This article appeared in a journal published by Elsevier.

The attached
copy is furnished to the author for internal non-commercial research
and education use, including for instruction at the authors institution
and sharing with colleagues.
Other uses, including reproduction and distribution, or selling or
licensing copies, or posting to personal, institutional or third party
websites are prohibited.
In most cases authors are permitted to post their version of the
article (e.g. in Word or Tex form) to their personal website or
institutional repository. Authors requiring further information
regarding Elsevier’s archiving and manuscript policies are
encouraged to visit:
http://www.elsevier.com/copyright
 Author's personal copy

Chemical Engineering Science 66 (2011) 6209–6219

Contents lists available at SciVerse ScienceDirect

Chemical Engineering Science

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

Design of continuous ion exchange process for the wastewater treatment

Roman Bochenek, Robert Sitarz, Dorota Antos n
Department of Chemical Engineering and Processing, Rzeszow University of Technology, ul. W.Pola 2, 35-959 Rzeszow, Poland

a r t i c l e i n f o abstract

Article history: A generic design procedure for the continuous ion exchange process is proposed. The procedure is
Received 18 May 2011 based on the optimized arrangement of parallel batch columns. The continuity of the process is
Received in revised form achieved by proper shifting of the inlet conditions. The method for the optimization of the process
28 August 2011
variables is presented.
Accepted 30 August 2011
Available online 7 September 2011
The concept is demonstrated on the example of ammonia removal from wastewaters. Two
flowsheet schemes are considered utilizing fresh or recycled regenerating agent.
Keywords: The superiority of the optimized process over the periodic operation is demonstrated.
Dynamic simulation & 2011 Elsevier Ltd. All rights reserved.
Ion exchange
Optimization
Pollution
Wastewater treatment
Zeolites

1. Introduction considered attractive because of the relative simplicity of applica-

tion (Blanchard et al., 1984) and in many cases is proven to be
Ion exchange has numerous applications for industry as well as economic and effective technique to remove ions from waste-
laboratory research. It is used in production of various acids, bases, waters, particularly from diluted solutions (Pintar et al., 2001;
salts, for industrial drying and treatment of gases, in biomolecular Valverde et al., 2006).
separations as well as in the food industry. Ion exhange processes is The operating cycle of ion exchange fixed bed (column)
also commonly used in treatment of drinking and wastewater in consists of two working modes: saturation and regeneration. In
commercial and industrial applications such as water softening, the first mode the feed stream containing wastewater is delivered
demineralization and decontamination (Spiro, 2009). into the column inlet and the effluent free of the polluting ion is
The ion exchange process comprises of the interchange of ions withdrawn at the column outlet. The saturation is interrupted
between a solution and an insoluble solid, i.e., polymeric or when the concentration of the polluting ion in the effluent
mineralic ion exchangers such as ion exchange resins (functiona- exceeds a permissible limit. The bed is regenerated in two stages:
lized porous or gel polymer), natural or synthetic zeolites, elution and re-equilibration. Elution consists of replacing the
montmorillonite, clay, etc. In a water treatment system undesir- retained ions with ions of a regenerant. It is followed by the
able ions in the water supply are replaced with more acceptable column re-equilibration where the regenerant is washed out and
ions. Water decontamination consists of removal of ionic pollu- the column is brought back to the initial conditions. The alternate
tans such as phosphate, nitrate, ammonia, which appear in modes of saturation and regeneration result in a periodic dis-
various types of agricultural, domestic and industrial wastewaters tribution of concentration in the effluent stream characteristic for
or heavy metals discharged in effluent from electroplating plants, processes of concentration swing adsorption. After a number of
metal finishing operations, as well as a number of mining and cycles the process attains the cyclic steady state (CSS) in which
electronics industries (Pintar et al., 2001; Da˛browski et al., 2004; distribution of concentration in each cycle is identical.
Spiro, 2009). The process can be realized using a single bed as well as a system
Although other physicochemical methods of water purification of interconnected beads based on variation of upstream and down-
such as chemical reactions, electro-flotation, reverse osmosis and stream conditions (Nilchan and Pantelides, 1998). The performance of
adsorption may be under given conditions more effective than such a periodic operation is strongly affected by a number of
ion exchange (Flores and Cabassud, 1999), the latter process is operating variables. Therefore, the design of effective process requires
solving an optimization problem aimed at minimizing operating costs
of the process subject to the constraints of the product purity
n
Corresponding author. (Nilchan and Pantelides, 1998; Cruz et al., 2003; Jiang et al., 2004;
E-mail address: dorota.antos@prz.edu.pl (D. Antos). Cruz et al., 2005; Agarwal et al., 2010).

0009-2509/$ - see front matter & 2011 Elsevier Ltd. All rights reserved.
doi:10.1016/j.ces.2011.08.046
 Author's personal copy

6210 R. Bochenek et al. / Chemical Engineering Science 66 (2011) 6209–6219

The design procedures for various periodic processes of ion simple evolutionary strategy (Storn and Price, 1997) for non-linear,
exchange have been discussed in several publications concerning a non-differentiable and non-convex problems with continuous
batch system based on a single column operating periodically (Sun variables.
et al., 1990; Lewus et al., 1998; Castel et al., 2000) as well as a The production rate of purified water (process product) was
continuous system utilizing multicolumn setups such as the carousel used as the performance criterion and the goal function of the
system (Softening et al., 1990; Ernest et al., 1997; Hritzko et al., 2000; optimization procedure.
Huckman et al., 2001; Rochette, 2006). The carousel process refers to It has been shown that the optimized continuous process can
the operation systems in which feed and regenerating streams are significantly outperform the periodic operation.
delivered subsequently through the columns mimicking counter-
current movement of the fluid flow and the columns. The movement
is realized by switching regenerant and feed ports in equal time 2. Principles of the process
intervals. However, the volume of the feed stream used for the
column saturation typically is much larger compared to that required 2.1. Cyclic process of ion exchange in a single column
in the regenerating stages. The equality of switching intervals may
cause that the flowrate of streams delivered in the regeneration mode In the simplest case, the cyclic ion exchange system comprises a
has to be set lower compared to that used during the saturation. Such single batch column operating in a periodic manner, which is based
a forced flowrate adjustment is a serious drawback of the carousel on alternate stages of the column saturation and regeneration.
process limiting the process performance. The flowsheet scheme of the process is presented in Fig. 1.
In this study an alternative procedure for the design of the During the saturation period (stage I) the column is loaded
continuous ion exchange process in a multicolumn setup has been with wastewater (i.e., with the feed stream), the ionic pollutants
proposed. The procedure is based on the parallel delivery of feed and are retained on the stationary phase and the purified water
regenerating streams into the column inlets in such a way that in the stream is withdrawn at the column outlet. The operation of
same time interval a certain number of columns operate in the saturation is stopped when the pollutant concentration in the
out,I
saturation mode while the rest of them are regenerated. The number purified water, C P , exceeds permissible upper limit, EpsI.
of columns operating in the saturation mode is time-invariant, which In the next stage the column has to be regenerated. The
assures the continuity of the process. The operating cycles of regeneration comprises two operations: elution (stage II) where the
consecutive columns are sequentially shifted in time. The length of regenerant is delivered into the column to elute the retained ions in
the time intervals for the saturation and regeneration can be adjusted the form of concentrated solutions, and the column re-equilibration
to assure optimal performance of the overall process. (stage III), where the column is rinsed out to remove the regenerant
To verify the concept suggested a model system of ion exchange and restore initial conditions. The elution stage is stopped when the
process has been selected, i.e., the removal of ammonia ions from concentration of the ionic pollutant in the column effluent, Cpout,II ,
wastewaters on a synthetic zeolite has been analyzed. Two flowsheet drops below a threshold concentration, EpsII.
schemes have been considered differing in the manner of the column The last stage, i.e., of re-equilibration is interrupted when the
regeneration—in the first one a fresh solution of the regenerant was concentration of the regenerant in the effluent stream, Crout,III ,
used for the column regeneration whereas in the second one the decreases below the threshold concentration, EpsIII.
exhausted regenerant was partly recycled and reused in the regen- The stream of the wastewater purified in the saturation mode
eration mode. For the process design and optimization a dynamic is partly consumed for the preparation of regenerating solutions
model was employed based on the equation of the linear driving force in the elution and re-equilibration stages (VwII , VwIII , see Fig. 1).
in the solid phase (i.e., LDF model). The Differential Evolution (DE) To improve the process performance, i.e., to reduce the water
procedure was adopted for the process optimization based on a consumption, the effluent from the elution stage (stage II)

Fig. 1. Flowsheet scheme for complete operating cycle on a single column without regenerant recycling, QI, QII, QIII are the flowrates of the inlet streams in the stages I, II,
III, respectively, Qrecov is the streams of purified water, VwII , VwIII are the volumes of the purified water consumed for supplementing the streams in the elution and the
re-equilibration stages, respectively.
 Author's personal copy

R. Bochenek et al. / Chemical Engineering Science 66 (2011) 6209–6219 6211

Fig. 2. Flowsheet scheme for complete operating cycle on a single column with regenerant recycling.

Fig. 3. Gantt chart for a setup of eight parallel column operating in a periodic node.

containing exhausted regenerant can be partly recycled (see the feed and product flowrates of the node are multiplication of
Fig. 2). The recycled part is collected and temporarily stored in the corresponding flowrates of the single column.
a buffer tank to be completely used up in the initial period of the The operating cycles for a setup of eight parallel columns are
elution stage II of the consecutive cycle. The exhausted regenerant illustrated in Gantt chart in Fig. 3.
is collected within a certain time interval, whose length is Large-scale processes, which require large amounts of product,
restricted by the threshold concentration of the pollutant in the are usually more cost-effective for continuous operations. There-
rec
buffer tank, C P rEpsrec . Subsequently, in the same stage, the fore, in this study a continuous process of the wastewater
column is regenerated with a portion of fresh regenerant. treatment was considered. In this process the operating cycles
The contribution of the exhausted regenerant to the whole of subsequent columns are shifted in time so that in each time
regeneration stream used in the elution stage is determined by interval a certain number of columns could operate in the
the recycling ratio U. saturation mode while remaining ones in the regeneration mode.
Since the last step of re-equilibration is designed for washing Because the time length necessary for accomplishing the satura-
out the regenerant as well as other impurities clogging the bed, tion and regeneration stages are different to assure the process
the effluent of the step III is discharged; it is not recycled to the continuity some of the columns have to wait for the use in a
process. subsequent cycle in the work stoppage. Therefore, a complete
operating cycle for each column of the continuous node com-
prises four stages: saturation, elution, re-equilibration and the
2.2. Continuous process of ion exchange work stoppage. Gantt chart for an eight-column continuous setup
is presented in Fig. 4. It can be observed that after the startup of
The throughput of the ion exchange process can be increased the operation in each time interval four columns are loaded with
by utilizing a system of several identical columns arranged in the feed stream, i.e., they work in the saturation mode, four are
parallel. If all these columns are simultaneously operated under regenerated or remain in the work stoppage. In this case the
the same conditions the node works periodically. The values of process conditions are optimized for the column operating
 Author's personal copy

6212 R. Bochenek et al. / Chemical Engineering Science 66 (2011) 6209–6219

Fig. 4. Gantt chart for a setup of eight column operating in the continuous mode; individual column optimization.

Fig. 5. Gantt chart for continuous system of eight-column system, node column optimization.

individually but not optimized for the column operating jointly in work stoppage. The sum of the productive and non-productive units
the multicolumn node. has to be equal to the number of the column in the node:
The work stoppages cause lengthening of the cycle time, which
NP þ NR ¼ Ncol ð1Þ
reduces the process effectiveness. Therefore, optimizing the
process parameters for individual columns (individual column The number of productive time units can be calculated as
optimization) and synchronizing their cycles to achieve contin- follows:
uous operation is not sufficient; to upgrade the process perfor-  I
t

mance, optimization of the process parameters for the column NP ¼ INT c Ncol ð2Þ
t
operating jointly in the system (node optimization) is required.
Gantt chart for an optimized eight-column node is presented where INT is an integer function rounding down; tc is the total
in Fig. 5. working cycle time given by the following equation:
The throughput of the optimized node is significantly higher
t c ¼ t I þt II þ t III ð3Þ
compared to that obtained for the individual column optimiza-
tion. In the case illustrated in Fig. 5, five columns are running in where tI, tII, tIII are the lengths of the time periods for saturation,
the saturation mode, three in the regeneration mode or in the elution and re-equilibration, respectively.
work stoppage. In this case the length of the work stoppages is so The length of the time unit is expressed as
short that they are not noticeable in the chart. tI
To synchronize the runs of columns assembled in the contin- tj ¼ ð4Þ
NP
uous system the shift time between the cycles of consecutive
columns has to be adjusted in such a way that the number of The stoppage time can be calculated as follows:
columns operating simultaneously in each of two modes (i.e., t stop
¼ NR tj ðt II þ t III Þ ð5Þ
saturation and regeneration) is time-invariant, which assures
continuity of the feed delivery and the product withdrawal. The For the total operating time in the node it holds:
shift time corresponds to the switching time of the inlet condi- tot
t ¼ t I þt II þ t III þ t stop ð6Þ
tions. It can be determined based on the concept of time unit,
which is elucidated as follows. Additionally, the following dependence is fulfilled:
Each stage consists of a certain number of the time units; the
Ncol tj ¼ t I þ t II þt III þt stop ð7Þ
total number of the time units for all the stages is equal to the
number of columns in the node. The length of the periods of saturation, elution and re-
The time units can be divided into two categories: productive units, equilibration, tI, tII, tIII, can be determined on the basis of
NP, corresponding to the saturation mode, and non-productive units, simulations of operation cycles for a node column running under
NR, corresponding to the periods of elution, re-equilibration and the CSS conditions.
 Author's personal copy

R. Bochenek et al. / Chemical Engineering Science 66 (2011) 6209–6219 6213

3. Process optimization determining the following process variables:

3.1. Optimization procedure  volumetric flowrates Q in the stages I, II, III, which are related
to the time lengths: tI, tII, tIII. Increasing the flowrates results in
To solve the optimization problem the Differential Evolu- reducing the operation time, which improves the process
tion (DE) method was used. It utilizes the evolutionary performance. On the other hand, it alters the HETP value, i.e.,
algorithm belonging to the class of stochastic population based it can reduce the column efficiency due to worsening condi-
methods that proved their efficacy in solving hard nonlinear tions for the mass transport. Therefore, determining the values
global optimization problems involving noncontinuous, of flowrates for each stage requires solving an optimization
nondifferentiable and nonconvex functions in the optimiza- problem.
tion models (Michalewicz and Fogel, 2002; Onwubolu and Typically, the regenerant is not adsorbed; therefore, in the re-
Babu, 2004). equilibration stage (stage III) the mass transport kinetics does
The evolutionary techniques had found wide application for not affect the process efficiency. In such a case the flowrate QIII
solving various problems of chemical and process engineering can be limited only by the pressure drop in the system or the
such as batch operation scheduling, designing various subsys- pump capacity;
tems, e.g., heat exchanger network or water network, molecular  shift time of cycles of subsequent columns (i.e., the switching
design, designing processes and apparatus (Deb, 2001; Babu et al., time) and the number of columns operating in individual
2005; Jeżowski et al., 2005, 2007; Babu and Angira, 2006; modes in each time unit. These variables can be calculated
Bochenek et al., 2007; Srinivas and Rangaiah, 2007). on the basis of the model simulations if the flowrates Qk (k¼I,
The DE method is a simple evolutionary strategy proposed by II, III) are known;
Storn and Price (1997) for solving nonlinear, nondifferentiable  threshold concentration of the polluting ion in the column
and nonconvex optimization models with continuous variables. It effluent in the saturation stage, i.e., the water purity, EpsI. This
can handle both unconstrained and constrained optimization value can be set arbitrarily according to the permissible
problems and was proved for its high performance for solving concentration limit;
multi-modal problems (Price et al., 2005). The algorithm of the  threshold concentration of the polluting ion at the end of the
method is illustrated in Fig. 6. elution stage, EpsII.
The value EpsII should be considered as an indicator of the
regeneration efficiency. Complete regeneration, i.e., complete
removal of the ionic pollutant from the column (EpsII  0)
3.2. Formulation of optimization problem usually requires long elution time, which reduces the process
effectiveness. On the other hand incomplete regeneration
The problem of the optimal design of the water purification causes reduction of the loading capacity of the bed. Therefore,
node that consists of a given number of columns relies on the value of EpsII should be optimized;

Fig. 6. Algorithm of the DE method, see details in (Price et al., 2005).

 Author's personal copy

6214 R. Bochenek et al. / Chemical Engineering Science 66 (2011) 6209–6219

 threshold concentration of the regenerant at the end of the re- phase concentration, respectively; u is the superficial velocity; ee and
equilibration stage, EpsIII; et are the external, internal and total porosity; DL is the dispersion
the threshold concentration, EpsIII, can be selected on a level coefficient, and km,i is the overall mass transport coefficient.
that mimics the effect of the hardness of the wastewater The coefficient km,i lumps the contribution of external and
contained in the feed stream; internal mass transport kinetics to the overall mass transport
 additionally, for the system involving recycling: the recycling mechanism. It also depends on the isotherm course as follows
ratio U and the threshold concentration of the pollutant in the (Kaczmarski et al., 2001; Antos et al., 2003):
recycled stream, Esprec; !
d2p
2
1 k00i 1 ee Fe dp

The threshold concentration Epsrec should be selected properly; ¼ þ 1 þ ð16Þ
low value of Epsrec results in higher water consumption whereas kmi ð1 þ k0 Þ2 k1i 0i
et 6keff ,i 60Dintra,i
too high Epsrec causes reduction of the regeneration efficiency.
Finding the best values of Epsrec and U is an optimization problem. dqni
k00i ¼ Ft ð17Þ
dCi
As a criterion of the process performance the productivity of
dqn
 
pure water can be taken into account. The productivity can be k1i ¼ Fe ep þ ð1ep Þ i ð18Þ
dCi
defined as the volume (or mass) flow of the recovered water,
Qrecov: where Fe ¼ 1ee =ee and Ft ¼ 1et =et are the phase ratio:
Q I t I VwII VwIII dqni
Pr ¼ Qrecov ¼ w ð10Þ Dintra, i ¼ ep Dpeff , i þ ð1ep ÞDseff , i ð19Þ
t tot dCi
where VwII and VwIII are the volumes of the purified water consumed where keff is the effective external mass transport coefficient in
in the stages II and III, respectively, to supplement the streams QII ionic solutions; Dpeff, Dseff are the effective coefficients of diffusion
and QIII (see Figs. 1 and 2). in macropores and micropores, respectively; ep is the internal
To optimize the process performance the following problem porosity. For a linear isotherm dqni =dCi is the Henry constant.
can be solved: For self-sharpening fronts the equations presented above can
max ½PrðvÞ ð11Þ be used, provided that the isotherm slope dqni =dCi is replaced with
the isotherm chord Dqni =DCi (Lapidus and Amundson, 1952;
where v is a vector of the decision variables.
Kaczmarski et al., 2001; Antos et al., 2003).
The optimization problem is solved subject to the constraint
The model is solved with the boundary and initial conditions:
concerning the purity of the recovered water:
Initial conditions:
out,I
CP rEpsI ð12Þ Ci ðt ¼ 0,xÞ ¼ Ci0 ð20Þ
Z tI Cpout,I dt
out,I
Cp ¼ qi ðt ¼ 0,xÞ ¼ q0i ð21Þ
0 t
Typically, the column is free of the polluting ions in the initial
In the recycling mode additional constraint should be added
state Ci0 ¼ 0; q0i ¼ 0 (i¼p).
that is related to the concentration of pollutant in the recycled
Boundary conditions
stream collected at the column outlet:
At the column inlet the Danckwerts-type boundary conditions
rec
C p rEpsrec ð13Þ can be assumed:

The following decision variables can be taken into account: @Ci ðt,x ¼ 0Þ
uðCiF ðtÞCi ðt,x ¼ 0ÞÞ ¼ ee DL ð22Þ
v ¼[QI, QII, EpsII, Epsrec, U] @x
The two last variables Epsrec, U are active for the process where CiF ðtÞ is the profile of concentration at the column intlet. In
involving the regenerant recycling. the saturation stage CiF ðtÞ stands for the feed concentration for the
The optimization problem should be solved for the system eluite ion as well as the regenerant ions while in the regeneration
operating under CSS conditions, which correspond to repeatabil- stage it describes the regenerant profile. For step changes of the
ity of the concentration profiles at the column outlet for each inlet concentration in k stage it holds:
consecutive cycle.
CiF ðtÞ ¼ Ciin,k k ¼ I, II, III ð23Þ
3.3. Modeling of the process dynamics where Ciin, I
¼ CiF
At the column outlet continuity of the concentration profile is
To predict the process performance a simplified dynamic assumed:
model based on the equation of linear driving force (LDF model) @ Ci ðt,x ¼ LÞ
was used (Ruthven, 1984). ¼0 ð24Þ
@x
The LDF model consists of the mass balance equation in the
mobile phase: The model has to be coupled with the isotherm equation, e.g.,
for the Donnan isotherm model it holds:
@Ci @q @2 C @C
et þ ð1et Þ i ¼ ee DL 2i u i ð14Þ qni
 n zi =zc
q gm gsc zi =zc
 
@t @t @x @x ¼ Ki c , Ki ¼ is ð25Þ
Eq. (14) is coupled with the following kinetic equation:
Ci Cc gi gm
c

@qi Eq. (25) is combined with the conditions of electroneutrality in

¼ km,i ðqni qi Þ ð15Þ the mobile phase:
@t
where i denotes the system component, e.g., eluite (pollutant) ion zi Ci þ zc Cc þ zco Cco ¼ 0 ð26Þ
i¼p, and co-ion i¼co, t, x are the time and the axial coordinates, and in the solid phase:
respectively; Ci(t, x) is the concentration in the mobile phase; qi(t, x),
qni ðt,xÞ are the solid phase concentration and the equilibrium solid zi qni þzc qnc þ zm G ¼ 0 ð27Þ
 Author's personal copy

R. Bochenek et al. / Chemical Engineering Science 66 (2011) 6209–6219 6215

where z is the ionic charge, m denotes the solid matrix (zm ¼  1 Dispersion coefficient:
for the cation exchanger); and G is the exchange capacity. the dispersion coefficient was calculated according to the
The coefficient K lumps contribution of activity coefficients of Gunn equation (Gunn, 1987).
ions in the mobile gm and the solid phase gs. All the model parameters were assumed to be time invariant,
Typically, the coefficient K is a function of the regenerant i.e., gradual bed destruction or deactivation due to clogging,
concentration, r, i.e., salt supplying counter- and co-ions. The ageing, etc., was not taken into account.
same can holds for G, which can decrease with the increase of the The remaining operating variables were set as described
ionic strength due to the reduction of the electrostatic potential of below:
the solid matrix. Therefore, additional dependencies should be
quantified (Górka et al., 2007). - feed composition, i.e., the inlet composition in the stage I:

K ¼ f ðCr Þ ð28Þ CpF ¼ NH þ ¼ 0:0153mol=L

4

CrF ¼ NaCl ¼ Cco

F
¼ Cl ¼ 0:0017 mol=L
G ¼ f ðCr Þ ð29Þ

A relatively low inlet concentration of the ionic pollutant,

4. Results and discussion CpF ¼ NH þ , was selected for the simulations. It corresponded to
4
weakly non-linear range of the isotherm for which ion-
4.1. Model parameters exchange process can be effective; the use of concentrated
solutions is unprofitable due to the fast saturation of the bed
The model parameters used for the simulations were the same loading capacity. Nevertheless, the principles of the optimiza-
as determined in our previous study for the system of ammonia tion procedure presented here are valid for any initial
removal on a zeolite ion exchanger with NaCl salt as the concentration.
regenerant (Górka et al., 2007). - concentration of the regenerant (salt NaCl) in the stage II:
The bed parameters were assumed as follows: Crin,II ¼ 0:5mol=L ð3%wt=vÞ
the column dimensions: column length L¼25 cm, column dia-
meter dcol ¼10 cm; the bed porosities were et ¼0.7, ee ¼0.5, ep ¼0.3.
The following values and dependencies of the isotherm and - threshold concentrations:
kinetic coefficients were used: pollutant concentration in the collected stream (i.e., permis-
sible limit):
– empirical dependencies of the isotherm and kinetic coefficients EpsI ¼ 1  1024 mol=L
for the eluite ion, p, on the concentration of counter-ions, Cc,
supplied by the regenerant, i.e.,
threshold concentration of the regenerating agent:
 for the isotherm coefficients:
EpsIII ¼ CRF ¼ 0:0017 mol=L
GðCc Þ ¼ 2:45 þ3:07 expðCc =0:02Þ

- flowrate of the water stream in the re-equilibration stage III:

Kp ðCc Þ ¼ 16:4215:41 expðCc =0:07Þ Q III ¼ 6 L=min

To quantify these concentration dependencies the isotherm The value of QIII was set as maximal regarding the system
course was determined for different contents of the regen- capacity limitation.
erant in the mobile phase. For each set of equilibrium data
the isotherm coefficients G and Kp were estimated. Next, G
and Kp were appropriately correlated with the regenerant 4.2. Process simulation
concentration (see details in Górka et al., 2007).
Because co-ions were not adsorbed, the coefficient, Kco was To simulate the process dynamics the model described by Eqs.
set equal to zero (Kco ¼0). (14)–(29) was employed with parameters reported above in
 for the effective external mass transport coefficient: Section (5.1). The model was solved for the eluite ion, i.e.,
ammonia ions, NH4þ , and for co-ions, Cl  . The concentration of
keff ,p ¼ 1:0  105 ðCc Þ0:37
the counter-ions was calculated using the electroneutrality con-
ditions (Eqs. (26) and (27)). Note that because the ionic charge of
The influence of flowrate variations on the keff value was co-ions was 9zN9 ¼1 the inlet co-ion concentration was equal to
in,k
accounted for by the relation of Wilson and Geankoplis the concentration of the regenerant salt Cco ¼ Crin,k (k¼I, II, III).
(1966), i.e.: Local values of the isotherm and kinetic coefficients were
 calculated according to the temporal concentration profile of
keff ,p ¼ f ðwÞ0:333 the counter-ion using Eqs. (28) and (29).
To simulate the breakthrough profiles in the saturation mode
For co-ions the mass transport resistances were neglected. the isotherm chord was used in Eqs. (17)–(19), whereas for
Note that for non-absorbable species the mass transport predicting the regenerating profiles the isotherm slope was
resistances are negligible (see Eq. 16 at k00i -0). calculated. Details of the procedure can be found elsewhere
Diffusion coefficients: (Górka et al., 2007).
The simulation of consecutive stages of the process relied on
Dseff ,p ¼ 4:97  10211 ½m2 =s
proper exchanging the boundary conditions (Eqs. (22) and (23)).
The contribution of the pore diffusion Dpeff,p to the mass To simulate the process including regenerant recycling the
transport mechanism could be neglected (Górka et al., 2007). additional sub-stages in the regeneration mode had to be taken
 Author's personal copy

6216 R. Bochenek et al. / Chemical Engineering Science 66 (2011) 6209–6219

into account. The effluent containing exhausted regenerant was When U is set the outlet concentration CpU can be calculated
assumed to be discharged to the waste within the period (tI, according to Eq. (31) and the corresponding time for the elution of
tcollect). At the time point, tcollect, the effluent started to be collected concentration CpU , i.e., trec, can be found.
in a buffer tank. The whole storage is used up in stage II of a next As mentioned above the optimization problem should be
cycle. The value of tcollect was calculated according to the average solved for the system operating in the cyclic steady state. The
rec
concentration C P using the following mass balance equation: CSS was attained after few cycles, i.e., after second cycle for the
column operating without recycling and after fourth for the
Z t collect recycling mode. The simulation of the start up of the process is
rec
Q II C p t rec ¼ Q II Cpout,II ðtÞdt þQ II EpsI ðt II t collect Þ ð30Þ demonstrated in Figs. 7 and 8.
t II

where trec is the recycling duration. The last term on the left hand 4.3. Optimization results
side of the equation corresponds to the mass of the pollutant in
the stream of purified water coming from stage I (see Fig. 2). 4.3.1. Single column optimization
The recycling duration trec is determined by the choice of the The cyclic process of ion exchange was optimized for an single
recycling ratio U that was defined as follows: column operating according to the flowsheet scheme involving
the regenerant recycling as well as without recycling. The
Cpout,II,max CpU ðt rec Þ objective function was defined by Eq. (10), the consumption of
U¼ ð31Þ
Cpout,II,max EpsII water VwII , VwIII was determined by the water demand in the
regenerating and re-equilibrating streams.
where Cpout,II,max is the maximum of pollutant concentration in the The optimization results are summarized in Table 1. It com-
regeneration profile, CpU is the concentration corresponding to the prises the parameters of the optimization procedure, the searching
end of the recycling period. range of the decision variables and the optimal solution. It is evident

Table 1
Optimization results for the single column optimization.

Without recycling With recycling

Parameters of the optimization procedure

CRa 0.3 0.3
Fa 0.8 0.8
Number of generation 120 150
Number of individuals 10 15

Searching range of the decision variables

QI (L/min) 0.3–10 0.3–5
QII (L/min) 0.3–10 0.3–5
II
Eps (mol/L) 10  4–3  10  3 10  4–4  10  3
rec
Eps (mol/L) – 10  4–4  10  3
U – 0–1

Best solution
Objective function OF (L/min) 1.14 1.46
QF ¼QI (L/min) 2.16 2.67
QII (L/min) 1.38 2.88
EpsII (mol/L) 8.06  10  4 5.06  10  4
Epsrec (mol/L) – 1.33  10  3
Fig. 7. Concentration profiles of ammonia at the column outlet after two cycles, U – 0.710
for the column without recycling; the re-equilibrating stage III is very short,
a
therefore not visible on the plot. See Fig. 6.

Fig. 8. Concentration profiles of ammonia in CSS for first four cycles (column without recycling): (a) complete concentration profile and (b) concentration profile for the
saturation stage I.
 Author's personal copy

R. Bochenek et al. / Chemical Engineering Science 66 (2011) 6209–6219 6217

that the process involving recycling performs better compared to that Two optimization approaches for the optimization were com-
realized without recycling. The performance improvement for the pared: the operating parameters were optimized for a single
process being investigated was ca. 28%. column operating individually in the node (individual column
optimization) and for the column operating jointly in the node
(node column optimization).
4.3.2. Node column optimization The optimization results are summarized in Table 2.
As described above, to increase the throughput a multicolumn It can be observed that for the node column optimization the
node can be used. The node can operate periodically—it then system performs markedly better compared to that for the
comprises a setup of parallel columns operating periodically individual column optimization. The performance differences
under the same conditions (Fig. 3) or continuously, where the result from unproductive gaps in the operating cycles, i.e., the
inlet conditions for each column are sequentially changed in time work stoppages occurring in the continuous operation. The node
(Figs. 4 and 5). column optimization allows minimizing gaps between cycles. The
The productivity, Pr, of the periodic node is just a multi- superiority of the performance of the optimized mode is visua-
plication of that for a single column presented in Table 1. lized in Fig. 9a and b.
The performance optimization for the continuous multicolumn An important observation is the upgrade of the process
node was done for three setups comprising 3, 6 and 8 columns. efficiency for the system based on the regenerant recycling (even

Table 2
Optimization results for the process of ion exchange in a multicolumn node. Parameters of the optimization procedure and the searching range of the decision variables are
the same as reported in Table 1.

Without recycling With recycling

3 Columns 6 Columns 8 Columns 3 Columns 6 Columns 8 Columns

QI (L/min) 2.01 1.99 2.15 2.26 2.27 2.49

QII (L/min) 1.59 1.54 1.39 2.63 2.79 2.80
EpsII (mol/L) 7.97  10  4 8.24  10  4 8.19  10  4 5.14  10  4 4.88  10  4 4.75  10  4
Epsrec (mol/L) – – – 1.12  10  3 1.25  10  3 1.36  10  3
U – – – 0.679 0.715 0.744
OF (L/min), node column optimization 3.39 6.78 9.12 4.19 8.54 11.2
OF (L/min) individual column optimization 1.85 5.54 7.39 2.41 7.22 9.62

Fig. 9. Comparison between the optimization results (a) productivity obtained for the individual column optimization and the node column optimization for the flowsheet
scheme without regenerant recycling; (b) productivity obtained for the individual column optimization and the node column optimization for the flowsheet scheme with
regenerant recycling; (c) comparison between productivity of the nodes with and without recycling.
 Author's personal copy

6218 R. Bochenek et al. / Chemical Engineering Science 66 (2011) 6209–6219

up to 83% for the three columns system) as demonstrated in of such additional empirical correlations into the dynamic model
Fig. 9c. and the optimization procedure is straightforward. In such a case
The productivity achieved in the continuous system with node the operating parameters can be periodically adjusted according
column optimization is slightly lower than that for the periodic to the predicted changes of the bed properties and the process can
node (compare Tables 1 and 2). It results from the necessity of be interrupted to discard the adsorbent bed as it approaches
adjustment of the process parameters to achieve the continuous saturation with the contaminant. If an unpredictable bed destruc-
operation, which reduces the number of degree of freedom in the tion is expected, the process controlling has to be used based on
optimization procedure. Increase of the column number benefits the monitoring of the concentration profiles and adjusting accord-
in a reduction of the time of the work stoppages. Therefore, the ingly the operating variables. The procedure of the process
differences between productivity for the periodic and the con- modeling and optimization developed in this work could be useful
tinuous setups disappear with the increasing column number. for the process design also in this case. However, a detailed
Nevertheless, the observed superiority of the periodic setup analysis of this problem is outside the scope of this study.
over the continuous one is only apparent. The periodic system
operates more efficiently only if the throughput is also periodic,
i.e., water to be treated is stored in a large amount and is 5. Conclusions
processed in the purification node in any time.
If the wastewater delivery is continuous the periodic node A design procedure of the continuous ion exchange process for
performs much worse compared to the continuous system. To the wastewater treatment has been proposed. The node for the
illustrate the problem one can consider a certain wastewater wastewater treatment consisted of a multicolumn setup arranged
stream (throughput) to be processed in a purification node. Accord- in such a way that the time-invariant number of columns could
ing to the given throughput the number of column has to be operate in the productive mode (saturation). The operating cycles
adjusted so that they could operate under optimal conditions. For of consecutive columns in the node were shifted in time; proper
instance, for the throughput (feed flowrate) 4 L/min a set of three adjustment of the shift time assured continuity of the operation.
columns can be used, for 8 L/min six columns while for 10 L/min For the process design a model of the process dynamic was
eight columns. In the continuous node the given throughput stream used, i.e., the LDF model, which was incorporated into the
can be processed without storage, i.e., in 3-columns system two optimization procedure based on the differential evolution
columns are always in the saturation mode, in 6-column system method.
four columns while in 8-column system five columns that are The process conditions were optimized subject to the con-
processing required stream of wastewater. straint of the pollutant concentration in the output water stream
For a periodic node a part of the throughput has to be stored in for the periodic as well as continuous system.
a buffer tank and can be processed in a next cycle. In the example The best results in terms of the process performance were
being studied the given throughput can be is processed without achieved in case of optimization of the operating conditions for
storages only in cycles with even numbers. For cycles with odd the columns operating jointly in the purification node.
numbers the throughput delivered into the node is in an excess The optimization results indicated that continuous process can
with respect the column capacity and has to be partly stored (e.g., significantly outperform the periodic operation.
the throughput 4 L/min can be split into the three columns Additionally, the possibility of the improvement of the process
working parallel into the amount of ca 1.33 L/min per column, productivity was analyzed for the flowsheet scheme involving
which can be processed within the saturation period making ca. recycling of the exhausted post-regenerating stream. Introducing
62% of the total cycle time that gives ca. 2.5 L/min of effective feed the regenerant recycling resulted in reduction of the consumption
flowrate for each odd cycle). The amount of feed stored during of purified water in the regeneration mode and increased the
non-productive period of the cycle can be consumed in the next process performance.
(even) cycle with maximal productivity. The procedure of the process design suggested in this study
It is obvious that the operation of storing reduces the effective was developed for an ideal ion-exchange system for which
throughput of the system. The difference of productivity between gradual changes of bed properties were neglected. However, after
the periodic and optimized continuous node is demonstrated for proper adjustments, it can potentially be used for real systems.
the system without recycling in Table 3.
Similar results can be achieved for the columns operating in a
setup involving recycling (not shown here). Nomenclature
As it can be observed, the advantage of the continuous systems
enhances with increase of the column number (see Table 3). C concentrations in the mobile phase (mol/L)
It should be noted that for the simulations presented an ideal DL dispersion coefficient (m2/s)
ion-exchange system was assumed; no changes of bed properties Dpeff effective coefficients of diffusion in macropores (m2/s)
progressing during a long-time operation was assumed. In a real Dseff effective coefficients of diffusion in micropores (m2/s)
system a gradual loss of the loading capacity due to the bed keff effective interparticle mass transport coefficient (m/s)
contamination can often be observed. If the isotherm coefficients km lumped mass transport coefficient (1/s)
are changing from cycle to cycle in a systematic manner, their K isotherm coefficient (dimensionless)
values could be correlated with the cycle number. Implementation L column length (m)
N number of time units
Table 3 OF objective function (mol/(Lmin))
Performance of different continuous multicolumn setups. Pr productivity
q solid phase concentration (mol/Lsolid)
40 80 100
(L/min) (L/min) (L/min) qn solid phase concentration in equilibrium with the liquid
phase (mol/Lsolid)
OF, optimized continuous node 3.39 6.78 9.12 Q volumetric flow rate (L/min)
OF, periodic node/periodic throughput 3.42 6.84 9.12 u superficial velocity (m/s)
OF, periodic node/continuous throughput 2.74 5.48 7.31
V column volume (m3)
 Author's personal copy

R. Bochenek et al. / Chemical Engineering Science 66 (2011) 6209–6219 6219

x axial coordinate (m) Cruz, P., Santos, J.C., Magalhaes, F.D., Mendes, A., 2003. Cyclic adsorption separa-
z ion charge tion processes: analysis strategy and optimization procedure. Chem. Eng. Sci.
58 (14), 3143–3158.
Da˛browski, A., Hubick, Z., Podkościelny, P., Rubens, E., 2004. Selective removal of
Greek letters the heavy metal ions from waters and industrial wastewaters by ion-exchange
method. Chemosphere 56, 91–106.
Deb, K., 2001. Multi-objective Optimization Using Evolutionary Algorithms. John
g activity coefficient Wiley, Singapore.
G exchange capacity (mol/Lsolid) Ernest, M.V., Bibler, J.P., Whitley, R.D., Linda- Wang, N.H., 1997. Development of a
ee external porosity carousel ion-exchange process for removal of Cesium-137 from alkaline
nuclear waste. Ind. Eng. Chem. Res. 36, 2775–2788.
ep internal porosity Flores, V., Cabassud, C., 1999. A hybrid membrane process for Cu(II) removal from
et total porosity industrial wastewater, comparison with a conventional process system.
Desalination 126, 101–108.
Górka, A., Bochenek, R., Warcho", J., Kaczmarski, K., Antos, D., 2007. Ion exchange
Subscripts kinetics in removal of small ions. effect of salt concentration on inter- and
intraparticle diffusion. Chem. Eng. Sci. 63, 637.
Gunn, D., 1987. Axial and radial dispersion in fixed beds. Chem. Eng. Sci. 42,
c counter-ions
363–373.
co co-ions Hritzko, B.J., Walker, D.D., Linda- Wang, N.H., 2000. Design of a carousel process
p eluite (pollutant) ion for cesium removal using crystalline silicotitanate. AIChE J. 46 (3), 552–564.
r regenerant Huckman, M.E., Latheef, I.M., Anthony, R.G., 2001. Designing a commercial ion-
exchange carousel to treat DOE wastes using CST granules. AIChE J. 47 (6),
1425–1431.
Superscripts Jiang, L., Fox, V.G., Biegler, L.T., 2004. Simulation and optimal design of multiple-
bed pressure swing adsorption systems. AIChE J. 50 (11), 2904–2917.
Jeżowski, J., Bochenek, R., Ziomek, G., 2005. Random search optimization approach
0 initial conditions for highly multi-modal nonlinear problems. Adv. Eng. Software 36 (8),
in the column inlet 504–517.
Jeżowski, J., Bochenek, R., Poplewski, G., 2007. On application of stochastic
k stage index
optimization techniques to designing heat exchanger- and water networks.
F feed Chem. Eng. Process.: Process Intensification 46, 1160–1174.
m mobile phase Kaczmarski, K., Antos, D., Sajonz, H., Sajonz, P., Guiochon, G., 2001. Comparative
out column outlet modeling of breakthrough curves of bovine serum albumin in anion-exchange
chromatography. J. Chromatogr. A. 925, 1.
rec recycling Lapidus, L., Amundson, N.L., 1952. Mathematics of adsorption in beds. the effect of
s solid phase longitudinal diffusion in ion exchange and chromatographic column. J. Phys.
Chem. 56, 984.
Lewus, R.K., Hakan Altan, F., Carta, G., 1998. Protein adsorption and desorption on
gel-filled rigid particles for ion exchange. Ind. Eng. Chem. Res. 37, 1079–1087.
Michalewicz, Z., Fogel, D.B., 2002. How to Solve it: Modern Heuristics. Springer
Verlag, Berlin.
Acknowledgment Nilchan, S., Pantelides, C.C., 1998. On the optimisation of periodic adsorption
processes. Adsorption 4 (2), 113–147.
The financial support of Polish Scientific Committee within the Onwubolu, G.C., Babu, B.V., 2004. New Optimization Techniques in Engineering.
Springer-Verlag, Heidelberg.
grant no. N N209 033538 is greatly acknowledged. Pintar, A., Batista, J., Levec, J., 2001. Integrated ion exchange/catalytic process for
efficient removal of nitrates from drinking water. Chem. Eng. Sci. 56,
1551–1559.
References Price, K.V., Storn, R.M., Lampinen, J.A., 2005. Differential Evolution—A Practical
Approach to Global Optimization. Springer Verlag, Berlin.
Rochette, F., 2006. Novel technology: fresh approach to countercurrent ion
Agarwal, A., Biegler, L.T., Zitney, S.E., 2010. A superstructure-based optimal exchange. Filtration and Separation 43 (7), 18–19.
synthesis of PSA cycles for post-combustion CO2 capture. AIChE J. 56 (7), Ruthven, D.M., 1984. Principles of Adsorption and Adsorption Process. John Willey,
1813–1828. New York.
Antos, D., Kaczmarski, K., Piatkowski, W., Seidel-Morgenstern, A., 2003. Concen- Softening, Y.S., Grevillot, G., Tondeur, D., 1990. Modelling and optimization of the
tration dependence of lumped mass transfer coefficients—linear vs. nonlinear cyclic regime of an ion-exchange process for sugar juice. Chem. Eng. J. 43,
chromatography and isocratic vs. gradient operation. J. Chromatogr. A. 1006, B53–B66.
61. Spiro, D.A., 2009. Ion exchange resins: a retrospective from industrial and
Babu, B.V., Pallavi, G.C., Syed Mubeen, J.H., 2005. Multiobjective differential engineering chemistry research. Ind. Eng. Chem. Res. 48, 388–398.
evolution (MODE) for optimization of adiabatic styrene reactor. Chem. Eng. Srinivas, M., Rangaiah, G.P., 2007. Differential evolution with tabu list for solving
Sci. 60 (17), 4822–4837. non-linear and mixed-integer non-linear programming problems. Ind. Eng.
Babu, B.V., Angira, R., 2006. Modified differential evolution (MDE) for optimization Chem. Res. 46, 7126–7135.
of nonlinear chemical processes. Comput. Chem. Eng. 30, 989–1002. Storn, R., Price, K., 1997. Differential evolution—a simple and efficient heuristic for
Blanchard, G., Maunaye, M., Martin, G., 1984. Removal of heavy metals from global optimization over continuous spaces. J. Global Opt. 11, 341–359.
waters by means of natural zeolites. Water Res. 18, 1501–1507. Sun, Y., Grevillot, G., Tondeur, D., 1990. Modelling and optimization of the cyclic
Bochenek, R., Jezowski, J., Bartman, S., 2007. Optimization of heat exchanger regime of an ion-exchange process for sugar juice softening. Chem. Eng. J. 43,
network with fixed topology by genetic algorithms. Chem. Eng. Trans. 12, 53–66.
195–200. Valverde, J.L., De Lucas, A., Carmona, M., Pérez, J.P., González, M., Rodrı́guez, J.F.,
Castel, C., Schweizer, M., Simonnot, M.O., Sardin, M., 2000. Selective removal of 2006. Minimizing the environmental impact of the regeneration process of an
fuoride ions by a two-way ion-exchange cyclic process. Chem. Eng. Sci. 55, ion exchange bed charged with transition metals. Sep. Purif. Technol. 49,
3341–3352. 167–173.
Cruz, P., Magalha~ es, F.D., Mendes, A., 2005. On the optimization of cyclic Wilson, E.J., Geankoplis, C.J., 1966. Liquid mass transfer at very low Reynolds
adsorption separation processes. AIChE J. 51 (5), 1377–1395. number in packed beds. Ind. Eng. Chem. Fund. 5, 9–14.