Accepted Manuscript

Removal of amoxicillin from contaminated water using NH4Cl-activated car‐

bon: Continuous flow fixed-bed adsorption and catalytic ozonation regeneration

Kamyar Yaghmaeian, Gholamreza Moussavi, Ahamd Alahabadi

PII: S1385-8947(13)01171-6

DOI: http://dx.doi.org/10.1016/j.cej.2013.08.118

Reference: CEJ 11223

To appear in: Chemical Engineering Journal

Please cite this article as: K. Yaghmaeian, G. Moussavi, A. Alahabadi, Removal of amoxicillin from contaminated

water using NH4Cl-activated carbon: Continuous flow fixed-bed adsorption and catalytic ozonation regeneration,

Chemical Engineering Journal (2013), doi: http://dx.doi.org/10.1016/j.cej.2013.08.118

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers

we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and

review of the resulting proof before it is published in its final form. Please note that during the production process

errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

1

Removal of amoxicillin from contaminated water using NH4Cl-

activated carbon: Continuous flow fixed-bed adsorption and

catalytic ozonation regeneration

Kamyar Yaghmaeian1, Gholamreza Moussavi2*, Ahamd Alahabadi2

1. Center for Water Quality Research, Institute for Environmental Research, Tehran

University of Medical Sciences, Tehran, Iran

2. Department of Environmental Health Engineering, Tarbiat Modares University,

Tehran, Iran

*Corresponding author: Tel: +98 21 82883827; Fax: +98 21 82883825;

E-mail address: moussavi@modares.ac.ir (G. Moussavi)

2

Abstract

This study investigates the elimination of amoxicillin from water by adsorption onto

NH4Cl-activated carbon (NAC) and standard activated carbon (SAC) in the fixed-bed

columns with ozone-regeneration of the saturated carbon. Breakthrough curves for

amoxicillin (AMX) adsorption were determined at various empty bed contact times

(EBCTs) ranging from 2 to 10 min. Results determined breakthrough times of 5.5 to 103 h

for the NAC bed and 10 to 98 h for SAC bed. The breakthrough-point adsorption capacity

of NAC was much greater than that of SAC at each of the tested EBCTs. Based on the

Thomas model, the maximum adsorption capacity of NAC increased from 2.5 to 5.1 g/g

and that of SAC increased from around 1.4 to 3.0 g/g with an increased EBCT from 2 to 10

min. The saturated NAC bed was completely regenerated in-situ through a catalytic

regeneration process with an ozone dose of 0.25 g-O3/g-NAC and the regenerated NAC

demonstrated adsorption behavior similar to that of fresh NAC. The NAC-bed improved

the quality of the contaminated real water from background contaminants and completely

removed AMX. Therefore, the NAC fixed-bed adsorption and subsequent ozonation

presents a promising and efficient process for the removal of AMX as an emerging

contaminate from polluted water.

Keywords: Adsorption, chemically activated carbon, fixed-bed, breakthrough curve,

catalytic ozonation regeneration

3

1. Introduction

Pharmaceutical antibiotics are becoming increasingly problematic contaminants of water

resources, particularly in surface and ground water sources located around industrial and

residential communities. They enter water sources mostly through discharge from

pharmaceutical industries and from municipal wastewater treatment plants [1].

Consumption of water contaminated with antibiotics can have several adverse effects on

humans including acute and chronic toxicity [2]. Another critical concern regarding

antibiotics in water sources is the development of bacterial resistance to medicinal

treatment of bacterial infections [3,4]. Amoxicillin (AMX) is one of the most widely used

β-lactam antibiotics used to treat human and animal infections [4,5]. Water containing this

class of chemicals needs to be treated using an efficient process in order to protect human

health and the environment against the adverse affects of contamination by antibiotics

including AMX. There are currently various biological, chemical, and physical methods

available to remove AMX from contaminated water including biodegradation, advanced

oxidation processes (AOPs) [4,6-10] and adsorption [5, 11, 12].

Although bioprocesses are considered as the best option to treat biodegradable

contaminants, they are often inefficient in complete degradation of the antibiotics because

most antibiotics are biorecalcitrant [13]. AOPs are an efficient technique to degrade

synthetic contaminants including antibiotics. However, some of the intermediates

generated during the partial oxidation process of complex compounds might be even more

toxic than the parent contaminant. Moreover, AOPs are very expensive and operationally

complex for the complete degradation of recalcitrant compounds [11]. Adsorption is an

efficient process which is simple to design and operate and is thus considered a superior

4

technique for accumulation of toxic organic contaminants from contaminated streams onto

a solid material particularly activated carbon [14-16]. A few types of activated carbon have

been tested for adsorption of AMX [11,17] among which the NH4Cl-activated carbon

(NAC) has demonstrated a superior performance. Nevertheless, the performance of NAC

in the fixed-bed column and its potential to regenerate needs to be established before

practical application of the NAC to remove AMX.

Accordingly, this study was designed (a) to evaluate the effect of empty bed contact times

(EBCT) on the performance of an NAC fixed-bed column in comparison with standard

activated carbon (SAC) for adsorption of AMX, and (b) to investigate carbon regeneration

capacity of the saturated bed using the process of in-situ catalytic ozonation. The first

phase of this work involved testing the effects of various AMX-contaminated solution flow

rates (3.34 to 6.67 mL/min) corresponding to EBCT ranging from 2 to 10 min on AMX

adsorption in the NAC fixed-bed. In this phase, behavior of the NAC fixed-bed column in

adsorption of AMX was modeled using the Thomas model [18] and the Yoon-Nelson

model [19]. Moreover, performance of the NAC column was examined for removal of

AMX from a sample of real contaminated water. The second phase of the study was aimed

at evaluating the ozonation process for regeneration of the saturated NAC and SAC beds

and reusability of the regenerated carbon beds. Spent activated carbons are conventionally

regenerated using thermal, physicochemical or biological processes; and then the carbon is

reactivated in thermal reactors [20], all of these processes are complex, costly and have

high energy-demand.

Recently various alternative methods including ozonation [21-23], microwave assisted

regeneration [24], thermal and acid washing [25], H2O2-based oxidation [26], iron oxide

nanocatalyst [27], wet peroxide oxidation [28], UV/H2O2 oxidation [29], and

5

electrochemical process [30], have been assessed for the regeneration of different activated

carbons. The process of ozonation is an emerging option among the newly developed

alternatives for regeneration of saturated activated carbon [21,23]. In an ozone

regeneration process, spent carbon can be regenerated in-situ at atmospheric pressure and

temperature and thus it cuts costs from transportation and other expensive off site

processes [21]. Using ozone regeneration processing of spent carbon takes advantage of

the generation of oxidative species that are more reactive than molecular ozone,

particularly hydroxyl radical from the interaction of ozone with the surface of the

carbon particles [31]. It facilitates efficient catalytic ozonation regeneration. Therefore, the

second phase of the study investigated the performance of ozonation at various doses for

the process of in-situ regeneration of the saturated NAC bed and reusability of the

regenerated carbon.

2. Materials and methods

2.1. Materials

Analytical grade AMX was obtained from Sigma Aldrich Co., and used as received; its

characteristics are given in Table 1. NAC and SAC were used as adsorbent beds in the

experiments. Briefly, the NAC and SAC were both mesoporous activated carbons with

specific surface areas of 1029 and 1024 m2/g, respectively. The pHzpc of NAC and SAC

was 6.6 and 7.4, respectively. The preparation and characterization of the NAC are

explained in detail elsewhere [11]. The activated carbon used as the SAC was purchased

from Merk Co. to compare with the recently developed NAC. The mesh of NAC and SAC

particles packed in columns was 24/16. A stock solution of AMX was prepared by

dissolving 1 g of AMX in 1 L of distilled water (1000 mg/L). Two types of contaminated

6

water (aqueous solution and real contaminated water) were prepared and examined in this

study. The aqueous solution was made from mixing aliquots of AMX stock solution with

distilled water. The real contaminated water sample was prepared by spiking the samples

taken from a ground water aquifer with aliquots of AMX stock solution. The reason of

selecting ground water for preparation AMX-contaminated water is that it usually contains

most of the ionic species which may interfere with the AMX adsorption onto activated

carbons as well as with catalytic ozonation regeneration process. All other chemicals used

in this study were analytical grade.

2.2. Adsorption column experiments

The adsorption experiments were conducted on a bench-scale column setup as shown in

Fig. 1. The experimental set up consisted of two glass columns each with an internal

diameter of 1.1 cm and a total height of 30 cm equipped with an inlet and sampling valves.

One column was packed with SAC and another was packed with NAC. Carbon beds were

14 cm in both columns, giving a bed volume of 13.3 cm3 in each column. A perforated

glass plate located at a distance of 5 cm from the bottom supported the adsorbent packed in

the column. The columns were operated in parallel mode by injections of the aqueous

solution from the inlet tank by a peristaltic pump (Heidolph Co.). Effluent was collected in

a separated tank for each column. Effluent from the columns was periodically sampled and

analyzed for AMX concentrations and other parameters. The performance of each column

was evaluated for treating aqueous solutions at various empty bed contact times (EBCT)

ranging from 2 to 10 min through regulating the inlet flow between 1.34 and 6.67 mL/min

with a constant bed volume of 13.3 cm3. At each EBCT, the columns were operated until

7

the saturation condition was reached. Columns were operated at room temperature

(25±3 °C). The pH level of the contaminated water at inlet was fixed at 6.8.

2.3. Regeneration of the AMX-saturated bed using ozonation

Regeneration of the AMX-saturated NAC and SAC beds was conducted in-situ using

ozonation at 1.4 mg-O3/min for various times of 1 to 3 h. For regeneration with each ozone

dose, the inlet AMX solution to the saturated column was stopped and regeneration was

started via injecting the ozone-gas stream at room temperature and atmospheric pressure.

The process of ozonation was continued for the specified times, the solution was

discharged from the column, the bed was washed with distilled water, and then AMX

solution (50 mg/L) was injected to the regenerated-bed column at the flow rate of 6.7

mL/min corresponding to an EBCT of 2 min. Operation of the regenerated bed column was

continued until breakthrough point was reached.

2.4. Analysis

Concentrations of AMX in the water were measured using a Knauer HPLC (C18 ODS

column; 250×4.6×5) with a UV detector 2006 at a wavelength of 190 nm. The mobile

phase was a mixture of buffer phosphate with pH=4.8 and acetonitrile with a volumetric

ratio of 60/40 with an injection flow rate of 1 mL/min. The pH level was measured for the

samples using a pH meter (Sense Ion 378, Hack). Temperatures of solutions were taken

using a mercury thermometer. Tests at each EBCT were conducted in duplicate to ensure

reproducibility of results; the mean of these two measurements was used to determine

evaluations.

8

The adsorption performance capability of SAC and NAC was determined in terms of

breakthrough time (tbk), breakthrough-point capacity (BC), and carbon usage rate (CUR) as

the following relations:

(1)

(2)

BC = quantity of adsorbate fed to the adsorber divided by mass of adsorbent at

breakthrough time, mg/g

Q = water flow rate to the adsorber, L/h

C = inlet concentration of AMX, mg/L

tbk = breakthrough time (is considered as the point on the breakthrough curve where no

AMX was detected in the effluent), h

Madsorbent = mass of adsorbent in column, g

3. Results and discussions

3.1. Adsorption of AMX in the fixed-bed column

The obtained breakthrough curves of AMX adsorption are shown in Fig. 2. As seen in Fig.

2, the service time (time to reach breakthrough point) time at EBCTs of 2, 4, 6 and 10 min

was 6, 18, 50 and 103 h for NAC, respectively, and 10, 20, 40 and 98 h for SAC,

respectively. It is found that NAC column had significantly greater (20%) service time than

SAC in adsorption of AMX at EBCT of 6 min. Accordingly 6 min can be selected as the

optimum EBCT for adsorption of AMX in NAC column. It is also observed in Fig. 2 that

steeper breakthrough curves was achieved at lower EBCTs (higher solution flow rates) for

both NAC and SAC columns. Greater service time and steeper breakthrough curves at

9

lower EBCTs can be justified considering the effect of contact time on the dominated

mechanism controlling the adsorption process. Recent study indicated that diffusion

through the boundary liquid layer surrounding the carbon particles was the main step

controlling the rate of AMX adsorption onto SAC and NAC [11]. More contact time in the

adsorption bed resulted in a higher rate of AMX molecule diffusion through the boundary

layer around the adsorbent, which in turn stimulated a higher rate of adsorption. This in

turn caused more effective adsorption and thus a smaller mass transfer and finally extended

the service time of the bed. Moreover, an increased EBCT provided a better opportunity

for adsorption of the solute onto the adsorbent and the adsorption equilibrium was reached

at a higher EBCTs [32,33]. Other researchers [15,34-38] have also reported an increased

breakthrough time with increased EBCT for the adsorption of other contaminants in other

various adsorbent fixed-bed columns.

The capability of the NAC and SAC fixed-bed columns for dynamic adsorption of AMX

was compared in Fig. 3 in terms of breakthrough-point capacity and carbon usage rate. As

observed in Fig. 3, the breakthrough-point capacity increased from 73.3 to 274.1 mg/g for

NAC and from 31.6 to 65.2 mg/g for SAC with an increase of EBCT from 2 to 10 min. An

increased adsorption capacity associated with the increase of EBCT can be attributed to

increased mass loading of AMX applied to a constant mass of the adsorbent bed in the

column with an increased inlet solution flow rate. This resulted in an increased mass

transfer of AMX molecules to the adsorbent bed and thereby increased its adsorption

capacity [11]. Furthermore, Fig. 3 shows that the adsorption capacity of AMX was higher

for the NAC bed than that for SAC bed at all tested EBCTs under similar operational

conditions. The better performance of the NAC column compared to the SAC column in

terms of AMX adsorption is attributed to the higher density of surface functional groups

10

(including hydroxyl, aliphatic C-H, C=O, carboxylic, and carbonyl groups) on the surface

NAC, which enhanced the interaction of AMX molecules through an increased availability

of adsorption sites on the surface of NAC compared to SAC [11]. The greater capacity of

NAC than SAC in adsorption capacity of AMX determines that NAC has a lower rate of

usage than SAC (Fig. 4) and hence less bed change out is required for treating specific

volumes of contaminated water, making it more cost effective.

3.2. Adsorption columns modeling

Performance predictions were made for columns packed with NAC and SAC and analyzed

for practical application using the Thomas model [18] to describe adsorption kinetics and

maximum adsorption capacity (q0). The linear form of the Thomas model is given in Eq.

(3):

(3)

where C0= inlet AMX concentration (mg/L); Ct = AMX concentration at time t (mg/L); kT

= Thomas rate constant (L/h.mg); q0 = maximum adsorption capacity (mg/g); M = mass of

the adsorbent (g); Q = solution flow rate (L/h); and t = sampling time (h). The values of kT

and q0 can be determined from plotting versus t.

The Thomas model (Eq. (3)) was well-fitted with the data from tests of column adsorption

capacity (Fig. 2) with a high degree of correlation coefficient (R2>0.9) for all tested

EBCTs. The obtained information is summarized in Table 2. According to data provided in

Table 2, the Thomas rate constant (kT ) decreased while the maximum adsorption capacity

increased with increased EBCT for both NAC and SAC. The decrease in kT with the

increase of EBCT (decrease of flow rate) can be related again to the decrease of boundary

layer diffusion limitation related to increased contact time that served to reduce mass

11

transfer resistance [38]. This facilitated the diffusion of AMX molecules onto the NAC and

SAC fixed-bed [39]. The results obtained in this study support the findings of related

research such as Cr(VI) adsorption by waste acorn of Quercus ithaburensis [33],

ammonium ion adsorption by zeolite 13X [40], and fluoride adsorption by Kanuma mud

[38]. Information shown in Table 2 clearly indicates the better performance of the NAC

bed for adsorption of AMX from contaminated water compared to an SAC bed as

discussed above.

Moreover, to predict the breakthrough behavior of the bed, the data from the experiment

were fitted with the Yoon-Nelson model [19] given as Eq. (4).

(4)

where C0 is inlet AMX concentration (mg/L); Ct is AMX concentration at time t (mg/L);

is the Yoon-Nelson rate constant (1/h), and denotes the time (h) needed for 50%

breakthrough point. The values of and can be determined from slope and intercept of

the plot of versus t. The information of the Yoon-Nelson model fitted to data

from the experiment is summarized in Table 3. As seen in Table 3, the Yoon-Nelson model

could be well fitted (R2>0.9) with the data obtained on breakthrough adsorption. Table 3

also indicates that the value of decreased from 0.036 to 0.007 1/h with an increased

EBCT from 2 to 10 min (decrease of inlet flow rate from 4 to 0.8 L/h).

The times needed for 50% AMX breakthrough from the experiments (Fig. 2) was close to

those obtained from the Yoon-Nelson model, implying that Yoon-Nelson model

adequately described the behavior of the NAC and SAC fixed-bed column in treating

AMX-contaminated water. No fixed-bed column experiments were found in the related

literature to make a direct comparison with results obtained in this present study. However,

other researchers have also observed an association between a decrease of and an

12

increase of with a decrease in inlet flow rate in fixed-bed columns packed with different

adsorbents for the removal of other contaminants [15,33,41-43].

3.3. Regeneration of the AMX saturated-carbons

As stated in section 2.3, the AMX-saturated beds were regenerated by ozonation at various

doses. The breakthrough curves of the activated carbon beds regenerated with different

ozone doses are given in Fig. 5. As seen in Fig.5a, the regenerated NAC bed could attain a

breakthrough time that coincided with that of fresh carbon (5.5 h). Regeneration efficiency

of the saturated carbon using ozone given by Eq. (6) was calculated as 36%, 82% and

100% for ozone doses of 56, 168, and 252 mg-O3/g-C, respectively. The regeneration

efficiency of the SAC bed was 79% at the ozone dose of 252 mg-O3/g-C. It can therefore

be determined that the AMX-saturated NAC column could be efficiently regenerated and

reactivated with ozonation at a relatively low dose and that the adsorption capacity of NAC

could be completely restored. It should be noted that the AMX saturation-ozone

regeneration of the NAC bed was repeated for 4 consecutive cycles and results (not shown)

indicated no adverse effects from selected processes on the adsorption capacity

performance of regenerated carbon compared with the original carbon under those

conditions.

(6)

Therefore, it is clearly depicted that the NAC adsorption-ozonation can be used as an

efficient technique to treat water contaminated with the toxins. Alvárez et al. [23] reported

an optimum dose of 0.4 g-O3/g-GAC to recover most of the original GAC adsorption

capacity for gallic acid. Valdés and Zaror [22] also showed effective regeneration of

benzothiazole saturated-activated carbons by an ozonation process. Álvarez et al. [21] also

13

reported regeneration efficiency of 79.2% of phenol-saturated activated carbon using an

ozone dose of 0.45 g-O3/g-GAC. They study [21] also indicated that regenerated carbon

preserved most of its adsorption capacity when regenerated through a number of

adsorption-ozone regeneration cycles. It is seen that NAC could more efficiently be

regenerated by ozone than by other carbons. The difference in optimum doses of ozone and

the regeneration efficiency attained in the present study compared to rates cited in other

publications can be related to the difference in the adsorption system (batch versus

continuous fixed-bed), properties of the carbons, and characteristics of the adsorbates.

The mechanism for regeneration of the AMX saturated-NAC can be explained by both the

direct oxidation of the AMX molecules adsorbed onto the NAC (Eq. (1)) and catalytic

ozonation through decomposition of ozone on the surface of the activated carbon particles

(Eq. (7)) and thereby the generation of reactive oxidative radicals in particular hydroxyl

radicals (Eq. (9)) [22]. The generated hydroxyl radicals then react with the AMX

molecules and oxidize them into final products (Eq. (10)) thus regenerating carbon for

reuse.

(7)

(8)

(9)

(10)

To elucidate the contribution of the proposed mechanisms in the regeneration process, a set

of regeneration tests was carried out with 1 g of scavenger of tert-butanol and 1 g of

phosphate with an optimum ozone dose of 252 mg-O3/g-C (1.4 mg-O3/min for 3 h).

Regeneration efficiency was reduced from 100% at the absence of phosphate and tert-

butanol to 59% and 65% at the presence of them, respectively.

14

The reduced regeneration efficiency in the presence of phosphate can be justified by the

fact that this ion has a high affinity for the Lewis sites [44,45] on the surface of the carbon.

In fact, the presence of phosphate ions has limits the capacity for ozone adsorption and

decomposition (Eq. 8) at the Lewis sites and thus inhibited propagation of generation

reaction (Eq. 9). This defect has resulted in reduction of the rate of the AMX radical

oxidation (Eq. 10) and subsequently inhibited the process of regeneration. In addition,

considering the fact that tert-butanol is a well-known hydroxyl radical scavenger [46], it

suggests that oxidation of AMX adsorbed onto carbon using molecules (Eq. 10) has

been the main mechanism for carbon regeneration. Indeed, the presence of tert-butanol has

likely competed with the MAX molecules in reacting with the generated (Eq. 10) and

thus reduced the rate of AMX oxidation and subsequently inhibited the process of

regeneration. The considerable reduction of the regeneration performance in the presence

of tert-butanol suggests that regeneration of the saturated NAC has been mainly conducted

through radical oxidation.

3.4. Treatment of the real water contaminated with AMX in the NAC fixed-bed column

To understand the effect of the water and its constituents on the efficacy of the NAC fixed-

bed column (which had better performance than SAC), a real ground water sample was

spiked with AMX at the concentration of 50 mg/L and fed to the column with EBCT of 10

min where the maximum tbt could be attained. The obtained breakthrough curve comparing

the treatment of contaminated real water to that of synthetic water is shown in Fig. 6. It

demonstrates that the value of tbt reduced from 103 h for treating synthetic contaminated

water to 78 h when treating MAX-contaminated real water. The main characteristics of

feed water and effluent at breakthrough point are given in Table 4. Results given in Table 4

15

indicate that in addition to complete adsorption of MAX, the NAC fixed-bed improved the

quality of the contaminated water, particularly in terms of adsorption of DOC and nitrate,

which are two major water contaminants. It can therefore be deduced that constituents of

the water adsorbed onto the surface of NAC occupied some of the adsorption sites. This in

turn, reduced the availability of sites for AMX adsorption and thus decreased the

breakthrough time. Accordingly, quality of the feed water to the NAC column can affects

its performance, which implies that optimum operational conditions need to be obtained

from any specific water, before full scale practical application.

4. Conclusion

The performance of bench-scale NAC and SAC fixed-bed columns was studied on the

removal of AMX from aqueous solution and real contaminated water samples. The NAC-

bed column performed considerably better than the SAC-bed column treating a 50 mg/L

AMX solution. Breakthrough time and adsorption capacity increased with increasing

EBCT for both tested beds. Ozonation proved to be an efficient and viable process for in-

situ regeneration of the NAC saturated beds. The ozone-regenerated NAC bed could be

reused for 4 consecutive cycles without significant decrease in AMX adsorption capacity.

The treatment of an AMX-contaminated real water sample was also examined by same

fixed-bed procedure. The NAC bed was more efficient than SAC one in the treatment of

AMX-contaminated real water. It can therefore be concluded that intermittent adsorption in

the NAC-bed column and regeneration of the saturated bed with ozone is an efficient and

promising technology for treating antibiotic-contaminated waters.

16

References

[1] E. Zuccato, S. Castiglioni, R. Bagnati, M. Melis, R. Fanelli, Source, occurrence and fate of

antibiotics in the Italian aquatic environment, J. Hazard. Mater. 179 (2010) 1042–1048.

[2] E.R.E. Mojica, D.S. Aga, Antibiotics pollution in soil and water: potential ecological

and human health issues, Encycl. Environ. Health 28 (2011) 97-110.

[3] V. Homem, L. Santos, Degradation and removal methods of antibiotics from aqueous

matrices e a review, J. Environ. Manage. 92 (2011) 2304-2347.

[4] D. Dimitrakopoulou, I. Rethemiotaki, Z. Frontistis, N.P. Xekoukoulotakis, D. Venieri, D.

Mantzavinos, Degradation, mineralization and antibiotic inactivation of amoxicillin By

UVA/TiO2 photocatalysis, Environ. Manag. 98 (2012) 168-174.

[5] W.S. Adriano, V. Veredas, C.C. Santana, L.R.B. Gonçalves, Adsorption of amoxicillin on

chitosan beads: kinetics, equilibrium and validation of finite bath models, Biochem. Eng. J.

27 (2005) 132–137.

[6] D. Klauson, J. Babkina, K. Stepanova, M. Krichevskaya, S. Preis, Aqueous photocatalytic

oxidation of amoxicillin, Catal. Today 151 (2010) 39-45.

[7] R. Andreozzi, M. Canterino, R. Marotta, N. Paxeus, Antibiotic removal from wastewaters:

the ozonation of amoxicillin, J. Hazard. Mater. 122 (2005) 243-250.

[8] A.G. Trovo, S.A. Santos Melo, R.F. Pupo Nogueira, Photodegradation of the

pharmaceuticals amoxicillin, bezafibrate and paracetamol by the photo-Fenton process e

17

application to sewage treatment plant effluent, J. Photochem. Photobiol. A 198 (2008) 215-

220.

[9] C. Mavronikola, M. Demetriou, E. Hapeshi, D. Partassides, C. Michael, D. Mantzavinos,

D. Kassinos, Mineralisation of the antibiotic amoxicillin in pure and surface waters by

artificial UVA- and sunlight-induced Fenton oxidation, J. Chem. Technol. Biotechnol. 84

(2009) 1211-1217.

[10] F. Ay, F. Kargi, Advanced oxidation of amoxicillin by Fenton’s reagent treatment, J.

Hazard. Mater. 179 (2010) 622-627.

[11] G. Moussavi, A. Alahabadi, K. Yaghmaeian, M. Eskandari, Preparation,

characterization and adsorption potential of the NH4Cl-induced activated carbon for the

removal of amoxicillin antibiotic from water, Chem. Eng. J. 217 (2013) 119-128.

[12] B. Ghauch, A. Tuqan, H. Abou Assi, Antibiotic removal from water: elimination of

amoxicillin and ampicillin by microscale and nanoscale iron particles, Environ. Pollut. 157

(2009) 1626–1635.

[13] K. Kümmerer, The presence of pharmaceuticals in the environment due to human use

e present knowledge and future challenges, J. Environ. Manage. 90 (2009) 2354-2366.

[14] R. Han, D. Ding, Y. Xu, W. Zou, Y. Wang, Y. Li, L. Zou, Use of rice husk for the

adsorption of congo red from aqueous solution in column mode, Bioresource Technol. 99

(2008) 2938-2946.

[15] S-H. Lin, R.-S. Juang, Adsorption of phenol and its derivatives from water using

synthetic resins and low-cost natural adsorbents: a review, J. Environ. Manage. 90 (2009)

1336–1349.

[16] A.A. Ahmad, B.H. Hameed, Fixed-bed adsorption of reactive azo dye onto granular

activated carbon prepared from waste, J. Hazard. Mater. 175 (2010) 298-303.

18

[17] E.K. Putra, R. Pranowo, J. Sunarso, N. Indraswati, S. Ismadji, Performance of

activated carbon and bentonite for adsorption of amoxicillin from wastewater: mechanisms,

isotherms and kinetics, Water Res. 43 (2009) 2419-2430.

[18] H.C. Thomas, Heterogeneous ion exchange in a flowing system, J. Am. Chem. Soc. 66

(1944) 1466-1664.

[19] Y.H. Yoon, J.H. Nelson, Application of gas adsorption kinetics. Part 1. a theoretical

model for respirator cartridge service time, Am. Ind. Hyg. Assoc. J. 45 (1984) 509-516.

[20] J.C. Crittenden, R.R. Trussell, D.W. Hand, K.J. Howe, G. Tchobanoglous, Water

Treatment: Principles and Design, second ed., Wiley, New Jersey, 2005.

[21] P.M. Álvarez, F.J. Beltrán, V. Gómez-Serrano, J. Jaramillo, E.M. Rodrı�guez,

Comparison between thermal and ozone regenerations of spent activated carbon exhausted

with phenol, Water Res. 38 (2004) 2155-2165.

[22] H. Valdés, C.A. Zaror, Ozonation of benzothiazole saturated-activated carbons:

influence of carbon chemical surface properties, J. Hazard. Mater. B137 (2006) 1042–

1048.

[23] P.M. Alvárez, F.J. Beltrán, F.J. Masa, J.P. Pocostales, A comparison between catalytic

ozonation and activated carbon adsorption/ozone-regeneration processes for wastewater

treatment, Appl. Catal. B. 92 (2009) 393–400.

[24] D. Xin-hui, C. Srinivasakannan, W.-W. Qu, W. Xin, P. Jin-hui, Z. Li-bo, Regeneration

of microwave assisted spent activated carbon: process optimization, adsorption isotherms

and kinetics, Chem. Eng. Process. 53 (2012) 53– 62.

[25] S.W. Nahm, W.G. Shim, Y.-K. Park, S.C. Kim, Thermal and chemical regeneration of

spent activated carbon and its adsorption property for toluene, Chem. Eng. J. 210 (2012)

500–509.

19

[26] A. Anfruns, M.A. Montes-Morán, R. Gonzalez-Olmos, M.J. Martin, H2O2-based

oxidation processes for the regeneration of activated carbons saturated with volatile

organic compounds of different polarity, Chemosphere 91 (2013) 48–54.

[27] C.-A. Chiu, K. Hristovski, S. Huling, P. Westerhoff, In-situ regeneration of saturated

granular activated carbon by an iron oxide nanocatalyst, Water Res. 47 (2013) 1596-1603.

[28] K. Okawa, K. Suzuki, T. Takeshita, K. Nakano, Regeneration of granular activated

carbon with adsorbed trichloroethylene using wet peroxide oxidation, Water Res. 41

(2007) 1045 –1051.

[29] R.S. Horng, I.C. Tseng, Regeneration of granular activated carbon saturated with

acetone and isopropyl alcohol via a recirculation process under H2O2/UV oxidation, J

Hazard. Mater. 154 (2008) 366-72.

[30] R. Berenguer, J.P. Marco-Lozar, C. Quijada, D. Cazorla-Amorό, E. Morallόn,

Electrochemical regeneration and porosity recovery of phenol-saturated granular activated

carbon in an alkaline medium, Carbon 48 (2010) 2734-2745.

[31] G. Moussavi, A. Khavanin, R. Alizadeh, The investigation of catalytic ozonation and

integrated catalytic ozonation/biological processes for the removal of phenol from saline

wastewaters, J. Hazard. Mater. 171 (2009) 175-181.

[32] S. Ghorai, K.K. Pant, Investigations on the column performance of fluoride adsorption

by activated alumina in a fixed-bed, Chem. Eng. J. 98 (2004) 165–173.

[33] E. Malkoc, Y. Nuhoglu, Y. Abali, Cr(VI) adsorption by waste acorn of Quercus

ithaburensisin fixed beds: prediction of breakthrough curves, Chem. Eng. J. 119 (2006)

61–68.

20

[34] V.C. Taty-Costodes, H. Fauduet, C. Porte, Y.S. Ho, Removal of lead (II) ions from

synthetic and real effluents using immobilized Pinus svlvestris sawdust: adsorption on a

fixed column, J. Hazard. Mater. 123 (2005) 135-144.

[35] J. Wu, H.-Q. Yu, Biosorption of 2,4-dichlorophenol from aqueous solutions by

immobilized Phanerochaete chrysosporium biomass in a fixed-bed column, Chem. Eng. J.

138 (2008) 128-135.

[36] J.M. Salman, V.O. Njoku, B.H. Hameed, Batch and fixed-bed adsorption of 2,4-

dichlorophenoxyacetic acid onto oil palm frond activated carbon, Chem. Eng. J. 174

(2011) 33-40.

[37] S. Chen, Q. Yue, B. Gao, Q. Li, X. Xu, K. Fu, Adsorption of hexavalent chromium

from aqueous solution by modified corn stalk: A fixed-bed column study, Bioresource

Technol. 113 (2012) 114-120.

[38] N. Chen, Z. Zhang, C. Feng, M. Li, R. Chen, N. Sugiura, Investigations on the batch

and fixed-bed column performance of fluoride adsorption by Kanuma mud, Desalination

268 (2011) 76-82.

[39] M.A. Acheampong, K. Pakshirajan, A.P. Annachhatre, P. N.L. Lens, Removal of

Cu(II) by biosorption onto coconut shell in fixed-bed column systems, J. Ind. Eng. Chem.

19 (2013) 841–848.

[40] H. Zheng, L. Han, H. Ma, Y. Zheng, H. Zhang, D. Liu, S. Liang, Adsorption

characteristics of ammonium ion by zeolite 13X, J. Hazard. Mater. 158 (2008) 577–584.

[41] K.Y. Foo, L.K. Lee, B.H. Hameed, Preparation of tamarind fruit seed activated carbon

by microwave heating for the adsorptive treatment of landfill leachate: A laboratory

column evaluation, Bioresource Technol. 133 (2013) 599-605.

21

[42] S. Kundu, A.K. Gupta, As(III) removal from aqueous medium in fixed bed using iron

oxide-coated cement (IOCC): experimental and modeling studies, Chem. Eng. J. 129

(2007) 123–131.

[43] B. Cheknane, M. Baudu, J.-P. Basly, O. Bouras, F. Zermane, Modeling of basic green

4 dynamic sorption onto granular organo–inorgano pillared clays (GOICs) in column

reactor, Chem. Eng. J. 209 (2012) 7–12.

[44] G. Moussavi, R. Khosravi, N. Rashidnejad Omran, Development of an efficient

catalyst from magnetite ore: characterization and catalytic potential in the ozonation of

water toxic contaminants, Appl. Catal. A. 445–446 (2012) 42-49.

[45] J. Nawrocki, B. Kasprzyk-Hordern, The efficiency and mechanisms of catalytic

ozonation, Appl. Catal. B 99 (2010) 27–42.

[46] N. Ma, N.J.D. Graham, Degradation of atrazine by manganese-catalyzed

ozonation-influence of radical scavengers, Water Res. 34 (2000) 3822-3828.

22

Table and Figure Legends

Table Legends

Table 1: Main properties of the amoxicillin

Table 2: Information on the Thomas model for adsorption of amoxicillin onto NAC and

SAC fixed-beds

Table 3: Information of the Yoon-Nilson model for the adsorption of amoxicillin onto

NAC and SAC fixed-beds

Table 4: Characteristics of the AMX-contaminated real water in the feed and effluent

streams of the NAC-bed column

Figure Legends

Figure 1: The schematic setup of the experiment

Figure 2: Breakthrough point curves representing AMX adsorption onto (a) NAC bed and

(b) SAC bed columns at various EBCTs ranging from 2 to 10 min (AMX concentration:

50 mg/L; pH= 6.8)

Figure 3: Values for breakthrough-point capacity at various EBCTs (2-10 min) for both

NAC and SAC fixed-bed columns (AMX concentration: 50 mg/L; pH= 6.8)

Figure 4: Values for carbon usage rate at various EBCTs (2-10 min) for both NAC and

SAC fixed-bed columns (AMX concentration: 50 mg/L; pH= 6.8)

23

Figure 5: Breakthrough point curves of the regenerated (a) NAC and (b) SAC beds at

various ozonation doses compared with fresh carbons

Figure 6: Breakthrough point curves representing the NAC bed for the treatment of

aqueous solution and real contaminated water

24

Research highlights

• Amoxicillin removal was investigated by NAC and SAC fixed-bed columns.

• The NAC-bed attained a longer service time and better adsorption capacity than the

SAC-bed

• The saturated NAC-bed was efficiently regenerated with ozonation at 0.25

gO3/gNAC

• The regenerated NAC-bed preserved its adsorption capacity during repeated cycles

• The real contaminated water sample was efficiently treated in the NAC-bed column

25

Figure 1

1. Inlet tank 2. Peristaltic pump 3. Control valve 4. Sampling port 5. Glass column 6. Adsorbent bed 7. Effluent line 8. Effluent tank 4

1

2 3 3

4

5

6 7 7

8 8

26

Figure 2

a

b

27

Figure 3

28

Figure 4

29

Figure 5

a

b

30

Figure 6

31

Character/value Parameter

Molecular structure

26787-78-0 CAS number

C16H19N3O5S Molecular formula

365.4 g/mol Molar mass

3430 mg/L at 20 °C Solubility in water

2.4 (carboxyl), 7.4 (amine), and 9.6

(phenol)

Dissociation constant (pKa)

2.73x10-19

at 20 °C Henry's law constant

Table 1

32

Table 2

NAC column SAC column

Q (L/h) EBCT (min)

kT (L/g.h) q0 (mg/g) R2 kT (L/g.h) q0 (mg/g) R2

3.996 2 0.73 2518 0.977 0.84 1417 0.959

1.998 4 0.22 2733 0.972 0.20 2581 0.952

1.332 6 0.16 4185 0.973 0.18 2890 0.965

0.7992 10 0.13 5107 0.951 0.12 2956 0.955

33

Table 3

NAC column SAC column

Q (L/h) EBCT (min)

kYN (1/h) τ (h) R2 kYN (1/h) τ (h) R2

3.996 2 0.036 19.20 0.977 0.042 42.60 0.939

1.998 4 0.011 41.10 0.972 0.011 150.10 0.944

1.332 6 0.008 94.30 0.973 0.009 261.20 0.965

0.7992 10 0.007 178.00 0.931 0.006 443.80 0.945

35

Parameter Unit Feed water NAC effluent at breakthrough

point Reduction (%)

Amoxicillin mg/L 50 0 100

Total dissolved solids

mg/L 1386 1250 9.8

Alkalinity mg/L as CaCO3 106 100 5.7

Total hardness mg/L as CaCO3 140 136 2.9

Chloride mg/L 760 685 9.9

Sulfate mg/L 332 328 1.2

Nitrate mg/L 16 2.7 83.1

pH -- 6.9 6.8 ---

Dissolved organic carbon

Adsorption at 254 nm 0.119 0.006 95