See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/241014656

Adsorption equilibria of O2, Ar, Kr and Xe on activated carbon and zeolites:

single component and mixture data

Article in Adsorption · April 2011

DOI: 10.1007/s10450-011-9337-3

CITATIONS READS

88 1,451

5 authors, including:

Moises Bastos-Neto Andreas Moeller

Universidade Federal do Ceará 3P Instruments GmbH & Co. KG
95 PUBLICATIONS 1,150 CITATIONS 41 PUBLICATIONS 1,185 CITATIONS

SEE PROFILE SEE PROFILE

Frieder Dreisbach
TA Instruments
61 PUBLICATIONS 1,742 CITATIONS

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Gas separation/purification and storage View project

Capture and conversion of CO₂: from nanoscale to process project View project

All content following this page was uploaded by Moises Bastos-Neto on 30 March 2016.

The user has requested enhancement of the downloaded file.

Adsorption equilibria of O2, Ar, Kr and
Xe on activated carbon and zeolites:
single component and mixture data

Adsorption
Journal of the International
Adsorption Society

ISSN 0929-5607
Volume 17
Number 2

Adsorption (2011) 17:371-383

DOI 10.1007/
s10450-011-9337-3

1 23
Your article is protected by copyright and
all rights are held exclusively by Springer
Science+Business Media, LLC. This e-offprint
is for personal use only and shall not be self-
archived in electronic repositories. If you
wish to self-archive your work, please use the
accepted author’s version for posting to your
own website or your institution’s repository.
You may further deposit the accepted author’s
version on a funder’s repository at a funder’s
request, provided it is not made publicly
available until 12 months after publication.

1 23
 Author's personal copy
Adsorption (2011) 17: 371–383
DOI 10.1007/s10450-011-9337-3

Adsorption equilibria of O2 , Ar, Kr and Xe on activated carbon

and zeolites: single component and mixture data
R.E. Bazan · M. Bastos-Neto · A. Moeller ·
F. Dreisbach · R. Staudt

Received: 2 February 2010 / Accepted: 8 February 2011 / Published online: 23 February 2011
© Springer Science+Business Media, LLC 2011

Abstract This work provides a set of experimental data on separation process (The Linde Group 2009). Their natural
the adsorption of pure component, binary and ternary mix- abundances decrease as the atomic numbers increase and in-
tures on activated carbon sample and two different zeolites fluence directly their prices on the gas market. Considering
at 303 K and moderate pressures (up to 10 bar for mix- only the abundance and the difficulty in producing these no-
tures). Pure component data were measured by gravimetry ble gases, to give a notion of their prices on the market as an
and mixture data by volumetry coupled with chromatogra- example, Argon can be 5 times more expensive than dry Ni-
phy. Results encourage more research on new materials and trogen. Krypton is usually about 100 times more expensive
enhancement of adsorption-based separation processes with than Argon and Xenon 10 times more than Krypton. The
the proposed target. price per gram of Xenon can be comparable or even higher
than the price of Gold.
Keywords Adsorption · Carbon · Zeolite · Noble gas · During the last years adsorptive air separation processes
Oxygen · Argon · Krypton · Xenon · Separation have been increasingly applied with the purpose of improv-
ing the product purity with lower operational costs. How-
ever, the low efficiency of the adsorptive separation of Kr
1 Introduction and Xe from air prevents the production of these noble gases
simultaneously in adsorptive air separation units. The key
Noble gases have very specific properties, such as the lack aspect for the separation efficiency dwells in the adsorption
of chemical reactivity, low solubility, very low boiling and selectivity—which depends on the used adsorbent material.
melting points and very low conductivity. These proper- Literature reports regarding the adsorption of noble gases
ties are key factors in lighting, medical, space and many are scarce. Although some are concerned about the mater-
other applications, for instance: to provide inert atmosphere ial adsorption properties and selectivity (Greathouse et al.
for the production of air-sensitive compounds; in excimer 2009; Mueller et al. 2006), most of them are associated
lasers, which are commonly used in laser surgery and mi-
to other subjects like the study of structural and theoret-
crofabrication; and as filler gases for the production of high-
ical aspects of surface adsorption (Simonyan et al. 2001;
efficiency windows.
Da Silva and Stampfl 2008); the treatment of nuclear off-gas
Argon, Krypton and Xenon are obtained from air through
(Munakata et al. 2003) and trace analysis for geophysical
liquefaction, followed by fractional distillation—cryogenic
purposes (Marrocchi et al. 2005) for example. The purifica-
tion of noble gases is also a problem of industrial interest
R.E. Bazan · M. Bastos-Neto · A. Moeller · R. Staudt ()
and the development of an adsorption process for separation
Institut für Nichtklassische Chemie e.V., University of Leipzig, of O2 /Ar/Kr/Xe mixtures would increase productivity and
Permoserstr. 15, 04318 Leipzig, Germany the profit of adsorptive air separation units.
e-mail: [email protected] The objective of this work is to provide a set of experi-
F. Dreisbach mental data on adsorption of noble gases and Oxygen and to
Rubotherm GmbH, Universitätstr. 142, 44799 Bochum, Germany encourage further research and development on this topic.
 Author's personal copy
372 Adsorption (2011) 17: 371–383

Table 1 Textural properties of the adsorbent samples 4 Experimental

Property B3 4AK 13X-K2
4.1 Sample preparation
Specific surface area (m2 g−1 )a 778 c 479
Specific surface area (m2 g−1 )b 1012 477 c,d 640 The specific mass loss for each adsorbent material during
Total pore volume (cm3 g−1 ) 0.441 0.198 c,d 0.356 sample preparation was determined gravimetrically (Bazan
Average pore size (nm) 0.8 0.4 0.9 et al. 2008).
Form Granular Bead Bead The AC sample SorboNorit B3 (from 120 to 130 g) was
a ISO
regenerated at 373 K (2 K min−1 ) under vacuum (10−7 bar)
9277 (1995)
for 4 hours. The zeolites 13X-K2 and 4AK (∼230 g and
b Rouquerol et al. (1999) ∼150 g, respectively) were regenerated at 673 K (heating
rate of 1 K min−1 ) under vacuum (10−7 bar) for 6 hours.
c Couldnot be determined with N2 at 77 K due to the low kinetics on
the sample
d Evaluated with CO2 isotherm at 273.15 K (Lowell et al. 2006)
4.2 Sample characterization

Adsorption isotherms of O2 , Ar, Kr and Xe and some of The samples were analyzed for their textural characteristics
their mixtures on activated carbon and zeolite samples were by measuring N2 isotherms at 77 K. The data handling for
measured. Pure-component data were evaluated by gravime- the calculation of BET specific surface area, pore volume
try. Mixture data were obtained by volumetry coupled with and average pore size was accomplished in accordance with
chromatography. literature reports (ISO 9277 1995; Rouquerol et al. 1999).

4.3 Pure component data

2 Materials
The adsorption isotherms of pure components were mea-
sured gravimetrically using a high-precision magnetic sus-
Pure and mixture adsorption equilibrium of O2 , Ar, Kr and
pension balance (Rubotherm, Germany). After the sample
Xe were measured on three different samples at 30°C. One
preparation the temperature was set to 303 K and pure
activated carbon (AC) SorboNorit B3 provided by the com-
gas was introduced in the sample cell up to a given pres-
pany Norit (The Netherlands) and two zeolites: Koestrolith
sure. The sample weight was observed and the equilibrium
13X-K2 and Koestrolith 4AK supplied by CWK Chemiew-
state was marked when no mass variation was detected.
erk Bad Köstritz GmbH (Germany). These samples were
Each measurement at constant pressure corresponds to an
chosen due to their commercial availability and their differ-
experimental equilibrium point. The handling of the mea-
ent characteristics. The main textural properties of the car-
sured weight increase data in order to obtain excess adsorp-
bon sample are summarized in Table 1.
tion isotherms can be found elsewhere (Dreisbach 1998;
Herbst and Harting 2002; Dreisbach et al. 1999, 2002;
Keller and Staudt 2005; Staudt et al. 1997; Do and Do 2003;
3 Theory Bastos-Neto et al. 2005).

3.1 Models for pure and multicomponent adsorption 4.4 Mixture data

Pure component data were handled with the Langmuir Adsorption equilibrium data for mixtures were measured in
model in order to obtain parameters for further comparisons a volumetric device, as shown in Fig. 1. It consists basically
and to evaluate multicomponent adsorption (Do 1998). of two vessels with known volume (V1 ) used to store a spe-
The Ideal Adsorbed Solution Theory (IAST), originally cific amount of the gas (or gases) that is opened to the ad-
proposed by Myers and Prausnitz (1965), was used for sorption chamber, with volume V2 , which contains the ad-
the analysis of multicomponent equilibrium adsorption. The sorbent material in analysis (volume VS ). Temperature and
model does not require data for mixtures and is thermody- pressure conditions are controlled through sensors, whose
namically consistent. Applying IAST requires the calcula- characteristics are listed in Table 2.
tion of an integral equation to obtain the reduced spreading All internal volumes were precisely determined by ex-
pressure. The binary data were evaluated by a procedure de- pansion experiments. The density of pure gases was ob-
scribed by Do (1998) using the single component parameter tained according to Schmidt and Wagner (1985), Tegeler et
set fitted with Langmuir model. al. (1999) and Lemmon and Span (2006). The density of
 Author's personal copy
Adsorption (2011) 17: 371–383 373

Fig. 1 Scheme of the volumetric apparatus

Table 2 Sensors specifications After preparation, the gas mixture was expanded to the
Sensor Specification Accuracy adsorption chamber and, with the aid of the circulation
pump, the system was kept homogenized. The adsorption
Pressure 0.8 . . . 1, 2 bar abs. ±0.08% equilibrium was reached within 3 hours, when no pressure
0 → 323 K variation was detected (Bazan et al. 2008) and the composi-
0 . . . 25 bar ±0.03% tion of the gas phase, monitored with GC analyses, remained
263 → 353 K
constant after at least three measurements.
0 . . . 160 bar ±0.09%
263 → 353 K
4.5 Selectivity
Temperature 223 → 673 K ±0.1 K

GC-HP 5980 343 K ±0.05% (mol) The selectivity (SAB ) of the component A in a mixture A/B
Detector: TCD 303 → 323 K, was calculated from the ratios of gas mole fractions in the
373 K gas phase (y) and in the adsorbed phase (x) according to the
4.5 ml min−1
following equation (Greathouse et al. 2009):
xA yB
SAB = (1)
mixtures, which are also dependent on the pressure, tem- xB yA
perature and composition were evaluated using the Redlich-
Kwong-Soave equation of state (Sandler 1999). The binary
interaction coefficient was set zero and the (pseudo)critical 5 Results and discussion
temperature of the mixture was evaluated according to Reid
et al. (1987). This work is focused on the experimental determination of
The mixture preparation was carried out using the whole partial loadings from measurements of multicomponent ad-
volume V1 . The desired partial pressures were previously sorption employing a volumetric/chromatographic method
estimated for the given temperature, considering the volume following the guidelines reported by Talu (1998). The partial
needed for GC composition analysis. Then, each component loadings were determined directly through mass balances
was stepwise added to the system following the order of in- from an analysis of the composition of the gas mixture be-
creasing component concentration (Bazan et al. 2008). fore and after adsorption equilibrium.
 Author's personal copy
374 Adsorption (2011) 17: 371–383

5.1 Pure component face areas and pore volumes, the isotherm profiles are quali-
tatively the same for all three adsorbents. This indicates that
The obtained experimental data with the respective Lang- the adsorptives’ thermophysical properties play the major
muir fit are presented in Fig. 2, Tables 3 and 4. role on the equilibrium of adsorption. However it was de-
From those data it is possible to observe that even though tected that the zeolite Koestrolith 4AK adsorbs much less
the samples have different polarities, pore sizes, specific sur- than the other samples, certainly due to its low pore vol-
ume. Furthermore, the low kinetics related to the small pore
size leads to extremely long times for the equilibrium to be
reached.
In every case Xe was the most strongly adsorbed, fol-
lowed by Kr, then O2 and Ar. The similarity between O2
and Ar adsorption by each adsorbent is very remarkable.
For a better sample comparison Fig. 3 was created with
the plots of each gas using the Langmuir model fits at the
experimental pressure ranges.
Comparing the Langmuir fits with data reported on liter-
ature, it is possible to conclude that the results presented in
this work are in conformity with the order of magnitude.
One can note that the adsorption of Ar is also quantita-
tively very similar to the adsorption of O2 . For both gases
the AC SorboNorit B3 shows much better performance than
the zeolites. Additionally, the sample SorboNorit B3 pre-
sented adsorption capacity for Ar similar to the Ag-ETS-10
reported by Ansón et al. (2008) and the IRMOF-1 reported
by Greathouse et al. (2009) for pressures up to 10 bar. Ad-
sorption capacities of O2 on our zeolites are comparable to
literature data, which presented performance inferior to the
SorboNorit B3.
Data on Kr adsorption are also in good agreement with
literature reports. The 5A-CMS studied by Mahajan and
Walker (1969) showed however, a more improved adsorp-
tion capacity for Kr in comparison with our samples. AC
SorboNorit B3 and Koestrolith 13X-K2 compete to each
other for the best adsorption capacity. The carbon shows a
slightly better performance than the zeolite up to ca. 12 bar,
when both isotherms cross each other and the last one turns
to show higher adsorption capacity.
In the case of Xe, the zeolite Koestrolith 13X-K2 clearly
showed the best adsorption capacity followed by our car-
bon sample. The poor adsorption performance of the sam-
ple Koestrolith 4AK indicates that its textural character-
istics may not be favorable for the adsorption of this
gas.
Textural characteristics are known to play an important
role on the adsorption properties (Bastos-Neto et al. 2007).
By comparing only these characteristics displayed in Ta-
ble 1, it is possible to understand why the sample Koestrolith
4AK, for example, presented the least favorable results with
all gases since it is expected that lower pore volumes tend to
lead to inferior adsorption capacities. Additionally, the sam-
ple presents the smallest average pore size, which consists
Fig. 2 Adsorption isotherms of Ar, Kr, Xe and O2 on (a) Sor- of a barrier for the diffusion of relatively bigger molecules
boNorit B3; (b) Koestrolith 13X-K2 and (c) Koestrolith 4AK such as Xe.
 Author's personal copy
Adsorption (2011) 17: 371–383 375

Table 3 Experimental pure

component isotherm data Ar Kr O2 Xe
Pressure Adsorbed Pressure Adsorbed Pressure Adsorbed Pressure Adsorbed
[bar] amount [bar] amount [bar] amount [bar] amount
[mmol/g] [mmol/g] [mmol/g] [mmol/g]

SorboNorit B3
0.000 0.0000 0.000 0.000 0.000 0.000 0.000 0.000
2.222 0.5097 0.472 0.583 2.123 0.527 0.269 1.705
3.967 0.8468 0.949 0.985 3.934 0.888 0.485 2.340
7.057 1.3206 2.743 1.961 7.057 1.388 0.941 3.174
10.116 1.7091 3.910 2.404 9.966 1.756 1.367 3.721
13.235 2.0278 6.817 3.210 12.905 2.075 1.691 4.047
15.874 2.2595 9.760 3.782 15.904 2.384 2.510 4.654
19.563 2.5423 12.790 4.256 19.026 2.639 3.158 4.995
21.993 2.7059 15.820 4.600 21.777 2.834 4.003 5.335
26.701 3.0088 18.910 4.895 24.962 3.038
30.570 3.2005 22.660 5.175 28.774 3.252
24.820 5.298 31.494 3.392
28.060 5.467
30.160 5.562

Koestrolith 13X-K2
0.000 0.0000 0.000 0.0000 0.000 0.000 0.000 0.0000
0.102 0.0089 0.103 0.0596 0.943 0.086 0.283 1.5984
0.263 0.0223 0.260 0.1501 1.923 0.175 0.470 2.7185
0.503 0.0420 0.518 0.2980 2.952 0.264 0.899 4.8124
0.962 0.0802 0.962 0.5505 4.008 0.358 2.393 7.7858
2.603 0.2117 2.492 1.3843 5.535 0.479 4.117 8.7701
4.027 0.3205 4.072 2.1287 7.617 0.639 7.297 9.5284
6.967 0.5330 6.907 3.1875 11.514 0.910 10.296 9.8626
10.116 0.7372 10.236 4.0112 15.060 1.116
15.964 1.0537 15.994 4.8875 18.870 1.297
21.903 1.3083 22.143 5.4221 22.980 1.519
27.840 1.727

Koestrolith 4AK
0.000 0.0000 0.000 0.0000 0.000 0.0000 0.000 0.0000
0.643 0.0565 0.676 0.1994 0.613 0.0588 0.319 0.1132
0.988 0.0859 1.321 0.4228 1.027 0.0981 0.544 0.2550
2.485 0.2083 2.323 0.6257 2.422 0.2234 0.820 0.4448
4.084 0.3320 3.748 0.9180 3.982 0.3548 1.420 0.7363
6.976 0.5308 6.760 1.3552 6.955 0.5798 1.666 0.9048
9.490 0.6818 10.120 1.6762 10.000 0.7775 2.383 1.1646
12.640 0.8490 12.850 1.8449 12.940 0.9431 3.802 1.4658
15.880 0.9989 16.240 1.9985 16.450 1.1261
19.000 1.1228 18.970 2.0927 18.910 1.2330
22.510 1.2445 21.970 2.1779 22.120 1.3448
24.940 1.3196 25.240 2.2441 24.940 1.4302
27.880 1.4015 27.760 2.2897 27.790 1.5196
30.640 2.3317
 Author's personal copy
376 Adsorption (2011) 17: 371–383

Table 4 Langmuir fit

parameters for the single Sample Ar Kr O2 Xe
component isotherms qmax b qmax b qmax b qmax b
qmax : specific saturation
adsorption capacity in the SorboNorit B3 5.555 0.044 6.811 0.137 5.817 0.044 6.335 1.135
Langmuir isotherm [mmol g−1 ] Koestrolith 13X-K2 4.118 0.021 8.430 0.085 4.839 0.020 11.344 0.773
b: Langmuir adsorption constant Koestrolith 4AK 3.106 0.030 2.981 0.122 3.308 0.031 3.918 0.165
[bar−1 ]

Fig. 3 Adsorption isotherms on the three samples. (a) Argon; (b) Oxygen; (c) Krypton (Jee et al. 2002) and (d) Xenon (Fomkin et al. 1975)

It is important to note that all pure component isotherms Such behavior is also observed from the data reported in Ta-
presented Type-I feature, which is typical for the adsorption ble 1.
mechanism of supercritical gases in low pressure ranges. Intriguingly, with only those arguments it is not possi-
The sequence of adsorption preference and capacity of the ble to elucidate the competition between Koestrolith 13X-
studied gases for all adsorbents is directly proportional to K2 and SorboNorit B3. Their average pore sizes are practi-
the sequence of the critical temperatures of the gases. Re- cally the same and apparently play no important role on the
ferring to Table 5, one can conclude that the higher the differential adsorption. The carbon has higher specific sur-
critical temperature, the larger the amount adsorbed inde- face area and total pore volume and this explains the higher
pendently on the studied adsorbent material. Another ex- uptakes of Ar and O2 . With Kr and Xe, for which the exper-
pected consequence of working with adsorption of supercrit- imental conditions are closer to the critical temperature, an-
ical fluids is that for a given gas, the larger the specific sur- other factor significantly interferes on the adsorption affin-
face area, the higher the adsorption capacities are reached. ity of the zeolite. Carbons are known to be apolar as zeo-
 Author's personal copy
Adsorption (2011) 17: 371–383 377

Table 5 Some properties of the used gases lites polar and the reason for the overcoming performance
of Koestrolith 13X-K2 might be attributed to an electrosta-
Gas Van-der Heat of Boiling Critical Polarizability
tic related factor.
Waals radiia evaporation point temperature α/10−30
[nm] [kJ mol−1 ] [K] [K] [m3 ] The polarizabilities (Table 5) from Ar and O2 differ about
6% from each other and are not supposed to considerably
O2 0.152 6.8b 90.2 154.6 1.541c influence the adsorption by all samples. This difference is
Ar 0.188 6.447 87.3 150.8 1.642d raised to more than 50% for Kr, comparing with Ar and
Kr 0.202 9.029 119.9 209.4 2.486d O2 . As a result of the pressure (and consequently density)
Xe 0.216 12.636 165.1 289.7 4.047d increase, a growing number of Kr molecules would be af-
fected by the higher polarity of the zeolite surface turn-
a Bondi (1964) ing them more susceptible to polarization. Taking into ac-
b Dobrinski et al. (2003) count the affinity of species from the same nature (polar
c Hettema et al. (1994) or apolar), those polarized molecules would be more eas-
d Pershina ily attracted by the zeolite rather than by the carbon sam-
et al. (2008)
ple. Xe has a polarizability value about 60% higher than
Kr, therefore its probability to be attracted by a polar sur-

Table 6 Binary adsorption data

on SorboNorit B3 Mixture yA xA nA N P SAB
exp IAST exp IAST exp IAST

A: Kr 0.055 0.224 0.188 0.429 0.322 1.917 1.710 8.65 4.95

B: Ar 0.030 0.173 0.110 0.318 0.178 1.838 1.622 8.61 6.77
A: Xe 0.014 0.331 0.307 0.675 0.585 2.040 1.902 8.34 34.44
B: Ar 0.006 0.180 0.160 0.337 0.271 1.875 1.698 8.52 36.33
A: Xe 0.021 0.167 0.140 0.542 0.475 3.245 3.391 6.40 9.49
B: Kr 0.008 0.070 0.057 0.224 0.189 3.214 3.298 6.52 9.84
A: O2 0.905 0.904 0.910 1.485 1.429 1.714 1.570 8.48 1.00
P : total pressure [bar] B: Ar 0.950 0.952 0.953 1.616 1.523 1.697 1.597 8.66 1.05
N : total adsorbed amount A: Kr 0.060 0.221 0.195 0.423 0.345 1.912 1.775 8.53 4.48
[mmol g−1 ], nA : partial B: O2 0.025 0.128 0.088 0.233 0.148 1.816 1.671 8.58 5.75
adsorbed amount [mmol g−1 ]
exp: experimental data A: Xe 0.011 0.294 0.245 0.577 0.459 1.965 1.871 8.25 37.39
B: O2 0.006 0.168 0.153 0.311 0.274 1.853 1.789 8.67 36.54

Table 7 Binary adsorption data

on Koestrolith 13X-K2 Mixture yA xA nA N P SAB
exp IAST exp IAST exp IAST

A: Kr 0.060 0.274 0.351 0.091 0.135 0.333 0.386 3.33 5.91

B: Ar 0.030 0.146 0.207 0.046 0.069 0.317 0.335 3.38 5.54
A: Xe 0.020 0.371 0.805 0.154 0.594 0.413 0.738 3.22 28.94
B: Ar 0.010 0.216 0.560 0.077 0.299 0.359 0.534 3.36 27.21
A: Xe 0.035 0.188 0.325 0.141 0.600 0.747 1.847 2.32 6.42
B: Kr 0.016 0.101 0.181 0.075 0.307 0.742 1.697 2.49 6.74
A: O2 0.900 0.901 0.911 0.284 0.275 0.315 0.301 3.38 1.01
B: Ar 0.949 0.951 0.957 0.302 0.293 0.317 0.306 3.42 1.04

P : total pressure [bar] A: Kr 0.064 0.245 0.349 0.087 0.146 0.354 0.418 3.33 4.73
N : total adsorbed amount B: O2 0.032 0.128 0.205 0.044 0.075 0.340 0.365 3.39 4.42
[mmol g−1 ], nA : partial A: Xe 0.023 0.349 0.729 0.149 0.609 0.427 0.836 3.17 22.76
adsorbed amount [mmol g−1 ]
B: O2 0.012 0.200 0.573 0.074 0.348 0.369 0.608 3.34 20.32
exp: experimental data
 Author's personal copy
378 Adsorption (2011) 17: 371–383

Table 8 Binary adsorption data

on Koestrolith 4AK Mixture yA xA nA N P SAB
exp IAST exp IAST exp IAST

A: Kr 0.075 0.259 0.243 0.234 0.197 0.903 0.813 9.81 4.31

B: Ar 0.040 0.161 0.141 0.131 0.108 0.809 0.761 9.80 4.62
A: Xe 0.075 0.248 0.374 0.212 0.355 0.855 0.951 9.92 4.07
B: Ar 0.028 0.171 0.173 0.143 0.138 0.836 0.800 9.89 7.17

P : total pressure [bar] A: O2 0.900 0.900 0.910 0.739 0.696 0.821 0.764 9.85 1.00
N : total adsorbed amount B: Ar 0.950 0.950 0.956 0.770 0.738 0.811 0.772 9.92 1.00
[mmol g−1 ], nA : partial A: Kr 0.068 0.224 0.205 0.213 0.176 0.949 0.858 9.73 3.96
adsorbed amount [mmol g−1 ]
B: O2 0.035 0.139 0.114 0.120 0.092 0.865 0.811 9.70 4.44
exp: experimental data

face is much higher and that can be the reason for the better
adsorption by Koestrolith 13X-K2 in comparison with the
SorboNorit B3.

5.2 Mixtures

All measured binary data are summarized in Table 6 (Sor-

boNorit B3), Table 7 (Koestrolith 13X-K2) and Table 8
(Koestrolith 4AK). The experimental error for mixture data
is expected to be about 5% (±2.5%) and depends mainly on
the determination of gas phase composition (chromatogra-
phy).
The binary mixture Ar/O2 deserves a special attention
due to the technological challenges involved on the sepa-
ration of both species. The similarity between the adsorp-
tion of O2 and Ar could be verified before. Additionally, in
Table 5, we can see that argon and oxygen have other sim-
ilar properties like molecular size, boiling point as well as
adsorption affinity, which make them very hard to be sep-
arated. Developments on this subject are desirable for both
oxygen purification and argon production processes.
A graph of the Ar/O2 system was plotted in Fig. 4 for the
zeolite Koestrolith 13X-K2. The experimental data at two
different equilibrium concentrations are compared with the
curve generated by the IAST model at an average value of
the experimental pressure conditions, using the Langmuir
fit parameters for pure component equilibrium. Figure 4(a)
presents the amount adsorbed against the Argon gas phase
equilibrium concentration. The similarity is here confirmed
by the symmetry of the lines. Good agreement between
model and experiment can be observed. Figure 4(b) exhibits
the x–y diagram for the same system, indicating almost lin-
ear relationship between gas-phase and adsorbed-phase con-
centrations in equilibrium. The selectivity for O2 in this sys- Fig. 4 Binary adsorption of Argon and Oxygen. (a) Adsorbed amount
versus Ar concentration in the gas phase and (b) x–y diagram
tem was 1.01 at yAr = 0.1 and 1.04 at yAr = 0.05, which
indicates slightly preferable adsorption of Oxygen over Ar-
gon. trolled process should be more feasible and not energy-
Rege and Yang (2000) brought up the problem of sepa- intensive like cryogenic distillation/adsorption. Jin et al.
rating Ar/O2 by PSA and concluded that a kinetically con- (2006) made a further investigation on kinetically controlled
 Author's personal copy
Adsorption (2011) 17: 371–383 379

Fig. 5 Binary adsorption of Argon and Krypton. (a) Adsorbed amount Fig. 6 Binary adsorption of Krypton and Xenon. (a) Adsorbed amount
versus Kr concentration in the gas phase and (b) x–y diagram versus Xe concentration in the gas phase and (b) x–y diagram

Oxygen-Argon separation for Argon production with differ-

tal amount adsorbed increases with the concentration of Kr
ent adsorbents and found out that such process is signifi-
cantly affected by the operating conditions. on the gas phase. Satisfactory matching is also found be-
Ansón et al. (2008) carried out adsorption measurements tween experimental data and IAST predictions, especially
of Ar, O2 and N2 in silver exchanged zeolites, which demon- for Ar. The slight discrepancies are attributed to the IAST
strated selectivity for Argon over Oxygen. The authors sug- input data of the pure component equilibrium, which corre-
gested that those adsorbents could be useful for the produc- sponds to the Langmuir fit parameters. In this case, one can
tion of high purity Oxygen (>99%) from a previously en- infer that the description of the experimental results for Kr
riched Oxygen stream (from PSA air separation) containing by the Langmuir model is inadequate.
approximately 95% O2 and 5% Ar. The binary mixture Kr/Xe in SorboNorit B3 is presented
Either by kinetic or equilibrium controlled processes, this in Fig. 6. The relevance of studying this mixture is regarded
matter should be dealt with more research and development by the high commercial value of xenon and its increasing de-
on new materials capable of efficiently make use of the small mand for current applications. Figure 6(a) shows also good
differences on the physical properties of both species to po- agreement between model (IAST) and experimental data.
tentialize the separation. The total amount adsorbed tends to increase with the Xe
Similar plots for the mixture Ar/Kr on the zeolite Koe- concentration. The x–y diagram in Fig. 6(b) reveals the big-
strolith 13X-K2 are presented on Fig. 5. The preferential ger affinity for Xe in comparison with Kr. The selectivity
adsorption for Kr may be verified in both plots and the to- for Xe reaches 9.49 at yXe = 0.02 and 9.84 at yXe = 0.008.
 Author's personal copy
380 Adsorption (2011) 17: 371–383

Fig. 7 Binary adsorption of Oxygen and Xenon. (a) Adsorbed amount Fig. 8 Binary adsorption of Krypton and Oxygen. (a) Adsorbed
versus O2 concentration in the gas phase and (b) x–y diagram amount versus Kr concentration in the gas phase and (b) x–y diagram

dencing imprecision on the description of the pure compo-

Those values are notably larger than the ones calculated by nent isotherms on the Koestrolith 4AK.
Greathouse et al. (2009) with GCMC simulations of gas ad- Further comparisons between experimental data and
sorption by IRMOF-1. IAST predictions for the binary mixtures might be observed
As mentioned before, both species Kr and Xe are very directly from Tables 6–8.
important for many applications today. Several materials The data corresponding to the ternary mixtures are pre-
could fulfill the requirements to separate them as they can be sented in Table 9 for the AC sample and Table 10 for the
more easily adsorbed. The problem, in the viewpoint of pro- zeolite Koestrolith 4AK. For a better visualization of the ad-
duction, is their very low concentration in the main source sorption equilibrium, Figs. 9 and 10 were plotted for Sor-
(air). Therefore, adsorption plays important role on the de- boNorit B3 and Fig. 11 for Koestrolith 4AK.
velopment of energy-efficient processes. The ternary diagrams indicate the direction, in which
Other mixtures of industrial interest are the Xe/O2 and there is preferable adsorption. Each point in those graphs
Kr/O2 . The corresponding binary data are plotted in Fig. 7 represents the coordinates of the concentration of each
for the AC SorboNorit B3 and Fig. 8 for the zeolite Koe- species in a given mixture. Circles are used for the concen-
strolith 4AK, respectively. The plots reveal the preferential tration in the gas phase and triangles for the concentration
adsorption of Xe and Kr over O2 . IAST predictions were in the adsorbed phase. Stars indicate the IAST predictions
very accurate for the first mixture but less for the last, evi- for the concentration in the adsorbed phase. The line that
 Author's personal copy
Adsorption (2011) 17: 371–383 381

Table 9 Ternary adsorption

data on SorboNorit B3 Mixture i yi xi ni N P
exp IAST exp IAST exp IAST

O2 /Ar/Kr Kr 0.030 0.111 0.105 0.206 0.176 1.849 1.682 8.59

Ar 0.100 0.100 0.088 0.185 0.148
O2 0.870 0.789 0.807 1.459 1.359

O2 /Ar/Xe Xe 0.007 0.176 0.169 0.331 0.298 1.886 1.766 8.35

Ar 0.109 0.074 0.086 0.139 0.153
O2 0.884 0.751 0.745 1.416 1.315

O2 /Kr/Xe Xe 0.006 0.142 0.125 0.289 0.245 2.038 1.963 8.33

Kr 0.076 0.215 0.209 0.439 0.409
O2 0.918 0.643 0.666 1.309 1.308
P : total pressure [bar]
N : total adsorbed amount Ar/Kr/Xe Xe 0.007 0.153 0.150 0.307 0.289 2.008 1.931 8.42
[mmol g−1 ], ni : partial adsorbed Kr 0.072 0.222 0.203 0.446 0.392
amount [mmol g−1 ]
Ar 0.921 0.625 0.647 1.256 1.251
exp: experimental data

Table 10 Ternary adsorption

data on Koestrolith 4AK Mixture i yi xi ni N P
exp IAST exp IAST exp IAST

O2 /Ar/Kr Kr 0.038 0.136 0.120 0.108 0.098 0.797 0.811 9.77

Ar 0.112 0.058 0.093 0.046 0.075
O2 0.851 0.806 0.787 0.643 0.638
P : total pressure [bar]
N : total adsorbed amount O2 /Ar/Xe Xe 0.035 0.150 0.188 0.126 0.163 0.837 0.870 9.80
[mmol g−1 ], ni : partial adsorbed Ar 0.108 0.048 0.082 0.040 0.071
amount [mmol g−1 ]
O2 0.858 0.801 0.730 0.671 0.636
exp: experimental data

connects a circle (origin) to a triangle (end) symbolizes the

equilibrium of adsorption and points toward the species with
more adsorption affinity. The longer the line is, the higher
the affinity.
The blank symbols representing the mixture Ar/Kr/Xe in
Fig. 9 show the AC sample’s high preferential adsorption of
Xe and less significantly of Kr as their concentrations in the
adsorbed phase are considerably increased when compared
to the gas phase while the concentration of Ar is notably
reduced. The filled symbols corresponding to the mixture
Ar/O2 /Kr in the same plot also shows a small trend for O2
and even more for Kr. As visible in Table 9, the concen-
tration of Ar is practically the same in the gas and in the ad-
sorbed phase, revealing its weaker adsorption by SorboNorit
B3. Figure 10 also displays the preferential adsorption for
Xe in both mixtures, confirming previous comments about
its properties. In both figures, the matching between experi-
ments and IAST predictions was exceptionally good.
Similar comments can be used for the mixtures presented
in Fig. 11. In both cases there is a “decrease in the con-
centration” of Ar and O2 as Kr or Xe exhibit better ad- Fig. 9 Ternary diagram for adsorption of Ar/O2 /Kr and Ar/Kr/Xe on
sorption capacities, although O2 shows a slighter reduction. AC SorboNorit B3 at 303 K
 Author's personal copy
382 Adsorption (2011) 17: 371–383

Fig. 10 Ternary diagram for adsorption of O2 /Ar/Xe and O2 /Kr/Xe

Fig. 11 Ternary diagram for adsorption of O2 /Ar/Kr and O2 /Ar/Xe on
on AC SorboNorit B3 at 303 K
zeolite Koestrolith 4AK at 303 K

IAST predictions were somehow inaccurate, especially for

polarization of Kr molecules, achieving higher adsorption
the mixture containing Xe. This might be again an indication
capacities. Koestrolith 13X-K2 was also the best sample for
of unsatisfactory description of the isotherms data by the
adsorption of Xe in this study. This could be explained by
Langmuir model, since only a small amount of experimental
the sample’s favorable textural characteristics and the polar-
points for xenon on Koestrolith 4AK are available (Fig. 2).
Those experimental ternary data meet the expectations izability of Xe. The zeolite Koestrolith 4AK presented much
that one can make about the selectivity of the gases just smaller pore sizes and lower specific surface area compared
observing the single component adsorption isotherms. This to the other samples, which explain the lower adsorption ca-
observation implies that no significant interactions between pacities.
species are found in the studied systems. Experiments on binary mixtures confirmed the challenge
of separating Ar/O2 due to their very similar physical-
chemical properties. Binary measurements with Kr/Xe have
6 Conclusions shown better results than simulations with new materials re-
ported in literature, but more investigations on new materials
This work has presented series of experimental data on sin- are also desirable to improve the selectivity.
gle component, binary and ternary adsorption equilibria of Ternary data confirmed the preferential adsorption of Xe
noble gases (Ar, Kr and Xe) and O2 in one activated carbon over the other gases as well as of Kr over O2 and Ar. Those
sample (SorboNorit B3) and two zeolites (Koestrolith 13X- experiments are also evidence for the absence of significant
K2 and Koestrolith 4AK). All of the samples have shown interactions between molecules of the different gases on the
qualitatively similar adsorption capacities. Xe was the most adsorption of their mixtures.
adsorbed gas, followed by Kr, then O2 and Ar, which corre- Predictions with the IAST model were generally satis-
sponds to the expected behavior of the adsorption of super- factory. The mismatching between experiment and model
critical fluids at lower pressures. The adsorption capacity of prediction found in some cases are attributed to inadequate
all studied gases was found to be proportional to their criti- description pure component data by the Langmuir isotherm
cal temperatures. model.
AC SorboNorit B3 revealed higher affinity with Ar and
O2 in comparison with the zeolites. In the case of Kr, the Acknowledgements The authors acknowledge financial support
from Conselho Nacional de Desenvolvimento Científico e Tecnológico
carbon exhibits higher adsorbed amounts at pressures up to
—CNPq (Brazil), Deutscher Akademischer Austausch Dienst—DAAD
ca. 12 bar. For higher pressures the higher polarity of ze- (Germany), Chemiewerk Bad Koestritz—CWK (Germany) and The
olite Koestrolith 13X-K2 surface appears to influence the Norit Group (The Netherlands).
 Author's personal copy
Adsorption (2011) 17: 371–383 383

References Jin, X., Malek, A., Farooq, S.: Production of argon from an oxygen-
argon mixture by pressure swing adsorption. Ind. Eng. Chem. Res.
Ansón, A., Kuznicki, S.M., Kuznicki, T., Haastrup, T., Wang, Y., Lin, 45, 5775–5787 (2006)
C.C.H., Sawada, J.A., Eyring, E.M., Hunter, D.: Adsorption of ar- Keller, J., Staudt, R.: Gas Adsorption Equilibria: Experimental Meth-
gon, oxygen, and nitrogen on silver exchanged ETS-10 molecular ods and Adsorption Isotherms. Springer, New York (2005). ISBN:
sieve. Microporous Mesoporous Mater., 109, 577–580 (2008) 0-387-23597-3
Bastos-Neto, M., Torres, A.E.B., Azevedo, D.C.S., Cavalcante, C.L., Lemmon, E.W., Span, R.: Short fundamental equations of state for 20
Jr.: Methane adsorption storage using microporous carbons ob- industrial fluids. J. Chem. Eng. Data, 51(3), 785–850 (2006)
tained from coconut shells. Adsorpt., J. Int. Adsorpt. Soc. 11, Lowell, S., Shields, J.E., Thomas, M.A., Thommes, M.: Characteriza-
911–915 (2005) tion of Porous Solids and Powders: Surface Area, Porosity and
Bastos-Neto, M., Canabrava, D.V., Torres, A.E.B., Rodriguez- Density. Particle Technology Series. Springer, The Netherlands
Castellón, E., Jiménez-López, A., Azevedo, D.C.S., Cavalcante, (2006). ISBN: 1-4020-2302-2
C.L., Jr.: Effects of textural and surface characteristics of micro-
Mahajan, O.P., Walker, P.L., Jr.: Krypton adsorption on microporous
porous activated carbons on the methane adsorption capacity at
carbons and 5A zeolite. J. Colloid Interface Sci., 29(1), 129–137
high pressures. Appl. Surf. Sci. 253, 5721–5725 (2007)
Bazan, R.E., Bastos-Neto, M., Staudt, R., Papp, H., Azevedo, D.C.S., (1969)
Cavalcante, C.L., Jr.: Adsorption equilibria of natural gas compo- Marrocchi, Y., Razafitianamaharavo, A., Michot, L.J., Marty, B.: Low-
nents on activated carbon: pure and mixed gas isotherms. Adsorp. pressure adsorption of Ar, Kr, and Xe on carbonaceous materials
Sci. Technol., 26(5), 323–332 (2008) (kerogen and carbon blacks), ferrihydrite, and montmorillonite:
Bondi, A.: van der Waals volumes and radii. J. Phys. Chem., 68(3), Implications for the trapping of noble gases onto meteoritic mat-
441–451 (1964) ter. Geochim. Cosmochim. Acta, 69(9), 2419–2430 (2005)
Da Silva, J.L.F., Stampfl, C.: Trends in √ adsorption
√ of noble gases Mueller, U., Schubert, M., Teich, F., Puetter, H., Schierle-Arndt, K.,
He, Ne, Ar, Kr, and Xe on Pd(111)( 3X 3)R30◦ : all-electron Pastre, J.: Metal-organic frameworks—prospective industrial ap-
density-functional calculations. Phys. Rev. B, 77, 1–13 (2008), plications. J. Mater. Chem. 16, 626 (2006)
045401 Munakata, K., Kanjo, S., Yamatsuki, S., Koga, A., Ianovski, D.: Ad-
Do, D.D.: Adsorption Analysis: Equilibria and Kinetics. Series on sorption of noble gases on silver-mordenite. J. Nucl. Sci. Tech-
Chemical Engineering, vol. 2. Imperial College Press, London nol., 40(9), 695–697 (2003)
(1998). ISBN: 1-86094-130-3 Myers, A.L., Prausnitz, J.M.: Thermodynamics of mixed-gas adsorp-
Do, D.D., Do, H.D.: Adsorption of supercritical fluids in non-porous tion. AIChE J., 11(1), 121–126 (1965)
and porous carbons: analysis of adsorbed phase volume and den-
Pershina, V., Borschevsky, A., Eliav, E., Kaldor, U.: Adsorption of
sity. Carbon 41(9), 1777–1791 (2003)
inert gases including element 118 on noble metal and inert sur-
Dobrinski, P., Krakau, G., Vogel, A.: Physik für Ingenieure.
faces from ab initio Dirac–Coulomb atomic calculations. J. Chem.
Teubner, Leipzig (2003). ISBN-10: 3-519-46501-9, ISBN-13:
Phys. 129, 144106 (2008)
97835194650110
Dreisbach, F.: Untersuchung von Adsorptionsgleichgewichten Reid, R.C., Prausnitz, J.M., Poling, B.E.: The Properties of Gases and
methanhaltiger Gasgemische an Aktivkohle als Grundlage Liquids, 4th edn. McGraw–Hill, New York (1987). ISBN: 0-07-
zur Auslegung technischer Adsorptionsanlagen. PhD Thesis, 051799-1
Fortsch.—Ber. VDI, Reihe 3 Nr. 547. VDI Verlag, Düsseldorf Rege, S.U., Yang, R.T.: Kinetic separation of oxygen and argon using
(1998) molecular sieve carbon. Adsorpt., J. Int. Adsorpt. Soc. 6, 15–22
Dreisbach, F., Staudt, R., Keller, J.U.: High pressure adsorption data (2000)
of methane, nitrogen, carbon dioxide and their binary and ternary Rouquerol, F., Rouquerol, J., Sing, K.: Adsorption by Powders &
mixtures on activated carbon. Adsorpt., J. Int. Adsorpt. Soc. 5, Porous Solids. Academic Press, San Diego (1999). ISBN: 0-12-
215–227 (1999) 598920-2
Dreisbach, F., Lösch, H.W., Harting, P.: Highest pressure adsorption Sandler, S.I.: Chemical and Engineering Thermodynamics., 3rd edn.
equilibria data: measurement with magnetic suspension balance Willey, New York (1999). ISBN: 0-471-18210-9
and analysis with a new adsorbent/adsorbate-volume. Adsorpt., J. Schmidt, R., Wagner, W.: A new form of the equation of state for pure
Int. Adsorpt. Soc., 8(2), 95–109 (2002) substances and its application to oxygen. Fluid Phase Equilib. 19,
Fomkin, A.A., Serpinskii, V.V., Bering, B.P.: Investigation of the ad- 175–200 (1985)
sorption of xenon on NaX zeolite within a broad range of pres-
Simonyan, V.V., Johnson, J.K., Kuznetsova, A., Yates, J.T., Jr.: Mole-
sures and temperatures. Russian Chemical Bulletin. Translated
cular simulation of xenon adsorption on single-walled. J. Chem.
from Izvestiya Akademii Nauk SSSR, Seriya Khimieheskaya 6,
Phys., 114(9), 4180–4185 (2001)
1244–1248 (1975)
Greathouse, J.A., Kinnibrugh, T.L., Allendorf, M.D.: Adsorption and Staudt, R., Bohn, S., Dreisbach, F., Keller, J.U.: Gravimetric and volu-
separation of noble gases by IRMOF-1: grand canonical Monte metric measurements of helium adsorption equilibria on different
Carlo simulations. Ind. Eng. Chem. Res., 48(7), 3425–3431 porous solids. In: McEnaney, B., et al. (eds.) Characterization of
(2009) Porous Solids 4, pp. 261–266. Royal Society of Chemistry, Lon-
Herbst, A., Harting, P.: Thermodynamic description of excess don (1997)
isotherms in high-pressure adsorption of methane, Argon and ni- Talu, O.: Needs, status, techniques and problems with binary gas ad-
trogen. Adsorpt., J. Int. Adsorpt. Soc. 8, 111–123 (2002) sorption experiments. Adv. Colloid Interface Sci. 76–77, 227–269
Hettema, H., Wormer, P.E.S., Jørgensen, P., Jensen, H.J.Aa., Hel- (1998)
gaker, T.: Frequency-dependent polarizabilities of O2 and van Tegeler, Ch., Span, R., Wagner, W.: A new equation of state for argon
der Waals coefficients of dimers containing O2 . J. Chem. Phys., covering the fluid region for temperatures from the melting line
100(2), 1297–1302 (1994) to 700 K at pressures up to 1000 MPa. J. Phys. Chem. Ref. Data,
ISO 9277: Determination of the specific surface area of solids by gas 28(3), 779–850 (1999)
adsorption using the BET method (1995) The Linde Group: Cryogenic Air Separation: History and Techno-
Jee, J.G., Park, M.K., Yoo, H.K., Lee, K., Lee, C.H.: Adsorption and logical Progress. http://linde-anlagenbau.de/process_plants/air_
desorption characteristics of air on zeolite 5a, 10x, and 13x fixed separation_plants/documents/L_2_1_e_09_150dpi.pdf. Accessed
beds. Sep. Sci. Technol., 37(15), 3465–3490 (2002) 25 November 2009

View publication stats