method. The outgassing flux Q is expressed as
Q = CAP	(2)
where C is the conductance and AP is the pressure difference.
In both methods, the outgassing rate q of the sample is obtained by dividing the measured outgassing flux Q by the surface area A of the sample: q = Q/A. The measured outgassing rate is the net rate of the difference
between the intrinsic outgassing rate of the surface and the readsorption rate.15
A common problem of both methods is that the test sample is placed in a vacuum measurement chamber whose walls also outgas during the measurement. Thus, the background outgassing flux of the empty chamber must be measured and subtracted from the measured outgassing flux of the sample in the chamber.
The presence of a hot cathode residual gas analyser (RGA) or ion gauge in the test chamber can cause additional outgassing, and the residual gas can react with the hot filament. Because they exhibit a certain pumping speed, RoR measurements can only be obtained with an inert vacuum gauge, such as a spinning rotor gauge (SRG).1
2 EXPERIMENTAL SETUP AND SAMPLE DESCRIPTION
We measured the time dependence of the outgassing rate of SS during bake-outs at 250 °C and 350 °C. Our main goal was to determine the time needed to reach the outgassing rate below 10-13 mbar L s1 cm-2 at these bake-out temperatures.
We made two simple cylindrical chambers to serve as the outgassing samples (Figure 1). Each chamber was assembled from a CF 16 flange and had a short connection tube with an inner diameter of 16 mm and a wall thickness of 1.5 mm. The circular end plates had a thickness of 2 mm, and the cylindrical body had an inner diameter di = 108.7 mm and a wall thickness of 2.62 mm. The only difference between the two chambers was the cylindrical body length. The length of the body of
chamber V1 was h = 10 mm, and the length of the body of chamber V2 was l2 = 210 mm.
The chambers were connected to a vacuum measurement system, as shown in Figure 2. The system was pumped by a turbomolecular pump and was equipped with a quadrupole mass spectrometer (QMS), a Bayard-Alpert gauge (BAG) and an SRG. The SRG was connected to a small chamber V0 that was made of a standard CF16 4-way cross and a CF16 T-piece. The chamber V0 can be isolated from the pump by a valve (E0). The chambers V1 and V2 were symmetrically connected to V0 through the bake-proof CF16 all-metal valves E1 and E2, respectively. The calibration gas was introduced into V0 through valve E3.
For the modelling of the outgassing behaviour of SS by DLM and RLM, the thickness of the sample is an important parameter. The same reduction in the outgassing rate of a plane sheet takes 4 times longer time if the sample is 2 times thicker. The entire measurement sample chamber was constructed from AISI type 304 L SS, but we were unable to make the whole chamber from the same sheet of SS. Additionally, each part had a different thickness. However, the measurement procedure was designed so that we could easily calculate the outgassing rate of the cylindrical body (which has a well-defined thickness) from the separate measurements of the outgassing flux of each chamber.
The chambers V0, V1 and V2 were placed in an oven where they could be uniformly heated up to 350 °C. Additional vacuum measurement instruments (BAG and QMS) were mounted outside the oven due to the
Figure 1: A photo of two SS cylindrical vacuum chambers used for the outgassing rate study
Slika 1: Fotografija dveh cilindričnih vakuumskih posod iz nerjavnega jekla, namenjenih za študij razplinjevanja
Figure 2: A schematic of the outgassing measurement system: QMS -quadrupole mass spectrometer; SRG - spinning rotor gauge; BAG -Bayard-Alpert gauge; E0, E1, E2 and E3 - bake-proof vacuum valves; V0, V1 and V2 - measuring volumes
Slika 2: Shema vakuumskega sistema za meritve razplinjevanja: QMS - kvadropolni masni spektrometer; SRG - viskoznostni merilnik z lebdečo kroglico; BAG- Bayard-Alpertova trioda; E0, E1, E2 in E3 -pregrevni vakuumski ventili; V0, V1 in V2 - merilni volumni
limitation of their operational temperature. During the bake-out, the SRG suspension head was removed, and only a thimble with a rotor was baked with the other parts.
The outgassing flux measurements at ambient temperature were performed using the RoR method. The outgassing flux measurements while baking the vacuum system (at temperatures of either 250 °C or 350 °C) were performed using the throughput method, which is described in Section 2.2.
2.1 Measuring of outassing flux using the RoR method
By closing valves E0, E2 and E3 and opening valve E1, we measured the pressure increase in chambers V0 + V1 with the SRG and used this reading to calculate the outgassing flux Qi = Qvg + 0v1. With the same SRG, we also measured the pressure increase in chambers V0 + V2 when valve E1 was closed and E2 was opened. This measurement gave Q2 = Qvg + Qv2. In both cases, Qvg was the same because of the symmetrical configuration of the connection with V1 and V2. By subtracting the two measured outgassing fluxes, the outgassing flux of VG can be eliminated: Q2 - Qi = Qv2 - Qvi. The difference is equal to the outgassing flux from the major part of the cylindrical body of chamber V2 having a surface area A = ndiX (I2 - li).
To calculate the outgassing flux by the RoR method following Equation i, we need to know the volume of the measurement system. The volumes Vi and V2 can be calculated from geometrical measurement because they are composed of simple cylindrical parts. The volume VG has a more complicated geometrical shape because it includes the inner volume of the CF16 valves. Its volume can be determined by the static gas-expansion method. The calculated surface areas of Vi and V2 are accurate. However, we can only estimate the surface area of VG. The volumes and the surface areas for the 3 parts of the vacuum system are given in Table 1.
Table 1: The geometrical surface areas and volumes of particular parts of the vacuum measurement system, including their ratios Tabela 1: Geometrijska površina A in volumen V posameznih delov vakuumskega merilnega sistema, vključujoč njihovo razmerje
	Geometrical surface area A (cm2)	Volume V (L)	Ratio (V/A) (L cm-2)
VG	475 (rough estimate)	G.12G6	2.54 x IG-4
Vi	247	G.1G46	4.23 x IG-4
V2	931	1.962	2.11 x IG-3
Vtot	1653	2.1872	1.32 x IG-3
2.2 Measuring of outassing flux using throughput method
To measure the outgassing flux of the test chambers during baking, the use of the SRG is not possible because the operating temperature for the measuring head of the SRG only ranges from 10 °C to 50 °C.16 As a
consequence, the RoR method cannot be used. Therefore, an adapted throughput measurement method that differs from the orifice method was used.
The steady-state outgassing flux from chamber Vi can be calculated by measuring the pressure difference APi, as measured by the BAG or QMS while opening and closing valve Ei. By knowing the effective pumping speed S at the point where the BAG or QMS are connected to the vacuum system, we can calculate the corresponding outgassing flux Qi = S x APi. Similarly, by opening and closing E2, we can determine AP2, and the outgassing flux Q2 = S x AP2. Also, the difference between the two fluxes is the outgassing flux from the major part of the cylindrical body of chamber V2, which has a surface area A = ndi x (I2 - li), as in the case of RoR measurements.
We were primarily interested in the outgassing of hydrogen from SS, so we used a QMS tuned to the mass number 2 (H2 peak) for the AP measurements. The basic output signal of the QMS is an ion current of a particular gas. The ion current Ii+ as a function of the partial gas pressure Pi is given by Ii+ = gi x Pi, where gi is the sensitivity coefficient, which depends on the gas type, the QMS geometry and the settings of various operational parameters. For accurate measurements, gi must be determined experimentally (calibrated in situ) for each instrument and each gas. Therefore, to measure the outgassing flux of hydrogen, the sensitivity coefficient of the QMS gH2 and the effective pumping speed SH2 had to be determined. For hydrogen, the outgassing flux Q equals
Q = -
S H, AI H 2
gH,
(3)
To more easily determine the outgassing rate q = Q/A, we combined the effective pumping speed and the sensitivity coefficient into a single calibration coefficient Kh2:
KH, = SH, / gH,	(4)
The calibration coefficient Kh2 can be determined experimentally in a straightforward manner and is described in the following section.
2.3 Determination of calibration coefficient
We can introduce a known flow of a calibration gas through valve E3, which is an adjustable leak valve. By closing valves EG, Ei and E2, the flow rate of the calibration gas Qcal into the volume VG can be measured by the RoR method using the same SRG as for the outgassing flux measurement. After the flow of calibration gas is determined, we can measure the difference in the hydrogen ion current AIh2+ by opening valve EG. Because the calibration gas flow was introduced to the QMS from the same direction as the outgassing flux from the measuring chambers, the calibration coefficients for the calibration gas flow and the outgassing flux
have similar values. The calibration coefficient can be calculated by rearranging Equation 3:
Öcal
K =
A/ h,
(5)
The measurements of A/h2+ have been performed for a wide range of calibration gas flows ßcal. Figure 3 shows that when the Qcal (measured by the SRG) increases, the A/h2+ (measured by the QMS) increases proportionally.
From the data in Figure 3 and with Equation 5, we have calculated the calibration coefficient. The calculated values are given in Table 2. The mean value of the calibration coefficient K of the QMS for hydrogen was 1.12 X 105 mbar L A-1 s-1, and the relative standard deviation was a(K)/K = 3.61 %.
Table 2: The calibration coefficients of QMS for hydrogen while measuring outgassing flux by the throughput method Tabela 2: Kalibracijski koeficient QMS pri merjenju pretoka raz-plinjevanja vodika s pretočno metodo
/+(A)
1.21 X 10-
1.10 X 10-
3.63 X 10-
2.37 X 10-11
1.47 X 10-11
1.42 X 10-
1.41 X 10-
5.54 X 10-12
1.83 X 10-12
1.24 X 10-12
1.21 X 10-
1.20 X 10-
Kh2 (mbar L A-1 s-1)
116998
118505
114295
115234
113737
110694
110739
106750
107483
107763
108467
108439
The final equation to calculate the outgassing rate is K
q = t (a/ v2-a/ v1)
(6)
Figure 3: The difference in the hydrogen ion current, measured by QMS (using the throughput method), as a function of the hydrogen gas flow, measured by SRG (using the RoR method) Slika 3: Razlika v ionskem toku vodika (ki je bil izmerjen s pretočno metodo in uporabo QMS) v odvisnosti od pretoka vodika (izmerjenega z metodo hitrosti naraščania tlaka in uporabo SRG)
where A/yi and A/y2 are the changes in the hydrogen ion current when closing valves E1 and E2, respectively.
3 RESULTS AND DISCUSSION
3.1 Time dependence of H2 outgassing rate during bake-out at 250 °C
The measuring chambers were first baked at 250 °C for 380 h. The bake-out temperature was increased to 350 °C, and baking continued for an additional 140 h. The chambers were baked by keeping the outer side of the chamber wall in air while the inner side was under vacuum.
During the bake-out at 250 °C, we performed several measurements of the outgassing rate at the bake-out temperature using the throughput method. The heat treatment intensity expressed as a Fourier number, the baking time of the system and the outgassing rates during the heat treatment are given in Table 3.
Table 3: The measured hydrogen outgassing rate of the SS vacuum chamber during the bake-out. All of the outgassing rates were measured at a temperature of 250 °C.
Tabela 3: Izmerjena gostota pretoka razplinjevanja vodika vakuumske komore iz nerjavnega iekla med pregrevanjem pri 250 °C
^Fo
0.20
0.77
0.77
0.78
0.88
0.89
0.91
1.54
1.59
1.71
1.72
1.72
1.87
2.06
2.07
2.08
2.26
2.27
2.84
2.85
3.06
3.07
t (h) at 250 °C
24
95
95.2
96
108
110
112
190
196
211
211.5
212
227
228.5
230
254
255
256
278
279
350
351
376
378
q (mbar l s-1 cm-2)
2.6 X 10
1.10 X 10-
1.12 X 10-8
1.19 X 10-8
8.20 X 10
8.40 X 10
8.80 X 10
4.20 X 10-9
3.90 X 10
2.90 X 10
2.95 X 10-9
3.00 X 10
2.50 X 10-9
2.40 X 10
2.30 X 10
2.00 X 10
2.05 X 10-9
2.08 X 10
1.80 X 10
1.85 X 10-9
1.00 X 10
1.05 X 10
6.90 X 10
7 30 X 10-10
The Fourier number (or dimensionless time) Fo = 4Dtd-2 is a characteristic quantity for describing and modelling diffusion. It was used to compare the outgassing rates for different experiments.17 To determine the dimensionless time Fo, the following must be known: the diffusion constant D for hydrogen at temperature T, the processing time t and the wall thickness d.
The parameters t and d are measured directly. The temperature dependence of D is given by
D = D0 exp
- E^
kT
(7)
Typical values for the diffusion pre-exponential factor (Do = 0.012 cm2 s-1) and the activation energy (Ea = 0.57 eV)18 were used to calculate the Fo values.
The bake-out at 250 °C was interrupted a few times to cool the chambers and to measure the outgassing rates at RT. The time course was as follows: bake-out at 250 °C for 24 h, interrupt baking for the RT measurement, bake-out at 250 °C for 73 h, interrupt baking for the RT measurement and bake-out at 250 °C for 284 h followed by an RT measurement. The temperature was then raised to 350 °C, and baking was performed for additional 140 h. The system was then cooled again, and the last RT measurement was performed. The results of the hydrogen outgassing rate measurements at RT, which were performed using the RoR method, are summarised in Table 4.
Table 4: The hydrogen RT outgassing rates of SS after the indicated bake-out treatment
Tabela 4: Gostota pretoka razplinjevanja vodika iz nerjavnega jekla pri sobni temperaturi po navedenem postopku pregrevanja
T (°C)
250
250
250
350
Incremental time ^t (h)
24
73
284
140
Fo (At)
0.19
0.59
2.31
8.66
^Fo
0.19
0.78
3.09
11.75
q
(mbar L s-1 cm-2)
2.10 X 10-12
A comparison of our results of the RT outgassing rate with the results obtained by Park et al.19 is shown in Figure 4. The Fourier numbers for the data from Park et al. were recalculated using our activation energy for the diffusion Ea = 0.57 eV (instead of Ea = 14.5 kcal mol-1 — 0.622 eV, which was used by Park et al.). Our results
Figure 5: The hydrogen outgassing rate of SS as a function of the dimensionless time Fo during the heat treatment at 250 °C Slika 5: Gostota pretoka razplinjevanja vodika iz nerjavnega jekla v odvisnosti od brezdimenzijskega časa Fo med postopkom pregrevanja pri 250 °C
agree and are within a factor 3 to 4 of the results of Park et al.
A graphical representation of the results (taken from Table 3) of the outgassing rates measured at 250 °C as a function of dimensionless time Fo is shown in Figure 5.
In Figure 6, the same measured outgassing rates as a function of the bake-out time are presented and compared with the calculations based on the DLM.16 The outgassing rate predicted by diffusion theory for t = 380 h is q = 5 x 10-11 mbar L s-1 cm-2, which is 15 times lower than the measured value q = 7 x 10"10 mbar L s-1 cm-2. In the early stage of bake-out, until approximately 100 h at 250 °C, the predicted curve according to the DLM and the measured data agree. However, discrepancies start to emerge after 100 h. For Fo > 0.8, the atomic hydrogen recombination on the SS surface is assumed to become a rate-limiting step. When a hydrogen atom has moved from the bulk to the surface/vacuum interface, it needs to find another hydrogen atom on the surface to recombine.
Figure 4: A comparison of the hydrogen outgassing rates between our measurements and the measurements of Park et al.17, which were performed after the system was cooled to RT
Slika 4: Primerjava gostote pretoka razplinjevanja vodika iz nerjavnega jekla med našimi meritvami in meritvami Parka et al.17, ki so bile izvedene po ohladitvi sistema na sobno temperaturo
Figure 6: A comparison of the DLM-predicted hydrogen outgassing rate of SS and the measured data during the heat treatment at 250 °C as a function of the bake-out time
Slika 6: Primerjava gostote pretoka razplinjevanja vodika iz nerjavnega jekla, predvidenega na osnovi modela DLM, z merilnimi rezultati v odvisnosti od časa pregrevanja pri 250 °C
As the number of freely diffusing hydrogen surface atoms is reduced, recombination becomes less likely. Thus, when the diffusion of hydrogen atoms to the surface become small, the recombination of the hydrogen atoms might be the rate-limiting step, i.e., outgassing from the surface is described by the rate at which the hydrogen atoms can be recombined, not the bulk diffusion rate.
4 CONCLUSIONS
The determination of the calibration coefficient K for the BAG and QMS allowed us to measure the outgassing rate during the heat treatment without stopping the heating process. The DLM governs the initial removal of hydrogen from SS, and it is assumes that hydrogen surface recombination plays an important role in the outgassing rate at lower hydrogen concentrations. To compare the different outgassing treatments, the Fo number appears to be a good choice because it can be accurately calculated for any processing time and temperature. The Fo number is a parameter that shows the level of heat treatment of SS, and it is related to the diffusion of hydrogen atoms in SS.
An outgassing rate q = 2.86 x lO13 mbar L s1 cm 2 at RT was achieved for AISI type 304 L SS after baking the system at 250 °C for 380 h (conversion to dimensionless time gives Fo = 3.09). This outgassing rate was further reduced to q = 5.7 x 1014 mbar L s1 cm 2 by baking for another 140 h at T = 350 °C (Fo = 8.66, total dimensionless time ^Fo = 11.75).
Acknowledgement
S. Avdiaj acknowledges partial financial support by the European Union, European Social Fund. The study
was implemented in the framework of the Operational Programme for Human Resources Development for the period 2007-2013.
5 REFERENCES
I	K. Jousten, Thermal outgassing, Proceedings of the CERN Accelerator School, Snekersten, Denmark, CERN report, edited by S. Turner, 1999, 111-125
2Y. Ishikava, V. Nemanic, Vacuum, 69 (2003), 501-512
3	M. Leich, Hydrogen outgassing of stainless steel, our present knowledge, Proceedings of the 1st Vacuum Symposium, UK, 2010 (http://www.ss.dsl.pipex.com/rgaug/pdfs/ vs1/Leisch.pdf)
4	J. Lafferty, Foundations of Vacuum Science and Technology, John Wiley & Sons, New York 1998, 625-652
5	P. A. Redhead, Extreme high vacuum, Proceedings of the CERN Acceleration School, Snekersten, Denmark, CERN report, edited by S. Turner, 1999, 213-227
6M. Bernardini et al., J. Vac. Sci. Technol., A 16 (1998), 188-193
7B. Zajec, V. Nemanic, Mater. Tehnol., 35 (2001) 5, 259-264
8 R. Reid, Outgassing Surface Conditioning, 2009, (www.stfc.ac.uk/ resources/pdf/ronreid25209.pdf)
9C. Benvenuti, P. Chiggiato, F. Cicoira, V. Ruzinov, Vacuum, 50 (1998), 57-63
10 R. Calder, G. Lewin, Br. J. Appl. Phys., 18 (1967), 1459
II	B.C. Moore, J. Vac. Sci. Technol., A 13 (1995), 545-548
12	V. Nemanic, J. Setina, J. Vac. Sci. Technol., A 17 (1999), 1040-1046
13	V. Nemanic, T. Bogataj, Vacuum, 50 (1998), 431-437
14	V. Nemanic, B. Zajec, J. Setina, J. Vac. Sci. Technol., A 19 (2001), 215-222
15	P. A. Redhead, Hydrogen in vacuum systems; an Overview, Proceedings of the First International Workshop on Hydrogen in Materials and Vacuum Systems, AIP Conference Proceedings, Volume 671, 2003, 243-254, (http://adsabs.harvard.edu/abs/2003AIPC..671.. 243R)
16www.mksinst.com
17	V. Nemanic, T. Bogataj, Kovine zlitine tehnologije, 32 (1998) 3-4, 249-253
18	D. M. Grant et al., J. Nucl. Mater., 149 (1987), 180-191
19	Park et al., J. Vac. Sci. Technol., A 26 (2008), 1166-1171