Membrane Gas Separation

Table 13.5 CO
2
and N 
2
permeability (  P  ), diffusion (  D  ) and sorption (  S  ) coeffi cients 
obtained for three Pebax ® 1657 – PEGEPI blends by using the gas permeation method in 
time - lag mode under an upstream pressure of 4 bars and at 25 ° C 
1657/10% PEGEPI
1657/20% PEGEPI
1657/50% PEGEPI
P
CO2
(a)
104
±
1
90
±
10
30
±
1
P
N2
(a)
1.78
±
0.02
1.8
±
0.5
0.5
 
α
 
CO2/N2
58.5
50
60
D
CO2
(b)
4.8
±
0.3
5.0
±
0.6
4.4
±
0.6
D
N2
(b)
14
0.16
±
0.05
0.08
 
α
 
DCO2/DN2
0.3
31
55
S
CO2
(c)
220
±
13
180
±
32
69
±
7
S
N2
(c)
1
129
±
64
60
 
α
 
SCO2/SN2
220
1.4
1.2
(a) Barrer or 10 
−
10
cm 
3
(STP) cm cm 
−
2
s 
−
1
cmHg 
−
1
.
(b) 10 
−
7
cm 
2
s 
−
1
.
(c) 10 
−
4
cm 
3
(STP) cm 
−
3
cmHg 
−
1
.

---

CO
2
 Permeation with Pebax®-based Membranes for Global Warming Reduction
271
Pebax ® 1657 to its 50:50 blend with PEG 200. The CO 
2
permeability increase was 
attributed to a solubility increase with the increase in ethylene oxide (EO) units, but also 
to a diffusivity increase due to an increase in free volume (either total fractional free 
volume or individual free volume space): at high PEG content, there would be an increase 
of Pebax ® segment motion, since low molecular weight PEG would act as a plasticizer 
in inserting its molecules between chain segments to screen out polymer/polymer inter-
actions, and generating additional intermolecular space for small Brownian motion [43] . 
Our sorption and diffusion data indicated an apparently more complex behaviour 
of the blends. We did not observe a clear increase in CO 
2
solubility coeffi cient in the 
membrane with increasing PEG content but an increase in the CO 
2
diffusivity. As no free 
volume measurements were performed in the present study, we can only suggest a similar 
mechanism of diffusivity enhancement by PEG additive in Pebax ® 1657 as that proposed 
by Yave et al. [43] . 
The infl uence of the nature of the polymer component on the permeability of blend 
membranes is worth discussing. Based on DSC and AFM data, Car et al. [10] reported a 
good miscibility of Pebax 
® 
1657 with PEG200 in blends up to 50 
wt.% PEG. For 
PEG300, we observed a similar behaviour of the blends in DSC, i.e. for the polyether 
phase, unique glass transition temperature and melting temperature, whose values decrease 
when the PEG content increases (Table
13.6 
). However, in optical microscopy we 
observed that at a low PEG content, the crystalline PA phase was fi nely dispersed in the 
blend, even more than in pure Pebax ® 1657; at higher PEG300 contents, the dispersed 
crystalline phase became coarser. At 50 wt.% of PEG300, the phase was very coarse, and 
a slight exudation of PEG was observed on the membrane faces (Figure 13.8 ).
These features suggest a phase separation in the matrix, followed by a release of liquid 
PEG to the surface, although no PEG300 melting peak was observed in DSC trace of the 
50% PEG membrane, probably because the expelled PEG was easily removed in handling 
operations. We assume that the phase separation limits the improvement in CO 
2
diffusiv-
ity by a saturation of the polymer matrix, leading to a stagnancy of the blend membrane 
performances. 
20%
40%
Content of PEG300 in the blend (%)
60%
0%
0
0
10
20
30
40
50
60
70
80
90
20
40
60
80
100
120
140
160
CO
2
permeability/Barrer
Ideal selectivity CO
2
/N
2
180
200
Figure 13.7 CO 
2
permeability ( • ) and CO 
2
 /N 
2
ideal selectivity coeffi cient ( 
䊐
 ) of the 
different PEG300 - Pebax ® 1657 blend membranes, 25 ° C, 4 bars

---

272
Membrane Gas Separation
0% (a)
0
100
m
m
0
100
m
m
0
100
m
m
0
100
m
m
0
100
m
m
0
100
m
m
5% (b)
10% (c)
15% (d)
20% (e)
50% (f)
Figure 13.8 Optical microscopy pictures of the different Pebax ® 1657/PEG300 blends
Table 13.6  Values of the glass transition temperatures of PE and the melting temperature 
of PE and PA blocks for the different Pebax ® grades, determined from DSC measurements 
(in the fi rst heating scans). The PA crystallinity is obtained with the values found in the 
literature for the melting enthalpy of PA 6 
T
m
polyamide 
block ( ° C)
T
m
polyether 
block ( ° C)
T
g
polyether 
block ( ° C)
wt.% PA 
crystallinity 
in Pebax ®
wt. % PA 
crystallinity 
in PA block
1657
207
14
−
55
15
30
1657/PEG300 5%
205
9
−
56
12
24
1657/PEG300 15%
204
7
−
65
–
–
1657/PEG300 20%
203
7
−
65
–
–
1657/PEG300 30%
200
1
–
–
–
1657/PEG300 50%
196
0
–
–
–

---

CO
2
 Permeation with Pebax®-based Membranes for Global Warming Reduction
273
The clearly lower miscibility of PEG300 compared with that PEG200 is attributed to 
the difference in their chemical structure: PEG200 has four EO units for two hydroxyl 
end groups, while PEG300 has six EO units. The gains in mixing entropy and hydrogen -
bonding interactions should be much larger for PEG200 than for PEG300. 
The larger increase in CO 
2
permeability for PEG200 blend membranes compared with 
the PEG300 one can be explained by the larger positive contribution of PEG200 to the 
gas diffusion (larger free volume effect), and the higher PEG content obtainable in the 
blend membranes. 
In summary, blending of Pebax ® 1657 with PEG300 resulted in membranes of higher 
CO 
2
permeability and selectivity due to enhanced CO 
2
diffusivity. However, a phase 
separation that occurred at high PEG300 contents led to a stagnancy of performance. The 
performance of Pebax ® membranes are compared with those obtained for the Pebax ®
blend membranes on a CO 
2
/N 
2
Robeson ’ s plot [5,44] (Figure 13.9 ).