Study of vacuum properties of Non Evaporable Getters NEG as a function of elemental composition

1
Study of vacuum properties of Non Evaporable
Getters (NEG) as a function of elemental
composition
S. Patel1, O.B. Malyshev1, K.J. Middleman1, J.S.
Colligon2, R. Valizadeh2,, V. M. Vishnyakov2 1
The Cockcroft Institute, Science and Technology
Facility Council, Daresbury Science and
Innovation Campus, Keckwick Lane, Daresbury,
Warrington, WA4 4AD, UK 2 Department of
Chemistry, Manchester Metropolitan University, UK
Abstract The internal surface of 50 cm long and
34 mm diameter stainless steel tubes were coated
with TiV, TiZrV and TiHfV getter films using a
cylindrical magnetron. The NEG coated tubes were
baked for 24 hours at various temperatures
ranging from 150C to 300C. The TiZrV film had
the lowest activation temperature which was found
to be 180C. The TiV film had the highest
activation temperature of 250C and the highest
saturation coverage for CO of about 3 monolayers.
The evolution of surface chemical composition
during the activation was determined by x-ray
photoelectron spectroscopy (XPS) and the
functional properties were evaluated by pumping
speed measurements. Film composition was
determined by Rutherford back scattering and was
found to be non homogenous in depth. Surface
morphology was studied by SEM which revealed a
columnar structure.
Determination of Pumping Capacity and Initial
Sticking Probability The pumping capacity of
various NEG coatings was determined using a
dedicated test system available in the laboratory
(For more details see poster reference
VSTP4-304). The tubes were activated for a period
of 24 hours prior to CO injection. From the
ratio of the top to bottom RGA CO partial
pressure, it was possible to determine the
initial sticking probability by referring to Test
Particle Monte Carlo simulations of the system.
These results are shown in the table below. Figs
7, 8 and 9 show the changes in CO partial
pressure ratio as a function of surface coverage
for different compositions and activation
temperatures.
Fig 7
XPS Results XPS analysis shows that in addition
to dominant Zr, Ti, V and O peaks, a very small C
1s peak was also present. This was used to
calibrate the energy scale (See Fig 1). High
resolution XPS measurements of the core levels of
V, O, Ti, Zr and C (for a given TiZrV sample)
were also carried out at 180, 250 and 300C. As
shown in Fig 2, after heating to 300C the two
vanadium characteristic peaks are shifted from
524 and 517 eV, corresponding to vanadium oxide,
to 521 and 512 eV, corresponding to metallic
vanadium. The Ti at the surface behaves very
similarly to vanadium after annealing at 300C
(Fig 3). However, the reactivation of Zr was less
successful and, even after annealing at 300C,
there is still a considerable amount of zirconium
oxide present at the surface (Fig 4). The
residual oxygen peak observed at 532.5 eV is
mostly associated with surface zirconium oxide.
Fig 9
Fig 8
SEM Results The SEM images show that the
morphology of the film depends on substrate
topography, deposition conditions (e.g. pressure,
pulsed vs. dc magnetron sputtering) and film
composition. In pulsed magnetron where the
surface of the film is bombarded with energetic
ions the grain size are of order of 5 to 10nm
(Fig 10) Tertiary alloys (e.g. TiZrV) result in a
smaller grain size relative to binary alloys
(e.g. TiV), (See Figs 11 and 12). The porosity
of the film can be controlled by deposition
pressure, with higher pressure resulting in
greater porosity. Figs 13 and 14 show the film
on silicon and stainless steel substrates
respectively.
Fig 1
Fig 2
Fig 11
Fig 12
Fig 10
Fig 13
Fig 14
Fig 4
Fig 3
XRD Results XRD patterns of a film deposited on
stainless steel, silicon, and float glass are
shown in Fig 5. The only peak associated with the
film is a very broad peak at 2? 37.7.
RBS Results Fig 6 shows that the film is non
homogeneous in depth as a result of (1) surface
to target geometry during deposition and (2) due
to the magnetron coil length which was 1/5 of the
deposition tube .

• Conclusions
• The RBS data suggests that the pumping capacity
  and sticking probability is
• not dependent on a particular composition but
  is instead dependent on the
• individual alloy composition, as well surface
  topography, film morphology
• and grain size (as shown by the SEM results).
• The XPS results show the change in composition
  from an oxide rich to
• metallic surface as a result of annealing to
  300C.
• The average grain size calculated from the XRD
  results is 5-6nm for TiZrV,
• and is independent of the substrate.
• The highest initial sticking probability was
  obtained using a TiZrV coating
• activated at 300C.

Fig 5
Fig 6