Journal of Membrane Science 333 (2009) 30–37
Contents lists available at ScienceDirect
Journal of Membrane Science
journa l homepage: www.e lsev ier .com
Absorp zin
in a pla er
Su-Hsia L ha
a Department o aiwan
b R&D Center fo ian Un
a r t i c l
Article history:
Received 15 D
Received in re
Accepted 23 Ja
Available onlin
Keywords:
Carbon dioxid
Piperazine
2-Amino-2-me
Polypropylene
Membrane contactor
Wetting
Plasma treatment
Surface modification
ent
gas a
Expe
The C
iffusi
(AMP
ary l
d rou
tio of
d AM
absorption flux was observed for plasma-treated PP hollow fibers in comparison to non-treated fibers
when using 1.0 mol/dm3 AMP as the absorbent. Membrane durability was greatly improved by the treat-
ment of plasma with CF4 (20 W/5 min). The membrane mass transfer coefficient km for plasma-treated
PP membranes was comparable to that of polytetrafluoroethylene (PTFE) membranes.
© 2009 Elsevier B.V. All rights reserved.
1. Introduc
Carbon d
gas, and thu
surface tem
burning ele
a separation
emission so
packed tow
to remove C
process for
because of t
include floo
ment, new
(GAM) proc
streams. Ho
larger conta
columns. Th
and no foam
∗ Correspon
∗∗ Correspon
E-mail add
(K.-L. Tung).
0376-7388/$ –
doi:10.1016/j.mtion
ioxide is the most emitted anthropogenic greenhouse
s is believed to be responsible for increasing the earth’s
perature. Half of CO2 emissions come from fossil fuel-
ctric power plants [1]. Therefore, the development of
process for removal and recovery of CO2 from these
urces is badly needed. In general, bubble columns,
ers, venturi scrubbers, and/or sieve trays can be used
O2 from process streams. The best-known commercial
CO2 separation is the packed column system. However,
he disadvantages of the packed column system, which
ding, channeling, and the need for large-scale equip-
technologies are needed. A gas absorption membrane
ess is one alternative to recovery of CO2 from waste gas
llow fiber membrane contactors (HFMCs) offer a much
ct area per unit volume compared to tray and/or packed
ey have the advantages of no flooding, no entrainment,
ing-limited flow rate restrictions [2,3].
ding author. Tel.: +886 3 2654129; fax: +886 3 2654199.
ding author. Fax: +886 3 4652040.
resses: sslin@nanya.edu.tw (S.-H. Lin), kuolun@cycu.edu.tw
The additional mass transfer resistances of membranes limit
the CO2 absorption rate in the membrane contactor module even
though the interfacial area of the membrane is greater than in con-
ventional absorbers [4]. Therefore, minimization of the membrane
mass transfer resistance is an important goal in gas absorption
processes that use membrane contactors. The success of GAM
technology is thus highly dependent on the wetting relationship
between the membrane and the liquid solvent used in the system.
It is preferable to use hydrophobic membranes for CO2 absorp-
tion processes. Hydrophobic membranes display less wetting and
swelling phenomena, which drive large increases in the mass trans-
fer coefficients for the membrane.
Several studies have investigated CO2 absorption in hollow
fibers. Yeon et al. [5] investigated CO2 absorption in poly(vinylidene
fluoride) (PVDF) and PTFE hollow fiber membranes using a sin-
gle absorbent (monoethanolamine, MEA). They reported that the
absorption rate per surface area was higher in PVDF membrane
than that in PTFE membrane because of the non-wetted condition
of membrane pore. Kim and Yang [6] used aqueous MEA, AMP, and
diethanolamine (DEA) solutions as absorbents to separate CO2–N2
mixtures in PTFE hollow fiber membrane contactors. They found
that the removal of carbon dioxide increased with the increase of
flow rate of an absorbent. As temperature rose, the absorption rate
of CO2 increased as well. At high temperature, vaporized water filled
membrane pores and shell side, which deteriorated the separation
see front matter © 2009 Elsevier B.V. All rights reserved.
emsci.2009.01.039tion of carbon dioxide by mixed pipera
sma-modified polypropylene hollow fib
in a,∗∗, Kuo-Lun Tung b,∗, Wei-Jie Chen b, Hao-Wei C
f Chemical and Material Engineering, Nanya Institute of Technology, Chung-Li 32091, T
r Membrane Technology and Department of Chemical Engineering, Chung Yuan Christ
e i n f o
ecember 2008
vised form 22 January 2009
nuary 2009
e 5 February 2009
e absorption
thyl-1-propanol
hollow fiber
a b s t r a c t
This paper examines the enhancem
polypropylene (PP) hollow fibers in a
of piperazine (PZ) and alkanolamine.
rates, and absorbent concentrations.
and absorbent concentrations. The d
using 2-amino-2-methyl-1-propanol
diffusion resistance in the gas bound
processes. The water contact angle an
in CF4. We show that the wetting ra
plasma treatment, when using mixe/ locate /memsci
e–alkanolamine absorbent
contactor
ng b
, ROC
iversity, Chung-Li 32023, Taiwan, ROC
of hydrophobic plasma treatment on the performance of
bsorption membrane (GAM) process using aqueous solutions
riments were conducted at various gas flow rates, liquid flow
O2 absorption flux increased with increasing gas flow rates
on resistance of the liquid phase was only important when
) or methyldiethanolamine (MDEA) as the sole absorbent. The
ayer is always dominant in the hollow-fiber CO2 absorption
ghness of the PP membrane increased by plasma treatment
the PP membranes decreased from 0.0674% to 0.027% after
P–PZ solution as the absorbent. An 8% increase in the CO2

S.-H. Lin et al. / Journal of Membrane Science 333 (2009) 30–37 31
efficiency. Kumar et al. [7] experimentally and theoretically stud-
ied a novel absorbent called CORAL for CO2 capture from dilute gas
streams. Feron and Jansen [8] applied CORAL absorbents in poly-
olefin membrane contactors for CO2 separation. They reported that
the mass tra
process is d
brane mass
decreased a
price of PTF
in industry.
chemical m
for comme
Plasma trea
modificatio
ification wi
reported th
upon CH4 p
in gas diffus
In this s
tion flux in
using aqueo
influence of
trations on
resistances
2. Theoret
2.1. Film mo
The mas
fiber memb
membrane,
at any cross
J(ndi) = K
=
(
where kL, k
of the liqui
an overall
ber of the
of the fiber
(kPa) of CO2
gas–membr
CO2 (m3 kP
resistance i
hollow fibe
1
KL
=
(
di
dok
If the film m
membrane
obtained by
KLa =
R
PM
where R is
of the conta
Table 1
Characteristics of the hollow fiber module.
UMP-0047R (PVDF) LM-2P18 (PP)
Module o.d. (mm) 16 18
. (mm
. (mm
. (mm
gth (c
of fib
pore s
rosity
area (n
divid
indi
pred
1.62
divid
ted t
0.0
embr
he ef
n in
2D
di ln
D−1k +
he v
esti
e of
CO2
10−7
be
ation
km,liq
erim
ateria
MDE
emb
sup
), and
of t
was
ation
asma
plasma reactor system was made in the USA (Dressler, PFG-
odifications of the hollow fibers by plasma treatment were
ed under the following working conditions: 10–50 W work-
er, 300 Torr working pressure, a reaction time of 5–30 min,
4 sccm mass flow. The hollow fiber (Do = 2.7 mm, Di = 1.8 mm,
cm) was rotated during plasma treatment (5 rpm). Plasma-
hollow fibers (number of fibers = 16) were arranged to form
ar triangle and housed in a polycarbonate shell.nsfer resistance in the CORAL membrane gas absorption
ominated by the liquid phase. The measured CO2 mem-
flux increased with increasing CO2 partial pressure, but
s cyclic liquid loading is increased. However, the current
E hollow fiber limits the application of these techniques
It is thus important to seek alternative methods, such as
odification of membranes that are cheaper than PTFE,
rcial realization of GAM processes for CO2 absorption.
tment is considered to be one feasible method for the
n of membrane surface characteristics. Membrane mod-
th plasma has been well-studied [9–11]. Pai et al. [9]
at the hydrophobic properties of carbon fibers improved
lasma treatment. They measured a water contact angle
ion layers of 132.8◦.
tudy, we experimentally investigated the CO2 absorp-
a plasma-treated PP hollow fiber membrane contactor
us solutions of mixed alkanolamines as absorbents. The
liquid flow rates, gas flow rates, and absorbent concen-
CO2 absorption fluxes were investigated. Mass transfer
are also discussed.
ical modeling
del
s transfer between gas and liquid through the hollow
rane contactor occurs in three parts: the gas film, the
and the liquid film [5]. The CO2 flux per unit fiber length
-section, J, can be expressed by Eq. (1):
L(ndi)
( Pg
He
)
= kg(ndi)(Pg − Pm)
km
RT
)
(ndi)(Pm − Pi) = kLE(ndi)
( Pi
He
)
(1)
m, and kg indicate the mass transfer coefficients (m/s)
d phase, membrane, and gas phase, respectively; KL is
liquid phase mass transfer coefficient; n is the num-
fibers; di and do are the inside and outside diameter
, respectively; Pg, Pi, and Pm, are the partial pressures
in the bulk gas phase, membrane–liquid interface, and
ane interface, respectively; He is Henry’s constant for
a/kmol); and E is an enhancement factor. The overall
n the liquid phase mass transfer through the porous
r membrane contactor can be expressed by Eq. (2) [12]:
g
)
( 1
He
)
+
( di
dokm
)(RT
He
)
+
1
EkL
(2)
odel for mass transfer is across an ideal non-wetted
[5], the overall mass transfer coefficient, KL, can be
Eq. (3):
/He
(3)
the CO2 absorption rate (kmol/(m3 s)) per unit volume
ctor.
Shell i.d
Fiber o.d
Fiber i.d
Fiber len
Number
Average
Fiber po
Surface
2.2. In
The
can be
kLdi
DL
=
The in
correla
kgde
Dg
=
The m
Here, t
as show
km = 
D−1e =
Using t
(6), we
the cas
that of
2.62 ×
km can
penetr
1
km
=
3. Exp
3.1. M
PZ,
The m
was a
0047R
Details
water
purific
3.2. Pl
The
600). M
perform
ing pow
and CF
L = 24.5
treated
a regul) 14 14
) 2.2 2.7
) 1.4 1.8
m) 31.4 24.5
ers 21 16
ize (m) 0.2 0.2
, ε (%) 50 40
ominal m2) 0.02 0.02
ual gas-film mass transfer coefficients
vidual mass transfer coefficient of the liquid phase kL
icted using Eq. (4) [2,13]:
(
d2i vL
LDL
)1/3
(4)
ual mass transfer coefficient of the gas phase has been
o the expression in Eq. (5) [14]:
23
(
4Hvg
g
)0.83( g
Dg
)0.44
(5)
ane mass transfer coefficient is predicted using Eq. (6).
fective diffusivity is calculated from the harmonic mean,
Eq. (7):
eε
(do/di)
(6)
D−1g (7)
alues for the fiber used in our hollow fiber module in Eq.
mated km,gas filled as approximately 2.69 × 10−3 m/s. In
liquid-filled pores, the effective diffusivity is similar to
in solution (1.49 × 10−9 m2/s), and km,liquid filled is about
m/s. If the pores are partially filled with liquid, then
calculated from the average fractional depth of liquid
ˇ, as in Eq. (8):
ˇ
uid filled
+
1 − ˇ
km,gas filled
(8)
ental
ls
A, and AMP were purchased from Aldrich Chemicals.
rane contactor used as a CO2 absorber in this study
ported PVDF hollow fiber unit from Pall Co. (UMP-
a PP hollow fiber unit from MICRODYN Co. (LM-2P18).
he hollow fiber units are listed in Table 1. Deionized
used. All chemicals were used without any further
.
treatment