pubs.acs.

org/IECR Article

Measurements and Modeling of Water Adsorption Isotherms of

Zeolite Linde-Type A Crystals
Yu Wang*
Cite This: https://dx.doi.org/10.1021/acs.iecr.9b06891 Read Online

ACCESS Metrics & More Article Recommendations *

sı Supporting Information
See https://pubs.acs.org/sharingguidelines for options on how to legitimately share published articles.

ABSTRACT: A systematic study of water adsorption isotherms of zeolite 3A

and 4A crystals over a pressure range of 10−5 to 10−1 bar and temperatures
Downloaded via UNIV OF NEW SOUTH WALES on April 17, 2020 at 15:17:43 (UTC).

from 25 to 250 °C was carried out. This work focuses on water isotherms of
3A and 4A zeolite crystals, not pelletized zeolites containing binders. Zeolite
3A crystals show approximately 12% less swing working capacity for partial
pressure swing processes at room temperature for deep dehydration, when
compared to 4A crystals. However, the 3A crystal can be regenerated at a
lower temperature than the 4A crystal. The 4A crystal has a steeper water
isotherm and higher heat of adsorption at zero coverage than 3A. Our results
on 3A and 4A crystals are compared to literature data for pelletized 3A and
4A zeolites which contain binders. A new triple-site Langmuir model has
been proposed and provides a satisfactory description when compared to
other models, such as the Toth model and dual-site Langmuir model. The
triple-site Langmuir model is consistent with the Linde-Type A crystal having three different sites for water adsorption. The
improvement in fitting is significant in the region where water loadings approach a maximum. This new model was also successfully
applied to existing literature data sets for the water isotherm at multiple temperatures for 3A and 4A beads. These results provide
fundamental water isotherms and a simple model to evaluate 3A and 4A in deep dehydration with temperature-swing, or partial
pressure-swing adsorption processes, or their combination using nontraditional monolith contactors.

■ INTRODUCTION
Adsorption processes are ubiquitous in nature as well as of great
2.8 Å, and excludes larger molecules, such as alcohols and heavy
hydrocarbons. In the late 1980s a large scale pressure swing
technological importance for gas separation and purification. adsorption process utilizing zeolite 3A for dehydration of fuel
The oil and gas industry uses adsorption dehydration units to grade ethanol was developed to replace the azeotropic
reduce water content to very low (subppm) levels upstream of distillation process. In addition to dehydration, 3A has been
liquefied natural gas (LNG) plants and deep natural gas liquid used for hydrogen gas in acetylene conversion, and alkylation
(NGL) extraction plants. The gas temperatures are reduced to and isomerization of feedstock in the oil refining and
less than −160 °C for LNG and −100 °C for NGL, respectively. petrochemical industries.
Because of these extremely low temperatures, thorough removal
Only a few water adsorption isotherm data sets for 3A are
of water is essential to prevent freezing in cryogenic operations.
Some other applications for which deep dehydration is reported in the literature.2−9 Most well accepted data are
important include feed stream drying in air-separation plants published by commercial manufacturers such as Union Carbide
operating at −195 °C, and in plants extracting helium from and Grace Davison for their products. Recently Lin et al.
natural gas at processing temperatures around −170 °C. measured water adsorption on the UOP 3A beads at 25, 40, 60,
Molecular sieves or zeolites are the preferred adsorbent of and 80 °C using a continuous-flow adsorption system. The K+/
choice for natural gas dehydration in cryogenic processing Na+ molar ratio was approximately 2:3. The data, as well as data
standards (less than 0.1 ppmv water or −101 °C dew point).1 from Grace Davison, have been accurately described with the
Zeolite A, typically 4A, is the most commonly used molecular generalized statistical thermodynamic adsorption (GSTA)
sieve employed for dehydration of gases because of its strong
adsorption of water at low concentration levels.
In some dehydration processes, molecular sieve 3A is Received: December 17, 2019
preferred over 4A for minimizing coadsorption of CO2 and Revised: April 1, 2020
H2S, and small hydrocarbon molecules. This is due to the Accepted: April 6, 2020
characteristically narrow structural pore openings of 3A, which is Published: April 6, 2020
approximately 3 Å. 3A only allows adsorption of smaller
molecules, such as water with an average dynamic diameter of

© XXXX American Chemical Society https://dx.doi.org/10.1021/acs.iecr.9b06891

A Ind. Eng. Chem. Res. XXXX, XXX, XXX−XXX
Industrial & Engineering Chemistry Research pubs.acs.org/IECR Article

isotherm model with eight parameters, suggesting adsorbent clays used as binders have larger pores and therefore negligible
beads having four energetically distinct adsorption sites.5 water capacity. Given the manufacturers use of binders of
Equilibrium isotherms of water and ethanol vapors have been different type and composition, the reported water capacities
measured and compared to show that 3A zeolite adsorbed a could be intrinsically different in the formulated samples
significant amount of water while very weak ethanol adsorption procured from different manufacturers.
was observed.6 Adsorption breakthrough runs by Simo et al. To exclude the effects of binders on water adsorption, one
were used to obtain equilibrium at 100, 145, 167, and 200 °C for must measure adsorption on zeolite crystals without any clays.
3A pellets. A very low ethanol loading of 0.03 mol/kg was This is especially important for the design of nontraditional beds
obtained while the water capacity was observed to be more than with structured contactors, in which different formulations can
5 mol/kg for mixture breakthrough at 167 °C. The adsorbent be used compared to commercial pellets or beads used in packed
selectivity for the “ethanol-water” pressure swing adsorption beds.
(PSA) process was measured to be ∼900 suggesting the Looking forward, the development of a structured contactor
molecular sieving mechanism for ethanol−water separation. utilizing adsorbent microchannels appears to have great promise
Gabrus et al. conducted full temperature swing adsorption in reducing cycle time and improving process capacity.13−19
(TSA) cycles, including dewatering and regeneration stages of Conventional gas separation systems are usually performed
the sorbent, for three water−alcohol systems (ethanol, n- using adsorbents in the form of beads or granules. However,
propanol, and n-butanol) using zeolite 3A and 4A pellets.8 mass transfer and pressure drop drawbacks in these conventional
Multitemperature Langmuir−Freundlich and dual-site Lang- packed beds impose limitations in operating the process at
muir−Langmuir equilibrium models were used to correlate optimum conditions. An alternative approach is to use
experimental equilibrium data of water adsorption on 3A pellets structured adsorbents in monolith to yield high mass transfer
at temperatures between 100° and 250 °C. Thermal desorption kinetics with large film and solid mass transfer coefficients and
studies revealed that water removal from 3A takes place up to low pressure drop.18,19 Therefore, isotherm data based on
400 °C, for 4A pellets it takes up to 500 °C. However, the water commercial pellets is not useful for describing the water
desorption rate decreased for the temperature higher than 240 adsorption on coated monolith bed. With the water adsorption
°C. Zeolites 3A and 4A show structural stability up to 500 °C. isotherm measured on zeolite crystals the water adsorption on
Heating above 500 °C leads to an irreversible collapse of monoliths can be reasonably derived for various monoliths with
structure and a loss of adsorptive capacity for 3A and 4A. different composition of formulations.
For industrial dehydration processes it is important to control The objective of this work is to measure accurate water
product purity at the end of the bed to meet the designed adsorption on zeolite in crystal forms over a wide range of
specifications which can be in the ppmv range. This requires temperatures and pressures to provide water isotherms for the
water adsorption isotherms to be measured at low concen- design of industrial LNG and NGL dehydration processes.
trations. However, the majority of papers describing water Adsorption isotherms were measured from 25 to 250 °C to allow
adsorption isotherms are reported on moderate and high vapor temperature swing adsorption process design. The isotherms
concentrations close to saturation. Additionally, significant also cover the low water concentrations required to meet
discrepancies were found in various studies on the measured subppm water specifications for deep dehydration. To the best
data.10 of our knowledge this is the first report of water isotherms over a
The level to which a gas is dried also depends on the wide range of temperatures and pressures based on zeolite 3A
regeneration conditions−temperature, pressure (or partial crystals. Multiple isotherm models were investigated to find the
pressure of water), and time. The concentration of water best description for the experimental data with a simple form.
remaining on the bed at the end of the regeneration cycle The triple-site Langmuir (TL) model has successfully described
determines the performance of the bed. This is because the all the data over the temperature and pressure ranges measured.
outlet gas will be in equilibrium with the last adsorbent at the Thus, this explicit pressure-based model has a merit to be
applied in modeling TSA and PSA processes.

■
end of the bed. The choice of regeneration temperature and
partial pressure of water depend on the isotherm to ensure
complete water removal from the zeolite bed to avoid reduced EXPERIMENTAL SECTION
water capacity in the subsequent adsorption step of the TSA or Material. Adsorbents used in this study were the commercial
pressure temperature swing adsorption (PTSA) process. To zeolites: Zeolite 3A crystals (Lot No. 10177046), from Alfa
optimize TSA or PTSA processes for the dehydration of gases, Aesar. Zeolite 4A crystals (Lot No. MKBS1477 V), purchased
accurate water adsorption equilibria are also required over a from Sigma-Aldrich. Also used in this work was the 13X pellets
wide range of temperatures, usually with adsorption at room (Sylobead MS S 544), purchased from Grace Davidson for
temperature and desorption at higher temperature with dry gas validation of the instrument. Distilled water was used in this
purging. study.
Most water adsorption data reported in the literature result Adsorbent Characterization. The elemental compositions
from the use of zeolite pellets or beads forms. Often the zeolite of all molecular sieve samples were analyzed by inductively
powder is mixed with an inorganic and organic binder, then coupled plasma (ICP) and performed by a third party. The
shaped into the desired geometry using processing techniques elemental analysis results from the ICP of all zeolite samples are
including extrusion, spray-drying, or granulation, and finally shown in Table 1. The commercial 3A crystals contain 6.93 wt %
thermally treated to remove the organic binder and impart Na and 8.64 wt % K from the total weight, corresponding cation
mechanical strength by the inorganic binder. Traditionally, the composition is 58 mol % Na+ and 42 mol % K+. The 4A crystal
information on structuring of adsorbents is kept as in-house and 13X pellet contain mostly sodium cation and little to no
know-how or only disseminated in patents.11 Various kinds of potassium. Zeolite 3A and 4A crystals have size distribution from
silica or clays are often used as the binders, their content being 0.5 to 10 μm. The samples are highly crystalline with no
approximately 16−20% w/w.12 It is generally assumed that the detectable extra framework Al by 27Al nuclear magnetic
B https://dx.doi.org/10.1021/acs.iecr.9b06891
Ind. Eng. Chem. Res. XXXX, XXX, XXX−XXX
Industrial & Engineering Chemistry Research pubs.acs.org/IECR Article

Table 1. Elemental Analysis of Zeolite Samples approximately the partial water pressure in the LNG and NGL
product at a pressure around 70 bar. The high end of pressure is
elemental composition (wt %)
chosen to be near the saturated water pressure at room
zeolite Al Na Si K Si/Al molar ratio temperature. Zeolite activation is carried out using the same
3A crystal 14.9 6.9 16.9 8.6 1.1 procedure throughout this work: specifically in situ regeneration
4A crystal 14.3 12.8 15.1 N/Aa 1.0 at 350 °C for 6 h under high vacuum prior to the adsorption
13X pellet 12.2 9.5 18.2 0.1 1.4 measurement. The system vacuum can reach 1 × 10−8 mbar with
a
N/A, not applicable. a high vacuum turbo molecular pump station.
The IGA application software additionally displays the kinetic
record in real time, that is, an uptake curve, on the screen
resonance (NMR). Scanning electron microscopy (SEM) alongside the isotherm data. This feature allows adjustments to
images and X-ray diffraction (XRD) patterns information are equilibration time based on uptake curves to ensure full
provided in the Supporting Information (Figures S1 and S2) for equilibrium is reached at each pressure measurement. An
3A and 4A crystals studied in this work. example of uptake curves is shown in Supporting Information
Adsorption Isotherm Measurements. The isotherm Figure S3. Equilibrium time for each point typically varies from
measurements were conducted using a Hiden Isochema 10 h to 30 min depending on the water concentrations; generally
Intelligent Gravimetric Analyzer (IGA-002), a gravimetric unit longer time was needed for isotherm measurements at lower
using a microbalance with resolution of 0.1 μg. The system is concentrations.
capable of measuring vapor adsorption from 0 to 0.99 relative As shown in Figure 1, the sample is put into the stainless steel
saturated pressure (P/P0) with P0 at temperatures of up to 50 °C pan hanging on the wire connected to the balance. Depending
with built-in anticondensation protection. The system has three on the temperature range, a furnace or water bath is used to
pressure sensors with one having a maximum range of 10 mbar control sample temperature. The weight change is monitored by
for active vapor pressure regulation, and two with maximum the thermostated microbalance. Corresponding capacity can be
ranges of 1 and 20 bar to allow measurements for gas adsorption. calculated from the data obtained from the system pressure and
The sample temperature was automatically controlled using a clean sample weight after regeneration.
recirculating bath for temperatures between 0 to 80 °C with
regulation accuracy of ±0.05 °C, and using a cryofurnace for
temperatures from −150 to 500 °C with regulation accuracy
±0.05 to ±1 °C depending on operating temperature. Since
■ RESULTS AND DISCUSSIONS
Equilibrium Data Validation on 13X Pellets. Zeolite 13X
water measurements are carried out at very low pressures, pellets from Grace Davidson were used as the control sample to
buoyance corrections can be negligible and were not measure water isotherms at two temperatures: 25 and 150 °C.
incorporated for water isotherm measurements. The latter temperature was chosen because it is the highest
Water adsorption isotherms have been measured for various temperature with reported isotherm data from the manufacturer.
zeolites over the temperature range of 25−250 °C and pressure The cation composition of 13X is shown in Table 1.
range from 1 × 10−5 to 3 × 10−2 bar. The low end of pressure is Figure 2 shows experimental data measured using the IGA-
chosen to meet the product specification of 0.1 ppmv of deep 002 gravimetric unit, compared with two sets of data from the
dehydration: 1 × 10−5 bar, that is, 0.1 ppmv at 1 × 102 bar, is manufacturer’s brochures. The blue dashed line was obtained

Figure 1. Schematic diagram of IGA-002 gravimetric unit.

C https://dx.doi.org/10.1021/acs.iecr.9b06891
Ind. Eng. Chem. Res. XXXX, XXX, XXX−XXX
Industrial & Engineering Chemistry Research pubs.acs.org/IECR Article

Figure 2. Water isotherm data of 13X pellets at 25° and 150 °C

compared to literature data in lines from Grace Davison and Union
Carbide. Lines are data from the manufacturers, symbols are data Figure 4. Isotherms for water adsorption on the Alfa Aesar 3A crystals
measured using the IGA-002 gravimetric unit. at 25, 50, 100, 150, 200 °C, compared with literature data from Grace
Davison [GD]2 for 3A pellets at 25 and 100 °C.
from Grace Davidson data,2 and the red solid line was obtained
from Union Carbide data.3 The water isotherms at 25 °C from
two manufacturers are very similar, and the measured data agree for dehydration applications. In the as-synthesized form, zeolite
very well with those from manufacturers at both temperatures. Na-A (4A) has the unit cell formulated of
The lowest pressure set point is routinely achieved at 1 × 10−5 Na12(AlO2)12(SiO2)12·27H2O.20 The Na+ (ionic radium =
bar. The data are plotted on a semilogarithmic plot, the pressure 0.93 Å) can be exchanged for K+ (ionic radium = 1.33 Å) to
in bar logarithmically in order to cover the four orders of have zeolite 3A with the pore aperture reduced to 3 Å from 4 Å.
magnitude. There appears to be excellent agreement between The aperture size of the sodium form can also be increased by
the experimental data and literature data from the manufacturer. exchanging Na+ with Ca2+ because 2 Na+ are replaced by one
Water Isotherm Data on Zeolite 3A and 4A Crystals. divalent cation. This less obstructed aperture is zeolite 5A which
Zeolite A (Linde-Type A, LTA) is widely used as an adsorbent has a pore size of 4.3 Å.

Figure 3. Comparison of water isotherms for Alfa Aesar zeolite 3A crystals at 25 and 50 °C (this work) with literature data of UOP 3A beads and UOP
3A powder5 and Grace Davison SYLOBEAD 3A2.

D https://dx.doi.org/10.1021/acs.iecr.9b06891
Ind. Eng. Chem. Res. XXXX, XXX, XXX−XXX
Industrial & Engineering Chemistry Research pubs.acs.org/IECR Article

Figure 8. Water isotherms of Alfa Aesar 3A crystals at 25, 50, 100, 150,
and 200 °C. Symbols are experimental points, solid lines are fit from
triple-site Langmuir model.

Figure 5. Isotherms for water adsorption on the 4A crystals at 25, 50,

100, 200, 250 °C, compared with literature data from Grace Davison2 for 3A zeolite pellets from the literature.2,5 The water isotherm is
with 4A pellets at 25° and 100 °C. nonlinear. To analyze the equilibrium at lower pressures, the
capacity is plotted in a semilogarithmic format to show the
capacity of all ranges, especially for low water concentrations and
to cover the four orders of magnitude.
The water capacity measured at 25 °C of Alfa Aesar 3A
crystals, shown in square filled symbols, is higher than that of
Grace 3A beads (pellets in sphere shape) shown in the dash blue
line. However, note that the shape of isotherms appear to be
similar. This is not surprising as zeolite in crystalline form has
higher water capacity; and the binder does not have much water
capacity compared to 3A crystals. However, an unexpected trend
is observed when compared to data of the UOP 3A beads at 25
°C,5 the 3A crystal measured in this work shows lower capacity
when water pressure is less than 1 × 10−4 bar as shown in Figure
3. Furthermore, the water capacity measured in this work at 25
°C is lower than the capacities at 40 °C for the UOP 3A crystals
reported in the same paper.5 The discrepancy cannot be
Figure 6. Comparison of water isotherms on zeolite 3A crystal and 4A explained by the difference of adsorbents as the potassium ion
crystal. The filled symbols are measured isotherm data of 3A crystals, content is similar, ∼40 mol % in these 3A samples. Potassium
empty symbols are measured data in 4A crystals. and sodium molar percentages for various 3A samples are Alfa
Aesar 3A crystals (42 mol % K and 58 mol % Na from ICP
analysis in Table 1), UOP 3A powder (49−53 mol % K and 1−
47 mol % Na), UOP 3A beads (37−41 mol % K and 58−62 mol
% Na), and Grace Davison 3A pellets (43 mol % K and 57 mol %
Na) elsewhere.5
There are several reported data sets for multitemperature
water adsorption on zeolite 3A pellets, including manufacturer
data on UOP 3A beads measured at 25, 40, 60, and 80 °C and
Grace Davison 3A beads measured at −20, 0, 25, 40, 60, 80, and
100 °C2, but none is found for zeolite 3A crystals without
binders. We have extended the temperature envelope to include
temperatures up to 200 °C for water isotherm measurement on
3A crystals. The isotherms of 3A crystals were measured at
temperatures of 25, 50, 100, 150, and 200 °C and pressures from
1 × 10−5 bar to 2 × 10−2 bar, shown in symbols in Figure 4, along
Figure 7. Water isotherms of Alfa Aesar 3A crystals at 25, 50, 100, 150, with Grace Davison published data on 3A pellets at 25 and 100
and 200 °C. Symbols are experimental points. Dashed lines are fit from °C. The capacities for water adsorption on the pellets are
dual-site Langmuir (DL) model, and solid lines are fit from the Toth consistently lower than that obtained in this study in crystals at
model. the same temperature. This is explained by binder content in the
pellets which behaves as filler with negligible water capacity.
Figure 3 is the water isotherm measured on commercial Alfa Zeolite 3A shows that water capacity approaches zero when the
Aesar zeolite 3A crystals, along with water vapor isotherms data temperature is up to 200 °C while the water pressure is less than
E https://dx.doi.org/10.1021/acs.iecr.9b06891
Ind. Eng. Chem. Res. XXXX, XXX, XXX−XXX
Industrial & Engineering Chemistry Research pubs.acs.org/IECR Article

Table 2. Fitted Triple-Site Langmuir Parameters for Water on 3A Crystals

ns1 (mol/kg) b10 (1/bar) E1 (kJ/mol) ns2 (mol/kg) b20 (1/bar) E2 (kJ/mol) ns3 (mol/kg) b30 (1/bar) E3 (kJ/mol)
9.37 1.26 × 10−06 59.75 1.06 4.67 × 10−03 48.37 3.35 1.15 × 10−07 51.42

example, at 2.5 × 10−2 bar which is close to the water saturation

pressure at 25 °C, the water capacity in the 3A crystals is about
13 mol/kg compared to 14.8 mol/kg in 4A, a 12% difference. At
1 × 10−5 bar the capacity is 3.36 mol/kg for 3A and 3.78 mol/kg
for 4A at 25 °C. Thus, for pressure swing adsorption with the
water partial pressure swing between 1 × 10−5 and 2.5 × 10−2
bar, the working capacity for 3A crystals is 12% lower compared
to the 4A crystals working capacity. However, 3A can be
regenerated at lower temperature compared to 4A due to its less
steep isotherm The water capacity for 3A at 1 × 10−4 bar is
negligible at 200 °C while 4A still has more than 1 mol/kg water
capacity, even at a higher temperature of 250 °C. This
Figure 9. Cation sites and structure of zeolite A. demonstrates that 4A needs a higher regeneration temperature
compared to 3A for similar dehydration conditions.
The saturation loading of water on the 4A zeolite crystals is
given by Breck20 as 2.89 × 10−1 g/g at 25 °C, which corresponds
to 16 mol/kg. The highest water capacity measured at 25 °C on
the 4A crystal in this study is 14.8 mol/kg at P/P0 = 0.8. For
zeolite A, as the total void space remains constant within the
rigid structure, the volume occupied by various cations will affect
the number of water molecules that can be adsorbed within the
zeolite. As a result, the water content increases with decreasing
ionic radius of the cation, from 27 (number of waters per unit
cell) for NaA (4A), to 30 for CaA (5A) which contains half the
number of ions for divalent ions.21 Zeolite 3A (KA) should allow
less water compared to 4A (NaA) as K+ has a larger ionic radius
than Na+. This is consistent with the trend observed when
comparing water capacity in zeolite 3A and 4A as shown in
Figure 10. Water isotherms of Sigma-Aldrich 4A crystals at 25, 50, 100, Figure 6.
200, and 250 °C. Symbols are experimental points. Solid lines are fit Multitemperature Isotherm Modeling. Temperature is a
from the triple-site Langmuir model. significant parameter affecting adsorption equilibrium capacity.
For modeling the TSA process, it is important to describe
isotherm data at multiple temperatures with one set of
10−4 bar. This suggests that 3A crystals can be reasonably parameters to quantify temperature dependence. The Toth
regenerated at this condition. model is a popular empirical description which has correct
To compare the water capacity of 3A with 4A crystals, water limiting behavior at low pressures and yields a finite adsorbate
isotherms on 4A crystals have been measured at temperatures of concentration at high pressures. It can accurately describe many
25, 50, 100, 200, and 250 °C. Figure 5 shows the water isotherms type I isotherms and has a simple form with parameters in the
measured in 4A crystals, along with literature data from Grace function of temperature.22
Davison in 4A pellets. The same trend has been observed for
comparison of water isotherms at 25 °C for 4A crystals and nsbp
n=
pellets as previously described for 3A crystals and pellets. The [1 + (bp)t ]1/ t (1)

i E yz
b = b0 expjjj
water capacity in crystal form is higher than that of the pellets
zz
k RT {
form over the entire pressure range. However, the comparison at
100 °C appears to be different because the water capacity in (2)
crystals is not always higher than that in pellets. Instead, the
literature data for pellets shows higher water capacity at the t = t0 + a/T (3)
intermediate pressure range from 5 × 10−4 to 1 × 10−2 bar.
Figure 6 shows a comparison of water capacity in 4A crystals where n is adsorption capacity at pressure of p, ns is the
and 3A crystals (isotherm data included in Supporting saturation capacity, b is an affinity parameter, E is the adsorption
Information Table S4 and S5). It is clear that the 4A zeolite heat, a is a constant parameter, and t is a heterogeneity
crystals have a higher capacity for water than the 3A zeolite parameter to count for surface heterogeneity; t is usually less
crystals over the entire pressure range at each temperature. For than unity.

Table 3. Fitted Triple-Site Langmuir Parameters for Water on 4A Crystals

ns1 (mol/kg) b10 (1/bar) E1 (kJ/mol) ns2 (mol/kg) b20 (1/bar) E2 (kJ/mol) ns3 (mol/kg) b30 (1/bar) E3 (kJ/mol)
11.28 1.36 × 10−06 58.61 4.06 1.87 × 10−06 41.40 1.74 2.62 × 10−06 92.98

F https://dx.doi.org/10.1021/acs.iecr.9b06891
Ind. Eng. Chem. Res. XXXX, XXX, XXX−XXX
Industrial & Engineering Chemistry Research pubs.acs.org/IECR Article

Table 4. Fitted Triple-Site Langmuir Parameters for Water on SYLOBEAD 3A

ns1 (mol/kg) b10 (1/bar) E1 (kJ/mol) ns2 (mol/kg) b20 (1/bar) E2 (kJ/mol) ns3 (mol/kg) b30 (1/bar) E3 (kJ/mol)
1.43 3.29 × 10−06 68.24 8.38 1.10 × 10−06 58.06 2.66 2.73 × 10−06 44.07

The isotherm data for the 3A crystals were fitted with the Toth
model. The results are shown as solid lines in Figure 7. It
provides a reasonable descriptionan agreement with data in
the middle range of water loadings for the region between 2 and
10 mol/kg, but lacking accuracy in both ends where water
loading is less than 2 mol/kg and higher than 10 mol/kg. The
model under predicts water capacities on 3A crystal in these two
regions. The discrepancy is more obvious for high water loadings
at 25 and 50 °C. A modified Toth model with temperature
dependence of saturated capacity has also been tried, but does
not provide significant improvement, and therefore is not
shown.
When heterogeneity parameter t equals 1 in the Toth model,
it reduces to the Langmuir model. The Langmuir model usually
considers all sites to be identical and cannot be successfully
applied to this complicated isotherm shape.22 Other conven-
tional isotherm models including Langmuir, Sips, Toth, etc. do
not allow correlation of the water adsorption data in 4A10 and
3A.23 Later Loughlin reexamined the Gorbach et al. data10 and
recommended considering a two-site model, one for the α cage
and one for the β cage in zeolite 4A.24 The resulting dual-site
Toth or dual-site Langmuir isotherm models fit the published
water isotherm data in 4A adequately.
The dual-site (process) Langmuir model can be written as
ns1b1p n bp
n= + s2 2
1 + b1p 1 + b2 p (4)

iE y
b1 = b10 expjjj 1 zzz
k RT {
Figure 11. Dual-site Langmuir and triple-site Langmuir modeling of
(5)

iE y
isotherm models to the water adsorption data in zeolite 3A pellet: (a)

b2 = b20 expjjj 2 zzz

dual-site Langmuir fit; (b) triple-site Langmuir fit. Data from Grace-

k RT {
Davison.2
(6)
where ns and b represent saturated loadings and fractional
equilibrium constants for two sites with subscript 1 and 2. The
model has six fitting parameters:ns1,ns2,b10,b20,E1, andE2.
In this work, the dual-site Langmuir model has been applied to
describe the multitemperature water isotherms in 3A and is
shown by the dashed lines in Figure 7. The fitting is good for
most loadings except for those higher than 11 mol/kg. When
compared to a previous Toth model, it is clear that the dual-site
Langmuir is a better fit for the low end region. This is consistent
with the hypothesis of two sites being more suitable than
homogeneous sites. However, the dual-site Langmuir model
appears to reach a plateau while experimental data shows
continuous increase of capacity suggesting an additional site to
accommodate more water. Thus, an extended multisite form,
triple-site Langmuir (TL) model, is proposed to describe water
Figure 12. Isotherms for water adsorption on Sylobead 4A at 0, 25, 40,
isotherm in the form of
60, 80, 100, 120, 140, and 160 °C. Symbols represent data from Grace
Davidson. Solid lines represent the dual-site Langmuir model fit. ns1b1p n bp n bp
n= + s2 2 + s3 3
1 + b1p 1 + b2 p 1 + b3p (7)

Table 5. Fitted Dual-Site Langmuir (DL) Parameters for Water on SYLOBEAD 4A

ns1 (mol/kg) b10 (1/bar) E1 (kJ/mol) ns2 (mol/kg) b20 (1/bar) E2 (kJ/mol)
9.57 2.02 × 10−06 56.86 1.59 3.24 × 10−03 53.50

G https://dx.doi.org/10.1021/acs.iecr.9b06891
Ind. Eng. Chem. Res. XXXX, XXX, XXX−XXX
Industrial & Engineering Chemistry Research pubs.acs.org/IECR Article

Figure 13. Isosteric heat of water on zeolite(a) 3A crystals and beads and (b) 4A crystals and beads, calculated from isotherms in Figure 7 (3A crystal),
Figure 11 (3A bead), Figure 8 (4A crystal), and Figure 12 (4A bead) using eq 10.

where b3 and ns3 correspond to the equilibrium constant and cage) and the remaining ion is located near the four-member
saturated loading for the third site. b3 has a similar temperature ring inside the cavity (Site III in the sodalite cage).32,33 The
dependence form as in b1 and b2 in eqs 5 and 6. Compared to the three cations located in Site II play a dominant role in the
frequently accepted and use of the dual-site Langmuir model in molecular-sieving action of A-type zeolites. Due to the cations
the literature, the application of triple-site Langmuir was found partially blocking the window, for 3 Na+ cations, only molecules
in very few publications, including two on adsorption with size smaller than 4 Å can diffuse through the eight-member
isotherms,25,26 and others on catalyst kinetics which can be aperture to access the α-cage.
linked to multiple active sites for adsorption.27−30 No papers Zhang et al.34 studied the water adsorption on thin zeolite 3A
were found using the triple-site Langmuir for describing water wafers by using mid-infrared spectroscopy coupled with
adsorption. multivariate data analysis. Their results demonstrated that
Figure 8 illustrates that the triple-site Langmuir model zeolite 3A has three water adsorption sites or environments that
describes all of the data very well over the entire range of are sequentially occupied. The water adsorbs on the zeolite to
concentration and temperatures measured. The fitted parame- form single, double, and triple water adsorption on single zeolite
ters are presented in Table 2. With the additional site, the triple- adsorption sites. Their observation is consistent with our fitting
site model accurately describes the high end of water results that require the triple site Langmuir model for data
concentration compared to the dual-site Langmuir model. The description.
square of error is calculated to be 0.93 for triple-site Langmuir Similarly, the isotherms on 4A crystals can be described well
model compared to 5.9 for Toth model and 8.4 for dual-site with the triple-site Langmuir model as shown in Figure 10. The
Langmuir model. To evaluate trade-off between the goodness of related fitting parameters are tabulated in Table 3. It is worth
fit of the models and the simplicity of the model, the Akaike mentioning that the heat of adsorption values E vary from ∼41
information criterion (AIC) is calculated as statistical regression to 93 kJ/mol for the TL model employed to 4A crystals,
metrics for measuring the performance of regression models. In compared to the range from ∼48 to 60 kJ/mol for 3A crystals. A
model comparison strategies, the model with the lowest AIC is high value heats of adsorption of 93 kJ/mol was fitted based on
preferred.31 The triple-site Langmuir model has the lowest AIC this model for 4A. Interestingly, this value aligns with the range
statistic, which is significantly lower than the DL and Toth of values from the literature, which reports the heat of
model, and therefore is to be regarded as the best fit. Detailed adsorption is high (100 to 125.5 kJ/mol) for a fractional loading
calculation is included in the Supporting Information. θ of ∼0.2 and decreases (<60 to 74 kJ/mol) for θ above ∼0.2.
The need of the three sites model to accurately describe water The high value of E is considered to be related to the site which
adsorption is considered to relate to the three cation sites in the has a strong bonding energy at low coverage. This is shown in
zeolite LTA structure. It is well-known that the strong anionic Figure 6 where 4A still maintains water loadings at high
nature of the zeolite framework and correspondingly strong temperature whereas 3A shows much less water loading at
cations bonded to the framework exert strong electric fields and similar conditions.
field gradients over the surface. Adsorbate molecules with Triple-Site Langmuir Modeling for Existing Water
permanent dipoles and quadruples can therefore have strong Isotherm Data Sets. There are published figures of water
interactions with exposed cations. Thus, the water molecule isotherms in zeolite 3A Sylobead by Grace Davidson, at seven
could have different bonding energy to the cations depending on temperatures in the range from −20 to 100 °C and over a range
the locations for zeolite 3A. Figure 9 illustrates the framework of of several decades of water vapor partial pressure.2 This
zeolite A in terms of two types of polyhedraa double 4-ring graphical format was digitized and the formulated data were
unit (D4R) and a truncated octahedron (β-cage). There is a replotted with loading (mol/kg) versus the water vapor partial
large cage (α-cage or cavity) with a diameter of 11.4 Å at the pressure (bar) in a logarithmic scale.
center of the unit, which can be entered through six eight- Llano-Restrepo and Mosquera35 used this data set to evaluate
member oxygen-ring apertures. There are 12 negative charges the most popular models including Langmuir, Sips, Toth,
that are balanced by cations in each unit cell. Three types of UNILAN, and Dubinin-Astakhov and found none of them being
cation sites are located in or near these rings. In the dehydrated capable of correlating the experimental adsorption isotherms.
zeolite 4A or 3A, eight monovalent ions are located in or near the They proposed a generalized statistical thermodynamic
six-member rings (Site I in β-cage), three ions are near the eight- adsorption (GSTA), a version based on Hill’s statistical
member aperture directly obstructing the entrance (Site II in α- thermodynamic adsorption model to successfully describe this
H https://dx.doi.org/10.1021/acs.iecr.9b06891
Ind. Eng. Chem. Res. XXXX, XXX, XXX−XXX
Industrial & Engineering Chemistry Research pubs.acs.org/IECR Article

data set. The model contains a total of 2 + m fitting parameters: explicit form for pressure dependence, it is well suited for
the theoretical maximum adsorption capacity (qmax), the process modeling for cycle processes such as PSA or TSA.
number of energetically different adsorption sites (m), and a Heat of Adsorption. Information related to the heat of
total of m dimensionless equilibrium parameters (k0n) as a adsorption (qst) and its variation with coverage can provide an
function of the enthalpies ΔH0n and entropies ΔS0n for each site. important guide to the energy required for regeneration in the
TSA process. qst(H2O) for zeolites can be calculated from its

iq y
n

q = jjj max zzz

water adsorption isotherms collected at different temperatures
m
()
∑n = 1 nK n0
p

k m { 1 + ∑m K 0 p ÅÄÅ ÑÉÑ
p0 using the Clausius−Clapeyron equation
ÅÅ Ñ
ÅÅ ∂ lnP ÑÑÑ
n

qst = −RÅÅÅ 1 ÑÑÑ

()
ÅÅ ∂ Ñ
n=1 n 0

ÅÅÇ T ÑÑÑÖ
p (8)

where () n (10)

ΔHn0 ΔSn0 By constructing a plot of ln p against the reciprocal of the

ln K n0 = − + temperature at constant loading, the value of qst is obtainable
RT T (9)
from the slope of these adsorption isosteres, shown in eq 10.
The best fit of all seven isotherms required six or seven Figure 13 depicts qst as a function of water loadings in zeolite 3A
parameters. The modeling analysis for both systems suggest to and 4A crystals, compared with the zeolite beads. These qst
have four or five sites for the best description of data even though values were calculated from water isotherms shown in Figures 7,
only three inflection points, are visible for the data.5,36 The 8, 11, and 12 respectively.
authors explain that the fourth inflection point is not visible The qst calculated from water isotherms shows similar trends
because the final inflection point only occurs at very high for crystal and bead forms. The range of values are comparable to
temperatures and pressures. Furthermore, this model has been the TL parameters (E1, E2, E3) extracted in Tables 2 and 3 for
demonstrated feasible for an assortment of different gas−solid both zeolite 3A and 4A crystals. For water loadings in the
adsorption systems test, but utilizing this model can be fairly medium range of 2−9 mol/kg, the qst values are constants of 58−
difficult due to its complex form.36 60 kJ/mol for zeolite 3A and 4A. However, both 4A crystals (this
Given that the multisite Langmuir model is simpler and not as work) and 4A Sylobead (literature data) show a higher heat of
parametrized as the GSTA, we applied both the dual-site adsorption over 90 kJ/mol at the limit of zero coverage (<0.5
Langmuir and triple-site Langmuir models to describe the 3A mol/kg). This indicates that water is bonded significantly more
Sylobead data set by Grace Davison. Similar to our observation strongly to zeolite 4A than 3A at low water loadings, which could
on the 3A crystal, the dual-site Langmuir model underestimates be related the high energy sites adjacent to the sodium cations.
the water capacity at high loadings, but triple-site Langmuir This qst value at the limit of zero coverage is consistent with the
provides an excellent description for all data with fitted one of the heat of adsorption E3 parameter, 92 kJ/mol extracted
parameters listed in Table 4. This is shown in Figure 11b. from the TL model for zeolite 4A crystals in Table 4. Zeolite 3A
This suggests that three sites could provide reasonable fit for all beads shows qst about 70 kJ/mol at the limit of zero coverage
data without the need for more sites, such as the four or five sites (<0.5 mol/kg), shown in Figure 13a, but no corresponding value
suggested by the GSTA model.5,35,36 available for 3A crystals due to lack of enough isotherms for
The second data set examined was the water isotherms of water loadings less than 0.5 mol/kg.
zeolite 4A SYLOBEAD, from Grace Davidson.2 The exper- At the other extreme, when water loadings increase and
imental data cover nine isotherms in a temperature range from approach the saturated loading in the 3A system, both crystals
0° to 160 °C. Similar to the approach for literature data with and beads show gradual decrease in heats of adsorption. qst
zeolite 3A beads, we have digitalized the graphic plot and fit the approaches 45 kJ/mol for a water loading of 11 mol/kg in 3A,
extracted data with the dual-site Langmuir model. It is which is comparable to the molar vaporization enthalpy of water
interesting that the dual-site Langmuir model can describe (∼44 kJ/mol). A similar observation has been shown that the
zeolite 4A data reasonably well without the need of an additional isosteric heat of adsorption at full loadings is just 3% greater than
site to improve the fit. This could relate to the data set where the heat of evaporation at 100 °C, and 7% less than the heat of
water capacity was only reported up to an intermediate pressure evaporation at 0 °C.35 This trend is not observed for either 4A
with water loading less than 11 mol/kg. As seen from previous crystals or beads with the water loading of 11 mol/kg. One
model fitting, triple-site Langmuir always gives a better explanation may be ascribed to the higher saturated loading for
water in 4A systems.

■
description at the high water loadings, but equivalent to dual-
site Langmuir for low and intermediate pressure range. Figure 12
shows the results of the temperature dependent dual-site CONCLUSIONS
Langmuir model with fitted parameters listed in Table 5. Water isotherms have been investigated over a wide range of
The successful application of the triple-site Langmuirin for water concentrations and temperatures for commercially
these LTA zeolites indicates this model might be suitable for available zeolite LTA crystals, 3A and 4A. A commercial
other cationic zeolites having multiple cation sites while gravimetric apparatus was validated by measuring water in Grace
interacting with polar adsorbates. By judiciously choosing the Davison 13X pellets at low and high temperature and found to
zeolite and polarizable molecules to be adsorbed, an analysis of be in agreement with the isotherm published by the vendor.
the isotherms with the dual-site Langmuir and triple-site Water isotherms on zeolite 3A crystals show consistently higher
Langmuir models and their goodness of fit may shed light on loading compared to literature data from Grace Davison on 3A
the number of accessibility sites. Because of its simplicity pellets.
compared to generalized statistical thermodynamic adsorption 4A zeolite crystals have a higher capacity for water than 3A
modeling (GSTA) for the description of water isotherms and zeolite crystals over the entire pressure range at each
I https://dx.doi.org/10.1021/acs.iecr.9b06891
Ind. Eng. Chem. Res. XXXX, XXX, XXX−XXX
Industrial & Engineering Chemistry Research pubs.acs.org/IECR Article

temperature. For pressure swing adsorption with water, a partial Professional Publishing, Elsevier: Waltham, MA, 2019; pp 307, Chapter
pressure swing between 1 × 10−5 to 2.5 × 10−2 bar, the working 9.
capacity for 3A crystals is 12% lower compared to 4A crystals (2) Sylobead Adsorbents for Process Applications; R. Grace & Co.-
working capacity. However, 3A can be regenerated at lower Conn.: Columbia, Maryland, 2010, https://grace.com/general-
temperature compared to 4A due to its less steep isotherm for industrial/en-us/Documents/sylobead_br_E_2010_f100222_web.
pdf.
similar dehydration conditions.
(3) Union Carbide Molecular Sieves Water and Air Data Sheets, F-43B-1;
Multitemperature isotherms of 3A crystals are analyzed to Union Carbide: New York, 1984.
determine the isotherm model that offers the best representation (4) Jović, S. a.; Laxminarayan, Y.; Keurentjes, J.; Schouten, J.; van der
of data. Both Toth and dual-site Langmuir model provide a Schaaf, J. Adsorptive Water Removal from Dichloromethane and
reasonable description, but fail in the region of high water Vapor-Phase Regeneration of a Molecular Sieve 3A Packed Bed. Ind.
loadings over 11 mol/kg. A simple triple-site Langmuir model Eng. Chem. Res. 2017, 56, 5042.
has been proposed and is proven to satisfactorily model (5) Lin, R.; Ladshaw, A.; Nan, Y.; Liu, J.; Yiacoumi, S.; Tsouris, C.;
multitemperature isotherms on 3A and 4A crystals as well as DePaoli, D. W.; Tavlarides, L. L. Isotherms for Water Adsorption on
literature data on 3A and 4A pellets. The accuracy of the three Molecular Sieve 3A: Influence of Cation Composition. Ind. Eng. Chem.
sites model (triple-site Langmuir) to describe water adsorption Res. 2015, 54, 10442.
is consistent with and related to the three cation sites (inside α- (6) Simo, M.; Sivashanmugam, S.; Brown, C. J.; Hlavacek, V.
cage. β-cage, and sodalite cage) in the zeolite LTA structure. Adsorption/Desorption of Water and Ethanol on 3A Zeolite in near-
Thus, the water molecules have different bonding energy to the Adiabatic Fixed Bed. Ind. Eng. Chem. Res. 2009, 48, 9247.
(7) Carmo, M. J.; Gubulin, J. C. Ethanol-Water Adsorption on
cations depending on the locations within the zeolite.
Commercial 3A Zeolites: Kinetic and Thermodynamic Data. Braz. J.
The heat of adsorption calculated from isotherms shows a
Chem. Eng. 1997, 14, 217.
loading dependence, and the range is consistent with the (8) Gabruś, E.; Nastaj, J.; Tabero, P.; Aleksandrzak, T. Experimental
corresponding fit parameters in the triple-site Langmuir models Studies on 3A and 4A Zeolite Molecular Sieves Regeneration in Tsa
for both zeolite 3A and 4A crystals. With an accuracy of Process: Aliphatic Alcohols Dewatering-Water Desorption. Chem. Eng.
describing isotherms and a simple formula based on pressure, J. 2015, 259, 232.
this triple-site Langmuir model can be conveniently applied to (9) Wang, K.-S.; Liao, C.-C.; Chu, R. Q.; Chung, T.-W. Equilibrium
temperature swing adsorption processes using adsorbents with Isotherms of Water and Ethanol Vapors on Starch Sorbents and Zeolite
heterogeneous energy sites to interact with water or other polar 3A. J. Chem. Eng. Data 2010, 55, 3334.
adsorbates. (10) Gorbach, A.; Stegmaier, M.; Eigenberger, G. Measurement and

■
Modeling of Water Vapor Adsorption on Zeolite 4AEquilibria and
ASSOCIATED CONTENT Kinetics. Adsorption 2004, 10, 29.
(11) Akhtar, F.; Andersson, L.; Ogunwumi, S.; Hedin, N.; Bergström,
* Supporting Information
sı
L. Structuring Adsorbents and Catalysts by Processing of Porous
The Supporting Information is available free of charge at Powders. J. Eur. Ceram. Soc. 2014, 34, 1643.
https://pubs.acs.org/doi/10.1021/acs.iecr.9b06891. (12) Tsibranska, I.; Assenov, A. Experimental Verification of the
XRD patterns and SEM images for zeolite 3A and 4A Model of Adsorption in Biporous Particles. Chem. Eng. Process. 2000,
crystals; example of water uptakes at several water 39, 149.
(13) Determan, M. D.; Hoysall, D. C.; Garimella, S. Heat- and Mass-
pressures during isotherm measurement; isotherm dta
Transfer Kinetics of Carbon Dioxide Capture Using Sorbent-Loaded
for water on 3A and 4A crystals at the studied Hollow Fibers. Ind. Eng. Chem. Res. 2012, 51, 495.
temperatures; statistical analysis for model selection (14) Lively, R. P.; Chance, R. R.; Kelley, B. T.; Deckman, H. W.;
using Akaike’s information criterion (PDF) Drese, J. H.; Jones, C. W.; Koros, W. J. Hollow Fiber Adsorbents for

■
CO2 Removal from Flue Gas. Ind. Eng. Chem. Res. 2009, 48, 7314.
AUTHOR INFORMATION (15) Rezaei, F.; Mosca, A.; Hedlund, J.; Webley, P. A.; Grahn, M.;
Mouzon, J. The Effect of Wall Porosity and Zeolite Film Thickness on
Corresponding Author the Dynamic Behavior of Adsorbents in the Form of Coated Monoliths.
Yu Wang − Corporate Strategic Research, ExxonMobil Research Sep. Purif. Technol. 2011, 81, 191.
and Engineering, Annandale, New Jersey 08801, United States; (16) Rezaei, F.; Webley, P. Structured Adsorbents in Gas Separation
orcid.org/0000-0002-3444-9207; Email: yu.wang@ Processes. Sep. Purif. Technol. 2010, 70, 243.
exxonmobil.com (17) Pahinkar, D. G.; Garimella, S.; Robbins, T. R. Feasibility of Using
Adsorbent-Coated Microchannels for Pressure Swing Adsorption:
Complete contact information is available at: Parametric Studies on Depressurization. Ind. Eng. Chem. Res. 2015, 54,
https://pubs.acs.org/10.1021/acs.iecr.9b06891 10103.
(18) Barnes, W.; Chen, C.-C.; Fowler, T.; McMahon, P.; Nagavarapu,
Notes A. K.; Shatto, D., Novel Adsorption-Based Gas Treating Technology
The author declares no competing financial interest. Platform for Upstream Gas Separations. In Offshore Technology

■ ACKNOWLEDGMENTS
The author gratefully acknowledges Barbara Carstensen and
Conference; Offshore Technology Conference: Houston, Texas, USA,
2018; p 11.
(19) Nagavarapu, A. K.; Fowler, T.; Barnes, W.; McMahon, P.;
Schoch, P.; Federici, J. Process Internsification of Upstream Gas
Mike Strasser for the SEM and XRD characterization for the
Separations through Adsorption-Based Technology. In Laurance Reid
materials studied in this work.

■
Gas Conditioning Conference; Norman, Oklahoma USA, 2018.
(20) Breck, D. W.; Eversole, W. G.; Milton, R. M.; Reed, T. B.;
REFERENCES Thomas, T. L. Crystalline Zeolites. I. The Properties of a New Synthetic
(1) Mokhatab, S.; Poe, W. A.; Mak, J. Y. Natural Gas Dehydration and Zeolite, Type A. J. Am. Chem. Soc. 1956, 78, 5963.
Mercaptans Removal. In Handbook of Natural Gas Transmission and (21) Breck, D. W. Zeolite Molecular Sieves: Structure, Chemistry, and
Processing, 4th ed.; Mokhatab, S., Poe, W. A., Mak, J. Y., Eds.; Gulf Use; John Wiley and Sons, Inc.: New York, 1974; p 771.

J https://dx.doi.org/10.1021/acs.iecr.9b06891
Ind. Eng. Chem. Res. XXXX, XXX, XXX−XXX
Industrial & Engineering Chemistry Research pubs.acs.org/IECR Article

(22) Do, D. D. Adsorption Analysis: Equilibria and Kinetics. Imperial

College Press: Singapore, 1998; Vol. 2.
(23) Llano-Restrepo, M.; Mosquera, M. A. Accurate Correlation,
Thermochemistry, and Structural Interpretation of Equilibrium
Adsorption Isotherms of Water Vapor in Zeolite 3A by Means of a
Generalized Statistical Thermodynamic Adsorption Model. Fluid Phase
Equilib. 2009, 283, 73.
(24) Loughlin, K. F. Water Isotherm Models for 4A (Naa) Zeolite.
Adsorption 2009, 15, 337.
(25) Son, K. N.; Cmarik, G. E.; Knox, J. C.; Weibel, J. A.; Garimella, S.
V. Measurement and Prediction of the Heat of Adsorption and
Equilibrium Concentration of CO2 on Zeolite 13X. J. Chem. Eng. Data
2018, 63, 1663.
(26) Pikaar, I.; Koelmans, A. A.; van Noort, P. C. M. Sorption of
Organic Compounds to Activated Carbons. Evaluation of Isotherm
Models. Chemosphere 2006, 65, 2343.
(27) Díez-Ramírez, J.; Díaz, J. A.; Dorado, F.; Sánchez, P. Kinetics of
the Hydrogenation of CO2 to Methanol at Atmospheric Pressure Using
a Pd-Cu-Zn/SiC Catalyst. Fuel Process. Technol. 2018, 173, 173.
(28) Vorontsov, A. V.; Dubovitskaya, V. P. Selectivity of Photo-
catalytic Oxidation of Gaseous Ethanol over Pure and Modified TiO2. J.
Catal. 2004, 221, 102.
(29) Ji, Y.; Koot, V.; van der Eerden, A. M. J.; Weckhuysen, B. M.;
Koningsberger, D. C.; Ramaker, D. E. A Three-Site Langmuir
Adsorption Model to Elucidate the Temperature, Pressure, and
Support Dependence of the Hydrogen Coverage on Supported Pt
Particles. J. Catal. 2007, 245, 415.
(30) Stakheev, A. Y.; Zhang, Y.; Ivanov, A. V.; Baeva, G. N.; Ramaker,
D. E.; Koningsberger, D. C. Separation of Geometric and Electronic
Effects of the Support on the CO and H2 Chemisorption Properties of
Supported Pt Particles: The Effect of Ionicity in Modified Alumina
Supports. J. Phys. Chem. C 2007, 111, 3938.
(31) Symonds, M. R. E.; Moussalli, A. A Brief Guide to Model
Selection, Multimodel Inference and Model Averaging in Behavioural
Ecology Using Akaike’s Information Criterion. Behavioral Ecology and
Sociobiology 2011, 65, 13.
(32) Kaushik, V. K.; Vijayalakshmi, R. P.; Choudary, N. V.; Bhat, S. G.
T. XPS Studies on Cation Exchanged Zeolite A. Microporous
Mesoporous Mater. 2002, 51, 139.
(33) Ohgushi, T.; Sakai, Y.; Adachi, Y. Abnormalities of Cation
Movements in 3A Zeolite. Microporous Mesoporous Mater. 2012, 151,
134.
(34) Zhang, A.; Zeng, W.; Niemczyk, T. M.; Keenan, M. R.; Haaland,
D. M. Multivariate Analysis of Infrared Spectra for Monitoring and
Understanding the Kinetics and Mechanisms of Adsorption Processes.
Appl. Spectrosc. 2005, 59, 47.
(35) Llano-Restrepo, M.; Mosquera, M. A. Accurate Correlation,
Thermochemistry, and Structural Interpretation of Equilibrium
Adsorption Isotherms of Water Vapor in Zeolite 3A by Means of a
Generalized Statistical Thermodynamic Adsorption Model. Fluid Phase
Equilib. 2009, 283, 73.
(36) Ladshaw, A.; Yiacoumi, S.; Tsouris, C.; DePaoli, D. Generalized
Gas−Solid Adsorption Modeling: Single-Component Equilibria. Fluid
Phase Equilib. 2015, 388, 169.

K https://dx.doi.org/10.1021/acs.iecr.9b06891
Ind. Eng. Chem. Res. XXXX, XXX, XXX−XXX