202
Membrane Gas Separation
carbon dioxide and water in a reducing environment. This is then further converted to
more hydrogen through the water gas shift reaction and CO
2
is then separated from
hydrogen before combustion. Finally, in oxy - fuel combustion air is replaced by oxygen
in the combustion process which produces
a fl ue gas of mainly H
2
O and CO
2
, which is
readily captured. For all three strategies, membrane technology has suffi ciently matured
to become commercially competitive with other separation technologies in the coming
decade.
Alongside the major gases, a range of minor gas components are present in all of these
applications [5] . For example, in natural gas sweetening the main separation is of CO
2
and H
2
S from CH
4
but water and a range of hydrocarbons are also present.
In post -
combustion capture the main separation is CO
2
from N
2
; however, the process also experi-
ences SOx (a combination of sulfur oxides), NOx (a combination of nitric oxides) and
water. In pre - combustion capture, separation is of CO
2
from H
2
and N
2
, with H
2
S, CO,
NH
3
, water and hydrocarbons also present. In addition, O
2
is often present in many of
these processes, along with argon if air is used. Depending on the fuel source, heavy
metals,
such as mercury and arsenic, may exist, as well as halogens, such as fl uorine and
chlorine, in their acidic form. Other processes than power generation with the possibility
for carbon capture, such as cement production, refi neries and blast furnaces, also have a
range of minor components present. The concentration of these minor components varies
considerably and is dependent on the process, fuel, temperature and pressure of reaction.
Potential concentrations of some of these minor components are provided in Table 11.1 .
For membrane gas separation processes the effects of these
minor components have to
be considered on a case by case basis.
The impact of some of these minor components can be signifi cantly reduced by a suit-
able pre - treatment process, such as a de - sulfurizer or an adsorbent guard bed. However,
remaining concentrations of these minor components in the membrane gas separation
process may compete with CO
2
(or other gas of interest) for separation and therefore
decrease the observed permeability, as well as degrade the membrane structure, altering
separation performance. Furthermore, the permeabilities of these minor components are
of
interest, because of the possibility of generating component - rich permeate or retentate
streams, dependent on membrane selectivity. Thus for example, it may be possible to
capture both acid gases (H
2
S and CO
2
) simultaneously from a pre - combustion stream.
Table 11.1 Composition of minor components in a range of industrial processes [6 – 8]
Pre - combustion
Post - combustion
Iron blast
furnace
Cement
production
CO
2
10 – 35 mol%
12 – 14 vol%
20 vol%
14 – 33 vol%
SOx
–
1000 – 5000 ppm
–
0 – 150 ppm
NOx
–
100 – 500 ppm
–
0 – 150 ppm
CO
300 – 4000 ppm
∼
10 ppm
22 vol%
0 – 130 ppm
H
2
S
500 – 1000 ppm
–
0 – 5000 ppm
–
NH
3
0 – 1500 ppm
–
–
–
H
2
∼
20 vol%
–
4 vol%
–
Water
Saturated
Saturated
4 vol%
12.8 vol%
Hydrocarbons
0 – 100 ppm
–
–
–
The Effects of Minor Components on the Gas Separation Performance
203
Hence, this provides the opportunity to reduce the number of separation processes the
feed gas may go through.
Gas permeation through non - porous polymeric membranes is generally described by
the solution - diffusion mechanism [2] . This is based on the solubility of specifi c gases
within the membrane and their diffusion through the dense membrane matrix. In turn, the
solubility of a specifi c gas component within a membrane is a function of its critical
temperature, as this is a measure of the gas condensability.
Critical temperatures for a
range of gas components are provided in Table 11.2 . Conversely, the diffusivity depends
upon the molecular size, as generally indicated by the kinetic diameter. Indeed, Robeson
et al. [9] have recently postulated that the relationship between the ideal permeability of
one species
P
i
and that of another
P
j
are related by a simple function:
P
kP
j
i
n
=
(11.1)
Where:
n
d
d
j
i
= ⎛⎝⎜
⎞
⎠⎟
2
(11.2)
However, to achieve the best fi t to this relationship they propose slight modifi cations to
the kinetic diameters provided by Breck [10] which are more generally used. Both values
are included in Table 11.2 where available.
Table 11.2 Kinetic diameter and critical temperature of common gases
Gas
Breck
kinetic
diameter ( Å ) [10]
Robeson kinetic
diameter ( Å ) [9]
Critical temperature
(K) [11]
H
2
2.89
2.875
33.2
N
2
3.64
3.568
126.2
CO
3.76
132.9
Ar
3.40
150.8
O
2
3.46
3.347
154.6
NO
3.17
180
CH
4
3.8
3.817
190.6
CO
2
3.3
3.325
304.2
HCl
3.2
324.6
C
3
H
8
4.3
369.8
H
2
S
3.6
373.2
NH
3
2.6
405.6
SO
2
3.6
430.8
NO
2
431.4
SO
3
491
C
6
H
14
507.4
C
6
H
6
5.85
561.9
H
2
O
2.65
647.3
Hg
1750
