Chemical Engineering Science ∎ (∎∎∎∎) ∎∎∎–∎∎∎

Contents lists available at ScienceDirect

Chemical Engineering Science

journal homepage: www.elsevier.com/locate/ces

Adsorption and diffusion of H2, N2, CO, CH4 and CO2 in UTSA-16
metal-organic framework extrudates
Vicente I. Agueda a, José A. Delgado a,n, María A. Uguina a, Pablo Brea a, Aud I. Spjelkavik b,
Richard Blom b, Carlos Grande b,n
a
Department of Chemical Engineering, Universidad Complutense de Madrid, 28040 Madrid, Spain
b
SINTEF Materials and Chemistry, Forskningsveien 1, 0373 Oslo, Norway

H I G H L I G H T S G R A P H I C A L A B S T R A C T

Measurement of adsorption equili-

brium of H2, N2, CO, CH4 and CO2 at
high pressure.

Measurement of diffusion of five
gases in UTSA-16 extrudates.

Simulation of a new PSA cycle for
hydrogen purification.

High purity H2 can be produced by
PSA process allied with high recov-
ery.

Interplay of new material and
new cycle for improved process
performance.

art ic l e i nf o a b s t r a c t

Article history: Recently, a new metal-organic framework, named UTSA-16, has been proposed for removing carbon dioxide
Received 10 June 2014 from gas mixtures by pressure swing adsorption in view of its high adsorption capacity and selectivity to this
Received in revised form gas. Fundamental adsorption and diffusion parameters are required for designing a pressure swing adsorption
13 August 2014
(PSA) process with this adsorbent, which are lacking in the literature. In this work, adsorption Henry's law
Accepted 14 August 2014
constants and reciprocal diffusion time constants of hydrogen, nitrogen, carbon monoxide, methane and
carbon dioxide on UTSA-16 extrudates have been measured. The pure gas adsorption isotherms of these gases
Keywords: at high pressures and different temperatures have also been measured.
Hydrogen production Based on the pure gas adsorption equilibrium and kinetic parameters, a PSA cycle for hydrogen purification
Carbon dioxide capture
from steam-methane reforming off-gas (containing hydrogen, carbon monoxide, methane and carbon dioxide)
Metal-organic framework
has been simulated. Introducing a rinse step, the process can yield hydrogen with 99.99–99.999% purity with
Adsorption
Diffusion 93–96% recovery and productivities between 2 and 2.8 mol kg  1 h  1.
Pressure swing adsorption & 2014 Elsevier Ltd. All rights reserved.

1. Introduction from the combustion of fossil fuels are a serious concern because of
their global warming effect on climate (IPCC, 2007).
Separations of gas mixtures containing carbon dioxide have The recovery of hydrogen from steam methane reforming
become quite important nowadays. Carbon dioxide emissions coming (SMR) off-gases is one of the largest points of industrial stationary
emissions of carbon dioxide. If pre-combustion techniques for
power production are adopted, the demand for hydrogen will
considerably increase. For this reason, advances in H2 purification
n and CO2 capture from this application are important.
Corresponding authors.
E-mail addresses: carlos.grande@sintef.no (Carlos Grande), In the H2 purification from SMR off-gas, hydrogen is recovered
jadeldob@ucm.es (J.A. Delgado). from a mixture saturated with water vapour containing hydrogen

http://dx.doi.org/10.1016/j.ces.2014.08.039
0009-2509/& 2014 Elsevier Ltd. All rights reserved.

Please cite this article as: Agueda, V.I., et al., Adsorption and diffusion of H2, N2, CO, CH4 and CO2 in UTSA-16 metal-organic
framework extrudates. Chem. Eng. Sci. (2014), http://dx.doi.org/10.1016/j.ces.2014.08.039i
2 V.I. Agueda et al. / Chemical Engineering Science ∎ (∎∎∎∎) ∎∎∎–∎∎∎

and carbon monoxide, methane and carbon dioxide as impurities. temperatures (298, 313 and 338 K), reaching pressures of
The impurity with highest concentration in the mixture is carbon 80 bar for hydrogen, 50 bar for nitrogen and methane, 5 bar
dioxide, with concentrations ranging between 15% and 25% for carbon monoxide, and 40 bar for carbon dioxide.
(Beaver and Sircar, 2010), while methane and carbon monoxide (iii) To evaluate the potential performance of the synthesized
content is lower than 5%. adsorbent on the recovery of hydrogen from SMR off-gases
Pressure swing adsorption (PSA) is mature technology for (composed of hydrogen, methane, carbon monoxide and carbon
purification of hydrogen (Stöcker et al., 1998). The unit is com- dioxide) by PSA using the measured adsorption properties.
posed by several columns (Baksh and Simo, 2013), each of them
comprising two or three layers of adsorbents to selectively remove
impurities: a first layer of alumina or silica gel to remove water
followed by a layer of activated carbon to remove CO2 and part of 2. Materials and methods
CH4 and a final zeolite layer to remove CO and other components
in the mixture (Ruthven et al., 1994; Chlendi and Tondeur, 1995; UTSA-16 was synthesized according to the method reported in
Waldron and Sircar, 2000). In this unit it is essential to maximize literature (Xiang et al., 2005, 2012). The crystals were agglomer-
hydrogen recovery while satisfying stringent purity, which can ated in the form of extrudates using polyvinyl alcohol (PVA, Alfa
be 499.999%. Aesar) as binder (Grande et al., this issue). Fig. 1 shows a SEM
Tuning the process and modifying or changing the adsorbent image of the final prepared extrudates. The UTSA-16 crystals used
materials may improve the unit performance. Several studies can in this study have a relatively wide particle size distribution (from
be found in literature dealing with optimization of the process and about 1 to 5 μm), as deduced from Fig. 1. However, this is not a
with evaluation of new material (Park et al., 1998, 2000; Lee et al., problem for obtaining cylindrical extrudates with regular size,
1999; Sircar and Golden, 2000; Yang et al., 2008). Most studies are very similar to commercial extrudates of zeolitic adsorbents.
related to optimization of the properties of activated carbons and Extrudate properties measured by mercury and nitrogen porosi-
zeolites (Bec, 2005; Lopes et al., 2010). The integrated CO2 capture metry are presented in Table 1.
and H2 production may bring new light to PSA technology for this The adsorption isotherms of the studied gases on UTSA-16
application where new materials and processes should be con- extrudates have been measured in a manometric equipment (BEL-
sidered (Rarig et al., 2002; Voss, 2005; Waade et al., 2012; Grande SORP HP, Japan). Prior to the experiments, the adsorbent was
et al., 2013). Metal-organic frameworks (MOFs) can potentially regenerated under vacuum overnight at 373 K. The values reported
result in significant improvements in the PSA process by selective in this work correspond to the absolute amount adsorbed, after
removal either of CO2 or of the other minor contaminants. Several transformation of the excess amount adsorbed which is measured
studies have considered MOFs as alternative materials (Wang experimentally (Grande et al., 2013).
et al., 2002; Jasra and Walton, 2008; Schell et al., 2012; Herm
et al., 2012; Moellmer et al., 2012; Banu et al., 2013; Silva et al.,
2013).
Recently, a novel microporous metal-organic framework, called
UTSA-16, has been proposed for capturing carbon dioxide from gas
mixtures by PSA in view of its high carbon dioxide adsorption
capacity and selectivity at ambient conditions (Xiang et al., 2005,
2012). This material has small cages of about 0.45 nm in diameter,
with pore openings of 0.33  5.4 nm2, and exhibits a carbon
dioxide adsorption capacity which is one of the highest among
MOF materials (Willis, 2010; Xiang et al., 2012). The high capacity
of this material with the zeolite-like density makes it a potential
candidate for being used in purification of hydrogen from SMR off-
gases.
The design and optimization of a PSA process requires at least
the measurement of fundamental adsorption and diffusion para-
meters of the components of the gas mixture to be separated. This
information must be incorporated into an appropriate theoretical
model to simulate the effect of the operation variables on the
separation performance (Park et al., 1998; Barg et al., 2000; Jiang
et al., 2004; Nikolic et al., 2009). Although the adsorption
equilibrium data of some gases involved in interesting carbon
dioxide separations (carbon dioxide, methane and nitrogen) on
UTSA-16 are available in the literature (Xiang et al., 2012), they Fig. 1. SEM image of a UTSA-16 extrudate.
cover a limited pressure range (up to1 bar). It is not possible to
find adsorption equilibrium data for other gases, particularly at
higher pressures required to design a H2-PSA. To our knowledge, Table 1
no diffusion parameters were reported in UTSA-16 MOF. Properties of synthesized UTSA-16 extrudates.
The objectives of this work are the following:
Extrudate Extrudate Particle Average pore BET surface
density dimensions porositya diameter (nm)b (m2 g  1)b
(i) To measure the adsorption Henry's law constants and reci- (g cm  3) (cylindrical)
procal diffusion time constants of hydrogen, nitrogen, carbon
monoxide, methane and carbon dioxide on UTSA-16 extru- 1.171 5 mm long  2 mm 0.49 1.48 805
in diameter
dates (Grande et al., this issue).
(ii) To measure the pure adsorption equilibrium isotherms of the a
Considering the void volume estimated by Hg porosimetry only.
cited gases in the prepared UTSA-16 extrudates at different b
Measured by nitrogen porosimetry.

Please cite this article as: Agueda, V.I., et al., Adsorption and diffusion of H2, N2, CO, CH4 and CO2 in UTSA-16 metal-organic
framework extrudates. Chem. Eng. Sci. (2014), http://dx.doi.org/10.1016/j.ces.2014.08.039i
 V.I. Agueda et al. / Chemical Engineering Science ∎ (∎∎∎∎) ∎∎∎–∎∎∎ 3

Pulse experiments were performed with a column packed with (KH) are calculated from the first moment of pulse responses,
UTSA-16 extrudates, placed inside the temperature controlled subtracting the delay time caused by the void volume in the
oven of a gas chromatograph (Agilent 6890) coupled with a installation, measured from helium pulses (considering helium as
thermal conductivity detector (TCD). Bed properties in the pulse a non-adsorbing tracer). A theoretical model was used to extract the
experiments are given in Table 2. A schematic drawing of the reciprocal diffusion time constants from the experimental pulse
experimental installation is shown in Fig. 2. In each pulse experi- responses, considering the flow pattern presented in Fig. 3. The
ment, a 20 μl loop filled with the studied gas is swept by the model is obtained by applying mass balances to the elements in the
carrier gas, crossing the adsorbent bed, and the pulse response is flow pattern, connecting them according to the direction of flow.
analysed by the TCD detector of the chromatograph. Helium was The resulting equations are given in Table S1 of Supplementary
used as carrier gas when hydrogen, nitrogen, methane, carbon Information. The model is fitted first to the helium pulses, to
monoxide and carbon dioxide are the adsorbates. Helium pulses determine the value of the tank volume (VT) parameter for different
were also recorded to measure the void volume and the global carrier gas flow rates and temperatures (it ranges from 0.55 to
dispersion in the installation, using nitrogen as carrier gas. Prior to 0.6 ml in all the experiments). Once this parameter is known, the
the pulse experiments, the adsorbent was regenerated at 373 K model is fitted to a pair of pulses of a determined adsorbing gas
overnight under helium flow. with different flow rates and constant temperature (minimizing the
All the gases employed in this study had purities higher than sum of squared residuals) to obtain the corresponding reciprocal
99.999% and were supplied by Praxair, Norway. diffusion time constant (Dc/r2c ).The fitting quality is evaluated with
the coefficient of determination, which is usually called r2. To avoid
confusion with radius, this parameter has been called COD in this
3. Results and discussion work. A comparison between representative theoretical and experi-
mental pulse responses at 298 K is presented in Fig. 4.
3.1. Chromatographic measurement of adsorption Henry's law The UTSA-16 extrudates may present resistance to diffusion
constants and reciprocal diffusion time constants within the micropores of the MOF crystals but also within the
macropores of the extrudates. The similarity between the mea-
A chromatographic method has been used to measure the pure sured values of the reciprocal diffusion time constants including
gas adsorption Henry's law constants and the reciprocal diffusion and removing the macropore resistance (Dc/r2c and (Dc/r2c )0 in
time constants of hydrogen, nitrogen, methane, carbon monoxide Table 3) indicates that micropore diffusion within the UTSA-16
and carbon dioxide on UTSA-16 extrudates from pulse responses. crystals is dominant within the studied conditions. For carbon
This method has been explained in detail elsewhere (Delgado et al., dioxide, the time of the peak maximum is much higher than for
2014), so only a brief summary will be given herein, indicating the the other gases because of the high value of KH for this gas. It can
changes made to adapt the method to the different experimental be observed that the value of KH correlates with the polarizability
installation used in this work. Adsorption Henry's law constants for H2, N2, CO and CH4, shown in Table 4, suggesting that
dispersion forces determine the adsorption affinities of these gases
Table 2 in UTSA-16. Electrostatic interactions seem to be weak for these
Bed properties in the pulse experiments. adsorbates. For carbon dioxide, although the polarizability is
higher than for the rest of gases, the much higher value of KH is
Bed length (cm) Bed mass (g) Porosity between particles Bed diameter (cm) due to the electrostatic interaction between its quadrupole
2.22 0.68 0.607 0.92
moment and coordinated water molecules in UTSA-16 cavities
(Xiang et al., 2012). However, carbon monoxide has a lower
adsorption affinity than methane, despite of the electrical
moments of carbon monoxide (Table 4). This adsorption behaviour
is also observed in activated carbon (Delgado et al., 2014), which
indicates that the interaction of the electrical moments of carbon
monoxide with the polar adsorption sites in UTSA-16 do not
overcome the higher dispersion interaction of methane, which
has higher polarizability.
Fig. 5 shows a plot of the values of Dc/r2c vs. KH for all gases. Similar
plots have been proposed previously (surface diffusivity against heat of
adsorption and boiling temperature, (Suzuki, 1990)). However, the
effect of the pre-exponential constant of KH is not considered in those
plots, and they were proposed for unrestricted diffusion in wide pores,
not for diffusion in pores with sizes similar to the one of the diffusing
molecule. It is observed that the reciprocal diffusion time constants in
UTSA-16 are inversely proportional to the residence time in the
adsorption sites. The activation energy of diffusion increases as K does.
This is indicative of a temperature activated diffusion mechanisms like
surface and crystal diffusion, which was also observed for carbon
monoxide, methane and carbon dioxide in BPL activated carbon and
13X zeolite (Delgado et al., 2014). It is noteworthy that molecular size
(Table 4) seems to have little effect on the diffusion rate, despite the
fact that one dimension of the pore opening (0.33 nm) is smaller than
the kinetic diameter of most adsorbates. Adsorption of the biggest
molecule (methane, with 0.38 nm) is quite significant (KH⪢1), with a
value of Dc/r2c one order of magnitude higher than the one of carbon
dioxide, indicating that methane can go easily through the pore
Fig. 2. Scheme of the setup for pulse measurements. windows.

Please cite this article as: Agueda, V.I., et al., Adsorption and diffusion of H2, N2, CO, CH4 and CO2 in UTSA-16 metal-organic
framework extrudates. Chem. Eng. Sci. (2014), http://dx.doi.org/10.1016/j.ces.2014.08.039i
4 V.I. Agueda et al. / Chemical Engineering Science ∎ (∎∎∎∎) ∎∎∎–∎∎∎

Fig. 3. Flow-pattern considered for reproducing experimental pulse responses.

0.003
0.6
CO

0.002
0.4 He

N.S.
N.S.

0.2 N 0.001

CO
CH

0.0 0.000
0 10 20 30 40 0 400 800 1200
t/s t/s
Fig. 4. Pulse responses from an bed packed with UTSA-16 extrudate. Grey lines are the experimental signal and black lines are the theoretical signal from the model. All the
pulses have a carrier gas flow rate of 1 ml min  1 at 298 K: (a) Pulses of helium, nitrogen, carbon monoxide and methane and (b) pulse of carbon dioxide.

Table 3
Experimental conditions for the pulse experiments with hydrogen, nitrogen, carbon monoxide, carbon dioxide and methane in UTSA-16 extrudates, and calculated
adsorption and diffusion parameters.

Gas Q, 10  6 m3 s  1 T (K) KH KH manometricc Dc/r2c (s  1) Ediff (kJ mol  1) (Dc/r2c )0 (s  1)b COD

a
H2 0.5 298 0.22 0.882 1  1 0.95
H2 1 298 0.17 0.882 1  1 0.95
N2 0.5 298 7.45 8.56 0.106 5.6 0.096 0.998
N2 1 298 7.78 8.56 0.106 5.6 0.096 0.998
N2 0.5 313 5.95 6.57 0.137 5.6 0.123 0.997
N2 1 313 6.49 6.57 0.137 5.6 0.123 0.997
N2 0.5 338 4.22 4.78 0.141 5.6 0.132 0.995
N2 1 338 3.91 4.78 0.141 5.6 0.132 0.995
CO 0.5 298 13.0 13.6 0.079 7.7 0.070 0.999
CO 1 298 13.3 13.6 0.079 7.7 0.070 0.999
CO 0.5 313 9.44 10.4 0.096 7.7 0.087 0.998
CO 1 313 10.3 10.4 0.096 7.7 0.087 0.998
CO 0.5 338 6.93 6.96 0.115 7.7 0.107 0.990
CO 1 338 6.11 6.96 0.115 7.7 0.107 0.990
CH4 0.5 298 24.3 27.9 0.049 12 0.043 0.998
CH4 1 298 25.6 27.9 0.049 12 0.043 0.998
CH4 0.5 313 18.6 20.4 0.063 12 0.056 0.997
CH4 1 313 19.3 20.4 0.063 12 0.056 0.997
CH4 0.5 338 12.2 13.3 0.088 12 0.079 0.997
CH4 1 338 12.3 13.3 0.088 12 0.079 0.997
CO2 0.5 298 1108 1416 0.0014 28 0.0012 0.996
CO2 1 298 1115 1416 0.0014 28 0.0012 0.996
CO2 1 338 291 312 0.0053 28 0.0046 0.993

a
The fitted pulse is insensitive to an increase of Dc/r2c .
b
Setting kmacro ¼ infinite.
c
From Dual Langmuir isotherm fitted to volumetric data.

In the absence of diffusion data for UTSA-16, it is not possible to with a similar average size of crystals as in this work (about 3 μm),
compare the measured diffusion parameters with other reported they are lower in UTSA-16 by a factor of 2, which can be attributed
data. Comparing the values of Dc/r2c of carbon monoxide and to the larger pore window of 13X zeolite (about 0.8 nm). For carbon
methane with data reported for 13X zeolite (Delgado et al., 2014), dioxide, the reciprocal diffusion time constant in UTSA-16 is higher

Please cite this article as: Agueda, V.I., et al., Adsorption and diffusion of H2, N2, CO, CH4 and CO2 in UTSA-16 metal-organic
framework extrudates. Chem. Eng. Sci. (2014), http://dx.doi.org/10.1016/j.ces.2014.08.039i
 V.I. Agueda et al. / Chemical Engineering Science ∎ (∎∎∎∎) ∎∎∎–∎∎∎ 5

Table 4 volumetric experiments, KH is calculated as follows:

Molecular properties of studied gases.
K H ¼ ðnmax 1 b1 þ nmax 2 b2 Þρp RT=ð1  εp Þ ð2Þ
Gas Polarizability Dipole moment, Quadrupole moment, Kinetic
(Å3)a Debyea Debye  (Å)a diameter (Å)b
The values of KH from pulse and adsorption equilibrium
experiments are quite similar, so the consistency of the results is
H2 0.79 0 0.52 2.9 verified.
N2 1.71 0 1.39 3.6 Fig. 6(f) shows a comparison between the adsorption isotherms
CH4 2.45 0 0 3.8
of CO2 at 298 K in UTSA-16 extrudates with the commercial
CO 1.95 0.11 2.84 3.6
CO2 2.51 0 4.28 3.3 adsorbents used in H2-PSA units: activated carbon (BPL granules)
and 13X zeolite pellets (Delgado et al., 2014). The adsorbed
a
NIST website. amounts are expressed in units of mol per unit volume of
b
Meulenberg et al. (2011).
adsorbent to take into account their different adsorbent densities.
It is observed that 13X zeolite has the highest carbon dioxide
adsorption capacity of the three adsorbents. However, the slope of
the isotherm is very steep at low pressures, which makes the
regeneration difficult, requiring low desorption pressures and/or
large purge gas volumes. On the contrary, activated carbon has the
lowest adsorption capacity, but the lowest slope at low pressures,
making the regeneration easier. UTSA-16 offers a higher carbon
dioxide adsorption capacity than activated carbon and an easier
regeneration than 13X zeolite. However, it is difficult to know
intuitively if these combinations of properties are good for using
UTSA-16 in the hydrogen purification from SMR-off gas by PSA.
The potential performance of UTSA-16 in this separation is studied
in the next section with a PSA simulation tool.
Fig. S1 in Supporting Information shows predictions of the
Fig. 5. Plot of Dc/r2c versus KH for the studied gases in UTSA-16 extrudate. The line is multicomponent adsorption behaviour of hydrogen, carbon mon-
a guide to the eye. oxide, carbon dioxide and methane using IAS theory, in some
representative conditions of the PSA simulations. In Fig. S1(a), the
adsorption capacities of the four components in equilibrium with a
gas having the feed composition of the PSA process are presented as
than in 13X zeolite (about ten times at 338 K), which means that a function of total pressure (from 0 to 16 bar). It is observed that the
diffusion in UTSA-16 is faster than in zeolite 13X, being beneficial selectivity towards carbon dioxide is very high in the whole
for H2-PSA operation. pressure range. The carbon monoxide/methane selectivity increases
as the partial pressure of carbon dioxide in the mixture does,
3.2. Adsorption equilibrium isotherms of hydrogen, nitrogen, reversing the selectivity at low pressures (given by the ratio
methane, carbon monoxide and carbon dioxide in UTSA-16 K HCO =K HCH4 ¼ 0:52) for partial pressures of carbon dioxide above
extrudates  0.7 bar ( 4 bar  0.17). The same behaviour is observed in the
absence of hydrogen in the mixture (Fig. S1(b)), where the carbon
Fig. 6 shows the adsorption isotherms of the studied gases in monoxide/methane selectivity reversal also occurs for partial pres-
UTSA-16 extrudates measured at three temperatures (298, 313 and sures of carbon dioxide about 0.7 bar (1 bar  0.7). The adsorption
338 K). It is shown that the adsorption capacity for a given behaviour is very different in the absence of carbon dioxide in the
adsorbate at a fixed pressure increases as the adsorption affinity mixture (Fig. S1(c)). This behaviour occurs in the bed zones near the
does, which is quantified with the corresponding value of KH light end, where carbon dioxide concentration is very low. In this
(Table 3). Fig. 6(e) shows a comparison between the adsorption case, the carbon monoxide/methane selectivity is lower than one.
isotherm of carbon dioxide in UTSA-16 crystals reported in the This results in a carbon monoxide concentration in the hydrogen
literature (Xiang et al., 2012), and the one measured in this work product higher than the one of methane.
for UTSA-16 extrudates. It is observed that the adsorption capacity
is about a 15% lower in the extrudates, which can be attributed to 3.3. Simulation of the hydrogen purification from SMR-off gas by PSA
the content of binder in the extrudates (reducing available surface with UTSA-16 as adsorbent
area and pore volume) and additionally some blocking effect of the
binder (Finsy et al., 2009). A four-column PSA cycle for hydrogen purification from SMR-
The experimental adsorption isotherms were fitted adequately off gas with UTSA-16 extrudates as adsorbent has been simulated.
with the Dual Langmuir model in the studied pressure range as The columns are 8 m long, with a diameter of 1.6 m. The high and
follows: low cycle pressures are set to 16 and 1 bar, respectively.

 Connectivity between the steps and pressure levels of the
nmax 1 b1 p nmax 2 b2 p ΔH 1
n¼ þ ; b1 ¼ b01 exp  ; different steps are presented in Fig. 7. The studied cycle is similar
1 þ b1 p 1 þ b2 p RT

 to the one studied in a previous work using BPL activated carbon
ΔH 2
b2 ¼ b02 exp  ð1Þ and 13X zeolite as adsorbents in a layered bed (Delgado et al.,
RT
2014). However, in this cycle a rinse step (a portion of tail gas rich
A comparison between the theoretical and the experimental in carbon dioxide is recycled as feed) has been introduced after the
isotherms is given in Fig. 6, and the obtained parameters are given feed step. The rinse step is a heavy component recycle and thus it
in Table 5. can be considered as a dual-reflux cycle (Yoshida et al., 2003;
To check the consistency of the adsorption equilibrium parameters Reynolds et al., 2006; Bhatt et al., 2013).The rinse step has been
calculated from pulse responses and from volumetric experiments, the included to take advantage of the higher carbon dioxide adsorp-
values of KH resulting from both methods are compared in Table 3. For tion capacity of UTSA-16 with respect to BPL activated carbon,

Please cite this article as: Agueda, V.I., et al., Adsorption and diffusion of H2, N2, CO, CH4 and CO2 in UTSA-16 metal-organic
framework extrudates. Chem. Eng. Sci. (2014), http://dx.doi.org/10.1016/j.ces.2014.08.039i
6 V.I. Agueda et al. / Chemical Engineering Science ∎ (∎∎∎∎) ∎∎∎–∎∎∎

Fig. 6. Experimental adsorption isotherms in UTSA-16 extrudates at different temperatures. Squares, 298 K. Circles, 313 K. Diamonds, 338 K. Solid lines are obtained with the
Dual Langmuir model: (a) hydrogen, (b) nitrogen, (c) carbon monoxide, (d) methane and (e) carbon dioxide. The dashed line depicts a carbon dioxide isotherm on UTSA-16
crystals at 298 K reported in the literature (Xiang et al., 2012). (f) Comparison of carbon dioxide isotherms (in mol l  1) at 298 K in 13X zeolite pellets, UTSA-16 extrudates and
BPL activated carbon granules.

Table 5
Dual Langmuir parameters of the studied gases in UTSA-16 extrudates.

Gas nmax1, (mol kg  1) nmax2 (mol kg  1) b01 (Pa  1) b02 (Pa  1) (  ΔH1) (kJ mol  1) (  ΔH2) (kJ mol  1)

H2 3.428 0 2.039  10  9 0 7.684 0

N2 1.326 1.773 2.154  10  8 1.669  10  12 8.558 30.28
CO 2.575 1.429 6.962  10  10 7.128  10  10 16.74 16.68
CH4 2.148 0.772 1.524  10  10 7.180  10  8 23.12 7.647
CO2 4.079 1.289 6.000  10  11 1.626  10  13 34.25 37.82

Please cite this article as: Agueda, V.I., et al., Adsorption and diffusion of H2, N2, CO, CH4 and CO2 in UTSA-16 metal-organic
framework extrudates. Chem. Eng. Sci. (2014), http://dx.doi.org/10.1016/j.ces.2014.08.039i
 V.I. Agueda et al. / Chemical Engineering Science ∎ (∎∎∎∎) ∎∎∎–∎∎∎ 7

Duration, s 30 180 30 30 180 30 30 180 30 30 210

STEP 1 2 3 4 5 6 7 8 9 10
COLUMN 1 ADS RIN DEQ1 PP DEQ2 BD RP PEQ2 PEQ1 BF
COLUMN 2 PEQ1 BF ADS RIN DEQ1 PP DEQ2 BD RP PEQ2
COLUMN 3 BD RP PEQ2 PEQ1 BF ADS RIN DEQ1 PP DEQ2
COLUMN 4 DEQ1 PP DEQ2 BD RP PEQ2 PEQ1 BF ADS RIN

16
ADS RIN
14
12
P at product end / bar

DEQ1
BF
10
8
PP
6 PEQ1

4 DEQ2

2 BD RP PEQ2

0
0 120 240 360 480 600 720 840 960
t/s
Fig. 7. (a) Connectivity between steps. F ¼ feed gas, L¼ light product, T ¼tail gas. (b) Cycle time schedule. Total cycle time ¼16 min. ADS ¼Adsorption, RIN ¼ rinse,
DEQ1 ¼depressurizing equalization 1, PP ¼ provide purge, DEQ2 ¼depressurizing equalization 2, BD¼ blowdown, RP ¼receive purge, PEQ2 ¼ pressurizing equalization 2,
PEQ1, pressurizing equalization 1, BF ¼backfill. (c) Pressure history for Simulation 4 in Table 6.

Table 6
Design parameters and performance results for the PSA cycle in Fig. 7 with UTSA-16 extrudates.

Simulation uF (m s  1) tRIN (s) PDEQ1 (bar)a PDEQ2 (bar)a Pur. (%) Rec. (%) Productivity, (molH2 kg  1 h  1)

1 0.0529 0 9.16 3.21 99.9902 80.3 3.71

2 0.0462 65 8.50 2.89 99.9947 94.8 2.78
3 0.0398 95 8.37 2.80 99.9980 96.7 2.03
4 0.0460 60 8.61 2.96 99.9991 93.0 2.80

a
Final pressure, also for the column receiving gas in the corresponding PEQ step.

which could lead to a higher carbon dioxide concentration in the The cycle time schedule is presented in Fig. 7(b), where the
waste gas, resulting in a higher hydrogen recovery in the light duration of the adsorption and rinse steps sum up 240 s. Fig. 7(c)
product and a consequently higher purity of CO2 in the tail gas. shows the pressure history for one simulation (Simulation 4 in
The feed gas is at 298 K and composed of hydrogen (76%), carbon Table 6). The final pressure of the provide purge step (step 4) has
monoxide (4%), carbon dioxide (17%) and methane (3%). been fixed to the value used in our previous work (5.73 bar). This

Please cite this article as: Agueda, V.I., et al., Adsorption and diffusion of H2, N2, CO, CH4 and CO2 in UTSA-16 metal-organic
framework extrudates. Chem. Eng. Sci. (2014), http://dx.doi.org/10.1016/j.ces.2014.08.039i
8 V.I. Agueda et al. / Chemical Engineering Science ∎ (∎∎∎∎) ∎∎∎–∎∎∎

value is reasonable for the provide purge step in industrial PSA is available elsewhere (Delgado et al., 2014). Adsorption equili-
processes for hydrogen purification (Batta, 1971). The adsorbent brium and kinetic parameters reported in this work were used
properties considered are the same as the ones corresponding to for the simulations. The multicomponent adsorption isotherm
the pulse experiments except for the porosity between particles, has been described with the IAS theory, introducing the Dual
which has been set to 0.459 (measured experimentally in a Langmuir isotherms deduced previously as the pure component
column with 2.5 cm in diameter). adsorption isotherms, and solving the corresponding equations
The cycle was simulated with a computer programme in with the Fast IAS algorithm proposed by Do (1998). An average
FORTRAN developed by Universidad Complutense de Madrid and isosteric heat has been considered for each adsorbate, calculated
commercialized by Inprocess Technology and Consulting (Spain). as follows:
Model equations in the programme are based on mass, energy
 
Z
and momentum conservation equations, shown in Table S2. 1 nF
∂ ln p 
Q st ¼ R dn ð3Þ
More information about the method used for solving the model nF 0 ∂ð1=TÞn

Fig. 8. Simulated spatial profiles in Simulation 4: (a) gas mole fractions at the end of the rinse step. (b) Adsorbed concentrations at the end of the rinse step. (c) Temperature
profile at the end of the rinse step. (d) Gas mole fractions at end of the DEQ2 step. (e) Gas mole fractions at end of the RP step. (f) Gas mole fractions at end of the PEQ1 step.

Please cite this article as: Agueda, V.I., et al., Adsorption and diffusion of H2, N2, CO, CH4 and CO2 in UTSA-16 metal-organic
framework extrudates. Chem. Eng. Sci. (2014), http://dx.doi.org/10.1016/j.ces.2014.08.039i
 V.I. Agueda et al. / Chemical Engineering Science ∎ (∎∎∎∎) ∎∎∎–∎∎∎ 9

where the integrand is deduced from the Dual Langmuir isotherm The column can be loaded with up to 4 mol kg  1 of carbon
of each component, and nF is the adsorption capacity calculated dioxide after the rinse step as shown in Fig. 8(b), significantly
at feed conditions. The following average isosteric heats are obtained: higher than the one obtained for a PSA cycle with similar
1 1
Q stH2 ¼ 7:78 kJ mol , QstCO ¼ 16.7 kJ mol  1, Q stCH4 ¼ 19:3 kJ mol , conditions and BPL activated carbon and 13X zeolite as adsorbents
1
and Q stCO2 ¼ 34:4 kJ mol . (up to 2.8 mol kg  1, Delgado et al., 2014). This results in a high
The process has been designed testing different combinations carbon dioxide concentration in the waste gas (about 60%), leading
of feed superficial velocity in the adsorption and rinse steps (uF, to high hydrogen recovery. Due to the high cyclic loading of carbon
the same value for both steps) and duration of the rinse step (tRIN) dioxide, a temperature variation should be expected within the
(the duration of the adsorption step is calculated as 240 s minus column as shown in Fig. 8(c). Two maximal temperatures are
tRIN),simulating the cycle until the cyclic steady state is reached. observed: one around 25 K at z¼ 2.7 m due to carbon dioxide
The following performance parameters are calculated: adsorption and another at z ¼3.8 m due to adsorption of other
components (CH4 and CO).
ðmoles H2 outÞADS þ ðmoles H2 outÞRIN ðmoles H2 inÞBF
Hydrogen purity ¼  100 Figs. 8(d) and (e) show the gas mole fraction profiles at the end
ðmoles outÞADS þ ðmoles outÞRIN ðmoles inÞBF
of the DEQ2 step (or beginning of the blowdown step) and at the
ð4Þ
end of the purge step, respectively, showing the change of these
profiles during the regeneration of the column, when waste gas is
ðmoles H2 outÞADS þ ðmoles H2 outÞRIN  ðmoles H2 inÞBF
Recovery ¼  100 ð5Þ obtained. It is observed that the previous co-current depressuriza-
ðmoles H2 inÞADS
tion step lead to a large increase of the concentration of impurities
in the column. Carbon monoxide and methane are concentrated in
ðmoles H2 outÞADS þ ðmoles H2 outÞRIN  ðmoles H2 inÞBF
Productivity ¼ ð6Þ the second half of the column, due to the previous displacement in
ðcycle timeÞðmass of adsorbent in one columnÞ
the adsorption and rinse steps by carbon dioxide, and carbon
The design specification was either a hydrogen purity of dioxide is concentrated in the first third of the column. After the
99.99 þ% or 99.999 þ%. The results are presented in Table 6. regeneration (Fig. 8(e)), the concentration of impurities in the
Simulations 1, 2 and 3 show that the inclusion of the rinse step column is greatly reduced, indicating that the regeneration is very
leads to a great increase in the hydrogen recovery, at the cost of a efficient.
lower productivity and some energy consumption for recompres- Fig. 8(f) presents the gas mole fraction profiles at end of the
sion of a portion of the tail gas. The reduction in productivity is PEQ1 step. It is observed that almost pure hydrogen fills the light
due to the reduction in the proportion of fresh feed to the cycle as end of the column after the RP, PEQ1, and PEQ2 steps, indicating
the amount of rinse gas increases, and to the reduction of that almost pure hydrogen is fed to the column through the light
maximum feed velocity allowable to get the desired purity, end in these steps.
because the velocity of the impurities concentration front is higher An important feature of UTSA-16 for this purpose is its high
for a higher proportion of impurities in the rinse step. It must be volumetric density and that the species diffuse relatively easy
noted that the achieved productivity is lower than one typically through the pores of the extrudates. For its real application in the
achieved by industrial polybed PSA process for hydrogen purifica- H2-PSA process however, a pre-layer of a water-selective adsor-
tion from SMR off-gas based on layered activated carbon/zeolite bent should be used to remove water before this adsorbent. The
columns, about 5 mol kg  1 h  1 (Stöcker et al., 1998). However, the results obtained from this work present a positive outcome of the
values of hydrogen recovery are quite high considering that utilization of metal-organic frameworks in H2-PSA utilization.
conventional PSA cycles for the studied separation typically
achieve recoveries ranging between 75% and 92% (Beaver and
Sircar, 2010). The economic benefit obtained from the higher
hydrogen recovery during all the process lifetime, can compensate 4. Conclusions
for the higher fixed cost of the adsorbent (Sircar and Golden,
2010). Another advantage of using UTSA-16 is that only one kind of The evaluation of a PSA process for H2 recovery and purification
adsorbent is used, which can withstand high concentrations of using metal-organic framework UTSA-16 was carried out. The
carbon dioxide without damaging its performance. This is good for evaluation was carried out by measuring fundamental adsorption
treating feeds with different composition, or for producing hydro- equilibrium and kinetic data that was not available in literature.
gen with different levels of purity depending on the destination of A chromatographic method has been used to measure the
the hydrogen product. In conventional activated carbon/zeolite adsorption Henry's law constants and the reciprocal diffusion time
columns, if the amount of carbon dioxide accumulated in the constants of H2, N2, CO, CH4 and CO2 on UTSA-16 extrudates. The
zeolite layer is too high, the performance drops abruptly because adsorption affinities of H2, N2, CO and CH4 correlate with their
carbon dioxide cannot be desorbed. Simulation 4 shows that the polarizability, indicating that dispersion forces have the major
process can produce hydrogen with 99.999 þ% purity with high contribution to the adsorption energy. For CO2, the affinity is much
recovery (93%). higher because of its strong quadrupole moment.
A detailed simulation of the PSA process is presented in Fig. 8 Reciprocal diffusion time constants are inversely proportional
(corresponding to Simulation 4 in Table 6). In Fig. 8(a) it is shown to the adsorption Henry's law constants, and steric hindrance
that at the end of the rinse step, a significant portion of the column seems not to be important in the diffusion of the studied gases,
(about 30%) is nearly unused, being filled mostly with hydrogen. suggesting a surface diffusion mechanism. The reciprocal diffusion
This is due to the low affinity of UTSA-16 for the lighter impurities time constant of carbon dioxide is significantly higher in UTSA-16
in the feed mixture (carbon monoxide and methane), so it is than in 13X zeolite.
necessary to have a portion of the column clean to get high The adsorption isotherms of the studied gases in UTSA-16
hydrogen purity. Moreover, such a portion should exist to allow extrudates have been measured until high pressures at different
concentration fronts to move forward in depressurization and temperatures. They can be described with the Dual Langmuir
provide purge steps. Although the amount of unused column has a isotherm in for a wide pressure range (up to 40 bar for carbon
negative effect on productivity, it is good for adapting the PSA dioxide). UTSA-16 presents a higher carbon dioxide adsorption
cycle to higher loads of impurities if less pure hydrogen product is capacity than activated carbon and a less steep carbon dioxide
acceptable. isotherm than 13X zeolite at low pressures.

Please cite this article as: Agueda, V.I., et al., Adsorption and diffusion of H2, N2, CO, CH4 and CO2 in UTSA-16 metal-organic
framework extrudates. Chem. Eng. Sci. (2014), http://dx.doi.org/10.1016/j.ces.2014.08.039i
10 V.I. Agueda et al. / Chemical Engineering Science ∎ (∎∎∎∎) ∎∎∎–∎∎∎

A four-bed PSA cycle for hydrogen purification from SMR-off using sorption enhanced reaction concept. In: Eguchi, K. (Ed.), Clean Energy
gas has been simulated. Introducing a rinse step in the cycle, Systems and Experiences. Intech, Luxembourg, ISBN: 978-953-307-147-3.
Bec, R.L., 2005. Method for purifying hydrogen-based gas mixtures using calcium
the process can yield hydrogen with 99.99–99.999% purity X-zeolite. U.S. Patent 6,849,106.
with 93–96% recovery and productivities between 2 and Bhatt, T.S., Storti, G., Rota, R., 2013. Optimal design of dual-reflux pressure swing
2.8 mol kg  1 h  1. adsorption units via equilibrium theory. Chem. Eng. Sci. 102, 42–45.
Chlendi, M., Tondeur, D., 1995. Dynamic behavior of layered columns in pressure
An important feature of the positive initial evaluation of UTSA- swing adsorption. Gas Sep. Purif. 9, 231–242.
16 for H2-PSA is its high volumetric density and that the species Delgado, J.A., Agueda, V.I., Uguina, M.A., Sotelo, J.L., Brea, P., Grande, C.A., 2014.
diffuse relatively easy through the pores of the extrudates. The Adsorption and diffusion of H2, CO, CH4, and CO2 in BPL activated carbon and
13X zeolite: evaluation of performance in pressure swing adsorption hydrogen
results obtained from this work present a positive outcome of the
purification by simulation. Ind. Eng. Chem. Res. 2014, http://dx.doi.org/10.1021/
utilization of metal-organic frameworks in H2-PSA utilization. ie403744u.
Do, D.D., 1998. Adsorption Analysis: Equilibria and Kinetics. Imperial College Press,
London.
Finsy, V., Ma, L., Alaerts, L., De Vos, D.E., Baron, G.V., Denayer, J.F.M., 2009.
Nomenclature Separation of CO2/CH4 mixtures with the MIL-53(Al) metal-organic framework.
Microporous Mesoporous Mater. 120, 221–227.
Grande, C.A., Blom, R., Möller, A., Möllmer, J., 2013. High-pressure separation of
b01,2 parameters in dual Langmuir model, Pa  1 CH4–CO2 using activated carbon. J. Chem. Eng. Sci. 89, 10–20.
Grande, C.A., Águeda, V.I., Spjelkavik, A., Blom, R., 2014. An efficient recipe for
Dc micropore diffusivity, m2 s  1
formulation of metal-organic frameworks. Chem. Eng. Sci., http://dx.doi.org/
ΔH1,2 parameters in dual Langmuir model, kJ mol  1 10.1016/j.ces.2014.06.048, this issue.
3
KH adsorption equilibrium Henry's law constant, m3gas msolid Herm, Z.R., Swisher, J.A., Smit, B., Krishna, R., Long, J.R., 2012. Metal-organic
kmacro combined mass transfer coefficient in the external film frameworks for hydrogen purification and precombustion carbon dioxide
capture. J. Am. Chem. Soc. 133, 5664–5667.
and the macropores, m s  1 IPCC, 2007. Climate Change 2007: Synthesis Report. Contribution of Working
n adsorbed concentration, mol kg  1; number of components Groups I, II and III to the Fourth Assessment Report of the Intergovernmental
nmax1,2 parameters in dual Langmuir model, mol kg  1 Panel on Climate Change. IPCC, Geneva, Switzerland.
Jasra, J.R., Walton, K.S., 2008. Effect of open metal sites on adsorption of polar and
p adsorptive pressure, Pa nonpolar molecules in metal-organic framework Cu-BTC. Langmuir 24,
P pressure, Pa 8620–8626.
Q volumetric flow rate, m3 s  1 Jiang, L., Fox, V.G., Biegler, L.T., 2004. Simulation and optimal design of multiple-bed
pressure swing adsorption systems. AIChE J. 50, 2904–2917.
Qst average isosteric heat, J mol  1 Lee, C.-H., Yang, J.a dn, Ahn, H., 1999. Effects of carbon-to-zeolite ratio on layered
R gas constant, J mol  1 K  1 bed H2 PSA for coke oven gas. AIChE J. 45, 535–545.
COD coefficient of determination Lopes, F.V.S., Grande, C.A., Ribeiro, A.M., Loureiro, J.M., Rodrigues, A.E., 2010. Effect
of ion-exchange on the adsorption of SMR off-gases on zeolite 13X. J. Chem.
rc half diffusion length in micropores, m
Eng. Data 55, 184–195.
T temperature, K Meulenberg, W., Hansch, R., Buchkremer, H.P., Stöver, D., 2011. Device for gas
v0 interstitial velocity, m s  1 separation and method for producing such a system. U.S. Patent 8,016,924.
z axial coordinate, m Moellmer, J., Lange, M., Moeller, A., Patzschke, C., Stein, K., Laessig, D., Lincke, J.,
Glaeser, R., Krautscheid, H., Staudt, R., 2012. Pure and mixed gas adsorption of
CH4 and N2 on the metal-organic framework Basolites A100 and a novel
Greek symbols copper-based 1,2,4-triazolylisophthalate MOF. J. Mater. Chem. 22, 10274–10286.
Nikolic, D., Kikkinides, E.S., Georgiadis, M.C., 2009. Optimization of multibed
pressure swing adsorption processes. Ind. Eng. Chem. Res. 48, 5388–5398.
εp particle porosity NIST website. 〈http://webbook.nist.gov/chemistry〉 (accessed May 2014).
ρp particle density, kg m  3 Park, J.-H., Kim, J.-N., Cho, S-H., 2000. Performance analysis of a four-bed H2 PSA
process using layered beds. AIChE J. 46, 790–802.
Park, J.-H., Kim, J.-N., Cho, S.-H., Kim, J.-D., Yang, R.T., 1998. Adsorber dynamics and
optimal design of layered beds for multicomponent gas adsorption. Chem. Eng.
Sci. 53, 3951–3963.
Rarig, D.L., Golden, T.C., Weist, E.L., 2002. Purification of CO2 from H2 PSA vent gas.
Acknowledgements
In: National AIChE Meeting, Indianapolis, United States.
Reynolds, S.P., Ebner, A.D., Ritter, J.A., 2006. Stripping PSA cycles for CO2 recovery
Financial support from the Ministry of Economy and Competitive- from flue gas at high temperature using a hydrotalcite-like adsorbent. Ind. Eng.
Chem. Res. 45, 4278–4294.
ness of Spain through project CTQ2012-34626 and of the Research
Ruthven, D.M., Farooq, S., Knaebel, K.S., 1994. Pressure Swing Adsorption. VCH
Council of Norway through the RENERGI programme by the project Publishers, New York.
Novel Materials for Utilization of Natural Gas and Hydrogen (Grant Schell, J., Casas, N., Blom, R., Spjelkavik, A.I., Andersen, A., Hafizovic Cavka, J.,
190980/S60) are gratefully acknowledged. Dr. Águeda thanks the Mazzotti, M., 2012. MCM-41, MOF and UiO-67/MCM-41 adsorbents for pre-
combustion CO2 capture by PSA: adsorption equilibria. Adsorption 18, 213–227.
financial support received from Ministry of Education, Culture and Silva, B., Solomona, I., Ribeiro, A.M., Lee, U.H., Hwang, Y.K., Chang, J.S., Loureiro, J.M.,
Sport of Spain through the “José Castillejo” Programme. Rodrigues, A.E., 2013. H2 purification by pressure swing adsorption using
CuBTC. Sep. Purif. Technol. 118, 744–756.
Sircar, S., Golden, T.C., 2000. Purification of hydrogen by pressure swing adsorption.
Sep. Sci. Technol. 35, 667–687.
Sircar, S., Golden, T.C., 2010. Pressure swing adsorption technology for hydrogen
Appendix A. Supporting information production. In: Liu, Ke, Song, Chunshan, Subramani, Velu (Eds.), Hydrogen and
Syngas Production and Purification Technologies. Wiley, New Jersey.
Supplementary data associated with this article can be found in Stöcker, J., Whysall, M., Miller, G.Q., 1998. 30 Years of PSA Technology for Hydrogen
Purification. UOP LLC, Des Plaines, IL.
the online version at: http://dx.doi.org/10.1016/j.ces.2014.08.039. Suzuki, M., 1990. Adsorption Engineering. Kodansha, Tokyo.
Voss, C., 2005. Applications of pressure swing adsorption technology. Adsorption
11, 527–529.
References Waade, W., Farnand, S., Hutchison, R., Welch, K., 2012. CO2 capture from SMRs: a
demonstration project. Hydrocarbon Process., 63–68.
Waldron, W.E., Sircar, S., 2000. Parametric study of a pressure swing adsorption
Baksh, M.S.A., Simo, M., 2013. Ten bed pressure swing adsorption process operating process. Adsorption 6, 179–188.
in normal and turndown modes. U.S. Patent 8,435,328. Wang, Q.M., Shen, D., Bülow, M., Lau, M.L., Deng, S., Fitch, F.R., Lemcoff, N.O.,
Banu, A.-M., Friedrich, D., Brandani, S., Düren, T., 2013. A multiscale study of MOFs Semanscin, J., 2002. Metallo-organic molecular sieve for gas separation and
as adsorbents in H2 PSA purification. Ind. Eng. Chem. Res. 52, 9946–9957. purification. Microporous Mesoporous Mater. 55, 217–230.
Batta, L.B., 1971. Selective adsorption process. U.S. Patent 3,564,816. Willis, R., 2010. Carbon Dioxide Removal from Flue Gases Using Microporous Metal-
Beaver, M.G., Sircar, S., 2010. Decentralized production of hydrogen for residential PEM Organic Frameworks. Final Technical Report. Project contract: DE-FC26-
fuel cells from piped natural gas by low temperature steam-methane reforming 07NT43092. Department of Energy, United States.

Please cite this article as: Agueda, V.I., et al., Adsorption and diffusion of H2, N2, CO, CH4 and CO2 in UTSA-16 metal-organic
framework extrudates. Chem. Eng. Sci. (2014), http://dx.doi.org/10.1016/j.ces.2014.08.039i
 V.I. Agueda et al. / Chemical Engineering Science ∎ (∎∎∎∎) ∎∎∎–∎∎∎ 11

Xiang, S., Wu, X., Zhang, J., Fu, R., Hu, S., Zhang, X., 2005. A 3D canted Yang, S.I., Choi, D-Y., Jang, S-C., Kim, S-H., Choi, D-K, 2008. Hydrogen separation by
antiferromagnetic porous metal-organic framework with anatase topology multi-bed pressure swing adsorption of synthesis gas. Adsorption 14, 583–590.
through assembly of an analogue of polyoxometalate. J. Am. Chem. Soc. 127, Yoshida, M., Ritter, J.A., Kodama, A., Goto, M., Hirose, T., 2003. Enriching reflux and
16352–16353. parallel equalization PSA process for concentrating trace components in air.
Xiang, S., He, Y., Zhang, Z., Wu, H., Zhou, W., Krishna, R., Chen, B., 2012. Microporous Ind. Eng. Chem. Res. 42, 1795–1803.
metal-organic framework with potential for carbon dioxide capture at ambient
conditions. Nat. Commun. 3, 954.

Please cite this article as: Agueda, V.I., et al., Adsorption and diffusion of H2, N2, CO, CH4 and CO2 in UTSA-16 metal-organic
framework extrudates. Chem. Eng. Sci. (2014), http://dx.doi.org/10.1016/j.ces.2014.08.039i