Chemical Engineering Journal 321 (2017) 510–520

Contents lists available at ScienceDirect

Chemical Engineering Journal

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

3D modeling of overall adsorption rate of acetaminophen on activated

carbon pellets
R. Ocampo-Perez a,⇑, C.G. Aguilar-Madera b, V. Díaz-Blancas a
a
Center of de Investigation and Postgraduate Studies, Faculty of Chemical Sciences, Autonomous University of San Luis Potosí, 78260 San Luis Potosí, Mexico
b
Earth Sciences Faculty, Universidad Autónoma de Nuevo León, Linares N.L, México 67700, Mexico

h i g h l i g h t s g r a p h i c a l a b s t r a c t

The overall adsorption rate of

acetaminophen on activated carbon
pellets was studied.

3D diffusional model was applied to
predict the concentration decay
curves.

Surface and pore volume diffusion are
important on the adsorption of
acetaminophen.

Desorption zones of acetaminophen
were identified from 3D modeling.

a r t i c l e i n f o a b s t r a c t

Article history: In this work the overall adsorption rate of acetaminophen on activated carbon pellets (ACP) was analyzed
Received 6 February 2017 in deep. The concentration decay curves were interpreted by a 3D diffusional model because the intra-
Received in revised form 27 March 2017 particle diffusion of acetaminophen inside ACP in radial and axial directions are important in this form
Accepted 28 March 2017
of activated carbon. The 3D diffusional model considers the external mass transfer, intraparticle diffusion
Available online 1 April 2017
(pore volume and surface diffusion) and the adsorption on an active site. The results demonstrated that
the application of 3D diffusional model based on pore volume diffusion interprets clearly the kinetic
Keywords:
curves, however values of effective diffusion coefficient (Dep) higher than molecular diffusivity are
Acetaminophen
Adsorption kinetics
obtained indicating superdiffusion phenomenon. On the other hand, the application of a general diffusion
3D modeling model evidenced that during the whole time interval, the acetaminophen diffuses consecutively by sur-
Surface diffusion face diffusion followed by pore volume diffusion. In short times the surface flux is higher than the pore
Poro volume diffusion volume flux, but at higher intervals of time, the relevance of both fluxes reverts. Finally, from 3D simu-
lation it is clear that at longer times, the solute mainly enters from the solution through the pellet bor-
ders, and acetaminophen desorption signs are even evident at the center of the pellet covers, due to the
inverse concentration gradient established between the pellet and the solution.
Ó 2017 Elsevier B.V. All rights reserved.

1. Introduction
Activated carbon (AC) is the most employed adsorbent for
industrial applications, especially for eliminating aromatic com-
⇑ Corresponding author.
E-mail address: raul.ocampo@uaslp.mx (R. Ocampo-Perez).
pounds from water due to its physicochemical and textural proper-
ties [1]. The activated carbon is mainly commercialized in the
granular (GAC), powdered (PAC), cloth, and pellets forms, and their
applications depend on the required residence time, available pres-
sure drop, capacity of regeneration, type and phase of process, and
the cost of the adsorbent.
Adsorption on activated carbon has been recommended by the
USA Environmental Protection Agency as the best available

http://dx.doi.org/10.1016/j.cej.2017.03.137
1385-8947/Ó 2017 Elsevier B.V. All rights reserved.
 R. Ocampo-Perez et al. / Chemical Engineering Journal 321 (2017) 510–520
technology for removing non-biodegradable toxic organic com-
pounds from drinking water and industrial wastewater [2], and
in the last decade has been extensively used to remove pharma-
ceutical compounds from aqueous solutions [3–7]. From literature,
it is well known that the adsorption mechanism of pharmaceutical
compounds on carbonaceous materials is governed by interactions
between p electron from aromatic rings of adsorbate and p elec-
tron of graphenic planes of surface of activated carbon [8–10]. Fur-
thermore, these interactions are favored depending on the content
of activating groups. Additionally, the electrostatic interactions and
formation of hydrogen bond can play and important role on the
adsorption of aromatic compounds [8].
To design appropriate packed adsorption columns, it is not only
necessary to know the adsorption mechanism; it is also critical to
obtain the adsorption rate and to identify the mass transfer mech-
anism that governs the overall adsorption rate. Empirical models
as pseudo first order [11] and pseudo second order [12] are gener-
ally the most common kinetics models used for describing adsorp-
tion kinetics of several pollutants onto various adsorbents duo to
their simplicity together with the good fit generally obtained
[13–15], however, the application of these models does not provide
information about the influence of operational variables on mass
transfer resistances. On the other hand, the diffusional models con-
sider the external mass transfer, intraparticle diffusion and adsorp-
tion on an active site, yielding a more realistic prediction of the
phenomena [16].
We want to stress that the diffusional model is more complex
when it is used to interpret experimental results. Its mathematical
solution is less trivial than kinetic models, and may be it is one of
the reasons for its unpopularity in the literature. As kinetic models
contain empirical coefficients, then their predictive abilities under
other operative conditions or scales, are diminished. The nature
and physics involved in the coefficients of kinetic models is
unclear, and furthermore their use for conditions different to those
used in the fitted experiment is compromised [17]. Several authors
have employed diffusional models to predict the adsorption rate of
pollutants on activated carbon [16,18–22]. Mendez-Diaz et al. [18]
investigated the adsorption kinetics of dimetridazole, metronida-
zole, ronidazole, and tinidazole onto activated carbon. The results
evidenced that pore volume diffusion governs the adsorption pro-
cess for four nitroimidazoles and the values of Dep varied from
5.84  107 to 20.3  107 cm2/s. Leyva-Ramos et al. [19] studied
the adsorption rate of pentachlorophenol onto granular activated
carbon, and the concentration decay data were predicted by a dif-
fusional model which takes into account adsorption, external mass
transfer, and intraparticle diffusion. The results indicated that the
overall rate of adsorption is mainly controlled by surface diffusion,
and the external mass transfer can be considered negligible.
In references cited above, the adsorbent particles have been
considered spherical and isotropic, and, therefore, the concentra-
tion of pollutant inside the particle is only a function of time and
radial position, which allows reducing the complexity of mathe-
matical modeling. However, it is evident that these assumptions
are not valid when the adsorbent particles have different shapes
such as activated carbon pellets. Thus, this work aims modeling
the overall adsorption rate of acetaminophen as a model com-
pound on activated carbon pellets in aqueous solution through
the application of a three-dimensional diffusional model that takes
into account the external mass transport, intraparticle diffusion,
and adsorption on active sites. A detailed analysis is given for the
intraparticle concentration profiles and the magnitude and direc-
tion of mass fluxes over time.
Acetaminophen was selected as a model pollutant since it is one
of the top prescription medicines worldwide [23]. Acetaminophen
cannot be completely removed in wastewater treatment plants,
and the detected residual concentrations in aqueous solutions
511
range from several hundred ng/L to 23.2 mg/L [24]. During the
chlorination processes significant amounts of acetaminophen are
transformed into 1,4-benzoquinone and N-acetyl-p-benzoquinone
imine increasing the toxicity of water. Furthermore, reports in
the literature suggest that acetaminophen-induced hepatotoxicity
was a major cause of acute liver failure in the United States and
that reactive metabolites of acetaminophen may have contributed
to the pathogenesis of drug-induced toxicity [25,26].
2. Experimental method and mathematical model
2.1. Adsorbate
Acetaminophen was used in this study, and was supplied by
Sigma-Aldrich. The physicochemical properties of this analgesic
are listed in Table 1. The concentrations of acetaminophen in aque-
ous solution were determined by a spectrophotometric method,
and the absorbance was assessed in a double-beam, UV 2600 Shi-
madzu spectrophotometer. The absorbance of the acetaminophen
in a water sample was measured at a wavelength of 244.1 nm.
2.2. Adsorbent
The activated carbon pellets (ACP) used in this work are dis-
tributed by Carbotecnia (Mexico), it is commercially available as
Carvapur and is produced from coconut shell. The average dimen-
sions of ACP were obtained by measuring one hundred pellets with
an optical microscopy, obtaining an average diameter of
3.6 ± 0.032 mm and height of 6.5 ± 0.212 mm. The ACP were previ-
ously washed several times with deionized water, dried in an oven
at 110 °C during 24 h, and stored in a plastic container.
The ACP was chemically and texturally characterized, determin-
ing their surface area, pore size distribution, solid density, void
fraction, surface groups, and pH of the point of zero charge.
Detailed descriptions of techniques and methods used for this
characterization were previously reported [27].
2.3. Method for obtaining the adsorption equilibrium data
The experimental data for the adsorption equilibrium of aceta-
minophen on ACP were obtained in a batch adsorber, which con-
sisted of a centrifugal vial of 50 mL. The procedure is described
as follows: Aqueous solutions of acetaminophen were prepared
in 50 mL volumetric flasks with concentrations ranging from 100
to 1000 mg/L. A portion of 10 mL of the solution was taken out
from each flask to corroborate the initial concentration of acetami-
nophen. A certain mass of ACP and 40 mL of the acetaminophen
solution were poured into the batch adsorber. Afterwards, the
batch adsorber was placed in a constant temperature water bath,
and the ACP and the solution were left in contact for 12 days. Once
equilibrium was attained, a sample of 10 mL was withdrawn and
analyzed to determine the concentration of the acetaminophen
at equilibrium.
2.4. Method for obtaining the rate of adsorption data
A rotating basket batch adsorber was used to obtain the exper-
imental concentration decay curves for the acetaminophen adsorp-
tion on ACP. This adsorber was composed of a 1 L three-neck
reaction flask and an impeller with its blades replaced with stain-
less steel baskets. An acetaminophen solution was poured into the
adsorber and the ACP particles were placed in the stainless steel
mesh baskets, which were attached to a shaft connected to a vari-
able speed motor. The adsorber was partially immersed in a con-
stant temperature water bath controlled by a recirculator. A
512 R. Ocampo-Perez et al. / Chemical Engineering Journal 321 (2017) 510–520
Table 1
Molecular structure and physicochemical properties of acetaminophen.
Compounds Molecular structure Molecular weight (g mol1)
Acetaminophen 151.17
C8H9NO2
a
Experimental value Ribeiro et al. [33].
mass of 2 g of ACP was placed in the baskets and put in contact
with 995 mL of a known concentration solution of acetaminophen
at pH 7. This pH was obtained by mixing proper volumes of 0.01 M
HCl or NaOH solutions. The concentration decay curves were
obtained at initial concentrations of acetaminophen from
100 mg/L to 700 mg/L, and a stirring speed of 200 rpm. The solu-
tion was periodically sampled (1 mL) and analyzed to determine
the acetaminophen concentration at different times.
2.5. Diffusional model
Fig. 1 shows a general scheme of ACP where it is evident that,
due to its dimensions and geometry, the mass transfer of acetami-
nophen in radial and axial directions should be taken into account.
Besides, the mathematical model includes the following assump-
tions: i) intraparticle diffusion occurs by pore volume diffusion
and surface diffusion, ii) the rate of adsorption on an active site
is instantaneous, iii) the particles are rigid, homogeneous and iso-
tropic, and iv) the convective transport inside the porous particle is
negligible in comparison with diffusion. The latter two considera-
tions imply that diffusion coefficients correspond for isotropic
effective diffusion tensors, while dispersion effects are discarded.
The model equations and initial and boundary conditions are as
follows:
Fig. 1. Scheme of ACP showing dimensions, domains of governing equations, and
specific nodes.
Size X, Y, Z (Å) Log(Kow) pKa DAB  106a (cm2/s)
X = 10.74 0.46–0.49 9.86 6.27
Y = 6.58
Z = 3.76
dCA

V ¼ mSkL CA  CAP jBps ð1Þ
dt
@CAP @q

ep þ qp ¼ r  Dep rCAP þ Ds qp rq ð2Þ
@t @t
t ¼ 0 min; CA ¼ 0 mg=L; CAP ¼ 0 mg=L ð3Þ



nps  Dep rCAP jBps þ Ds qp rq ¼ kL CA  CAP jBps ð4Þ
where CA and CAP are the solute concentration in the solution and
inside the particle, respectively, while q represents the mass
adsorbed of acetaminophen; Dep and Ds are the effective and surface
diffusion coefficients, respectively, ep is the void fraction of ACP and
qp represents the particle density. In Eq. (1), CAP jBps denotes the
solute concentration inside the pellet evaluated at the pellet-
solution (p-s) boundary. The Eq. (4) represents continuity condi-
tions of mass flux at the pellet-solution boundary. Finally, if the rate
of adsorption on an active site is instantaneous, then we assume
that local equilibrium is taking place between the acetaminophen
concentration in the solution inside the pore and the mass of acet-
aminophen adsorbed on the pore surface. This equilibrium is nor-
mally represented by the adsorption isotherm, which is the
mathematical relationship between q and CAP :
q ¼ f ðCAP Þ ð5Þ
It is important to clarify that, in our formulation, the solute con-
centration computed with Eq. (1) is apparently just time-
dependent. However, a deeper inspection will show that CA is also
surface-position-dependent since CAP is involved in Eqs. (1) and (4),
and it is a position and time-dependent variable. This is an impor-
tant feature of our formulation that is not commonly explored in
other works where spherical geometries for porous adsorbents
are considered [18,19]. For non-symmetrical particles, the
non-uniformity of intraparticle concentration provokes different
driving forces for mass flux around the porous particle, and fur-
thermore variations of solute concentration at the vicinity of solid
are expected.
The model represented by Eqs. (1)–(4) is the general diffusional
model (PVSDM). The general model can be simplified considering
that only the intraparticle diffusion mechanism may exist either
 
by pore volume diffusion (PVDM) Dep – 0; Ds ¼ 0 cm2 =s or sur-
 
face diffusion (SDM) Dep ¼ 0; Ds – 0 cm2 =s .
3. Results and discussion
3.1. Textural and chemical characterization of activated carbon pellets
According to the IUPAC classification, the N2 adsorption iso-
therm provides information about the type of porosity in materials
[28], as well as the form and geometry of pores. Fig. 2 shows the
adsorption isotherm of N2 at 77 K for ACP, which has a I(b)-type
behavior according to the IUPAC classification, indicating the pres-
ence of a wide distribution in the size of micropores, and also the
presence of narrow mesopores. The N2 adsorption data were used
 R. Ocampo-Perez et al. / Chemical Engineering Journal 321 (2017) 510–520 513

equilibrium to obtain the mathematical relation between q and

0.5 CAP must be evaluated (Eq. (5)).
The external mass transfer coefficient, kL , was assessed by the
procedure proposed by Furusawa and Smith [31], based on the fact
0.4 that when t ! 0 min, then CAP ! 0 mg=L and CA ! CA0 . Substitut-
0.8 ing these conditions in Eq. (1), the following equation can be
Vliq (cm3/g)

0.7
0.3 dV(r), cm3/(g nm)
0.6
0.5
0.4
0.2
0.3
0.2
0.1
0.1
0.0
Adsorption 0.0 0.4 0.8 1.2
Desorption Pore radious, nm
0.0
0.0 0.2 0.4 0.6
P/P0
Fig. 2. Adsorption isotherm of N2 at 77 K on ACP, and distribution of pore size using
DFT.
to estimate the ACP texture properties, listed in Table 2. ACP have a
specific area similar at those found for granular activated carbons
[3,6], with a pore volume of 0.53 cm3/g. Furthermore, the volume
of micropores, Vmic , obtained from the application of the
Dubinin-Radushkevich isotherm [28], demonstrated that ACP are
mainly composed of micropores, with their volume representing
86% of the total pore volume. The distribution of pore size was esti-
mated by applying the density functional theory (DFT), also shown
in Fig. 2, where the average radius of the ACP pore is between 0.4
and 1.6 nm. Finally, ACP had a value of pHPZC = 8.4, meaning their
nature is slightly basic due to the presence of a greater amount
of basic groups in the surface of ACP (See Table 2).
3.2. Concentration decay curves and estimation of mass transport
parameters
Concentration decay curves for acetaminophen on ACP were
obtained under the experimental conditions shown in Table 3.
These experiments were carried out at pH 7, and at rotating speeds
of 200 rpm. This rotating speed was chosen after previous experi-
ments demonstrated that external mass transport does not affect
the overall adsorption rate at rotating speeds greater than
200 rpm. It should be noted that the time required to reach equilib-
rium was a function of the initial concentration of acetaminophen,
ranging from 3000 min for an initial concentration of 105 mg/L up
to 10,000 min for initial concentrations above 400 mg/L. García-
Mateos et al. [29] and Lladó et al. [30] determined that the time
to reach equilibrium of acetaminophen on granular activated car-
bons was between 2000 and 3000 min at a temperature of 25 °C.
To solve the PVSDM, PVDM and SDM models, the values of kL , Ds
and Dep must be estimated. Furthermore, the adsorption
derived:
2
C 3
d A mSkL
4 CA0 5 ¼ ð6Þ
dt V
t¼0min
The term at the right end of Eq. (6) is the slope of the concentra-
tion decay at t ¼ 0 min, and was estimated by using the first two
1.6 2.0 data points of the concentration decay curve, at t ¼ 0 and
t ¼ 1 min. The values of kL estimated with Eq. (6) are given in
Table 3, and ranged from 2.30  103 to 0.88  103 cm/s. These
0.8 1.0 values are within the range of kL values reported by other authors
for rotating basket adsorbers [16,19].
The values of Dep and Ds were simultaneously estimated from
the diffusional model fit to the experimental data using one evolu-
tionary algorithm. Such algorithm tries to minimize the error func-
tion (Eq. (7)) through stochastic procedures. This method might
develop slower convergence in comparison with gradient-based
methods, but a larger lookup field can be set ensuring the finding
of global minimum. Besides, our first simulations showed that
gradient-based methods resulted in numerical errors caused by
rapid gradient changes of the objective function being minimized:
Z
tf  2
error ¼ CA;num  CA;lab dt ð7Þ
0
The mathematical relation q ¼ f ðCAP Þ was obtained from the
evaluation of the adsorption equilibrium of acetaminophen on
ACP at pH 7 and 25 °C. Fig. 3 shows this equilibrium, as well as
the fact that the adsorption isotherm had an ‘‘L”-type behavior,
according to the Giles classification [32], confirming that the acet-
aminophen aromatic rings are adsorbed in parallel on the ACP sur-
face, and that the solute and the water do not compete for the
active sites available on the surface of adsorbent. This mechanism
has been validated by Lladó et al. [30], who also establish that
interactions from hydrogen bonding play an important role in
the adsorption of acetaminophen on activated carbons with high
contents of phenol groups. This figure also shows that the adsorp-
tion capacity reached at an equilibrium concentration of 240 mg/L
was 190 mg/g, which is 25% lower than that obtained by Lladó
et al. [30] using F400 granular activated carbon.
Experimental data were interpreted using the Langmuir iso-
therm model, represented by the following equation:
qm KCAP
q¼ ð8Þ
1 þ KCAP
where qm represents the maximum adsorption capacity in mg/g, and
K is an isotherm constant in L/mg. The qm and K values were calcu-
lated using nonlinear regression, obtaining values of qm = 198.9
mg/g and K = 0.107 L/g. The prediction of the Langmuir model is also

Table 2
Chemical and textural properties of granular activated carbon.

Material SBET (m2 g1) VPa (cm3 g1) dPb (nm) Vmicc (cm3 g1) L0d (nm) epe Total acidic sites (meq g1) Total basic sites (meq g1) pHPZC
ACP 1096 0.53 2.2 0.456 1.38 0.49 0.287 0.304 8.5
a
Total pore volume.
b
Average pore diameter.
c
Micropore volume.
d
Average micropore width.
e
Void fraction of ACP.
514 R. Ocampo-Perez et al. / Chemical Engineering Journal 321 (2017) 510–520

Table 3
Experimental conditions for acetaminophen concentration decay curves during adsorption on pellets of activated carbon at pH = 7, T = 25 °C and 200 rpm.

Exp. No. CA0 (mg/L) m (g) CAe (mg/L) qe (mg/g) kL  103 (cm/s) PVSDM PVDM
8 2 6 2
Ds  10 (cm /s) Dep  10 (cm /s) Dep  105 (cm2/s)
1 105.5 2 1.22 52.16 2.30 0.59 1.94 6.12
2 209.5 2 9.68 99.9 1.80 0.79 6.31 2.16
3 322.4 2 38.19 142.11 1.90 1.39 4.64 1.91
4 417.7 2 71.06 173.33 1.40 1.09 6.29 1.40
5 501.1 2 131.02 185.03 1.20 1.18 5.71 1.10
6 633.3 2 240.6 196.35 0.90 1.29 3.95 0.69
7 708.1 2 292.86 207.63 0.88 1.36 3.09 0.77

PVDM model, using an optimum value of Dep = 7.75  106 cm2/s.

200 This value of Dep was obtained by minimizing Eq. (7) using the evo-
Mass of acetaminophen adsorbed, mg/g

lutionary algorithm method. On the other hand, Fig. 5 depicts the

160
120
80
40
0 nophen concentration in the solution and the concentration of
0 40 80 120
Concentration of acetaminophen at equilibrium, mg/L
Fig. 3. Adsorption isotherm of acetaminophen on ACP at pH 7 and 25 °C. The line
represents the prediction of the Langmuir model.
shown in Fig. 3, where, except for one experimental data, the Lang-
muir model adequately interprets the experimental data.
3.3. Application of pore volume diffusional model
The pore volume diffusion model assumes that Ds = 0, that is,
that the solute is exclusively diffused by pore volume diffusion both
in r and z directions. Fig. 4 exemplifies the experimental data of the
concentration decay curve for Exp. 7 and the prediction of the
Fig. 4. Concentration decay curves of acetaminophen on ACP. The lines represent
the prediction of PVDM using different positions on the particle.
variation of concentration profiles within the particle, CAP, at
2000, 4000, and 6000 min for Exp. 7.
Fig. 4 shows that the PVDM prediction drastically depends on
the position in the pellet-solution boundary where CA is being eval-
uated because, as mentioned before, CA = f(t,r,z). To analyze this
aspect in detail, nodes A, B, and C from Fig. 1 were chosen. The pre-
diction of the PVDM model in Node B did not interpret the data and
clearly overestimated the concentration decay curve, indicating
that a much higher value of Dep is required to obtain a better pre-
diction of experimental data. The slow adsorption rate at this point
is due to the solute is transported both in r and z directions (see
Fig. 5), causing a rapid increase of concentration at this point,
and drastically decreasing the driving force between the acetami-
160 200 240 acetaminophen in node B, resulting in a very slow adsorption rate
at this point. The evaluation in node C produces an adequate pre-
diction of experimental data for times of less than 2000 min, but
predicts an adsorption capacity higher than that observed experi-
mentally. Finally, the prediction in node A shows a two-stage
behavior, the first of which ranges from t = 0 to t = 4000 min, and
shows a gradual drop of concentration similar to that observed in
node C, until reaching a minimum closer to 4000 min; the second
stage shows an increase of the solution concentration, that is, the
solute initiates a phase of partial desorption from the pellet. These
results derive from the fact that, at times of less than 2000 min, the
solute reaches node A mainly from direction z, and, as time
increase, higher quantities of solute diffusing from the radial direc-
tion reach this point, causing an increase in concentration. Simul-
taneously, the concentration of solute in the solution decreases
significantly, reverting the mass flux direction from the pellet
towards the solution. The authors emphasize that these phenom-
ena have not been described previously because it is difficult to
measure it in the experiments, hence the importance of carrying
out rigorous mathematical modeling as well.
Based on these observations, the interpretation of experimental
data must be clearly carried out by averaging the value of CA
throughout the external surface of the pore. This average value
was obtained from the following equation:
R 2p R R R H R 2p
ðCA jz¼0 þ CA jz¼H Þrdrdh þ R CA jr¼R dhdz
Cav
A ¼
0 0 0 0
ð9Þ
2pR2 þ 2pRH
Fig. 4 also includes the prediction of the PVDM model with
Dep = 7.75  106 cm2/s using Cav A . This figure clearly reveals that
the model satisfactorily predicts the experimental data; however,
the value of Dep was higher than the experimental molecular diffusion
coefficient of acetaminophen on water (DAB = 6.27  106 cm2/s)
[33], which has no physical significance, as it means that acetamino-
phen diffuses faster within the pores than in the solution. To
verify this result, experiments 1–6 from Table 3 were also interpreted
with the PVDM model, and optimum Dep values are listed in Table 3,
 R. Ocampo-Perez et al. / Chemical Engineering Journal 321 (2017) 510–520
2000 min
Fig. 5. Evolution of the intraparticle profiles of acetaminophen on ACP as a function of time for Exp. 7.
and predictions shown in Fig. 6, highlighting that the PVDM model
notably interprets the concentration decay curves; however, all val-
ues of Dep showed that Dep > DAB. It is therefore concluded that the
PVDM adequately interprets the experimental data, but the value of
effective diffusion has no physical meaning.
3.4. Application of pore volume and surface diffusional model
The preceding results revealed that surface diffusion might play
an important role in the intraparticle diffusion of acetaminophen
on ACP. Also, considering that ACP have a high adsorption capacity,
concentration gradients in the solid phase may be more important
than those in the fluid phase within the pores, boosting surface dif-
fusion. To corroborate this hypothesis, experimental results were
then interpreted by applying the general diffusional model
(PVSDM). Dep and Ds values were optimized simultaneously until
the minimum global of the objective function was obtained.
Fig. 7 depicts the PVSDM prediction for Exp. 1–7, with practically
every decay curve being interpreted satisfactorily, and the opti-
mum Dep and Ds values are listed in Table 3. Clearly, the optimum
values of Dep proved to be an order of magnitude less than values
of DAB, while the values of Ds proved to be an order of magnitude
less than the values of Ds reported for the diffusion of pyridine
on granular activated carbon [16], but are, in turn, an order of mag-
nitude greater than those obtained in the diffusion of dyes on gran-
ular activated carbon [34]. Such difference in the values of Ds are
attributed mainly to the size of the pollutants molecules, because
the acetaminophen molecule is slightly larger than pyridine, but
smaller than methylene blue. It is also important to emphasize that
this is the first work where values of Ds are reported for the diffu-
sion of pharmaceutical compounds on activated carbon pellets.
The surface diffusion is the movement of the adsorbate mole-
cules along the pore surface of the AC, and its driving force is the
surface concentration gradient, q, whereas pore volume diffusion
refers to the movement of the adsorbate due to concentration gra-
dients in the fluid-phase (i.e., molecular mechanisms), and is
affected by the geometry of pore. To corroborate the mechanism
governing the intraparticle diffusion of acetaminophen on ACP,
the contribution of flux in the volume of pore, NAP, and the contri-
bution of flux corresponding to the surface diffusion, NAS, were
both calculated using the following equations:
NAP ¼ Dep rCAP
NAS ¼ Ds qp rq
Moreover, the point-wise norm of these mass fluxes are given
by the following equations:
515
mg/L
4000 min 6000 min
560.7
500
400
300
200
100
rffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi

2
2
kNAP k ¼ Dep rCAP jr-component þ rCAP jz-component ð12Þ
rffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi

2
2
kNAS k ¼ Ds qp rqjr-component þ rqjz-component ð13Þ
The values of ||NAP|| and ||NAS|| for experiment 7 are shown in
Fig. 8a and b, respectively. The arrows in both figures indicate
the direction of NAP and NAS, with colors representing the magni-
tude. From these figures, the following may be established: (i)
The direction of both fluxes along time converge at the center of
the pellet, as expected; (ii) During the whole time interval, the
acetaminophen diffuses consecutively by surface diffusion, fol-
lowed by pore volume diffusion; (iii) At short intervals
(t < 83.3 h), the surface flux is higher than the pore volume flux,
but at higher intervals, the relevance of both fluxes reverts; (iv)
At short intervals of time, the magnitude of NAP and NAS on the pel-
let corners is less when compared to the magnitude at other posi-
tions on the pellet; however, at longer intervals, both magnitudes
are higher than those shown on other positions of the particle; and,
(v) At long intervals of time, the direction of both fluxes at the cen-
ter of covers of pellet reverts, starting a desorption process towards
the solution.
From the values of NAP and NAS, it is possible to calculate the
surface diffusion contribution percentage (SDCP) with respect to
the total intraparticle diffusion by using the following equation:
kNAS k
SDCP% ¼  100 ð14Þ
kNAS k þ kNAP k
As an example, Fig. 9 shows the SDCP% variation as a function of
time and position for Exp. 7. This figure clearly shows that both
mechanisms of intraparticle diffusion are important, depending
of contact time. In short time intervals, the surface diffusion
reaches close to 100% of the total intraparticle diffusion, while
the pore volume diffusion is the mechanism governing the intra-
particle diffusion of acetaminophen on ACP at longer time inter-
vals; such result may be explained considering that, for Exp. 7,
90% of acetaminophen has been adsorbed during the first 83.3 h,
causing high concentration gradients in the solid phase, therefore
enhanced the driving force of surface diffusion. Finally, closer to
the equilibrium time, the solid phase concentration gradients
ð10Þ decrease, drastically boosting the pore volume diffusion. These
results confirm that both mechanisms of intraparticle diffusion
ð11Þ are important in the diffusion of acetaminophen on ACP.
Do [35] and Suzuki [36] pointed out that surface diffusion is the
most important mechanism of intraparticle transportation in
adsorbents with a large specific area, like activated carbons. They
516 R. Ocampo-Perez et al. / Chemical Engineering Journal 321 (2017) 510–520

Fig. 6. Concentration decay curves of acetaminophen on ACP at different experimental conditions. The lines represent the prediction of the PVDM model.

also found that the surface diffusion coefficient increases directly on sites requiring higher energy, and these adsorbed molecules do
when the mass of the adsorbed solute is increased. An explanation not have enough energy to be desorbed from a site and diffuse to
for this behavior is that, at the beginning, molecules are adsorbed another adsorption. Once the active sites with the higher adsorp-
 R. Ocampo-Perez et al. / Chemical Engineering Journal 321 (2017) 510–520 517

Fig. 7. Concentration decay curves of acetaminophen on ACP at different experimental conditions. The lines represent the prediction of the PVSDM model.

tion energy are occupied, the molecules will be adsorbed on the
sites with lower adsorption energy so more molecules can be des-
orbed and move from one site to another causing the Ds value is
increased with the adsorbed mass. Such phenomenon is evident
in the values of Ds reported in Table 3, where, in general, the
diffusion coefficient increases as the adsorbed amount of solute
is increased.
However, it is important to mention that an increase in the sur-
face diffusion coefficient does not necessarily mean that the sur-
face diffusion mechanism is more relevant than its pore volume
518 R. Ocampo-Perez et al. / Chemical Engineering Journal 321 (2017) 510–520

Fig. 8. Evolution of magnitude and direction of (a) NAP and (b) NAS during the adsorption of acetaminophen on ACP for Exp. 7.

16.7 h 83.3 h 133.3 h 183.3 h

100

90

80

70
SDCP%

60

50

40

30

20

10

r=0 r=R

Fig. 9. Evolution of SDCP% for acetaminophen transport inside ACP as a function of time.
 R. Ocampo-Perez et al. / Chemical Engineering Journal 321 (2017) 510–520
diffusion counterpart. As the adsorbed amount increases, the num-
ber of available adsorption sites decreases and, consequently,
higher amounts of solutes are transported through the pores of
the porous pellet. Such phenomenon would indicate that the mass
flux due to pore volume diffusion would be more relevant as the
adsorbed solute amount increases. To evidence this issue, Fig. 10
depicts the SDCP% variation for experiments 1, 2 and 7 at a contact
time of 83.3 h. In these three experiments, the adsorbed mass of
acetaminophen at 83.3 h were 52, 92, and 185 mg/g, respectively.
Fig. 10 clearly shows that, with an adsorbed mass of 52 mg/g, sur-
face diffusion governs the total intraparticle diffusion throughout
the whole particle. Conversely, when the adsorbed mass is
increased to 92 mg/g, areas of high contribution to surface diffu-
sion are visible at the center of the particle, as well as areas where
the pore volume contribution is increasingly important. Finally, at
high adsorbed amounts, the contribution front of surface diffusion
becomes more delimited, allowing the pore volume diffusion to
become more relevant. In general, these results prove that, as the
adsorbed mass increases, the surface flux increases too, allowing
for faster transportation of acetaminophen molecules towards
the inside of the ACP pores. Finally, the application of the PVSDM
evidenced that acetaminophen diffusion on ACP is both controlled
by surface diffusion and pore volume diffusion, depending on the
adsorbed amount and contact time. Therefore, the application of
an SDM that exclusively considers surface diffusion, and puts aside
pore volume diffusion, would derive in the incorrect calculation of
the values of Ds; for this reason the SDM was not applied to inter-
pret the adsorption rate of acetaminophen on ACP.
Exp. 1 Exp. 2
qe = 52 mg/g qe = 92 mg/g
r=0 r=R
83.3 h
Fig. 10. Evolution of SDCP% for acetaminophen transport inside ACP as a function of mass adsorbed of acetaminophen for Exp 1, 2 and 7 at a contact time of 83.3 h.
519
4. Conclusions
In this paper, we modeled the concentration decay curves of
acetaminophen on activated carbon pellets by using a three-
dimensional diffusion model, which considers the pore volume dif-
fusion and surface diffusion mechanisms. Mathematical models
based on pore volume diffusion, or surface diffusion, were found
to be equally capable of predicting the concentration decay curves.
However, such models calculate values of Dep and Ds that lack
physical relevance suggesting two scenarios: 1) values of Dep
higher than the molecular diffusivity, indicating superdiffusion
phenomena (still unreported in these types of systems), or 2) val-
ues of Dep equal to zero, indicating that the pore volume diffusion
is irrelevant when compared to the surface diffusion of
acetaminophen.
The application of a mathematical model including both diffu-
sion mechanisms corroborated that acetaminophen diffuses both
by pore volume diffusion and by surface diffusion, and that the rel-
evance of both diffusion mechanisms depends on the adsorbed
mass and contact time. Surface diffusion is favored under short
contact time conditions, because the solid phase concentration
gradients are more important than the fluid phase concentration
gradients within pores.
Finally, it is mentioned that the use of geometries representing
better the adsorbent material during the modeling stage, allows
establishing accurately the concentration profiles and direction of
intraparticle mass fluxes. The results of 3D-numeric simulations
indicate that there are internal delimited areas in the pellet domi-
Exp. 7
qe = 185 mg/g
100
90
80
70
SDCP%
60
50
40
30
20
10
520 R. Ocampo-Perez et al. / Chemical Engineering Journal 321 (2017) 510–520
nated by surface diffusion, as a form of advancing front. At longer
times, the solute mainly enters from the solution through the pel-
let borders, and acetaminophen desorption signs are even evident
at the center of the pellet covers, due to the inverse concentration
gradient established between the pellet and the solution.
Acknowledgement
This work was funded by Consejo Nacional de Ciencia y Tec-
nologia, CONACYT, Mexico, through grants Nos. CB-2013-01
221757, PN-I000/513/2016.
References
[1] J. Rivera-Utrilla, M. Sánchez-Polo, V. Gómez-Serrano, P.M. Álvarez, M.C.M.
Alvim-Ferraz, J.M. Dias, Activated carbon modifications to enhance its water
treatment applications. An overview, J. Hazard. Mater. 187 (2011) 1–23.
[2] USEPA, Granular Activated Carbon Treatment: Report EPA-540/2-91/024, US.
Environmental Protection Agency U.S. Government Printing Office,
Washington, DC, 1991.
[3] A.S. Mestre, R.A. Pires, I. Aroso, E.M. Fernandes, M.L. Pinto, R.L. Reis, M.A.
Andrade, J. Pires, S.P. Silva, A.P. Carvalho, Activated carbons prepared from
industrial pre-treated cork: sustainable adsorbents for pharmaceutical
compounds removal, Chem. Eng. J. 253 (2014) 408–417.
[4] B.N. Bhadra, P.W. Seo, S.H. Jhung, Adsorption of diclofenac sodium from water
using oxidized activated carbon, Chem. Eng. J. 301 (2016) 27–34.
[5] R. Mailler, J. Gasperi, Y. Coquet, C. Derome, A. Buleté, E. Vulliet, A. Bressy, G.
Varrault, G. Chebbo, V. Rocher, Removal of emerging micropollutants from
wastewater by activated carbon adsorption: experimental study of different
activated carbons and factors influencing the adsorption of micropollutants in
wastewater, J. Environ. Chem. Eng. 4 (2016) 1102–1109.
[6] F.J. García-Mateos, R. Ruiz-Rosas, M.D. Marqués, L.M. Cotorruelo, J. Rodríguez-
Mirasol, T. Cordero, Removal of paracetamol on biomass-derived activated
carbon: modeling the fixed bed breakthrough curves using batch adsorption
experiments, Chem. Eng. J. 279 (2015) 18–30.
[7] P. Iovino, S. Canzano, S. Capasso, A. Erto, D. Musmarra, A modeling analysis for
the assessment of ibuprofen adsorption mechanism onto activated carbons,
Chem. Eng. J. 277 (2015) 360–367.
[8] L.R. Radovic, C. Moreno-Castilla, J. Rivera-Utrilla, Carbon materials as
adsorbents in aqueous solutions, Chem. Phys. Carbon 27 (2001) 227–405.
[9] C. Moreno-Castilla, Adsorption of organic molecules from aqueous solutions on
carbon materials, Carbon 42 (2004) 83–94.
[10] R.W. Coughlin, F.S. Ezra, Role of surface acidity in the adsorption of organic
pollutants on the surface of carbon, Environ. Sci. Technol. 4 (1968) 291–297.
[11] Y.S. Ho, G. Mckay, Kinetic models for the sorption of dye from aqueous solution
by wood, Process Saf. Environ. Prot. 76 (1998) 183–191.
[12] G. Blanchard, M. Maunaye, G. Martin, Removal of heavy metals from waters by
means of natural zeolites, Water Res. 18 (1984) 1501–1507.
[13] S. Mondal, K. Sinha, K. Aikat, G. Halder, Adsorption thermodynamics and
kinetics of ranitidine hydrochloride onto superheated steam activated carbon
derived from mung bean husk, J. Environ. Chem. Eng. 3 (2015) 187–195.
[14] M.A. Chayid, M.J. Ahmed, Amoxicillin adsorption on microwave prepared
activated carbon from Arundo donax Linn: isotherms, kinetics, and
thermodynamics studies, J. Environ. Chem. Eng. 3 (2015) 1592–1601.
[15] G. Nazari, H. Abolghasemi, M. Esmaieli, Batch adsorption of cephalexin
antibiotic from aqueous solution by walnut shell-based activated carbon, J.
Taiwan Inst. Chem. Eng. 58 (2016) 357–365.
[16] R. Ocampo-Perez, R. Leyva-Ramos, P. Alonso-Davila, J. Rivera-Utrilla, M.
Sanchez-Polo, Modeling adsorption rate of pyridine onto granular activated
carbon, Chem. Eng. J. 165 (2010) 133–141.
[17] A.E. Rodriguez, C.M. Silva, What’s wrong with Lagergreen pseudo first order
model for adsorption kinetics, Chem. Eng. J. 306 (2016) 1138–1142.
[18] J.D. Méndez-Díaz, G. Prados-Joya, J. Rivera-Utrilla, R. Leyva-Ramos, M.
Sánchez-Polo, M.A. Ferro-García, N.A. Medellín-Castillo, Kinetic study of the
adsorption of nitroimidazole antibiotics on activated carbons in aqueous
phase, J. Colloid Interface Sci. 345 (2010) 481–490.
[19] R. Leyva-Ramos, L.A. Bernal-Jacome, J. Mendoza-Barron, M.M.G. Hernandez-
Orta, Kinetic modeling of pentachlorophenol adsorption onto granular
activated carbon, J. Taiwan Inst. Chem. Eng. 40 (2009) 622–629.
[20] R. Ocampo-Pérez, M.M. Abdel Daiem, J. Rivera-Utrilla, J.D. Méndez-Díaz,
Manuel Sánchez-Polo, Modeling adsorption rate of organic micropollutants
present in landfill leachates onto granular activated carbon, J. Colloid Interface
Sci. 385 (2012) 174–182.
[21] K.K.H. Choy, G. McKay, Sorption of cadmium, copper, and zinc ions onto bone
char using Crank diffusion model, Chemosphere 60 (2005) 1141–1150.
[22] K.K.H. Choy, J.F. Porter, G. McKay, Film–pore diffusion models—analytical and
numerical solutions, Chem. Eng. Sci. 59 (2004) 501–512.
[23] C. Liang, Z. Lan, X. Zhang, Y. Liu, Mechanism for the primary transformation of
acetaminophen in a soil/water system, Water Res. 98 (2016) 215–224.
[24] R. Rosal, A. Rodríguez, J.A. Perdigón-Melón, A. Petre, E. García-Calvo, M.J.
Gómez, A. Agüera, A.R. Fernández-Alba, Occurrence of emerging pollutants in
urban wastewater and their removal through biological treatment followed by
ozonation, Water Res. 44 (2010) 578–588.
[25] A.M. Larson, J. Polson, R.J. Fontana, T.J. Davern, E. Lalani, L.S. Hynan, J.S. Reisch,
F.V. Schiodt, G. Ostapowicz, A.O. Shakil, W.M. Lee, Acetaminophen-induced
acute liver failure: results of a United States multicenter, prospective study,
Hepatology 42 (6) (2005) 1364–1372.
[26] B.K. Park, N.R. Kitteringham, J.L. Maggs, M. Pirmohamed, D.P. Williams, The
role of metabolic activation in drug induced hepatotoxicity, Annu. Rev.
Pharmacol. Toxicol. 45 (2005) 177–202.
[27] A.I. Moral-Rodríguez, R. Leyva-Ramos, R. Ocampo-Pérez, J. Mendoza-Barron, I.
N. Serratos-Alvarez, J.J. Salazar-Rabago, Removal of ronidazole and
sulfamethoxazole from water solutions by adsorption on granular activated
carbon: equilibrium and intraparticle diffusion mechanisms, Adsorption 22
(2016) 89–103.
[28] F. Rodriguez-Reinoso, A. Linares-Solano, Microporous structure of activated
carbons as revealed by adsorption methods, in: Chemistry and Physics of
Carbon, New York, 1988.
[29] F.J. García-Mateos, R. Ruiz-Rosas, M.D. Marqués, L.M. Cotoruelo, J. Rodríguez-
Mirasol, T. Cordero, Removal of paracetamol on biomass-derived activated
carbon: modeling the fixed bed breakthrough curves using batch adsorption
experiments, Chem. Eng. J. 279 (2015) 18–30.
[30] J. Lladó, C. Lao-Luque, B. Ruiz, E. Fuente, M. Solé-Sardans, A.D. Dorado, Role of
activated carbon properties in atrazine and paracetamol adsorption
equilibrium and kinetics, Process Saf. Environ. Prot. 95 (2015) 51–59.
[31] T. Furusawa, J.M. Smith, Fluid-particle and intraparticle mass transport rates in
slurries, Ind. Eng. Chem. Fundam. 12 (1973) 197–203.
[32] C.H. Giles, D. Smith, A. Huitson, A general treatment and classification of the
solute adsorption isotherm. I. Theoretical, J. Colloid Interface Sci. 47 (1974)
755–765.
[33] A.C.F. Ribeiro, M.C.F. Barros, L.M.P. Veríssimo, C.I.A.V. Santos, A.M.T.D.P.V.
Cabral, G.D. Gaspar, M.A. Esteso, Diffusion coefficients of paracetamol in
aqueous solutions, J. Chem. Thermodyn. 54 (2012) 97–99.
[34] R. Ocampo-Pérez, R. Leyva-Ramos, M. Sanchez-Polo, J. Rivera-Utrilla, Role of
pore volume and surface diffusion in the adsorption of aromatic compounds
on activated carbon, Adsorption 19 (2013) 945–957.
[35] D.D. Do, Adsorption Analysis: Equilibria and Kinetics, first ed., Imperial College
Press, London, UK, 1998.
[36] M. Suzuki, Adsorption Engineering, first ed., Elsevier, Japan, 1990.