New lithium gas sorbents. I. The evaporable variant

Article abstract of DOI:10.1016/j.jallcom.2007.01.158

Solid solutions of lithium in Ag and Cu in the form of balls, wires, or strips are convenient sources for depositing lithium films as getters on the walls of vacuum vessels. Measurement of the O₂, CO and CO₂ sorption characteristics have shown that these films - e.g. those generated electrothermally from Ag/Li solid solutions - excel the best getters of the Ba-film type or the high porosity getters based on Ti, V and Zr alloys. It has been found that tight lithium films passivate as a sorption depth of approximately 100 ? is reached. Methods for further improvement of the sorption behavior of lithium coatings are discussed.

Full text of DOI:10.1016/j.jallcom.2007.01.158

Journal of Alloys and Compounds 455 (2008) 489–496
New lithium gas sorbents
I. The evaporable variant
K. Chuntonov , J. Setinab
Nanoshell Materials R&D GmbH, Primoschgasse 3, 9020 Klagenfurt, Austria
a,∗
a
b
Institute of Metals and Technology, 1000 Ljubljana, Slovenia
Received 8 September 2006; received in revised form 30 January 2007; accepted 30 January 2007
Available online 6 February 2007
Abstract
Solid solutions of lithium in Ag and Cu in the form of balls, wires, or strips are convenient sources for depositing lithium films as getters on the
walls of vacuum vessels.
Measurement of the O , CO and CO sorption characteristics have shown that these films – e.g. those generated electrothermally from Ag/Li
2 2
solid solutions – excel the best getters of the Ba-film type or the high porosity getters based on Ti, V and Zr alloys. It has been found that tight
˚
lithium films passivate as a sorption depth of approximately 100 A is reached. Methods for further improvement of the sorption behavior of lithium
coatings are discussed.
©
2007 Elsevier B.V. All rights reserved.
Keywords: Lithium alloys; Solid solutions; Thin films; Gas sorption; Passivation
1
. Introduction
necessary, because it provides safety and preserves functional
material. Together with the activation operation it is an instru-
ment for the sorption process control. Therefore, the availability
of a passivating gas sorbent with lithium as the active component
could mean an important progress in pertinent technology.
The favorable mechanical properties of solid solutions, which
allow comparatively easy shaping of the material into any desir-
able form, make the new gas sorbents even more attractive. The
material can be manufactured in the form of balls, wires, strips,
and foils with a dense compact structure, formed by pressing,
rolling or drawing the initial homogeneous ingots, which are
readily produced by vacuum metallurgy methods [7].
Getter technologies currently experience a rapid upsurge.
Within less than a decade a large number of metallic gas sor-
bents of new types have been developed. It is enough to mention
here as examples multilayer film coatings in which at least one
of the layers serves as a barrier for gases entering the vacuum
chamber through its walls [1,2] or protects the getter layer itself
from gas poisoning at the stage of the device assembly [3], or
gas sorbents with the structure of an isolated dendritic carcass
[
4] as well as high selectivity getters [5,6].
Here we describe new lithium gas sorbents on the basis of
solid solutions of Li in Ag, Al, Au, Co, or Cu [7], which not
only work – depending on the temperature – either as evaporable
As evaporation of Li from its solid solutions and also sorption
of gases by lithium and lithium alloys have their peculiarities,
in the present and in the following publications we describe
the behavior of lithium alloys in both getter applications, evap-
orable and non-evaporable, hoping that this will serve for a more
effective usage of the new material.
(
EGs) or non-evaporable getters (NEGs), but are also the first
representatives of passivating getters with a chemically highly
active component.
In many applications, where according to the operating con-
ditions temporary contacts between the getter material and the
atmosphere are inevitable, passivation is not just useful, but also
2. Experimental methods
The parent components, Ag (Alfa Aesar, 99.999%), Cu (Alfa Aesar,
9
9.999%), and Li (Chemetal, 99.9%), were remelted in vacuum to remove
∗
Corresponding author. Tel.: +43 463 31 93 29/0; fax: +43 463 31 93 29/7
E-mail address: k.chuntonov@nanoshell.at (K. Chuntonov).
volatile impurities and charged in the appropriate atomic ratio under argon
into thin walled steel tubes (12.7 mm in diameter), which were sealed off under
0
925-8388/$ – see front matter © 2007 Elsevier B.V. All rights reserved.
doi:10.1016/j.jallcom.2007.01.158

4
90
K. Chuntonov, J. Setina / Journal of Alloys and Compounds 455 (2008) 489–496
vacuum on both ends. To generate the solid solution, the metallic ampoule
with the components was heated in a vertical vacuum furnace to a temperature
higher then the melting point of the more refractory component and after an
◦
exposure for about half an hour was turned around its vertical axis by 180
several times for homogenization of the melt.
Without being taken out of the ampoules, the ingots were rolled with a press
in a hot state in order to obtain strips 2.5–3.0 mm thick. Subsequently, the outside
steel envelope was removed. The released semi-ready substrate was either rolled
into foil, or cut longwise to produce wires with a square cross-section, which
were finally brought to the required diameter by drawing.
Thin films of Li were evaporated from the new alloys onto a borosilicate
glass substrate in ultrahigh vacuum to evaluate their gettering properties. Evap-
oration was done in situ in a getter testing apparatus, which was an all-metal
vacuum system. During initial evacuation and evaporation it was pumped with
a turbomolecular pump.
A wire with a diameter of 0.5 mm was made from the alloy. It was placed
coaxially in a glass cylinder with an inner diameter of 27 mm. The ends of the
wire were connected to a power supply for resistive heating to the evapora-
tion temperature. The whole system was evacuated over night (without baking)
−
6
to a residual pressure below 1 × 10 Pa. During evaporation, which typically
Fig. 1. Phase diagram of Ag–Li.
−
4
lasted 10–15 min, the pressure in the system increased into the 10 Pa region.
This pressure increase was partly due to outgassing of the vapor source itself
and mainly due to thermal outgassing of power leads and surrounding parts,
which were indirectly heated. After switch-off of the heating power the pres-
sure dropped below 1 × 10 Pa within few minutes. Sorption tests were started
approximately half an hour after evaporation. This time interval was estimated
to be enough to cool down the glass substrate with the film to room temperature.
In ultrahigh vacuum, thermal energy can be exchanged by radiation only and
this is quite a slow process.
According to the phase diagram Ag–Li (Fig. 1), maximum
solubility of Li in solid Ag can reach 61 at.%, but at room tem-
perature the homogeneity boundary of solid solutions is close
to 46.5 at.% Li [11]. In the annealed state these alloys are very
ductile and are easily treated mechanically at room temperature
with the help of standard tools and techniques.
−6
The area of the evaporated getter films on the inner surface of the glass
cylinder was from 30 to 45 cm , depending on the length of the vapor source.
2
In the subsolidus range, and at average concentrations of Li
in the alloys, the partial vapor pressures of the components have
The area had to be kept small enough so that the initial getter pumping
speed was not larger than molecular conductance for a test gas to the film inside
the glass cylinder. Gas transport was limited mainly by the dimensions of the
substrate cylinder. In all cases where the gas transport was influencing the mea-
surements the appropriate corrections for the gas pressure above the Li film were
made.
The total amount of Li in the film was measured after the sorption test. The
glass cylinder was removed from the vacuum system and the film was sim-
ply washed from the surface with pure water. The lithium concentration in the
aqueous solution was determined with a Flame Atomic Absorption Spectrome-
ter (FAAS—Perkin-Elmer 2380). Lithium compounds contained in the reacted
film are non-volatile at room temperature and are readily soluble in water, and
therefore the calculated mass of Li dissolved in the water solution is equal to
the mass of Li previously condensed as a film onto the support. Concerning the
procedure of sorption measurements see Appendix A.
6
a ratio P /PAg ꢀ 10 . At this the value of P can vary with
Li
Li
−1
the temperature from a negligibly small value up to 10 mbar,
satisfying any needs which can appear in practice concerning the
deposition rate. More precisely, the vapor pressure of Li over the
Ag–Li solid solution with an fcc crystal lattice is described by
a formula:
0
Li
2
ln(P , bar) = ln(P ) + A + B(1 − x) + ln(x),
Li
where
ꢀ
ꢁ
T
0
Li
ln(P , bar) = 11.8019 + 0.7530
100
ꢀ
ꢁ
3
. Physicochemical basics
19.4268
T
−
− 2.0532 ln
T/1000
1000
Among the chemically most active metals, mainly the alkali
and alkaline-earth metals, lithium shows the highest capability
to interfuse with the metals of other groups in the Periodic Table,
and only lithium forms solid solutions in a wide concentration
range with some of the metals which themselves are stable in the
atmosphere at room temperature. Alloys of this kind, containing
the active metal in an inert matrix of a metal solvent appear to
be gas sorbents which are more convenient to handle than true
intermetallic compounds of the same active components [8] or
the so-called supported metals (referring to alkali metal films
on solid carriers, e.g., of K/Al2O3 [9]), as well as silicon- or
graphite-alkali metal clathrates or intercalates [10]. We consider
here the properties of solid solutions using the most studied
Ag–Li alloys as an example, but the conclusions are also valid
in the case of analogous systems.
is the vapor pressure of pure Li at 453.67 ≤ T (K) ≤ 3000 accord-
ing to [12], A = 374.5/T − 0.9533, and B = −0.5075.46/T +
0.34 ln(T) − 1.94 (T is the temperature in K, x is the mole fraction
of Li in the solid solution).
Fig. 2 shows in a graph form the temperature depen-
dence of the lithium vapor pressure for some Ag–Li solid
alloys.
In the system Cu–Li the solid solution range according to
[13] is limited to the interval 0–22 at.% Li. The vapor pressure
of Cu under the temperatures lower than 1300 K is by two orders
of magnitude lower than of Ag while the vapor pressure of Li
in binary alloys with these metals is approximately the same at
equal concentrations. Thus, the lithium vapor pressure above the
solid solution of Cu–15 at.% Li in the temperature range from

K. Chuntonov, J. Setina / Journal of Alloys and Compounds 455 (2008) 489–496
491
Fig. 2. Dependence of the vapor pressure of Li above solid solutions of Ag–Li
on temperature.
Fig. 3. Partial pressure of some gases in equilibrium with solid Li.
4. The gettering model
2
00 to 846 K is equal to
It is useful to generalize the known facts and ideas about
−
18731.3
ln(P , bar) =
+ 12.068.
getter materials in the form of a simple sorption model. Usu-
ally sorption is understood as a two stage process, in which
first the getter surface is saturated with gas molecules and then
the volume of the material becomes involved in the process
Fig. 4).
The sorption rate of the first stage is proportional to the
number of available sorption sites
Li
T
It follows from these data, taking into account the high activity
of metallic lithium, that solid Ag–Li alloys in the concentration
range 0–50 at.% Li as well as solid Cu–Li alloys in the concen-
tration range 0–22 at.% Li, can be used as evaporable getters
(
(
EG) with lithium film condensates as the active sorbent phase.
Basically, lithium, as well as barium, is able to sorb any gases
except noble gases. Lithium compounds with the elements of
all common residual gases or their components are known, and
these compounds are stable in vacuum at room temperature as
well as at higher temperatures [14,15].
Table 1 shows selected results of calculations using
FACTSAGE program [16] for the equilibria in the systems
Li–O–H–N–C, which give an idea about the affinity of the cor-
responding gaseous species towards lithium.
ꢂ
ꢃ
n
I
I
V = V0 1 −
,
(1)
n∗
I
where V ∼ εp; ε is a sticking coefficient, which for getters is
0
always smaller than 1, p is the pressure of the gaseous medium,
n = n(t) is the current density of adatoms, t is the time, and n*
is the surface gas density at saturation. The value of V = V (t)
changes with time from the maximum value of V at n = 0, to
zero at n = n*.
For the second stage, the sorption rate V = V (t) is deter-
mined by the rate of gas diffusion into solid bodies, and usually
Q ∼ t for the mass throughput of the sorbed gas, while the
mechanism of the volume sorption depends on the nature of a
getter. Chemically active metals (Fig. 4b), which form no solid
solutions with gases, bind gases in the form of compounds like
I
I
I
0
II
II
The same correlations are shown – in a more convenient form
for analyzing vacuum problems – in Fig. 3. On the whole it
follows from the thermodynamic estimations that in the tem-
1/2
◦
perature range from room temperature to ∼240 C lithium can
be considered as a potential gas sorbent for a particularly broad
application spectrum.
Table 1
Change of the Gibbs energy for reactions of Li metal with some gases in vacuum at the temperatures of 300, 400, and 500 К
Reaction
ꢀGr (kJ/mol)
00 K
3
400 K
500 K
Li(s) + (1/4)O2(g) = (1/2)Li2O(s)
−280.945
−68.325
−42.811
−150.874
−50.140
−274.744
−60.739
−38.737
−143.574
−45.680
−268.459
−52.962
−34.575
−136.230
−41.240
Li(s) + (1/2)H2(g) = LiH(s)
Li(s) + (1/6)N2(g) = (1/3)Li3N(s)
Li(s) + (1/3)CO(g) = (1/3)Li2O(s) + (1/6)Li2C2(s)
Li(s) + (1/5)CH4(g) = (4/5)LiH(s) + (1/10)Li2C2(s)
(s) and (g)—solid and gaseous state, respectively; the value ꢀGr is calculated for 1 mol Li.

4
92
K. Chuntonov, J. Setina / Journal of Alloys and Compounds 455 (2008) 489–496
Fig. 4. Two types of metallic gas sorbents: (a) the stage of adsorption of gases; (b) alkali, alkaline-earth and rare-earth metals; for simplification only gas diffusion
through the layer of reaction products is shown (absorption at room temperature); (c) transition metals, characterized by high solubility of gases (absorption at
elevated temperatures).
oxides, nitrides, hydrides, etc., formed on the surface of the
metal. Even at room temperature, the layer of products grows
fast enough due to diffusion of the reagents through this layer.
On the contrary, another group of getter materials like Hf,
Ti, V, and Zr forms with most of the gases solid solutions in a
wide concentration range (Fig. 4c). At room temperature, after
saturating the surface of the metal with gases, the surface is pas-
sivated, butonheating, withactivationofdiffusionprocesses, the
gases start to dissolve in the crystal lattice, and thereby sorption
acquires a volume character.
The specific sorption capacity of such a surface, imitating the
surface of a porous body, is equal to
C0[1 + k(1 + α)]
Cp =
,
(2)
1 + α
2
where Cp has dimensionality mol/cm , and C0 = n*.
It can be seen that the quantity of sorbed gases is directly pro-
portional to k and thus does not depend on α, namely, Cp ꢀ n*k.
These two types of behavior of getter materials prede-
termine the corresponding form of the sorption curves for
the case of a smooth atomic surface: continuously descend-
I
II
ing curves of the form (V + V ) for active metals, and more
I
abruptly descending curves of the form V for transition metals
(Fig. 5).
In order to single out the contribution of the structure of the
sorbent in gas sorption, let us complicate our model (Fig. 4) by
replacing the continuous phase boundary vacuum/sorbent by a
plane equal in area but perforated (S, Fig. 6). Let the new bound-
ary consist of a system of holes with the total area S1 distributed
in a certain way, and a remaining part of the surface S2 = S − S1,
which has the properties of transition metals, i.e. ε < 1. The den-
sity of adatoms at saturation is n*, and at room temperature
there is no diffusion into the volume. For holes ε = 1 (which is
correct till the equilibrium reaction sets in) and the equivalent
surface density of gases at saturation is equal to n*k (1 + α)/α,
where k = Sp/S. Sp is the total area of all the internal voids avail-
able to gases, which are situated below the holes (α = S1/S2 and
k ꢁ 1).
Fig. 5. Typical forms of sorption curves as a function of time.

K. Chuntonov, J. Setina / Journal of Alloys and Compounds 455 (2008) 489–496
493
Fig. 6. Structural aspects of adsorption: (a) continuous surface; (b) surface of a porous body: (I) A granule type; (II) “brush” type.
Unlike the capacity, the gettering rate depends on α, and the
the presence of a horizontal region, common for the sorption
more this applies the smaller is ε. Thus, for the initial gettering
rate we have
curvesofhighporositymaterials, inthegraphVp−t(Fig. 5). It
should be pointed out that the given sorption effect is based on
purely geometric reasons and does not depend on the nature
of the material, i.e. the result for a given material structure
will be the same for metals, ceramics or polymers.
I
V
V
α + ε
p
=
,
(3)
I
0
ε(1 + α)
I
0
3. Unlike k and α, porosity, the parameter which is often men-
tioned in getter certifications [17–19], cannot strictly be
correlated with the values V and C, i.e. it is not informative.
where V is the initial sorption rate of the smooth continuous
surface. The influence of α on the value of V is shown in Fig. 7
I
p
demonstrating the obvious fact that with the increase of the
fraction of the holes on the surface its pumping action grows
as a result of the increase of the averaged sticking coefficient
ε¯ = (α + ε)/(1 + α).
4
. The values of the coefficients α and k can be maximized,
if the same getter mass is distributed in an ordered way as
shown in Fig. 6b(II). This is a structure of a “brush” [20]
type consisting of a set of parallel isolated needles or rods of
the same diameter, the bases of which stand on the substrate
with the tops forming the apparent (virtual) boundary with
vacuum. The voids are united into a connected system of
slits, reaching towards the vacuum boundary of the voids,
which makes the transportation of gas molecules inside the
voids easier. The coefficient of the material utilization and
accordingly its lifetime increase by many times compared to
the getter films produced from powder particles (Fig. 6b(I)).
The present model takes into account only the surface
sorption but even with this limitation it leads to important con-
clusions:
1
. According to (2) the capacity of the porous is higher than
the capacity of a continuous body with the same surface S by
approximately k times.
2
. The sorption rate increases with the part of the surface
accounting for the holes (Eq. (3)). In cases of larger val-
ues for α and k, immediately after the filling of the area S2
5
. Experimental results and discussion
(Fig. 6) with gas particles, the process is stabilized, because
for holes ε = 1, and thus their capacity is high. This explains
Oxygen and oxygen containing gases (CO, CO2), known for their high affin-
ity to lithium, were selected for the first sorption tests. Dense lithium films from
1
80 to 350 A˚ thick, which correspond to 70–140 atomic layers, were studied.
A wire of the composition Ag–25 at.% Li with a diameter of 0.5 mm was
suspended along the axis of a glass or metal tube and served as a source of the
lithium film deposited on the inner surface of this tube. The wire was directly
−
6
heated with electric current in vacuum of about 10 mbar and lithium was
evaporated at the temperate of 450–600 C.
◦
TheresultsofthemeasurementsaregiveninFigs. 8–12. Asexpectedfromthe
general properties of lithium [21], thin lithium condensates show high sorption
activity at room temperature. If we assume, that sorption of O2, CO, and CO2
takes place according to the reactions
4Li + O2 = 2Li2O
6Li + 2CO = Li2C2 + 2Li2O
Fig. 7. Dependence of α on rate V at different values of ε.
10Li + 2CO2 = Li2C2 + 4Li2O,

4
94
K. Chuntonov, J. Setina / Journal of Alloys and Compounds 455 (2008) 489–496
Fig. 8. Sorption of O2 by a Li film at room temperature (the mass of the film is
Fig. 11. Dependence of the sorption of O2 by a Li film on time.
2
0
.040 mg, the area of the film is 30 cm ).
atmosphere this state is conserved for an indefinite time. Any further erosion
occurring is due to the moisture contained in the air.
This effect represents a case of passivation of the lithium surface in contact
with oxygen containing gases at room temperatures. The density of the passi-
vated layer increases approximately by a factor of four compared with the native
state of the film. The thickness of the reacted part of the film decreases from
1
20 A˚ to approximately 60–70 A˚ . Since the growth of the product layer takes
place at a very small rate, which is smaller than required for the growth of a com-
pact monocrystalline layer, the formed passivated coating is structurally almost
defectless.
Since it took at least 30 min from the end of the film deposition period to the
−
7
beginning of the sorption measurements, at a vacuum of 10 mbar, the surface
of the condensate had enough time to become covered with a monomolecular
layer of the residual gases. Therefore all the sorption curves given in Figs. 8–12
refer to absorption and not to adsorption. However, in the given context the form
ofthecurvesforthelithiumfilmsissurprising:curves V − Q in Figs. 8–10 clearly
show an extended horizontal region, while the curves Q−t in Figs. 11 and 12
evidently do not follow a parabolic but a linear sorption law. These phenomena
and also the high pumping rate require an explanation which can be based on
the following chemical arguments.
Fig. 9. Sorption of CO2 by a Li film at room temperature (the mass of the film
is 0.029 mg, the area of the film is 30 cm ).
2
All the products of the sorption reactions between lithium and reactive gases
can be considered as ionic compounds [22]. The Li cations are small, feature a
and taking into account the total amount of sorbed gases Q, it is not difficult to
define the mass of lithium, which has reacted in each concrete case (Figs. 8–10).
The calculations show that about one third of the total amount of lithium atoms
in the film took part in sorption, which corresponds approximately to a layer not
more than 120 A˚ thick.
With its sorption activity exhausted, observed from the outside of the glass,
the lithium film looked like the initial “mirror” and maintained its color and
metallic lustre for several days even after atmospheric air was admitted. In a dry
+
big Coulomb potential and therefore have a strong polarizing effect on the gas
molecules, facilitating their dissociation into atoms and with this preparing them
for diffusion into the film. In addition there are favorable transport conditions. It
is reasonable to suppose, that besides the usual diffusion of cations, considered
for the oxidation of metals [23,24], in the given case a counter flow of atoms of
the reacting gases also takes place and this diffusion of gases through the layer
of the takes place according to an interstitial mechanism. The O–O distance in
Fig. 10. Sorption of CO by a Li film at room temperature (the mass of the film
2
is 0.055 mg, the area of the film is 30 cm ).
Fig. 12. Dependence of the sorption of CO2 by a Li film on time.

K. Chuntonov, J. Setina / Journal of Alloys and Compounds 455 (2008) 489–496
495
Table 2
CO sorption by planar getters of different composition
Sorption characteristic
Ba–film
Li–film (this work)
0.28
1.8 × 10
HPTF strips
Gettering rate, GA (mbar L/s cm2)
Sorption capacity, Cm (mbar L/mg)
0.035 [26]
4.2 × 10 [26]
0.30 [18]
−
2
−1
−6
7.5 × 10 [18]
the structure of a monoxide Li2O, which is the basis of the growing layer of the
products, is 3.27 A˚ [21], from where we can derive that the volume of octahedral
voids ideally corresponds to the size of an oxygen atom (the radius of the latter
is 0.66 A˚ , whereas the radius of the inscribable sphere for the mentioned voids
pump, but with a more active gas sorbent. The thickness of the deposited
layers can be set in the range from a monoatomic layer to ∼120 A˚ , which
is the limit of gas permeability according to our measurements.
is 0.67 A˚ ).
◦
The low temperatures of lithium evaporation, from 450 to 600 C, and
This hypothesis still needs experimental confirmation.
also the high specific resistance of solid solutions of metals (for the 75 at.%
Ag/25 at.% Li alloy, the value of ρ is 31 ␮ꢁ cm) make expensive heavy-current
power supply units unnecessary. Further, if needed, lithium condensates and
products of their reactions with gases can be removed completely and without
any consequences for the metallic or other substrate by rinsing with water or
alcohol. And, finally, another advantage of lithium is the easy ionization of its
vapors accompanied by a strong increase of the reactivity of the gaseous species.
This is an aspect, which may be important for further developments of lithium
vacuum pumps.
Normalized sorption characteristics of lithium films are compared regarding
CO sorption as an example in Table 2 with the corresponding values for barium
films and for the so-called HPTF films based on Ti, V, Zr and other transition
metals [18,25].
The data show that according to the sum of the sorption parameters lithium
obviously performs best as a getter, having only a slightly lower sorption rate
than HPTF. However, this advantage of HPTF materials is only due to their
structure rather than to their chemical nature: high porosity layers of transition
metals are compared to dense lithium films.
The analysis of scanning electron microscopy photos of HPTF–materials
18] allow defining α = 1, and taking into consideration the data [27–29] it is
6. Conclusions
[
possible to use ε = 0.1 as an averaged sticking coefficient for the main residual
gases on a smooth surface of transition metals. Then according to (3) we get
1. Sorption of dilute oxygen or oxygen containing gases (like
CO, CO2) by lithium has a volume character and involves lay-
I
I
2
V /V ꢀ 6. Thismeansthatthevalueof0.30 mbar L/s cm mentionedincolumn
p
0
4
of Table 2 is practically determined by purely geometrical parameters: by the
˚
ers of a thickness of about 100 A, beyond which passivation
of the film takes place.
presence of holes on the surface with ε = 1 and with high sorption capacity (dense
2
films of the same alloys would have a gettering rate 0.30:6 = 0.05 mbar L/s cm ).
2
. Thesorptionrateisunusuallyhighevenattheinternalabsorp-
tionstagewhichcanbeascribedtothecatalyticactivityofLi
This also means that in the case of lithium there is also a substantial reserve
for enhancement of the sorption properties of film coatings by varying their
structure. Given the practical importance of this issue, let us discuss it in more
details.
There are three ways of increasing the effectiveness of film getters by means
of the structural factor. They are the usage of porous films of a granule type
+
cations and favorable diffusion characteristics of the sorption
products.
3. The factors which are limiting the depth of the passivated
layer are not clear and require further studies.
(
a), of ordered structures of a “brush” type (b), and of periodically renewed thin
layers (c).
4. One of the most promising mode of practical application of
lithium getter films is the periodical deposition of the metal
˚
in the form of dense layers not more than 120 A thick.
(
a) HPTF materials (Fig. 6b(I)) and getter films of group IIA metals have this
kindofstructureintheformofaconglomerateofpowderparticlesconnected
by diffusion bridges. Diffuse deposits of group IIA metals are obtained
by evaporation of the metal in dilute gas atmosphere, usually Ar or N2
5
. The analysis of the sorption process based on the simple
gettering model shows that the optimal getter film structure
is the “brush” type structure.
[
17,30,31].
A weak point of these films which is shortening their lifetime are diffu-
Acknowledgements
sion bridges, where saturation of the volume with gases, and, consequently,
embrittlement of the material with its further destruction take place long
before the theoretical limit of the sorption capacity is reached. Besides,
for active metals with their high values of ε, disintegration of the structure
does not produce such a strong effect as for transition metals with small ε,
because the increase of the sorption rate is determined by the ratio ε¯ /ε and
it is not large for large ε (Fig. 7). In fact, in accordance with (3) diffusive
films of barium compared with dense films show the increase of the initial
sorption rate for CO only by two times [31].
The authors would like to thank Prof. H. Schmidbaur, TU
M u¨ nchen, for his support and fruitful consultations and to Prof.
G. Voronin, Moscow State University, for his assistance in the
thermodynamic issues.
Appendix A. Measurement of sorption properties of
getter films
(
b) “Brush” structures in short represent a considerable step forward compared
to granule type structures. Besides the improvement of the sorption prop-
erties, these structures due to their dimensional uniformity defer the time
boundary for the appearance of loose particles and for the pealing off of the
material to such an extent, that this phenomenon is no more a problem.
Film structures of a “brush” type can be grown by crystallization from
a vapor or from a melt under a controlled heat dissipation from the sub-
strate through vacuum evaporation of a more volatile impurity [4,20], which
during solidification is pushed to the intercellular or interdendritic space.
The sorption properties of evaporated Li films were measured
in a getter test apparatus, which was constructed according to
ASTM Recommendations [32]. This document has been writ-
ten for non-evaporable getters (NEGs), but can be also applied
to evaporated getter films. According to the Recommendation
a getter is exposed to a well-known mass throughput Q of pure
(
c) Continuous or periodical deposition of thin lithium layers on the walls of
a vacuum vessel is a further development of the idea of a titanium vacuum
−
1
◦
test gas, measured in mbar L s at 23 ± 2 C. Equilibrium pres-

4
96
K. Chuntonov, J. Setina / Journal of Alloys and Compounds 455 (2008) 489–496
sure p above the getter surface is established according to the
instantaneous getter pumping speed G, measured in L s
[6] A.L. Kovacs, M.H. Peter, K.S. Ketola, J.F. Linder, US Patent 6,822,880
November 23, 2004).
7] K. Chuntonov, G. Voronin, O. Malishev, EP 05 025 992.8 (November 29,
005).
−
1
(
[
Q
G = p
2
.
[
[
8] R. Szwarc, US Patent 4,512,960 (April 23, 1985).
9] A. F u¨ rstner, Active Metals, VCH, Weinheim, 1996, pp. 381–426.
The gas sorption capacity C is a cumulative quantity of gas
sorbed by the getter while at operating temperature until the
pumping speed drops to 5% of the initial value (terminal pump-
ing speed GT). It can be calculated by integrating the mass
throughput:
[
10] S. Bobev, S. Sevov, J. Solid State Chem. 153 (2000) 92.
[11] H. Okamoto, Phase Diagrams for Binary Alloys, ASM International, Mate-
rials Park, OH, 2000.
[
12] P.I. Bistrov, D.N. Kagan, G.A. Krechetova, E.E. Spielrein, Zhidkokristal-
licheskije teplonositeli teplovih trub i energeticheskih ustanovok, Nauka,
Moscow, 1988.
ꢄ
τ
[13] A.D. Pelton, Bull. Alloy Phase Diagrams 7 (1986) 142.
[14] T.V. Talalaeva, K.A. Kocheshkov, Metodi elementno-organicheskoi khimii,
kn.1, Nauka, Moscow, 1971.
C =
Q dt,
0
[
15] L. Manceron, L. Andrews, in: A.M. Sapse, P.V.R. Schleyer (Eds.), Lithium
Chemistry. A Theoretical and Experimental Overview, John Wiley, New
York, 1995, pp. 1–43.
where τ denotes the time required until the terminal pumping
speed is reached. If the mass throughput is constant during the
test, then C is simply given by:
[
[
[
[
[
[
16] FactSage 5.4.1; www.factsage.com.
17] H.J.R. Perdijk, Suppl. Nuovo Cimento 5 (1967) 73.
18] Brochure SAES Getters: FED HTPF Getter, 2004.
19] P. della Porta, Vacuum 47 (1996) 771.
20] K. Chuntonov, H. Ipser, K. Richter, EP 06002762.0 (10 February 2006).
21] Gmelin’s Handbuch der Anorganische Chemie, 8. Auflage, Lithium, 20,
Verlag Chemie, GmbH, Bergstrasse/Weinheim, 1960.
22] H.U. Borgstedt, Ch. K. Mathews, Applied Chemistry of the Alkali Metals,
Plenum Press, New York, 1987.
23] O. Kubaschewski, B.E. Hopkins, Oxidation of Metals and Alloys, Butter-
worths, London, 1967.
24] K. Meyer, Physikalisch-chemische Kristallographie, Deutscher Verlag f u¨ r
Grundstoffindustrie, Leipzig, 1968.
C = Qτ.
It is convenient to normalize the getter pumping speed by divid-
ing G by the geometric area A of the film:
G
GA =
.
[
[
[
A
Similarly, the sorption capacity C of the film can be normalized
by the mass of active material m:
C
Cm =
.
m
[25] P. della Porta, “Gettering”, an Integral Part of Vacuum Technology, Tech-
nical Paper TP 202, SAES Getters, Proceedings of the American Vacuum
Society 39th National Symposium, 1992.
[
[
[
Such normalization of pumping speed and sorption capacity
allows comparison of sorption properties of various films having
different areas and different thicknesses.
26] F. Ricca, P. della Porta, Vacuum 10 (1960) 215.
27] R.A. Sallach, J. Phys. Chem. 76 (1972) 2156.
28] Y. Hatano, V. Alimov, A. Livshits, Y. Nakamura, N. Ohyabu, M.
Matsuyama, Annual Report 25-2, Hydrogen Isotope Research Center,
University of Toyama; http://www.hrc.toyama-u.ac.jp/annual/volume25/
volume25 e.html#25-2.
29] D.A. King, CRC Crit. Rev. Solid State Mater. Sci. 7 (1978) 175.
30] W. Espe, M. Knoll, M. Wilder, Electronics 23 (1950) 82.
31] B. Ferrario, Vacuum 47 (1996) 363.
32] ASTM F 798-97, Standard Practice for Determining Gettering Rate, Sorp-
tion Capacity, and Gas Content of Non-evaporable Getters in the Molecular
Flow Region, 2002.
References
[
[
1] C. Benvenuti, US Patent 6,468,043 (October 22, 2002).
2] N. Najafi, S. Massoud-Ansari, S. Tadigadapa, Y. Zhang, US Patent
[
[
[
[
6
,499,354 (December 31, 2002).
[
[
[
3] D.R. Sparks, US Patent 6,923,625 (August 2, 2005).
4] K. Chuntonov, EP 1 600 232 (November 30, 2005).
5] P. della Porta, C. Boffito, L. Toia, US Patent Appl. 20,040,101,686 (May
2
7, 2004).