Accepted Manuscript

High capacity polyethylenimine impregnated microtubes made of carbon nanotubes

for CO2 capture

Laura Keller, Burkhard Ohs, Jelena Lenhart, Lorenz Abduly, Philipp Blanke, Matthias
Wessling
PII: S0008-6223(17)31017-5
DOI: 10.1016/j.carbon.2017.10.023
Reference: CARBON 12460

To appear in: Carbon

Received Date: 4 July 2017

Revised Date: 6 October 2017
Accepted Date: 7 October 2017

Please cite this article as: L. Keller, B. Ohs, J. Lenhart, L. Abduly, P. Blanke, M. Wessling, High capacity
polyethylenimine impregnated microtubes made of carbon nanotubes for CO2 capture, Carbon (2017),
doi: 10.1016/j.carbon.2017.10.023.

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to
our customers we are providing this early version of the manuscript. The manuscript will undergo
copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please
note that during the production process errors may be discovered which could affect the content, and all
legal disclaimers that apply to the journal pertain.
 ACCEPTED MANUSCRIPT

N2
CNT polyethyleneimine

PT
RI
U SC
AN
M
D
TE
EP
C
AC
CNT microtubes impregnated with
polyethyleneimine CO2 & N2 CO2
 ACCEPTED MANUSCRIPT

High capacity polyethylenimine impregnated microtubes made of

carbon nanotubes for CO2 capture
**Laura Keller1,2 , **Burkhard Ohs1 , Jelena Lenhart1 , Lorenz Abduly1 , Philipp Blanke1 ,
*Matthias Wessling1,2
1

PT
RWTH Aachen University, Chemical Process Engineering, Forckenbeckstr. 51, 52074 Aachen, Germany
2
DWI - Leibniz Institute for Interactive Materials, Forckenbeckstr. 50, 52074 Aachen, Germany
** contributed equally to this work

RI
SC
Abstract

Anthropogenic carbon dioxide emissions cause global warming - one of the major challenges soci-

U
ety faces today. Carbon capture technologies can mitigate this development and stabilize the global
AN
average temperature. However, to reach their potential, energy efficient processes and new mate-
rials are required. In this work, we present a new type of hollow fiber sorbents. Highly porous
M
multi-walled carbon nanotube based (MWCNT) microtubes with a high internal surface area were
functionalized with polyethylenimine (PEI) via a wet impregnation method. Their tubular geom-
D

etry opens up the device design space to control gas distribution and residence time. For low PEI
fractions in the impregnation solution, the high surface area and porosity of the microtubes is pre-
TE

served, and the CO2 uptake increases with increasing PEI fractions. We observed a maximum CO2
uptake of 93 mg/gfiber at 20 wt.-% of PEI in the impregnation solution. At higher PEI loadings the
EP

CO2 uptake decreases. Due to the thermal and chemical stability of the PEI sorbent and the CNT
matrix, we anticipate the applicability of these microtubes for a wide range of separation processes
C

and operation conditions. Such tubular MWCNT geometries may serve as a scaffold platform to
immobilize a functional surface coating inside a controlled nanoporosity.
AC

1. Introduction

Globally, anthropogenic greenhouse gas emissions continue to increase. They were the highest
in human history from 2000 to 2010 and reached approximately 49 gigaton carbon dioxide equiv-

∗
Corresponding author. Tel: +49 241 80 95470. E-mail: Manuscripts.CVT@avt.rwth-aachen.de

Preprint submitted to Elsevier October 6, 2017

 ACCEPTED MANUSCRIPT

alent in 2010 [1]. In 2014, almost one-third of the world’s primary energy supply was provided by
coal while it accounted for almost half of the 32 Gt CO2 emissions by fuel combustion [2]. The en-
ergy sector is expected to continue to be the primary source of greenhouse gas and CO2 emissions
[1]. Mitigation strategies for greenhouse gas emissions are necessary to hold the global average
below +2°C of pre-industrial levels. Since especially CO2 emissions from the energy supply sector

PT
and the share of coal in the global fuel mix continue to increase [1] measures like the separation
and storage of carbon dioxide become indispensable [1, 3]. Carbon capture and storage (CCS)

RI
is considered a direct emissions mitigation option that buys enough time (around 50 years) for a
transition away from fossil fuels [4].

SC
Carbon dioxide can be separated from flue gas streams using sorption processes. Only absorp-

U
tion processes based on aqueous amine solutions are commercially used, but due to cost reasons
AN
mainly for CO2 separation from natural gas [5, 6]. Additional to high costs other disadvantages
of liquid amine absorption exist, e.g. equipment corrosion or degradation of the solvent [5–8].
Consequently, adsorption processes are considered as an alternative. Solid adsorbents like zeolites
M
[9, 10], metal-organic frameworks (MOF) [11, 12] or different amine-functionalized solid sorbents
(e.g. silica [13, 14], nanofibrillated cellulose [15], fly-ash or monolithic porous carbon [16–18], or
D

polymers [19]) are being investigated extensively. MOFs and zeolites are hydrophilic and thus the
TE

presence of water in flue gases leads to a decreased CO2 uptake [20]. Additionally, a large pressure
and/or temperature gradient is needed to ensure near complete desorption of CO2 and to achieve an
EP

efficient process performance [7]. Solid sorbents that consist of a supported polymeric amine can
have high sorption capacities, fast ad- and desorption cycles and low energy consumption [6, 7].
C

Due to their high selectivity towards CO2 in mixtures with nitrogen, methane, oxygen or hydrogen
even at low CO2 concentrations [21, 22], solid amines are particularly promising for carbon capture
AC

from low pressure streams such as flue gas or ambient air. Contrary to MOFs and zeolites, these sor-
bents exhibit an increased CO2 sorption capacity when water is present in the gas stream [15, 23].
Branched polyethylenimine (PEI) is a common choice because its branched chains have many CO2
capturing amino groups [24] and it thus offers high CO2 uptake capacities per molecule [25]. The
main reaction occurring between PEI and CO2 is believed to be the formation of a carbamate ion
[26]. Similar to absorption of CO2 to aqueous amine solutions, this formation is proposed to be a

2
 ACCEPTED MANUSCRIPT

two-stage reaction: two moles of amine react with one mole of CO2 via a zwitterion mechanism
which is followed by base catalyzed deprotonation of said zwitterion via a second amine group
[27–29]:
−−−*
CO2 + 2 R2 NH ) +
−− R2 NH2 + R2 NCOO
−
(1)

PT
−−*
CO2 + 2 R3 N −
)
+
−− R4 N + R2 NCOO
−
(2)

Typically, solid sorbents are regenerated using low pressures or high temperatures resulting in

RI
processes like pressure swing adsorption, temperature swing adsorption or vacuum swing adsorp-
tion [30]. In these processes, the gas mixtures usually flow through an adsorption column, i.e. a

SC
packed bed reactor, where the sorbent is present in the form of pellets or granules in the size range
of several millimeters. A high pressure drop resulting from high gas flow rates at decreasing pellet

U
diameter leads to high energy requirements. Furthermore, packed bed adsorption reactors suffer
from a significant mass transfer resistance and undefined flow conditions resulting in reduced prod-
AN
uct purities and/or recoveries [31]. The delicate balance between desired high amine loading and
depressed adsorption kinetics has recently been reported [32].
M
D
TE
C EP
AC

Figure 1: (a) Photo of CNT microtubes with a coin for size reference, (b) schematic representation of a hollow fiber
module with CNT microtubes, (c) photo of parallelized CNT microtubes, (d) schematic representation of a packed bed
with short pieces of CNT microtubes

3
 ACCEPTED MANUSCRIPT

So-called hollow fiber sorbents are a new modular solution to these limitations. They poten-
tially outperform granular structures due to high geometric surface areas, faster cycle times and
high packing densities [31]. Hollow fiber sorbents were first introduced by Feng et al. [33], who
showed their potential resulting from lower pressure drops and a lower mass transfer resistance.
Additionally, a better flow distribution leads to sharper breakthrough curves and thus higher re-

PT
coveries. Lively et al. [34] dispersed adsorption particles inside of a hollow fiber consisting of a
polymeric matrix. These fibers are produced by fiber spinning technology using both a polymer

RI
binder as well as high loadings of a sorbent filler such as zeolites. An additional sheath layer can
be formed by a simultaneous co-extrusion process. Low permeability polymers, such as PVDC,

SC
are used to prevent a loss of the gas mixture to the shell or the bore side of the hollow fiber sor-
bent [35]. Finally, a temperature moderator can be used to thermally balance the heat of adsorption

U
and desorption [36].
AN
The impregnation of a functional liquid on a highly porous support structure with a high inter-
nal surface area offers several advantages: the high surface area ensures fast adsorption while the
M
thin active adsorption layer leads to a lower mass transfer resistance. Amines have already been
immobilized in hollow fibers to increase their sorption capacity [20, 37]. However, these systems
D

always comprise a polymer matrix with a) a low internal surface area, b) low adsorption capacity,
TE

c) low thermal as well as chemical stability and d) limited thermal conductivity for temperature
control. To overcome these limitations and exploit the advantages mentioned before, a different
EP

support structure is needed.

C

We present the first microtubular sorbents based on carbon nanotubes (CNT) (see Figure 1(a)).
The CNT form a strong and flexible structure [38], very similar to the flat equivalent (so-called
AC

buckypapers). Due to this no polymeric matrix with low sorption capacities and a low surface
area, as in state-of-the-art hollow fiber sorbents, is required. CNT and thus the microtubes exhibit
a high thermal and chemical stability. A functional liquid, in this work polyethylenimine, can be
immobilized into the highly porous fiber wall to increase its sorption capacities. Bae et al. [39]
proposed the interactions between CNT and PEI to be hydrogen bonding and weak electrostatic
interactions. They determined the interactions to be spontaneous and stable under conventional

4
 ACCEPTED MANUSCRIPT

experimental conditions [39]. Different studies have shown good CO2 adsorption capacities with
amine-functionalized CNTs, using multi-walled CNT as well as single-walled CNT [40–42]. Re-
cently, a high sorption capacity of modified multi-walled CNT impregnated with tetraethylenepen-
tamine of 5 mmol/g (10 vol.-% CO2 in N2 , 1 vol.-% H2 O) was achieved [43]. Due to the lower cost
of multi-walled CNT and potentially similar sorption capacities, they were chosen for this work.

PT
In our amine impregnated hollow fibers, the small size of the CNTs ensures fast heat and mass
transfer, which are essential for efficient swing sorption processes. The CNT microtubes can be

RI
assembled into hollow fiber modules, see Figure 1(b), and when coated with a gas impermeable
layer a temperature moderator can be used on the shell side. Figure 1(c) shows a lab scale module

SC
of several microtubes. Additionally, shorter pieces of microtubes can potentially be arranged into
packed beds (Figure 1(d)). Thus, the shape of the microtubes (e.g. length, diameter) can be tailored

U
as well as the material itself can be functionalized and tailored to a specific application. In this
AN
work, the properties of the PEI-impregnated microtubes based on CNT, such as specific surface
area and the CO2 sorption capacity, are reported.
M
2. Experimental
D

2.1. Preparation of PEI-loaded CNT microtubes

As described in a previous study, we fabricated microtubes by filtrating a CNT suspension

TE

through a microfiltration hollow fiber membrane in dead-end filtration mode [44]. Suspensions of
multi-walled carbon nanotubes (MWCNT; outer diameter of 6-9 nm and 5 µm length; purity >95%,
EP

Sigma-Aldrich) were prepared by mixing 1 g of CNT, 10 g of Triton-X 100 (Sigma-Aldrich) and

1 L of deionized water (18 mOhm). The solution was magnetically stirred for 30 min and subse-
C

quently sonicated for 3 h (UP200S, Hielscher). For the production of the microtubes, one end of
AC

a polypropylene microfiltration hollow fiber membrane (PP S6/2, Accurel; 1.8 ± 0.15 mm inside
diameter and 0.45 ± 0.05 mm wall thickness) of 20 cm length was sealed with adhesive glue. The
CNT suspension was supplied at a constant pressure of 5 bar into the open end of the membrane us-
ing a pressure vessel. A homogeneous CNT layer was then formed by filtrating the CNT suspension
through the hollow fiber in inside-out mode. Next, 50 mL isopropanol (98% purity, Applichem)
were filtrated through the hollow fiber covered with CNT (inside-out) to remove the surfactant.
Due to shrinking of the CNT network formed on the bore side of the hollow fiber membrane, the

5
 ACCEPTED MANUSCRIPT

microtubes could easily be removed from the hollow fiber support after drying overnight at 30 °C
under vacuum (90 mbar).

The free-standing microtubes were used to form microtubular sorbents through the immobi-
lization of a polymeric amine, polyethylenimine (PEI). The PEI was loaded on the CNT hollow

PT
fibers using a common wet impregnation method [24]. The microtube was immersed for 6 hours
in a PEI/ethanol impregnation solution (PEI: Sigma-Aldrich, MW ≈ 800, ethanol: Overlack, pu-

RI
rity > 98.8 vol.-%) with 2.5, 5, 10, 20, and 30 wt.-% of PEI, respectively. The impregnation
solution was magnetically stirred for a homogeneous immobilization of the PEI. Afterwards, the

SC
PEI-impregnated CNT microtubes were dried at 30 °C under vacuum (90 mbar) for a minimum of
12 h to evaporate the remaining ethanol. Since boiling point of pure ethanol at 90 mbar is approxi-

U
mately 27 °C [45] and to ascertain that no water was present in the sorbent, additional drying steps
AN
were performed as described below (see section 2.2 and 2.3) before any analysis was conducted.

2.2. Adsorbent characterization

M
The BET specific surface area was determined using nitrogen adsorption and desorption at 77 K
(ASAP 2020, Micromeritics) [46]. All samples were degassed under vacuum (5 µmHg) at 80 °C
D

for 24 h. Thommes et al. [46] recommend the use of density functional theory based computa-
TE

tional procedures to assess a reasonably reliable pore size distribution of micro- and mesoporous
materials. Thus, pore size distribution and volume were calculated by applying non-local-density
functional theory (NLDFT) [47]. Ultra-high resolution images of the utilized carbon nanotubes in
EP

powder form were produced using brightfield scanning transmission electron microscopy (SU9000,
Hitachi). Samples were prepared as follows: the CNT suspension described above was diluted to a
C

CNT concentration of 0.2 g/L with deionized water and applied to a sample grid (CF300-Cu, 300
AC

mesh, copper with carbon film, Electron Microscopy Sciences). High-resolution images of the mi-
crotubes were produced using a scanning electron microscope (S-4800, Hitachi). All samples were
sputtered with a gold layer of approximately 3 nm thickness using a high vacuum sputter coater
(EM ACE600, Leica). The behavior of the hollow fiber adsorbent impregnated with 20 wt.-% PEI
in the impregnation solution over a range of temperatures (30 °C to 700 °C) was analyzed using
thermogravimetric analysis (TGA STA6000, Perkin Elmer). The heating rate was 10 K/min and
the measurement was conducted under a flow of 20 mL/min of nitrogen.

6
 ACCEPTED MANUSCRIPT

2.3. CO2 adsorption

Pure CO2 adsorption isotherms were measured at ambient temperature using pycnometry
(ASAP 2020, Micromeritics) for pressures ranging from 0.004 to 1 bar. For selectivity calculation a
pure N2 adsorption isotherm was recorded with a sorbent sample impregnated with 20 wt.-% PEI in

PT
ethanol using pycnometry. Degassing for pycnometry measurements was conducted under vacuum
(5 µmHg) at 80°C for 12 h. The equilibrium criterion was pn −pn+1 1
pn ∆t
< 4.5 · 10−7 1s . The CO2 uptake
after 45 minutes sorption time at 0.15 v.-% CO2 (mixed with N2 ; gas flow: 150 mL/min), 30 °C and

RI
ambient pressure was measured using thermogravimetric analysis (TGA STA6000, Perkin Elmer).

SC
Additionally, the kinetics of CO2 uptake with pure CO2 (150 mL/min) at 30 °C and 80°C at am-
bient pressure were recorded. This measurements enables to study the adsorption and desorption
behavior of the sorbent over time and at different temperatures. For TGA analysis the sample was

U
degassed for 2 h at 110 °C under a flow of 150 mL/min of pure nitrogen. Between adsorption steps
AN
of different temperatures, the sample was regenerated at degas conditions for 90 min.

2.4. Cyclic CO2 adsorption measurements

M
Cyclic adsorption measurements were conducted using thermogravimetric analysis (TGA
STA6000, Perkin Elmer) to analyze the regeneration performance of the fibers. Prior to the adsorp-
D

tion measurements, the PEI-impregnated fibers were degassed as described above (section 2.3).
TE

Then, the fibers were cooled down to 30 °C under nitrogen flow (150 mL/min, step 1). Subse-
quently, the CO2 adsorption step was conducted at a gas flow of 150 mL/min with 0.15 v.-% CO2
EP

(mixed with N2 ) for ten minutes (step 2). After the adsorption step, the sample was heated to 90 °C
under a nitrogen gas stream (150 mL/min, step 3) and held at 90 °C under the same gas flow for
20 min (step 4) to regenerate the sample. These four steps were repeated ten times.
C
AC

3. Results and discussion

3.1. Adsorbent characterization

The mass fraction of PEI loaded onto the microtubes using the wet impregnation method is rep-
resented in Figure 2. A linear increase of the mass fraction from 22% at 2.5 wt.-% PEI in ethanol to
196% at 30 wt.-% PEI is observed. TGA analysis of a sample impregnated with 20 wt.-% reveals a
weight loss of the samples from 30 °C to about 350 °C (see Figure S1). At low temperatures, up to

7
 ACCEPTED MANUSCRIPT

2.2

2.0

1.8

)
CNT
1.6

/g
PEI
PEI load on microtube (g
1.4

PT
1.2

1.0

RI
0.8

0.6

0.4

SC
0.2

0.0
0 5 10 15 20 25 30

PEI in ethanol (wt.-%)

U
Figure 2: Influence of the PEI fraction in the impregnation solution on the PEI load on the microtubes (error bars
represent the standard error of the mean)
AN
M
125 °C, the weight loss can be attributed to the removal of physi- and chemisorbed water molecules
[14, 48] and ethanol remaining from the impregnation step. In this first region, the weight decreases
D

by approximately 26%. This emphasizes the need for an efficient drying step before any adsorption
measurement. The second region of weight loss (about 38%) up to about 350 °C is predominantly
TE

caused by the volatilization and decomposition of PEI molecules [14]. Thus, the temperature for
regeneration of the sorbent should not exceed 125 °C.
EP

Ultra-high resolution pictures of the utilized CNT in powder form indicate that closed caps are
C

prevalent (see Figure S2). Thus, the inside of the CNT is not available as sorption sites. High-
resolution SEM pictures of the different PEI loadings emphasize that the morphology and thus the
AC

surface area of the PEI-impregnated CNT microtubes depend significantly on the PEI load. Figure 3
shows SEM pictures of pure as well as PEI-loaded CNT microtubes with 5 wt.-% and 30 wt.-%
PEI in the impregnation solution. At low PEI loadings, the high surface network structure of the
CNT microtubes is still present. However, at higher PEI loadings the PEI fills much of the voids
between the CNT and less of the formerly highly porous CNT network can be distinguished.

8
 ACCEPTED MANUSCRIPT

The pure CNT microtubes have a specific BET surface area of 203.6 m2 /g (see Table 1 and
Figure S3). The specific BET surface area of the PEI-loaded microtubes decreases with an increas-
ing PEI fraction in the impregnation solution. At 2.5 wt.-% and 5 wt.-% PEI in the impregnation
solution the specific surface area reduces from 203.6 m2 /g to 152.2 m2 /g and 139.8 m2 /g, respec-
tively. In comparison, the specific surface area decreases significantly at a PEI fraction of 10 wt.-%

PT
(75.4 m2 /g) and reaches a very low value of 4.7 m2 /g at PEI fractions of 30 wt.-%. Overall, the
total pore area decreases with increasing PEI load (cf. Table 1). Since the utilized CNT exhibit

RI
closed caps, the measured pores in the microtubes are the voids or textural pores between the CNT
in the hollow fiber wall. The PEI adsorbs to and thus covers the surface of the CNTs. This leads to

SC
increased outer diameters of the carbon nanotubes. At higher PEI loads the amount and thickness
of the surface coverage increases which leads to reduced pores or voids between the CNTs. This is

U
also visible in the SEM pictures (cf. Figure 3). For low PEI fractions (2.5 to 5 wt.-%), the total pore
AN
area increases slightly with increasing PEI content. Both are rather dilute PEI solutions with close-
lying weight fractions that could lead to similar properties of the impregnated sorbent. The pore
size distribution of a pure CNT microtube and CNT microtubes impregnated with 5 and 20 wt.-%
M
PEI in ethanol can be seen in Figure 4. The majority of the pore volume can be attributed to pore
widths of 25 to 60 nm. The decrease in pore volume with increasing PEI fraction is visible here,
D

too. Additionally, pore widths of 2 and 3 nm appear at 5 wt.-% PEI in the impregnation solution.
TE

The PEI is believed to coat the carbon nanotubes and thus narrow the voids between the CNT. This
would lead to a shift of pore widths to lower values.
EP

3.2. CO2 adsorption

C

The CO2 sorption isotherms at ambient temperature of pure CNT microtubes as well as micro-
AC

tubes with 2.5 wt-% and 10 wt.-% PEI in the impregnation solution are shown in Figure 5. The pure
microtubes show a linear increase of the sorption capacity with CO2 pressure. The total capacity at
1 bar is rather low with 8.3 mg/g. In comparison, the PEI-impregnated hollow fibers show a high
sorption capacity of 19.9 mg/g and 47 mg/g at 2.5 wt-% and 10 wt.-% PEI, respectively, at very low
CO2 pressures of 0.0035 bar or 350 ppm. For both PEI-impregnated microtubes, there is a sharp
increase in CO2 sorption capacity at very low pressures up to approximately 0.03 bar. This sharp
rise is followed by a linear increase of the amount of adsorbed CO2 to maximum values at 1 bar of

9
 ACCEPTED MANUSCRIPT

Table 1: BET surface area, total pore volume and total pore area calculated with NLDFT

PEI fraction in
BET surface
the impregna- Total pore area
area
tion solution

PT
wt.-% m2 /g m2 /g
0 203.6 146

RI
2.5 152.2 85.5
5 139.8 90.5

SC
10 75.4 43.1
20 34.7 23.4

U
30 4.7 1.65

AN
45.8 mg/g and 80.4 mg/g for 2.5 wt-% and 10 wt.-% PEI, respectively. These results show that the
PEI impregnation enhances the CO2 sorption capacity of the CNT microtubes and that the amount
M
of PEI in the impregnation solution influences the CO2 uptake. Further, sorbent regeneration by
lowering the pressure does not seem beneficial since high vacuum would be needed leading to high
D

costs. However, the high CO2 load at very low pressure suggests the applicability for CO2 capture
from air where it is currently present with about 400 ppm [49].
TE

To determine the selectivity of the PEI-impregnated CNT hollow fibers for CO2 over N2 a nitro-
EP

gen adsorption measurement was performed at 25°C using pycnometry (see Figure S5 for sorption
isotherms). As presented in other works on PEI-impregnated sorbents [50, 51], the selectivity was
C

calculated using the ideal adsorption solution theory [52]:

AC

qCO2 /qN2
S = (3)
pCO2 /pN2

Thus, the mole of CO2 adsorbed are divided by the mole of N2 adsorbed and normalized by the
corresponding partial pressures. For the CNT-PEI sorbent with 20 wt.-% PEI in the impregnation
solution and an assumed gas mixture of 15 vol.-% CO2 in N2 a selectivity of 196 was calculated.
For the application in post-combustion CO2 capture, the selectivity for CO2 over N2 is the most
relevant since these two gases are mainly present. Other possible applications include the upgrad-

10
 ACCEPTED MANUSCRIPT

ing of biogas or natural gas. Promising CO2 /CH4 selectivities have been reported for PEI on an
active carbon substrate (24) [53] and metal-organic frameworks impregnated with PEI (251) [54].
In future work, the selectivity of the PEI-impregnated CNT sorbents over CH4 will be identified.

Many isothermal models, such as Langmuir, Temkin, Freundlich, or Toth model, have been

PT
used to describe the adsorption behavior of amine-functionalized materials [55, 56]. The Freundlich
isotherm model:

RI
qe = K · pCO
1/n
2
(4)

SC
can be used to describe non-ideal adsorption on heterogeneous surfaces as well as multi-layer
sorption [57, 58]. It can also be applied for CO2 adsorption on amine-functionalized materials
[55]. Thus, we fitted our experimental data using the Freundlich isotherm model. The respective

U
fits are presented in Figure 5 and the model parameters are summarized in Table 2. Clearly, the
AN
Freundlich model is able to describe the adsorption isotherms precisely and the coefficient of de-
termination R2 is near unity.
M
Table 2: Freundlich isotherm constants
D

Parameter Unit 10 wt.-%

TE

2.5 wt.-%

KF [mg/g/bar1−1/n ] 42.35 76.25

n [-] 6.85 10.74
EP

R2 [-] 0.986 0.994

C

The PEI fraction in the impregnation solution is an important parameter for the fabrication
AC

process since it determines the amount of PEI immobilized on the fibers (cf. section 3.1). The
adsorption capacities of fibers with different PEI loads at 0.15 bar and ambient temperature, mea-
sured with TGA and pycnometry are presented in Figure 6. First, there is an increase in CO2 uptake
with increasing PEI load. The highest CO2 uptake is measured with TGA for 10 wt.-% PEI in the
impregnation solution: 65.8 mg/gfiber (TGA). For pycnometry, 20 wt.-% PEI in the impregnation
solution yield the highest CO2 uptake of 93.4 mg/gfiber . After this maximum, the adsorption ca-
pacity decreases to values as low as 13.4 mg/gfiber (for 30 wt.-% PEI in the impregnation solution,

11
 ACCEPTED MANUSCRIPT

measured with TGA). Differences in the measured CO2 uptake are believed to result from different
equilibrium times of the two measuring techniques. For pycnometry, the equilibrium time could
not be set but rather an equilibrium condition had to be met. This led to high sorption times of 35
to 65 hours. In contrast, the sorption time during TGA adsorption measurements was 45 minutes.
For high PEI loads, the difference in CO2 uptake measured by both techniques becomes more sig-

PT
nificant. This indicates that the equilibrium time has an increasing influence on the CO2 uptake
with an increasing PEI load. Both measuring techniques reveal the maximum in sorption capacity

RI
for a certain PEI load (10 wt.% for TGA and 20 wt.-% for pycnometry). This phenomenon has
been observed previously for different materials with immobilized PEI [20, 59]. Depending on the

SC
amine loading, the CO2 sorption onto PEI is believed to depend on several overlapping steps such
as film diffusion, intra-particle diffusion and adsorbate interaction with active sites (physisorption

U
and/or chemisorption with reactions of higher order) [59–62]. The observed maximum thus in-
AN
dicates a trade-off between high PEI content (i.e. available amine groups) and available internal
surface area or pores for diffusion. The CO2 sorption capacity increases up to 10 wt.-% (for TGA)
or 20 wt.% (for pynometry) PEI in the impregnation solution probably because more amine groups
M
are available for sorption. At higher PEI loads the PEI blocks the pores of the microtubes. PEI is
thus present in thicker layers blocking the free internal porosity of the MWCNT network imparting
D

the access of CO2 to the inner PEI coating layers and rendering them ineffective for CO2 sorption.
TE

The sorption isotherms (cf. Figure 5) indicated that a regeneration of the sorbent through a
EP

change in pressure would not be beneficial. Another possibility to desorb the CO2 is the use of a
higher temperature. In this work, the CNT-PEI sorbents could be fully regenerated under nitrogen
C

gas flow at elevated temperatures (cf. Figure 8), which enables the use of the full CO2 sorption
capacity shown in Figure 6. However, this does not lead to the desired separation. At 80°C the
AC

CO2 sorption capacity decreases as shown in Figure 7. Thus by increasing the temperature, about
60% of the adsorbed CO2 could be desorbed using pure CO2 as a purge gas. Higher temperatures
may lead to an increase in desorption efficiency, but a degeneration of PEI above 135°C has been
shown [29]. Steam has been proposed as a purge gas for efficient regeneration that does not impair
the sorbent [29, 63] and would have to be investigated in future studies. Furthermore, the sorption
behavior over time reveals that at 30°C for both 2.5 wt.-% and 5 wt.-% PEI in the impregnation

12
 ACCEPTED MANUSCRIPT

solution 50% of the final adsorption capacity are reached after 20 s. This indicates the feasibility
of short cycle times with the new microtube sorbent material.

Finally, we investigated the cyclic adsorption behavior of the PEI-impregnated CNT-based mi-
crotubes (cf. Figure 8). First, CO2 was adsorbed at 30 °C and a partial pressure of 0.15 bar.

PT
Subsequently, CO2 was desorbed under pure nitrogen at 110 °C. The adsorbed CO2 could fully
be desorbed within the desorption time of 20 min. The results also show that shorter cycle times

RI
than the tested ones are possible to increase the throughput of the process. However, the adsorption
capacity of the microtubes decreases about 10 % within ten cycles. This phenomenon has also been

SC
observed elsewhere, but can be mitigated if humidified gas streams are treated [7]. Furthermore,
the microtube weight loss over ten cycles is around 1%. This can be due to thermal instability of

U
the PEI or due water desorption, which was adsorbed on the microtubes in-between measurements.
AN
The presented work is not only significant for CO2 capture applications but may be beneficial
as a general scaffold for immobilization and functionalization applications:
M

• The inner porosity of microtubes can be functionalized with any adhesive or film-forming
D

non-volatile material.
TE

• The coated material may yet host other functional molecular species having a very specific
molecular recognition or conversion functionality.
EP

• The geometry of the MWCNT microtubes (length, diameter, wall thickness) can easily be
tailored to the application and the properties of its functional inner surface coating.
C

• Swift and controlled heating of the system may be achieved through the electrical conduction
AC

of the CNTs.

• Electronic activation of physico-chemical processes can be induced into the inner functional
coating layer through the CNT surface.

13
 ACCEPTED MANUSCRIPT

4. Conclusion

In this work, PEI-impregnated hollow fiber sorbents made of CNT are reported for the first
time. The production procedure proved to be a straight-forward way to make novel sorbents with
high CO2 sorption capacities. An optimal PEI load at 20 wt.-% of PEI in the wetting solution with

PT
a maximum in CO2 uptake of 93 mg/gfiber was identified. An increasing PEI load leads at the same
time to an increase in available amine groups and to a decreased internal surface area and pore size.

RI
We assume that a trade-off between the number of available amine groups and available internal
surface leads to this observed optimum. The CNT-PEI hollow fiber sorbents can be applied in the

SC
emerging field of direct air capture due to high amounts of adsorbed CO2 (47 mg/gfiber for 10 wt.-%
PEI in the impregnation solution) at very low partial pressures of 0.0035 bar. Regeneration using
pure CO2 as a purge gas at elevated temperatures of 80 °C was shown to be feasible and led to

U
the desorption of 60% of the adsorbed CO2 . However, this could be improved by using steam
AN
or vacuum-enhanced desorption at elevated temperatures. Cyclic adsorption measurements have
shown some loss of sorbent and decrease in CO2 uptake of 10% over ten cycles. In the future it
M
needs to be determined whether this decrease is prevented when humidified gas streams are treated.
Overall, these hollow fibers are expected to be highly effective in different separation processes and
D

under varying operation conditions due to their tubular shape, the thermal and chemical stability of
TE

both the PEI sorbent and the CNT support as well as the high CO2 sorption capacity.

Acknowledgement
EP

M.W. acknowledges the support through an Alexander-von-Humboldt Professorship. This work

was performed in part at the Center for Chemical Polymer Technology CPT, which is supported by
C

the EU and the federal state of North Rhine-Westphalia (grant no. EFRE 30 00 883 02).
AC

References

[1] IPCC, Climate Change 2014: Mitigation of Climate Change. Contribution of Working Group
III to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change, Cam-
bridge University Press, Cambridge, UK and New York, NY, USA, 2014.

[2] International Energy Agency, CO2 emissions from fuel combustion 2016 - Highlights (2016).

14
 ACCEPTED MANUSCRIPT

[3] International Energy Agency, 20 years of carbon capture and storage. Accelerating future
deployment (2016).

[4] R. S. Haszeldine, Carbon capture and storage: How Green Can Black Be?, Science 325 (May)
(2009) 1647–1652.

PT
[5] D. Aaron, C. Tsouris, Separation of CO2 from Flue Gas: A Review, Separation Science and
Technology 40 (1) (2005) 321–348.

RI
[6] Z. Sun, M. Fan, M. Argyle, Supported Monoethanolamine for CO2 Separation, Ind. Eng.

SC
Chem. Res 50 (2011) 11343–11349.

[7] A. Sayari, Y. Belmabkhout, Stabilization of amine-containing CO2 adsorbents: Dramatic ef-

U
fect of water vapor, Journal of the American Chemical Society 132 (18) (2010) 6312–6314.

AN
[8] E. Favre, Carbon dioxide recovery from post-combustion processes: Can gas permeation
membranes compete with absorption?, Journal of Membrane Science 294 (1-2) (2007) 50–
M
59.

[9] V. G. Gomes, K. W. Yee, Pressure swing adsorption for carbon dioxide sequestration from
D

exhaust gases, Separation and Purification Technology 28 (2) (2002) 161–171.

TE

[10] T.-H. Bae, M. R. Hudson, J. A. Mason, W. L. Queen, J. J. Dutton, K. Sumida, et al., Evaluation
of cation-exchanged zeolite adsorbents for post-combustion carbon dioxide capture, Energy
EP

Environ. Sci. 6 (1) (2013) 128–138.

[11] J. A. Mason, K. Sumida, Z. R. Herm, R. Krishna, J. R. Long, Evaluating metalorganic frame-

C

works for post-combustion carbon dioxide capture via temperature swing adsorption, Energy
AC

& Environmental Science 4 (8) (2011) 3030.

[12] K. Sumida, D. L. Rogow, J. A. Mason, T. M. McDonald, E. D. Bloch, Z. R. Herm, et al.,

Carbon Dioxide Capture in Metal-Organic Frameworks, Chemical Reviews (Washington, DC,
United States) 112 (2) (2012) 724–781.

15
 ACCEPTED MANUSCRIPT

[13] R. Sanz, G. Calleja, A. Arencibia, E. S. Sanz-Pérez, CO2 adsorption on branched

polyethyleneimine-impregnated mesoporous silica SBA-15, Applied Surface Science
256 (17) (2010) 5323–5328.

[14] K. Li, J. Jiang, S. Tian, F. Yan, X. Chen, Polyethyleneimine-nano silica composites: a low-

PT
cost and promising adsorbent for CO2 capture, J. Mater. Chem. A 3 (5) (2015) 2166–2175.

[15] C. Gebald, J. A. Wurzbacher, P. Tingaut, T. Zimmermann, A. Steinfeld, Amine-Based

RI
Nanofibrillated Cellulose As Adsorbent for CO2 Capture from Air, Environmental Science
& Technology 45 (20) (2011) 9101–9108.

SC
[16] A. Arenillas, K. Smith, T. Drage, C. Snape, CO2 capture using some fly ash-derived carbon
materials, Fuel 84 (17) (2005) 2204–2210.

U
AN
[17] M. L. Gray, Y. Soong, K. J. Champagne, J. Baltrus, R. W. Stevens, P. Toochinda, et al., CO2
capture by amine-enriched fly ash carbon sorbents, Separation and Purification Technology
35 (1) (2004) 31–36.
M
[18] L. Estevez, R. Dua, N. Bhandari, A. Ramanujapuram, P. Wang, E. P. Giannelis, A facile
D

approach for the synthesis of monolithic hierarchical porous carbons high performance ma-
terials for amine based CO2 capture and supercapacitor electrode, Energy & Environmental
TE

Science 6 (6) (2013) 1785.

EP

[19] M. L. Gray, J. S. Hoffman, D. C. Hreha, D. J. Fauth, S. W. Hedges, K. J. Champagne, et al.,

Parametric study of solid amine sorbents for the capture of carbon dioxide, Energy and Fuels
23 (10) (2009) 4840–4844.
C
AC

[20] Y. Labreche, R. P. Lively, F. Rezaei, G. Chen, C. W. Jones, W. J. Koros, Post-spinning infusion

of poly(ethyleneimine) into polymer/silica hollow fiber sorbents for carbon dioxide capture,
Chemical Engineering Journal 221 (2013) 166–175.

[21] Y. Belmabkhout, R. Serna-Guerrero, A. Sayari, Adsorption of CO2 -containing gas mixtures

over amine-bearing pore-expanded mcm-41 silica: Application for gas purification, Industrial
& Engineering Chemistry Research 49 (1) (2010) 359–365.

16
 ACCEPTED MANUSCRIPT

[22] Y. Belmabkhout, A. Sayari, Effect of pore expansion and amine functionalization of meso-
porous silica on CO2 adsorption over a wide range of conditions, Adsorption 15 (3) (2009)
318–328.

[23] P. J. E. Harlick, A. Sayari, Applications of pore-expanded mesoporous silica. 5. triamine

PT
grafted material with exceptional CO2 dynamic and equilibrium adsorption performance, In-
dustrial and Engineering Chemistry Research 46 (2) (2007) 446–458.

RI
[24] X. Xu, C. Song, J. M. Andrésen, B. G. Miller, A. W. Scaroni, Preparation and characteri-
zation of novel CO2 ”molecular basket” adsorbents based on polymer-modified mesoporous

SC
molecular sieve MCM-41 , Microporous and Mesoporous Materials 62 (1–2) (2003) 29 – 45.

[25] B. Dutcher, M. Fan, A. G. Russell, Amine-Based CO2 Capture Technology Development

U
from the Beginning of 2013 A Review, ACS Applied Materials & Interfaces 7 (4) (2015)
2137–2148. AN
[26] S. Satyapal, T. Filburn, J. Trela, J. Strange, Performance and properties of a solid amine
M
sorbent for carbon dioxide removal in space life support applications, Energy and Fuels 15 (2)
(2001) 250–255.
D

[27] M. Caplow, Kinetics of Carbamate Formation and Breakdown, Journal of the American
TE

Chemical Society 90 (24) (1968) 6795–6803.

[28] A. K. Saha, S. S. Bandyopadhyay, A. K. Biswas, Kinetics of absorption of CO2 into aque-

EP

ous solutions of 2-amino-2-methyl-1-propanol, Chemical Engineering Science 50 (22) (1995)

3587–3598.
C

[29] T. C. Drage, A. Arenillas, K. M. Smith, C. E. Snape, Thermal stability of polyethylenimine

AC

based carbon dioxide adsorbents and its influence on selection of regeneration strategies, Mi-
croporous and Mesoporous Materials 116 (1-3) (2008) 504–512.

[30] H.-J. Bart, U. von Gemmingen, Adsorption, in: Ullmann’s Encyclopedia of Industrial Chem-
istry, Wiley-VCH Verlag GmbH & Co. KGaA, 2012, p. 30.

[31] F. Rezaei, P. Webley, Structured adsorbents in gas separation processes, Separation and Pu-
rification Technology 70 (3) (2010) 243 – 256.

17
 ACCEPTED MANUSCRIPT

[32] P. Bollini, N. A. Brunelli, S. A. Didas, C. W. Jones, Dynamics of CO2 Adsorption on Amine

Adsorbents. 2. Insights Into Adsorbent Design, Industrial & Engineering Chemistry Research
51 (46) (2012) 15153–15162.

[33] V. Feng, C. Y. Pan, C. W. McMinis, J. Ivory, D. Gosh, Hollow-fiber-based adsorbers for gas

PT
separation by pressure-swing adsorption, AIChE Journal 44 (7) (1998) 1555–1562.

[34] R. Lively, R. R. Chance, W. J. Koros, H. Deckman, B. T. Kelley, Sorbent fiber compositions

RI
and methods of temperature swing adsorption, US Patent 8,133,308 (2012).

SC
[35] D. Bhandari, K. O. Olanrewaju, N. Bessho, V. Breedveld, W. J. Koros, Dual layer hollow fiber
sorbents: Concept, fabrication and characterization, Separation and Purification Technology
104 (0) (2013) 68 – 80.

U
AN
[36] R. P. Lively, R. R. Chance, B. T. Kelley, H. W. Deckman, J. H. Drese, C. W. Jones, et al.,
Hollow fiber adsorbents for CO2 removal from flue gas, Industrial and Engineering Chemistry
Research 48 (15) (2009) 7314–7324.
M
[37] Y. Fan, R. P. Lively, Y. Labreche, F. Rezaei, W. J. Koros, C. W. Jones, Evaluation of CO2
D

adsorption dynamics of polymer/silica supported poly(ethylenimine) hollow fiber sorbents

in rapid temperature swing adsorption, International Journal of Greenhouse Gas Control 21
TE

(2014) 61 – 71.
EP

[38] L. Dumée, K. Sears, J. Schütz, N. Finn, M. Duke, S. Gray, A Preliminary Study on the Effect
of Macro Cavities Formation on Properties of Carbon Nanotube Bucky-Paper Composites,
Materials 4 (12) (2011) 553–561.
C
AC

[39] J. Bae, S. H. Cha, J. Park, A new polymeric binder for silicon-carbon nanotube composites in
lithium ion battery, Macromolecular Research 21 (7) (2013) 826–831.

[40] Q. Liu, Y. Shi, S. Zheng, L. Ning, Q. Ye, M. Tao, et al., Amine-functionalized low-cost
industrial grade multi-walled carbon nanotubes for the capture of carbon dioxide, Journal of
Energy Chemistry 23 (1) (2014) 111–118.

18
 ACCEPTED MANUSCRIPT

[41] E. P. Dillon, C. A. Crouse, A. R. Barron, Synthesis, characterization, and carbon dioxide

adsorption of covalently attached polyethyleneimine-functionalized single-wall carbon nan-
otubes, ACS Nano 2 (1) (2008) 156–164.

[42] F. Su, C. Lu, H. S. Chen, Adsorption, desorption, and thermodynamic studies of CO2 with

PT
high-amine-loaded multiwalled carbon nanotubes, Langmuir 27 (13) (2011) 8090–8098.

[43] M. Irani, A. T. Jacobson, K. A. Gasem, M. Fan, Modified carbon nan-

RI
otubes/tetraethylenepentamine for CO2 capture, Fuel 206 (2017) 10–18.

SC
[44] Y. Gendel, O. David, M. Wessling, Microtubes made of carbon nanotubes, Carbon 68 (0)
(2014) 818 – 820.

U
[45] http://ddbonline.ddbst.com/AntoineCalculation/AntoineCalculationCGI.exe, Vapor Pressure

AN
Calculation by Antoine Equation,accessed on Sep. 5, 2017.

[46] M. Thommes, K. Kaneko, A. V. Neimark, J. P. Olivier, F. Rodriguez-Reinoso, J. Rouquerol,

M
et al., Physisorption of gases, with special reference to the evaluation of surface area and pore
size distribution (IUPAC Technical Report), Pure and Applied Chemistry 87 (9-10) (2015)
D

1051–1069.
TE

[47] J. Landers, G. Y. Gor, A. V. Neimark, Density functional theory methods for characterization
of porous materials, Colloids and Surfaces A: Physicochemical and Engineering Aspects 437
EP

(2013) 3–32.

[48] S. Loganathan, M. Tikmani, A. K. Ghoshal, Novel pore-expanded MCM-41 for CO2 capture:
C

Synthesis and characterization, Langmuir 29 (10) (2013) 3491–3499.

AC

[49] https://www.esrl.noaa.gov/gmd/ccgg/trends/global.html, ESRL Global Monitoring Division -

Global Greenhouse Gas Reference Network, accessed on Sep. 27, 2017.

[50] Y. Lin, Q. Yan, C. Kong, L. Chen, Polyethyleneimine Incorporated Metal-Organic Frame-

works Adsorbent for Highly Selective CO2 Capture, Scientific Reports 3 (1) (2013) 1859.

19
 ACCEPTED MANUSCRIPT

[51] F. Liu, K. Huang, C. J. Yoo, C. Okonkwo, D. J. Tao, C. W. Jones, et al., Facilely synthe-
sized meso-macroporous polymer as support of poly(ethyleneimine) for highly efficient and
selective capture of CO2 , Chemical Engineering Journal 314 (2017) 466–476.

[52] A. L. Myers, J. M. Prausnitz, Thermodynamics of mixed-gas adsorption, AIChE Journal

PT
11 (1) (1965) 121–127.

[53] L. Pino, C. Italiano, A. Vita, C. Fabiano, V. Recupero, Sorbents with high efficiency for CO2

RI
capture based on amines-supported carbon for biogas upgrading, Journal of Environmental
Sciences (China) 48 (2016) 138–150.

SC
[54] S. Xian, Y. Wu, J. Wu, X. Wang, J. Xiao, Enhanced Dynamic CO2 Adsorption Capacity and
CO2 /CH4 Selectivity on Polyethylenimine-Impregnated UiO-66, Industrial and Engineering

U
Chemistry Research 54 (44) (2015) 11151–11158.
AN
[55] E. R. Monazam, L. J. Shadle, D. C. Miller, H. W. Pennline, D. J. Fauth, J. S. Hoffman, et al.,
Equilibrium and kinetics analysis of carbon dioxide capture using immobilized amine on a
M
mesoporous silica, AIChE Journal 59 (3) (2013) 923–935.
D

[56] R. Serna-Guerrero, Y. Belmabkhout, A. Sayari, Modeling CO2 adsorption on amine-

functionalized mesoporous silica: 1. a semi-empirical equilibrium model, Chemical Engi-
TE

neering Journal 161 (1) (2010) 173 – 181.

EP

[57] G. Crini, H. N. Peindy, F. Gimbert, C. Robert, Removal of c.i. basic green 4 (malachite green)
from aqueous solutions by adsorption using cyclodextrin-based adsorbent: Kinetic and equi-
librium studies, Separation and Purification Technology 53 (1) (2007) 97 – 110.
C
AC

[58] M. Özacar, İ. A. Şengil, Adsorption of reactive dyes on calcined alunite from aqueous solu-
tions, Journal of Hazardous Materials 98 (1) (2003) 211 – 224.

[59] Z. Chen, S. Deng, H. Wei, B. Wang, J. Huang, G. Yu, Polyethylenimine-Impregnated Resin

for High CO2 Adsorption: An Efficient Adsorbent for CO2 Capture from Simulated Flue Gas
and Ambient Air, ACS Applied Materials & Interfaces 5 (15) (2013) 6937–6945.

20
 ACCEPTED MANUSCRIPT

[60] A. Heydari-Gorji, A. Sayari, CO2 capture on polyethylenimine-impregnated hydrophobic

mesoporous silica: Experimental and kinetic modeling, Chemical Engineering Journal 173 (1)
(2011) 72 – 79.

[61] P. Bollini, N. A. Brunelli, S. A. Didas, C. W. Jones, Dynamics of CO2 adsorption on amine

PT
adsorbents. 2. insights into adsorbent design, Industrial & Engineering Chemistry Research
51 (46) (2012) 15153–15162.

RI
[62] A. Sayari, Q. Liu, P. Mishra, Enhanced Adsorption Efficiency through Materials Design for
Direct Air Capture over Supported Polyethylenimine, ChemSusChem 9 (19) (2016) 2796–

SC
2803.

[63] W. Li, S. Choi, J. H. Drese, M. Hornbostel, G. Krishnan, P. M. Eisenberger, et al., Steam-

U
stripping for regeneration of supported amine-based CO2 adsorbents, ChemSusChem 3 (8)
(2010) 899–903.
AN
M
D
TE
C EP
AC

21
 ACCEPTED MANUSCRIPT

PT
RI
SC
(a)

U
AN
M
D
TE

(b)
C EP
AC

(c)

Figure 3: SEM images showing the morphology of (a) pure CNT-microtubes, (b) CNT-microtubes loaded with 5 wt.-%
PEI in impregnation solution and (c) CNT-microtubes loaded with 30 wt.-% PEI in impregnation solution
22
 ACCEPTED MANUSCRIPT

pure CNT microtube

7

dV/dlog(W) pore volume (cm³/g)

6

PT
5

RI
3 5 wt.-% PEI

SC
20 wt.-% PEI
1

0
1 10 100

U
Pore width (nm)

AN
Figure 4: Influence of the PEI fraction in the impregnation solution on the pore size distribution of the (PEI-
impregnated) microtubes.
M
D

80
TE
)

10 wt.-% PEI
fiber

60
uptake (mg/g

EP

40

2.5 wt.-% PEI

2
Specific CO
C

20
AC

0
pure CNT microtube

0.0 0.2 0.4 0.6 0.8 1.0 1.2

Pressure (bar)

Figure 5: Sorption isotherms at room temperature for pure as well as loaded microtubes with 2.5 and 10 wt.-% PEI in
the impregnation solution. The lines represents the fits of the Freundlich-model for the amine-functionalized micro-
tubes.

23
 ACCEPTED MANUSCRIPT

100
0.15 bar measured

with pycnometry
90

80

)
fiber
uptake (mg/g
70

PT
60 measured

with TGA
50

RI
2
Specific CO

40

30

SC
20

10

0 10 20 30

U
PEI load in ethanol (wt.-%)

AN
Figure 6: Influence of the PEI fraction in the impregnation solution on the specific sorption capacities at a partial CO2
pressure of 0.15 bar at ambient temperature measured with TGA and pycnometry (error bars represent the standard
pn −pn+1 1
error of the mean). For the pycometry measurements the equilibrium criterion was pn ∆t < 4.5 · 10−7 1s . For the
M
TGA measurements the CO2 uptake was taken after 45 minutes of sorption time.
D
TE

50 5 wt-% PEI, 30°C

)
Fiber

2.5 wt-% PEI, 30°C

40
uptake (mg/g

EP

30
C Specific CO
2

20 5 wt-% PEI, 80°C

AC

2.5 wt-% PEI, 80°C

10

0
0 10 20 30 40 50
Time (min)

Figure 7: CO2 sorption capacity over time for PEI loads of 2.5 wt.-% and 5 wt.-% at 30°C, 80°C and ambient pressure
measured with TGA

24
 ACCEPTED MANUSCRIPT

PT
RI
SC
60 120

50

U
100
)
fiber
uptake (mg/g

40
AN

Temperature (°C)
80

30
M
60
2
Specific CO

20

40
D

10
TE

0 20
0 100 200 300 400 500
Time (min)
EP

Figure 8: Cyclic CO2 sorption for a PEI load of 20 wt.-%. Adsorption takes place at 30°C and a CO2 gas fraction of
15 vol-%. Desorption takes place at 90°C and pure nitrogen.
C
AC

25