Membrane Gas Separation

14.3.5
CO
2
Capture 
The recovery of carbon dioxide from large emission sources is a formidable technological 
and scientifi c challenge which has received considerable attention for several years [58 –
60] . In particular, the identifi cation of a capture process which would fi t the needs of 
target separation performances, together with a minimal energy penalty, is a key issue. 
Currently, the main strategies for the carbon dioxide capture in a fossil fuel combustion 
process are the following [61] . 
• Oxy - fuel combustion: this option consists in performing the oxygen/nitrogen separa-
tion on the feed stream, so that a CO 
2
/H 
2
O mixture is produced through the combustion 
process. The advantage of feeding an oxygen 
- 
enriched gas mixture (95% oxygen) 
instead of air, is the achievement of a purge stream rich in CO 
2
and water with very 
low N 
2
content, therefore the CO 
2
can be easily recovered after the condensation of the 
water vapour.
• Pre - combustion capture: this solution is developed in two phases: (i) the conversion of 
the fuel to a mixture of H 
2
and CO (syngas mixture) through, e.g. partial oxidation, 
steam reforming or auto - thermal reforming of hydrocarbons, followed by water - gas 

---

296
Membrane Gas Separation
shift; (ii) the separation of CO 
2
(at 30 – 35%) from the H 
2
that is then fed as clean fuel 
to the turbines. In these cases, the CO 
2
separation could happen at very high pressures 
(up to 80 bar of pressure difference) and high temperatures (300 – 700 ° C).
• Post - combustion capture: in this case, the CO 
2
is separated from the fl ue gas emitted 
after the combustion of fossil fuels (from a standard gas turbine combined cycle, or a 
coal - fi red steam power plant). CO 
2
separation is realized at relatively low temperature, 
from a gaseous stream at atmospheric pressure and with low CO 
2
concentration (ca. 
5 – 15% if air is used during combustion). SO 
2
, NO 
2
and O 
2
may also be present in small 
amounts. This possibility is by far the most challenging since a diluted, low pressure, 
hot and wet CO 
2
/N 
2
mixture has to be treated. Nevertheless, it also corresponds to the 
most widely applicable option in terms of industrial sectors (power, kiln and steel 
production for instance). Moreover, it shows the essential advantage of being compat-
ible to a retrofi t strategy (i.e. an already existing installation can be, in principle, subject 
to this type of adaptation) [60] .
The conventional separation processes for CO 
2
capture are: absorption (with amines), 
adsorption (with porous solids with high adsorbing capacity such as zeolites or active 
carbon) and cryogenic separation [62,63] . Even though amine absorption is the most 
common technology for post - combustion capture, its use is by far the best available 
technology owing to the high energetic cost (in the range 4 – 6 GJ/ton CO 
2
recovered) 
related, in particular, to the signifi cant energy consumption in the regeneration step. 
Furthermore, this option requires large - scale equipment for the CO 
2
removal and chemi-
cals handling. Membranes are most often listed as potential candidates for their applica-
tion in post - combustion capture. However, the main problem related to their limited 
application is the low CO 
2
concentration and pressure of the fl ue gas, which requires the 
use of membranes with high selectivities (ca. 100) for fi tting the specifi cation delivered 
by the International Energy Agency, i.e. a CO 
2
recovery of 80% with a purity of at least 
80% [64] . The commercial membranes (CO 
2
/CH 
4
selectivity ca. 50), currently used to 
separate CO 
2
from natural gas at high pressures are not suited for one - stage operation, 
implying a large membrane area and high compression costs [64] . 
Favre [48] provides a critical comparison of the application of polymeric dense mem-
branes versus amine absorption in post 
- 
combustion capture. The energy requirement 
associated with membranes for post - combustion carbon capture for separating mixture 
containing 10% CO 
2
is much larger than that of absorption. Even very selective mem-
branes, showing selectivity above 120, require much more energy than absorption. 
However, for fl ue streams containing 20% or more CO 
2
(steel or kiln productions), rea-
sonable recoveries and permeate compositions can be attained with lower related cost 
than amine absorption. Scura et al. [65] proposed two different membrane system con-
fi gurations specifi cally for applications in which the species to be recovered is at a low 
concentration (the CO 
2
content in fl ue gas is around 10%) and low recoveries could be 
enough to meet process specifi cations. The fl ue gas stream compression (Figure 14.10 a) 
and the vacuum on the permeate stream (Figure 14.10 b), were compared for a permeate 
stream with the same CO 
2
concentration and recovery. In particular, the vacuum option 
is a valuable alternative instead of the more expensive feed compression.
At a fi xed pressure ratio, confi guration (a) requires a lower installed membrane area 
but a higher investment and operating compression cost (the whole feed stream must be 

---

Membrane Engineering: Progress and Potentialities in Gas Separations
297
pressurized). On the contrary, in confi guration (b), even if the total installed membrane 
area increases considerably, the compression cost strongly reduces (only the permeate 
stream must be sucked). The high size compressor in confi guration (a) is substituted by 
a high number of modules, easy control, low investment and operating cost membrane 
modules. For a 20% recovery from a fl ue gas stream with the 13% of CO 
2
and a pressure 
ratio of 15, the vacuum system (b) reduces the compression cost to less than 5% with 
respect to pressured system (a); in the meantime, the required membrane area in (b) 
increases up to more than 10 times. 
In 2008, Baker [66] proposed the possibility of using membranes with CO 
2
/N 
2
selectiv-
ity of ca. 50 (already commercial) as integrated multi - stage solutions. In this case, in fact, 
the appropriate choice of which kind of membrane can be used in each separation stage 
can make this application already feasible (Figure 14.11 ).
For the pre - combustion and oxyfuel capture processes, membranes based on alumina, 
zeolites, silica and carbon that show stability up to 300 ° C are generally proposed [27] . 
(a)
Flue gas
COMPRESSOR
POLYMERIC MEMBRANE
Retentate
Permeate
(CO
2
rich stream)
Flue gas
POLYMERIC MEMBRANE
VACUUM PUMP
Retentate
Permeate
(CO
2
rich stream)
(b)
Figure 14.10 Confi guration of the two membrane systems proposed for CO 
2
capture
13% CO
2
88% CO
2
2% CO
2
CO
2 
recovery = 90%
Figure 14.11 New applications: CO 
2
from coal power plant fl ue gas. Adapted from the 
oral presentation of Reference 66

---

298
Membrane Gas Separation
Materials which conduct
CO
3
2
−
ions (e.g. molten Li 
2
CO 
3
formed from the reaction of 
Li 
2
ZrO 
3
with CO 
2
) have also been studied for CO 
2
/CH 
4
separation up to 600 ° C [67,68] . 
These membranes may be economically effi cient, stable and robust in applications where 
excess heat/energy is readily available to melt the carbonate. Also their use in high -
temperature membrane reactors for integration in power generation cycles with CO 
2
capture has been proposed
[69] 
. However, signifi cant design optimization would be 
required in order to identify effi cient, feasible and environmentally sound technical solu-
tions. In addition, further development and validation of performance of these membranes 
in real applications are needed.