Desalination 228 (2008) 263276

Use of zeolites for treating natural gas co-produced waters

in Wyoming, USA
Hongting Zhaoa, George F. Vancea*, Girisha K. Ganjegunteb,
Michael A. Urynowiczc
a

Department of Renewable Resources, cDepartment of Civil & Architectural Engineering,

1000 E. University Avenue, University of Wyoming, Laramie, WY 82071, USA
Tel. +1 (307) 766-2297; Fax: +1 (307) 766-6403; email: gfv@uwyo.edu
b
Department of Soil and Crop Sciences, Texas Agricultural Experiment Station, Texas A&M University System,
1380 A&M Circle, El Paso, TX 79927, USA
Received 29 May 2007; Accepted 14 August 2007

Abstract
Fast development of the coalbed natural gas (CBNG) industry in many parts of the western US has resulted in
the co-production of potentially saline-sodic waters, hereafter referred to as CBNG water. Management of CBNG
water is a major environmental challenge because of its quantity and quality. In this study, the potential utilization
of three calcium (Ca2+)-rich zeolites were examined for removal of sodium (Na+) from CBNG waters. The zeolite
samples examined were from the St. Cloud (ST) mine in Winston, NM, the Bear River (BR) mine in Preston, ID,
and a mine in Arizona (AZ). The zeolite materials were used in adsorption and column experiments to evaluate
sorption as a function of water chemistry, particle size and flow rate. Regeneration and potential reuse of the media
were also examined. Surrogate CBNG waters that simulated the water chemistry of CBNG waters were used in the
different studies. Although the AZ-zeolite is a Ca2+-rich chabazite, results from column studies indicate there was
essentially no Na+ exchange, suggesting the Ca2+ is tightly held on chabazite adsorption sites. Results for ST-zeolite
and BR-zeolite, which are clinoptilolite zeolites, indicated that a Langmuir model fit the adsorption data well. The
maximum adsorption capacities from the adsorption curves for ST-zeolite and BR-zeolite were 9.6 and 12.3 (mg/g),
respectively, accounting for approximately 38% and 39% of their measured CEC values. Column studies indicated
that a metric tonne (1,000 kg) of ST-zeolite and BR-zeolite can be used to accumulatively treat 16,000 and 60,000 L
of CBNG water, respectively, in order to lower the sodium adsorption ratio (SAR, mmol1/2L!1/2) of the simulated
CBNG water from 30 to an acceptable level of 10. Based on the results of this study, Na+ removal with clinoptilolite
zeolite appears to be a potential water treatment technology for maximizing the beneficial use of poor-quality CBNG
water. Ongoing studies are evaluating water treatment techniques involving the direct application of zeolite to CBNG
waters and development of a field scale prototype.
Keywords:

Zeolite; Clinoptilolite; Chabazite; Sodic-saline water; Adsorption; Sodium; Calcium; Coalbed

methane co-produced water; Powder River Basin; Sodium adsorption ratio; SAR; Water treatment

*Corresponding author.
0011-9164/08/$ See front matter 2008 Published by Elsevier B.V.
doi:10.1016/j.desal.2007.08.014

264

H. Zhao et al. / Desalination 228 (2008) 263276

1. Introduction
Coalbed natural gas (CBNG) development in
Wyoming, Montana and other western US states
during the past 10 years has resulted in the
generation of a significant energy resource as
well as some serious environmental and regulatory concerns [13]. One of the most serious
concerns of the CBNG development is the coproduction of copious quantities of potentially
poor quality groundwater required to recover the
natural gas. Although the quality of CBNG water
varies within the western US, it is often not
suitable for direct irrigation. Within the Powder
River Basin (PRB) of Wyoming, CBNG water is
dominated by sodium (Na+) and bicarbonate
(HCO3!) ions, with pH ranging from 6.8 to greater
than 8.0, electrical conductivity (EC) from 0.4 to
4.9 dS m!1, sodium adsorption ratio (SAR) from
a low of 5 to a high of 70 mmol1/2 L!1/2 and total
dissolved solid concentrations from 270 to
2720 mg L!1 [1]. The SAR is a measure of the
relative preponderance of Na+ to calcium (Ca2+)
and magnesium (Mg2+) to water, and it is defined
as:
SAR (mmol1/2 L!1/2) = [Na+]/ [Ca2++ Mg2+]1/2 (1)
where Na+, Ca2+, and Mg2+ represent millimolar
concentrations (mmol L!1) of the respective ions.
The primary concern with CBNG co-produced
waters is the amount of Na+ and its influence on
the environments. Long-term irrigation of soils
with water of high SAR values can result in
deterioration of soil physical and chemical properties such as soil infiltration and permeability,
which can render soils unsuitable for plant
growth [13]. Therefore, cost-effective technologies are needed in order to allow the CBNG
waters to be used for beneficial purposes.
Natural zeolites have been effectively utilized
to remove toxic elements from polluted waters
[46]. Different zeolites exhibit distinctive cation
exchange capacities (CEC) and adsorption pro-

perties, and have cation preferences that result in

cation selectivity. Clinoptilolite and chabazite
zeolites have a preference for larger cations. For
example, clinoptilolite cation selectivity follows
the order of: Cs > Rb > K > NH4 > Ba > Sr > Na
> Ca > Fe > Al > Mg > Li, whereas chabazite
selectivity is Tl > Cs > K > Ag > Rb > NH4 >
Pb > Na = Ba > Sr > Ca > Li [5]. The slightly
higher preference of Na+ over Ca2+ of clinoptilolite zeolites makes them a better choice for
removing Na+ from CBNG waters.
Zeolites are hydrated aluminosilicates of alkaline and alkaline-earth metals. About 40 natural
zeolites have been identified during the past 200
years [7], with the most predominant consisting
of analcime, chabazite, clinoptilolite, erionite,
ferrierite, heulandite, laumontite, mordenite, and
phillipsite. More than 150 zeolites have been
synthesized, of which the zeolite types A, X, Y
and ZMS-5 are the most common [7]. Natural and
synthetic zeolites are used commercially because
of their unique adsorption, ion-exchange, molecular sieve and catalytic properties.
Commercial zeolite deposits in the US are
associated with the alteration of volcanic tuffs in
alkaline lake deposits and open hydrologic systems, and are located in Arizona, California,
Idaho, Nevada, New Mexico, Oregon, Texas,
Utah, and Wyoming. Zeolites in these deposits
are chabazite, clinoptilolite, erionite, mordenite,
and phillipsite [5]. Other components, such as
orthoclase and plagioclase feldspars, smectite,
opal, quartz and volcanic glass are present in the
different deposits.
Some of the advantages of utilizing natural
zeolite as a cation exchanger for CBNG water
treatment are the large number of deposits
throughout the western US, low mining cost and
their capacity to retain cations. The objectives of
this study were to (1) evaluate the potential of
naturally Ca2+-rich zeolites to reduce Na+ concentration in CBNG waters and (2) develop a costeffective treatment technology for utilizing
CBNG waters for beneficial purposes. A reduc-

265

H. Zhao et al. / Desalination 228 (2008) 263276

tion in the amount of Na+ and the release of

exchangeable Ca2+ from zeolites facilitates the
lowering of SAR levels, resulting in CBNG
waters that can be beneficially used by industry,
land owners and other downstream users. A
bench scale study was conducted to evaluate the
applicability of natural zeolites in order to
develop a cost-effective treatment system for
removing Na+ from CBNG waters.

CBNG water used in this study (Table 2) was

developed by compiling published data from
various sources [3,8] and prepared using appropriate amounts of inorganic chemicals and
tetraethyl orthosilicate as the silicon (Si) source.
2.2. Material characterization
X-ray analysis was conducted on the different
zeolite samples. Samples were first dried at
approximately 80C for several hours with water
loss ranging from <1% to 6%. Pulverized samples
were scanned using Cu K radiation at 1.54 .
Zeolite exchangeable cation concentrations and
cation exchange capacities (CEC) were determined by a modified method of Cerri et al. [9].
Briefly, 1.0 g of zeolite sample was reacted for
2 h with 33 ml of 1 M ammonium acetate
(NH4OAc) at room temperature followed by
centrifugation and filtration (Whatman filter #42)
of the supernatant. This process was repeated two
more times with the three extracts combined in a
100 ml volumetric flask that was brought to
volume using 1 M NH4OAc. The sum of major
exchangeable cation concentrations was used to

2. Experimental
2.1. Materials and chemicals
Upon consideration and comparison of the
cost, availability and potential performance, the
following natural zeolites were selected: two
commercially available natural clinoptilolite
samples obtained from the St. Cloud zeolite (STZeolite) mine in Winslow, NM and the Bear
River zeolite (BR-Zeolite) mine in Preston, ID,
and a commercial pre-activated Ca-rich chabazite
from Arizona (AZ-Zeolite) [Cabsorb Chabazite
Products (Cabsorb ZS500RW)], provided by
GSA Resources Inc. (http://www.gsaresources.
com) (Table 1). The water chemistry of surrogate

Table 1
Exchangeable cations (meq/100 g) and effective CEC (meq/100 g) of Ca-rich zeolites used for treatment of CBNG waters
Sample

Designation

Na+

Ca2+

Mg2+

K+

CEC

Zeolite-Clinoptilolite (St. Cloud, NM)

Zeolite-Clinoptilolite (Bear River, ID)
Zeolite-Chabazite (Arizona)a

ST-Zeolite
BR-Zeolite
AZ-Chabazite

4
10

77
67
Dominant

18
3

13
56

112
136
250

According to the product specification provided by the manufacturer (http://www.gsaresources.com).

Table 2
Water chemistry of synthetic CBNG waters (mg/L) used in the study
Parameters

Na

Ca

Mg

Cl

SO4

HCO3

SiO2

SAR

pH

CBNG Water I
CBNG Water II
CBNG Water III

205
411
822

10
10
10

2.5
2.5
2.5

1.25
2.5
2.5

45.4
65.8
65.8

5
10
10

505
1010
1010

5
10
10

15.5
30.1
60.2

8.20
8.90
8.90

266

H. Zhao et al. / Desalination 228 (2008) 263276

determine the effective cation exchange capacity

(ECEC).
2.3. Batch adsorption kinetic and equilibrium
studies
Adsorption kinetic studies were carried out
using 20.0 g zeolite samples diluted with the surrogate CBNG water to 1,000 ml at ambient
temperature. Aliquots of these samples (~5 ml)
were collected at various times using a syringe
with a GHP Acrodisc 0.2 m filter attached. The
samples were then acidified using a few drops of
concentrated nitric acid (HNO3) and stored until
analyzed.
Adsorption studies were conducted using the
batch equilibration technique [10]. A 5.0 g zeolite
sample was weighed into 50 ml polyethylene
bottles containing varying concentrations of
NaCl. The tubes were then filled with 45 ml of
sodium bicarbonate (NaHCO3) solution with a
concentration of 1,010 mg/L as HCO3!. The initial
Na+ concentrations in the tubes ranged from 0 to
5,000 mg/L. The tubes were shaken on a reciprocal shaker for 48 h, centrifuged, and the supernatant solution filtered and acidified as described
above.
2.4. Packed column and regeneration study
A continuous flow-through column apparatus
was utilized in order to predict zeolite performance under practical treatment system conditions. Column parameters were as follows: weight
of zeolite, 132.0 g; column diameter, 25.0 mm;
column media length, 250 mm (for ST-zeolite)
and 260 mm (for BR-zeolite); flow rates, 3.0, 6.0
or 16 ml/min; and bed volume, 125 cm3. Effluent
was collected using an automated fraction collector. For the Arizona chabazite, column set-up
parameters were: weight of zeolite, 32.2 g;
column diameter, 15.0 mm; flow rate, 5.0 ml/min;
and column media length, 34.0 cm. Regeneration
studies on the spent media were conducted using

0.1 M CaCl2 at a flow rate that was half of that

used for the CBNG water. The columns were
extensively rinsed with 34 L of DI water to
remove pore residual CaCl2 (checked by AgNO3
solution).
2.5. Instrumental
The concentrations of cations (Na+, Ca2+,
Mg2+) in solution were analyzed using inductively
coupled plasma (ICP) spectroscopy (Thermo
Jarrell Ash, model P300) with K+ analyzed using
atomic absorption (AA) spectroscopy (Perkin
Elmer, model 2380). The solutions were also
measured for pH and electrical conductivity (EC)
using pH and conductivity meters, respectively.

3. Results and discussion

3.1. Material characterization
X-ray diffraction (XRD) analysis (data not
shown) indicated a significant occurrence of
clinoptilolite in ST-zeolite samples (80% and
85% clinoptilolite, which was similar to the
manufacturers specifications). For ST-zeolite,
other minerals identified by XRD analysis
included smectite, quartz, cristobalite, feldspar
and illite. For BR-zeolite, XRD analysis also
confirmed a dominant presence of clinoptilolite
(8592%) with no apparent impurities, with the
balance of materials comprised of chalcedony.
Total chemical analysis indicated zeolite composition of ST-zeolite as SiO2 (66.9%), Al2O3
(12.0%), TiO2 (0.21%), Fe2O3 (1.44%), MnO
(0.07%), CaO (2.87%), Na2O (0.44%), K2O
(3.44%), P2O5 (0.06%), MgO (1.2%) and loss on
ignition values (LOI) (11.1%), and BR-zeolite as
SiO2 (66.5%), Al2O3 (10.4%), TiO2 (0.28%),
Fe2O3 (1.99%), CaO (2.33%), Na2O (0.49%),
K2O (4.32%), MgO (0.57%) and LOI (12.5%).
Exchangeable cations and effective CEC for
the zeolites are listed in Table 1. The measured
CECs were 112 and 136 meq/100g for ST-zeolite

H. Zhao et al. / Desalination 228 (2008) 263276

and BR-zeolite, respectively. Both ST-zeolite and

BR-zeolite samples are naturally enriched with
exchangeable Ca2+. It is interesting to note that
BR-zeolite contains considerable exchangeable
K+ (56 meq/100 g) as compared to that of
ST-zeolite (13 meq/100 g) (Table 1).
The Ca-rich Arizona zeolite (Cabsorb
ZS500RW) was used as received from the company, and has a composition of SiO2 (69.5%),
Al2O3 (16.6%), TiO2 (0.47%), Fe2O3 (4.33%),
CaO (4.49%), Na2O (2.40%), K2O (1.32%) and
MgO (0.89%). According to the manufacturer
this product originated from a natural high Na
chabazite that was thermally treated at low to
high temperatures. The upgrading processes
ranged from none to partial recrystallization with
the exchangeable cations in the final product
dominated by Ca2+.
3.2. Adsorption kinetic and equilibrium studies
3.2.1. Adsorption kinetics
Fig. 1 indicates that Na+ adsorption rates are
faster for BR-zeolite compared to ST-zeolite. In
addition, BR-zeolite removed more Na+ and
released greater amounts of Ca2+ than ST-zeolite.
Other major exchangeable cations (i.e., K+ and/or
Mg2+) were also replaced by Na+ and released into
the solution. More Ca2+ and K+ were released by
BR-zeolite than from ST-zeolite. Since K+ is an
important plant nutrient, increased levels in the
treated water would be beneficial for agricultural
irrigation purposes.
Adsorption kinetics is related to many factors
including particle size, pore structure and solute
diffusion. Generally, two kinetic models of
adsorption have been frequently used: (1) active
available site model and (2) chemical reaction
rate model, which is based upon the mass law
concept. For many adsorption processes occurring on heterogeneous materials, it has been
found that the pseudo-second-order kinetic equation agrees well with chemisorption as the ratecontrolling step. The pseudo-second-order kinetic

267

rate equation can be expressed as [11,12]:

2
d Qt
= k * (Qeq Qt )
dt

(2)

where Qeq is the sorption capacity at equilibrium,

Qt is the solid-phase loading of Na+ at time t
(min) and k (g/mmol*min) is the pseudo-secondorder rate constant.
Considering the boundary conditions of Qt =
0 (at t = 0) and Qt = Qt (at t = t), the following
linear equation can be obtained:

t
1
1
= +
t
Qt v0 Qeq

(3)

v0 = k * Qeq2

(4)

where v0 (mmol/g*min) is the initial adsorption

rate. Therefore, by plotting t versus t/Qt, v0 and
Qeq values for kinetic studies can be determined.
Adsorption results show that a 24-h equilibration period (1440 min) was sufficient to establish
steady state or equilibrium for future sorption
experiments using similar particle sizes and initial
Na+ concentrations. The results also suggest that
equilibrium may not be obtained in the pilot tests
where empty bed contact times (EBCT) are on
the order of 35 min [13].
Adsorption data (Fig. 1) were analyzed by the
pseudo-second-order kinetic model to estimate
rate constants, initial sorption rates, and zeolite
adsorption capacities for Na+. Relevant parameters are summarized in Table 3. High coefficient of determination values (R2 ~0.970.99)
indicated that the adsorption of Na+ on the tested
media were well described using the pseudosecond-order kinetic model. Approximately 22%
to 32% of the initial Na+ in solution (about
411 mg/L) was removed by the end of the 24-h
equilibration period. Rate constants and initial
adsorption rates decrease in the order of BRZeolite > ST-Zeolite with BR-Zeolite having the

268

H. Zhao et al. / Desalination 228 (2008) 263276

Fig. 1. Na+ adsorption and Ca2+, Mg2+ and K+ release kinetics for ST-zeolite (1440 mesh) and BR-zeolite (1440 mesh).
Results of pseudo-second-order model fitting are shown for Na+ adsorption. Note scale differences on the y-axis.
Table 3
Kinetic parameters for Na+ adsorption by zeolites in CBNG Water II using pseudo-second-order model
Media

R2

v0

Qeq (mmol/g)

Qeq (mg/g)

ST-Zeolite (1440)
BR-Zeolite (1440)

0.972
0.999

0.036
0.127

1.3610!3
1.0510!2

0.195
0.287

4.47
6.60

R2 = model-fitting coefficient. K = pseudo-second-order rate constant for the kinetic model (g/mmol*min); v0 = initial
adsorption rate (mmol/g*min); Qeq = sorption capacity at equilibrium (mmolAs/g) or mg/g.

highest equilibrium sorption capacity in the

kinetic tests.
Differences in adsorption by the two zeolites
can be attributed to variations in the physical and
chemical properties of the media (i.e., compo-

sition, pore size, surface area, surface charge,

sodium affinity). Adsorption kinetic parameters
are informative in predicting Na+ removal from
CBNG waters. Diffusion coefficients are also
important parameters for predicting the diffusion

H. Zhao et al. / Desalination 228 (2008) 263276

269

rate of adsorbate molecules in porous media [13].

In order to determine the effective diffusivity of
Na+ in the macropores of adsorbent media, a
diffusion equation for a macropore-controlled
system on a spherical shell element of the
adsorbent particle, proposed by Ruthven (1984)
[14], was used:

(1 ) qt +
p

2 c 2 c
c
= p Dp 2 +
(5)
t
R R
R

where p is adsorbent particle porosity, q(R,t) is

the adsorbed phase concentration (g/L of adsorbate), t (s) is time, c (g/L) is the adsorbate concentration in the supernatant solution, Dp (cm2/s)
is the macropore diffusivity of adsorbate in the
adsorbent, and R (cm) is the radial distance from
the center of the adsorbent particle. The macropore diffusivity (Dp) is assumed to be independent of concentration in the above equation. For
fractional adsorption uptake (Mt /Mmax), the solution for the equation with a 2% error is given by
[13]:

2 De t
Mt
6
1 2 exp
R 2
M max

(6)

where Mt (g/L) is the mass gain of adsorbent at

time t, Mmax (g/L) is the mass gain of adsorbent
at infinite time, and De (cm2/s) is the effective
diffusivity defined by:

De =

p Dp

p + (1 p ) K

(7)

A plot of ln[1!(Mt /Mmax)] vs. time should

generate a straight line with a slope of

2 De

2
R
p

Fig. 2. Adsorption of Na+ by ST-zeolite (1440) and

BR-zeolite (1440).

and an intercept of

6
ln 2

from which the effective diffusivity, De, for Na+
diffusion in the macropores of the media can be
calculated.
The values for effective diffusivity De (cm2/s)
calculated based on data points with Mt /Mmax
above 40% and assuming an average particle size
of 0.315 mm, are 2.0110!9 (R2 = 0.939) for STzeolite (1440) and 7.0310!9 (R2 = 0.948) for
BR-zeolite (1440), respectively. BR-zeolite has
a higher De value than that of ST-zeolite.
3.2.2. Adsorption equilibrium studies
The Na+ adsorption curves showed L-type
adsorption on ST-zeolite and BR-zeolite (Fig. 2).
The L-type adsorption curves are described well
using the Langmuir equation [10]:

q=

qmax KC
1 + KC

(8)

where q is the amount adsorbed (mg/g), qmax is the

maximum adsorption capacity of the solid phase,

270

H. Zhao et al. / Desalination 228 (2008) 263276

K is the Langmuir constant and C is the solution

Na+ concentration at equilibrium. Rearranging to
a linear form, Eq. (7) becomes:

C
1
C
=
+
q Kqmax qmax

(9)

Plotting C/q vs. C results in a slope of 1/(qmax)

and intercept of 1/(Kqmax).
The BR-zeolite had a greater adsorption
affinity for Na+ than did the ST-zeolite at the
same equilibrium solution Na+ concentrations.
The maximum Na+ adsorption capacities from the
adsorption curves (Fig. 2 and Table 4) for STzeolite and BR-zeolite are 9.6 and 12.3 (mg/g) or
0.42 and 0.54 (meq/g), which accounts for
approximately 38% and 39% of their measured
effective CEC values, respectively. The previously reported cation selectivity order indicated
Na+ is only slightly preferred over Ca2+ on
clinoptilolite, while K+ is more strongly adsorbed
[5]. The binding strength constant of BR-zeolite
is higher than that of ST-zeolite, indicating that
Na+ could be more strongly adsorbed by BRzeolite. The adsorption results are consistent with
the findings observed in the kinetic studies.
Cation adsorption behavior (i.e., adsorption
capacity and selectivity) by zeolites are related to
a variety of factors such as mineral purity, crystallinity, media purity, cation valence/hydration
degree and pre-existing cations, ion concentration
and multiple adsorption sites with varying degree
of accessibility as well as rate-limiting steps
(mass transfer and diffusion) [10]. Although STzeolite contains slightly higher exchangeable Ca2+
concentrations than that of BR-zeolite, the ease of
which Ca2+ exchanged from ST-zeolite is lower
than that of the BR-zeolite, suggesting that the
Ca2+ in ST-zeolite may occupy stronger adsorption sites. The higher K+ in BR-zeolite suggests
the preference for K+ on stronger adsorption sites.
This could result in Ca2+ bound to relatively weak
adsorption sites, making the Ca2+ in BR-zeolite

Table 4
Langmuir adsorption parameters for the adsorption of Na+
by ST-Zeolite (1440 mesh) and BR-Zeolite (14
40 mesh)
Sample

qmax (mg/g)

R2

ST-Zeolite
BR-Zeolite

9.6
12.3

872
1053

0.965
0.983

qmax = maximum adsorption capacity.

K = binding strength constant.

more easily replaced by Na+ than that in STZeolite. Overall, the adsorption results of this
study suggest that both zeolites can be used to
reduce Na+ level in CBNG waters by releasing
Ca2+, Mg2+ and K+, with the BR-zeolite having a
higher Na+ adsorption capability than ST-zeolite.
Considering charge balance, the amount of
Ca2+, Mg2+ and K+ released accounted for more
than 85% of the adsorbed Na+ on both ST-zeolite
and BR-zeolite, which suggests that the difference is probably due to the release of other
cations. Long-term experiments may provide
better evidence for the removal of Na+ from
CBNG waters.
3.3. Effect of water chemistry and particle size
The effects of water chemistry and zeolite
particle size on the CBNG water quality were
also studied. Similar to the previously described
adsorption studies, a one-point adsorption study
was carried out using BR-zeolite samples with
varying particle sizes in different CBNG waters
(Fig. 3). Results indicate the finer particle size
materials had a greater reduction in SAR of the
three CBNG waters, i.e., 1440 < 40100
100140 !140 !325. In porous materials
such as zeolites, the majority of reactive sites are
on internal surfaces. For example, ST-zeolite
(1440) has been characterized [15] to have an
average pore diameter of 4 , an internal CEC of

H. Zhao et al. / Desalination 228 (2008) 263276

271

Fig. 3. Effect of particle size and water chemistry

(Table 2) on SAR values by BR-zeolite (1440, 40100,
100140, !140 and !325 mesh, respectively).

about 0.8 meq/g, and an external CEC of about

0.1 meq/g. Therefore, diffusional resistance to
mass transport within larger particle-sized zeolites is generally higher than that of smaller
particle-sized zeolites, with some internal surfaces of the larger particles not utilized for
adsorption. As a result, smaller particle-sized
zeolite materials would generally give rise to
higher adsorption rates and capacities due to
shorter paths within and between particles.
3.4. Packed column study and data analysis
Column studies are useful for designing
scaled-up pilot projects for future applications.
Column performance depends upon a variety of
factors that include adsorbent type, particle size,
adsorption capacity, column diameter, water flow
rate, adsorbent bed depth, weight of adsorbent in

Fig. 4. Column breakthrough curves for ST-zeolite

(1440 mesh) in CBNG water II and corresponding
change in effluent SAR values (flow rate = 3.0 ml/min).

column, and contact time as well as water chemistry [13]. Results of column breakthrough studies
using 1440 ST-zeolite and BR-zeolite at a flow
rate of 3.0 and 6.0 ml/min are shown in Figs. 47.
The column performance results are consistent
with the trends observed in the batch adsorption
kinetic and equilibrium studies. Monitoring of the
effluent solution pH indicated that, as compared
to influent solution (pH 8.900.10), there was no
significant change in pH values of the leachates
for ST-zeolite (pH 9.000.30) and for BR-zeolite
(pH 8.900.10). BR-zeolite was found to remove
more Na+ from the synthetic CBNG waters and
that there was a longer break-through period for

272

H. Zhao et al. / Desalination 228 (2008) 263276

Fig. 6. Column breakthrough curves for AZ-chabazite

(1440 mesh) in CBNG water II and corresponding
change in effluent SAR values (flow rate = 5.0 ml/min).

Fig. 5. Column breakthrough curves for BR-zeolite

(1440 mesh) in CBNG water II and corresponding
change in effluent SAR values (flow rate = 3.0 ml/min).

the BR-zeolite compared to the ST-zeolite. It is

interesting to note that, for BR-zeolite significant
K+ was also released, which is an important plant
nutrient.
There was almost no removal of Na+ or Ca2+
released in the column study for the AZ-chabazite
(Fig. 6), though it has Ca2+ as the dominant
exchangeable cation and high CEC, suggesting
that the Ca2+ in the pre-activated Ca-rich chabazite is strongly adsorbed. This may be attributed
to the upgrading treatment process of the raw
materials, which, according to the manufacturer,
involves heating and recrystallization.
If one considers CBNG water with SAR
values equal to or less than 10 to be suitable for
land application, SAR breakthrough curves
(Figs. 4 and 5) can be used to estimate flow-

through treatment system processes that results in

effluent SAR of 10 mmol1/2 L!1/2. This would
allow for prediction of the amount of CBNG
water (SAR = 30 mmol1/2L!1/2) that can be treated
per metric ton (1,000 kg) of ST-zeolite and BRzeolite. Our results indicate about 8,000 and
30,000 L of CBNG water (SAR = 30) could be
treated by a ton of ST-zeolite (1440) and BRzeolite (1440), respectively.
In the Powder River Basin, WY, CBNG water
is generally pumped into reservoirs or ponds.
Therefore, if CBNG water were treated in a
column and continuously collected in a large
containment reservoir, CBNG waters processed
initially would have very low SAR values (n10)
with progressive treatment resulting in water with
higher SAR values (>10). Thus, if treated waters
with high SAR (>10) were mixed with the
initially treated waters with low SAR (n10), then
the overall SAR would be less. In order to estimate the overall water volume that can be treated
to a mean SAR value of 10, the area between the
SAR curve and SAR = 10 cutoff line (lower
shaded area in Figs. 4 and 5) was integrated. The
total amount of additional CBNG waters that
could be treated and combined with the initially
processed waters to reach a final SAR value of 10
would therefore be estimated using an equivalent

H. Zhao et al. / Desalination 228 (2008) 263276

273

Fig. 7. Column breakthrough curves (SAR) as a function of water chemistry, flow rate and particle size: (a) ST-zeolite
(1440 mesh), CBNG water II, flow rate 6.0 ml/min; (b) BR- zeolite (1440 mesh), CBNG water II, flow rate 6.0 ml/min;
(c) BR-zeolite (1440 mesh), CBNG water III, flow rate 3.0 ml/min; (d) BR-zeolite (40100 mesh), CBNG water II, flow
rate 16.0 ml/min.

area above SAR 10 (upper shaded area). This

concept of accumulative treatable water volume
is used throughout the following studies.
Extrapolating the results suggests that one
metric ton (1,000 kg) of ST-zeolite (1440) or
BR-zeolite (1440) can treat approximately
16,000 or 60,000 L of the synthetic CBNG
Water II (SAR = 30), respectively, to a final SAR
around 10, which is considered safe for land
application for agricultural use. Therefore, BRzeolite can treat nearly four times more water
than ST-zeolite. However, the maximum adsorption capacities obtained from the adsorption

curves were 12.3 and 9.6 (mg/g) for BR-Zeolite

and ST-Zeolite, respectively, suggesting BRzeolite can only treat 1.3 times more than that of
ST-zeolite. This large discrepancy could presumably be attributed to the lower effective
diffusivity (De) for Na+ diffusion in the macropores of the ST-zeolite than BR-zeolite as
observed in the kinetic studies.
3.4.1. Effect of flow rate, particle size and
water chemistry
Column study results (Figs. 4, 5, 7a,b) show
that when flow rates were increased from 3.0 to

274

H. Zhao et al. / Desalination 228 (2008) 263276

6.0 ml/min, the treatable water volume by one ton

of ST-zeolite (1440) and BR-zeolite (1440)
was reduced from 16,000 or 60,000 L to 11,000
and 50,000 L/ton, respectively. This would imply
that the reaction is controlled by an external diffusion process. Reduction in contact time results
in less Na+ removed and lower concentrations of
Ca2+ and Mg2+ released.
The breakthrough curve of CBNG Water III in
BR-zeolite (1440) column) is shown in Fig. 7c.
SAR and Na+ concentration in CBNG Water III
are twice as high as that of CBNG Water II. Thus
an increase in the initial Na+ concentration in the
CBNG Water resulted in a lower treatable water
volume, which was approximately 21,000 L/ton
compared to 60,000 L/ton for CBNG Water II.
Based upon results in Fig. 3, it is reasonable to
assume that more water can be treated if the
CBNG water contains less Na+. However, for
practical CBNG water treatment, the actual treatable water volume will be dependent upon its
composition chemistry and other factors.
A column study using 40100 BR-zeolite at a
flow rate of 16 ml/min (Fig. 7d) indicated a
volume of treatable water of approximately
60,000 L/ton, which is similar to that observed
for 1440 BR-zeolite at a flow rate of 3.0 ml/min
(Fig. 5). This suggests that the finer zeolite particle size and higher flow rate are capable of
producing a similar treatment as that using larger
particle size and a lower flow rate. Fig. 3 suggests that finer particle size zeolites could result
in a greater reduction in SAR; however, higher
flow rates could result in less Na+ adsorption and
Ca2+ release due to the limit of diffusion process
depending on zeolite particle size.
3.4.2. Regeneration of spent media and recycle
A preliminary regeneration study was conducted using BR-zeolite (1440). Fig. 8a indicates that a majority (90%) of the adsorbed Na+
can be replaced by Ca2+ after leaching 2 L 0.1 M
CaCl2 solution (Ca2+ = 4,000 mg/L), suggesting
the spent media can be recycled. In addition, the

Fig. 8. Regeneration of spent BR-zeolite column in 0.1 M

CaCl2 solution (a) BR-zeolite (1440 mesh) and (b) BRzeolite (40100 mesh).

BR-zeolite (40100) (flow rate = 16 ml/min) columns were also regenerated using a 0.1 M CaCl2
solution. Results of the regeneration effluent
solution composition (Fig. 8b) indicated that ~2 L
0.1 M CaCl2 solution should be sufficient to
remove a majority of the adsorbed Na+, which is
consistent with regeneration results obtained on
1440 BR-zeolite column study at a flow rate of
6.0 ml/min. Increasing the concentration of the
regeneration CaCl2 solution could result in less
CaCl2 solution used as well as a lower volume of
spent brine generated, thus reducing treatment
cost.

H. Zhao et al. / Desalination 228 (2008) 263276

Fig. 9. Column breakthrough curves (SAR) for the

regenerated BR-zeolite (40100 mesh) in CBNG water II
at a flow rate of 16.0 ml/min.

The regenerated BR-Zeolite (40100) column

was subjected to further CBNG Water II treatment (Fig. 9). As compared to the first treatment
using the BR-zeolite (40100), more CBNG
water was treated upon regeneration (76,000 L).
This can be attributed to a higher concentration of
Ca2+ (4,000 mg/L) in CaCl2 regeneration solutions
resulting in further replacement of un-exchanged
cations (i.e., Na+, K+, Mg2+) upon treatment in
CBNG water, thus making more Ca2+ available
for replacement and greater treatable water
volumes.
3.5. Economic feasibility
Prices for natural zeolite vary with zeolite
content, processing, and world markets [16,17].
Unit values for the 2005 zeolite market, obtained
through a US Geological Survey of domestic zeolite producers, ranged from $50 to $140 per
metric ton. The bulk of the tonnage sold ranged
from $80 and $140 per ton. The prices for industrial or agricultural applications ranged from $30
to $70 per ton for granular products down to
40 mesh and from $50 to $120 per ton for finer
(!40 to +325 mesh) ground material. Therefore,
not considering other expenses and the potentials

275

of regeneration and reuse of the media, the cost of

treating CBNG water with ST-zeolite is about
$0.801.00 per barrel (159 L), for BR-zeolite
approximately $0.200.40 per barrel. If repeated
use of the column is considered, the cost would
be reduced.
It is well known that zeolites have many
applications as adsorbents in agriculture, water
treatment, aquaculture, household products, industry and air/gas treatment [5,15]. For example,
zeolites are used for products such as pet litter,
fish tank media, or odor control applications with
prices ranging from $0.50 to $4.50/kg. Therefore,
by exploring applications of the spent zeolite
media for other locally, regional and/or nationwide uses, there could be a significant reduction
in the treatment cost, making the use of zeolite
treatment a more feasible, cost-effective technology for CBNG water treatment.
Acknowledgement
This research was supported by the US Geological Society, Wyoming Water Development
Commission, USDA National Research Initiative
grant (GFV) and a US Department of Energy
special grant to the State of Wyoming for investigating the treatment of CBNG waters using
zeolites. We appreciate the X-ray diffraction
analysis provided by Robert Gregory of the
Wyoming State Geological Survey.
References
[1] G.F. Vance, L.A. King and G.K. Ganjegunte,
Coalbed methane co-produced water: Management
options, Reflections, June (2004) 3134.
[2] G.K. Ganjegunte, G.F. Vance and L.A. King, Soil
chemical changes resulting from irrigation with water
co-produced with coalbed natural gas, J. Environ.
Qual., 34 (2005) 22172227.
[3] R.E. Jackson and K.J. Reddy, Geochemistry of
CBNG produced water in the Powder River Basin:
Salinity and sodicity. Water, Air, Soil Poll., On-line,
2007.

276

H. Zhao et al. / Desalination 228 (2008) 263276

[4] R.M. Barrer, ed., Zeolites and Clay Minerals as

Sorbents and Molecular Sieves, Academic Press,
London, 1978.
[5] F.A. Mumpton, La roca magica: Uses of natural
zeolites in agriculture and industry. Proc. Natl. Acad.
Sci. USA, 96 (1999) 34633470.
[6] U. Wingenfelder, C. Gerhardfurrer and R. Schulin,
Removal of heavy metals from mine waters by natural zeolites, Environ. Sci. Technol., 39 (2005) 4606
4613.
[7] V.J. Inglezakis, The concept of capacity in zeolite
ion-exchange systems, J. Colloid Interf. Sci., 281
(2005) 6879.
[8] Bureau of Land Management (BLM), Final Environmental Impact Statement and Proposed Plan Amendment for the Powder River Basin Oil and Gas
Project, Vols. 14, US Department of Interior, BLM,
Casper, WY, 2003.
[9] G. Cerri, A. Langella, M. Pansini and P. Cappelletti,
Methods of determining cation exchange capacities
for clinotilolite-rich rocks of the Logudoro region in
Northern Sardinia, Italy. Clays Clay Miner., 50
(2002) 127135.
[10] H.T. Zhao, Y.J. Deng, J.B. Harsh, M. Flury and
J. Boyle, Alternation of kaolinite to cancrinite and
sodalite by simulated Hanford Tank Waste and its

[11]

[12]

[13]

[14]
[15]

[16]
[17]

impact on cesium retention. Clays Clay Miner., 52

(2004) 113.
Y.S. Ho and G. McKay, Pseudo-second order model
for sorption processes, Process. Biochem., 34 (1999)
451465.
Y.S. Ho and G. McKay, The kinetics of sorption of
divalent metal ions onto sphagnum moss peat, Water
Res., 34 (2000) 735742.
M.D. Siegel, A.R. Aragon, H.T. Zhao, S.G. Deng,
M. Nocon and M.J. Aragon, Prediction of arsenic
removal by adsorptive media: Comparison of field
and laboratory studies, in: A. Satinder, ed., Arsenic
Contamination of Groundwater: Mechanism, Analysis, and Remediation, Wiley, Hoboken, NJ, 2007.
D.M. Ruthven, Principles of Adsorption and Adsorption Processes, Wiley-Interscience, New York, 1984.
G.S. Austin and R.S. Bowman, St. Cloud Mining
Companys Stone House Zeolite Mine and new uses
for clinoptilolite, Proc. Soc. for Mining, Metallurgy
and Exploration (SME) National Meeting, SME,
Littleton, CO, USA, 2002.
R.L. Virta, US Geological Survey Minerals Yearbook, 2004.
V.S. Gillerman, M.J. Weaver and E.H. Bennett,
Idaho Mining and Exploration, Idaho Geological
Survey, 2005.