Separation and Purification Technology 42 (2005) 237–247
Copper adsorption on chitosan–cellulose hydrogel beads:behaviors and mechanisms
Nan Li, Renbi Bai∗
Department of Chemical and Biomolecular Engineering, National University of Singapore, 10 Kent Ridge Crescent,Singapore 119260, Singapore
Received 26 May 2004; received in revised form 11 August 2004; accepted 12 August 2004
Abstract
The application of chitosan–cellulose hydrogel beads as an adsorbent for Cu adsorption from aqueous solutions was examined. Chitosan wasblended with cellulose to make chitosan–cellulose hydrogel beads and the hydrogel beads were crosslinked with ethylene glycol diglycidylether (EGDE). It was found that the addition of cellulose to chitosan made the hydrogel beads materially denser and the crosslinking reactionimproved the chemical stability of the chitosan–cellulose beads in solutions with pH values down to 1. Batch adsorption experiments indicated
Keywords:Chitosan–cellulose beads; EDGE crosslinking; Cu adsorption; Surface interaction; Mechanisms
1. Introduction
Heavy metal contamination of various water resources isof great concern because of the toxic effect to the human be-ings and other animals and plants in the environment. Themajor sources of heavy metal pollutants are usually frommany industries, including mining, metal plating, electricdevice manufacturing, and so on. For example, Cu is ex-tensively used in the electrical industry, and in the manu-facture of fungicides and anti-fouling paints [1]. AlthoughCu can be an essential trace element for the human beings,it could cause harmful, acute and even fatal effect when alarge dosage is ingested. Recent evidence also shows that Cucan be a human carcinogen and can cause severe harm to
the aqueous fauna when entering the water body [2]. Sinceheavy metal ions are not biodegradable, they are usually re-moved physically or chemically from the contaminated water.Conventional methods that have been used to remove heavymetal ions from various industrial effluents usually includechemical precipitation, membrane separation, ion exchange,evaporation, and electrolysis, etc. and are often costly or inef-fective, especially in removing heavy metal ions from dilutesolutions [2,3]. One of the new developments for metal re-moval in recent years is to use biosorption. Many materials ofbiological origins have been studied as adsorbents to removevarious heavy metal ions from water and industrial effluents[3]. Particularly, chitosan, a derivative from N-deacetylationof chitin—a naturally occurring polysaccharide from crus-tacean and fungal biomass, has been found to be capableof chemically or physically adsorbing various heavy metalions, including lead, vanadium, platinum, silver, cadmium,
238 N. Li, R. Bai / Separation and Purification Technology 42 (2005) 237–247
chromium, and copper [1,4–8]. Chitosan may also be cheaplyobtained from chitin which is the second most naturally abun-dant biopolymer (next to cellulose) and is readily availablefrom seafood-processing wastes.
Traditionally, chitosan has been used in the form of flakesor powder in metal adsorption. Because raw chitosan can becharacterized as a crystallized polymer, metal ions could onlybe adsorbed onto the amorphous region of the crystals [9].Progress has been made to produce chitosan hydrogel beadsto improve the adsorption capacity of chitosan by reducingthe crystallinity through the gel formation process [6], andalso provide the potential for regeneration and reuse of thehydrogel beads after metal adsorption [4,6,10,11]. The majormaterial limitation of the hydrogel beads is however in theirpoor acidic resistance and mechanical strength. Attemptshave been made to improve the chemical stabilities of thehydrogel beads in acidic conditions by chemical crosslink-ing of the surface with crosslinking agents, such as glutaric-dialdehyde (GA), ethylene glycol diglycidyl ether (EGDE),and epichlorhydrine [4]. Chemical crosslinking reaction wasfound to be able to reduce the solubility of chitosan hydro-gel beads in aqueous solutions of low pH values. On theother hand, polymer blending has increasingly been used asa method for providing polymeric materials with desirableproperties. Jin and Bai [4] studied lead adsorption with chi-tts
cTfmmaptabscamspaFp
2
2
d
glycol diglycidyl ether (EGDE), Cu(II) standard solution(1000 mg/l), and acetic acid were provided by Merck. Allother chemicals were of reagent grade purity and deionized(DI) water was used to prepare all solutions.
2.2. Preparation of chitosan–cellulose hydrogel beads
A 2 g amount of chitosan flake was added into 100 ml2%(w/w) acetic acid in a beaker and the contents in the beakerwere mixed on a hot plate stirrer at 70 ◦C and 200 rpm for 6 h.Then, a 2 g amount of cellulose powder was added into thechitosan solution and the mixing was continued for another6 h at room temperature (22–23 ◦C) and 200 rpm on the stirrer.The blended solution was then injected in droplets into a 1 MNaOH solution to form hydrogel beads through a vibrationnozzle system (Nisco Encapsulation Unit, LIN-0018, with anozzle size of 300 �m; see Fig. 1). The chitosan–cellulosehydrogel beads were allowed to stay in the NaOH solu-tion with slow stirring for another 12 h for hardening. Thehardened beads were finally separated from the NaOH solu-tion and were washed with DI water in a large beaker un-til the solution pH became the same as that of the fresh DIwater. Then, the beads were stored in DI water for furtheruse.
Fig. 1. The set-up of the granulation system and the macroscopic shape ofthe chitosan–cellulose hydrogel beads produced: (a) granulation system and(b) chitosan–cellulose hydrogel beads.
osan/polyvinyl alcohol (PVA) hydrogel beads and found thathe blending of PVA in chitosan improved the mechanicaltrength of the hydrogel beads.
In this paper, we report the study of usinghitosan–cellulose hydrogel beads for Cu adsorption.he interest in using cellulose as a blending polymer
or chitosan arises from two facts: (a) cellulose is theost abundant natural biopolymer with relatively strongechanical strength of up to 1 GN/m2 (10,000 MPa) [12],
nd (b) cellulose has similar chemical structures as chitosan,roviding the possibility of producing a homogeneous blendhat combine the unique properties of chitosan and the goodvailability of cellulose to make chitosan–cellulose hydrogeleads as an adsorbent for metals. The study included theynthesis of the chitosan–cellulose hydrogel beads, chemicalrosslinking of the chitosan–cellulose beads with EGDE,nd the examination of the adsorption performance andechanisms with the hydrogel beads for Cu removal. A
eries of batch adsorption experiments under various solutionH values and initial Cu concentrations was conductednd various analyses, such as zeta potential measurements,ourier transform infrared (FTIR) spectroscopy, and X-rayhotoelectron spectroscopy (XPS), were performed.
. Experimental work
.1. Materials and chemicals
Chitosan flakes (85% deacetylated) and cellulose pow-er (20 �m) were purchased from Sigma Co. Ethylene
N. Li, R. Bai / Separation and Purification Technology 42 (2005) 237–247 239
Some experiments involved in the crosslinking of thechitosan–cellulose hydrogel beads with EGDE. To do that,the beads in stock were taken out from DI water. The at-tached surface water was removed by placing them on a fewpieces of filter papers for several minutes. Then, 25 ml ofthe chitosan–cellulose beads was suspended into 25 ml ofDI water in a beaker, with the pH being adjusted to 12 bythe addition of 0.1 M NaOH solution. After 5 min, a 0.4 gamount of EGDE solution was added into the beaker andthe crosslinking reaction was allowed to proceed at 70 ◦C for6 h in a thermostatic water bath with continuous agitation.Finally, the mixture was cooled down to room temperature,and the crosslinked chitosan–cellulose beads were washed inan ultrasonic bath with sufficient DI water until the pH ofthe solution became around 6–6.5 (the same as the fresh DIwater). The beads were then stored in DI water for further use.
2.3. Dissolution and hydration rate test
The chitosan–cellulose hydrogel beads and thecrosslinked chitosan–cellulose hydrogel beads weretested for the dissolution property and hydration rate.For comparison purpose, similar chitosan hydrogel beadswithout cellulose were also prepared and crosslinked. In thedissolution tests, the various types of hydrogel beads werepwawstTtstobtptbb
H
wtdiss
2
r
in adsorption. To estimate the zeta potentials of thechitosan–cellulose hydrogel beads, about a 0.1 g amount ofthe dried chitosan–cellulose hydrogel beads (crosslinked ornon-crosslinked) was ground into powder and suspended into100 ml of DI water. The mixture was sonicated first for 4 h,followed by stirring for another 24 h, and then settled for12 h. Samples were taken from the supernatants (which hadcolloidal fragments from the chitosan–cellulose beads in it)and were used for zeta potential analysis. Before the analysis,each sample was distributed into several vials. The pH valuesof the sample in each of the vials were adjusted with 0.1 MHCl or 0.1 M NaOH solution to a desired level, and no back-ground electrolyte was added in the sample. A Zeta-Plus4instrument (Brookhaven Corp., USA) was used to measurethe zeta potentials of all the samples. Zeta potentials so de-termined from the fragments in the samples were assumed torepresent the zeta potentials of the chitosan–cellulose hydro-gel beads in solutions of the same pH values [13].
2.5. Adsorption experiments
2.5.1. Copper adsorption at different solution pHTo study the effect of solution pH on Cu adsorption on
chitosan–cellulose hydrogel beads, Cu solutions with an ini-tial concentration of 30 mg/l were prepared by diluting the1ttttrtaEebtnt4nctOnal
q
wc(te
laced in 0.1 M acetic acid, 0.1 M HCl, 0.1 M H2SO4, DIater, or 0.1 M NaOH to observe their solubility. About0.1 g amount of each type of the hydrogel beads waseighted and added into 50 ml of the different types of
olutions for a period of 24 h (with stirring). The weights ofhe beads after the test in each solution were determined.he differences in the weights before and after the solubility
ests give information on the stability of the beads in theolutions. In the hydration tests, a certain amount of eachype of the hydrogel beads stored in DI water was takenut from the stock and removed the surface water. Theeads were then weighed using a beam balance to obtainheir initial weights. The weighed beads were subsequentlylaced into a vacuum desiccator (at about 1 mTorr) at roomemperature for 3 days for drying. The weights of the driedeads were weighed again using the same beam balance asefore. The percentage of hydration was calculated by:
ydration rate = Wh − Wd
Wh(1)
hereWh is the weight of the hydrated beads weighed beforehe drying and Wd is the weight of dehydrated beads afterrying in the vacuum desiccator. The hydration rate is used tondicate the mass and water contents in the hydrogel beads. Amaller hydration rate is an indication of a greater mechanicaltrength of the hydrogel beads.
.4. Zeta potential measurement
Zeta potentials are often used as an important pa-ameter in analyzing the electrostatic surface interaction
000 mg/l standard Cu solution with DI water, and, then,he pH values of the solutions were adjusted to a value inhe range of 3–10 with 0.1 M NaOH or 0.1 M HCl. A cer-ain amount of chitosan–cellulose beads was taken out fromhe stock and placed on filter papers for a few minutes toemove the surface water. Then, about a 0.5 g amount ofhe chitosan–cellulose beads was weighed and added intonumber of 25 ml plastic bottles (which did not adsorb Cu).ach bottle contained 10 ml of the Cu solution of a differ-nt pH. The contents in the bottles were stirred in an or-ital shaker at 200 rpm and room temperature for adsorp-ion to proceed. The pH of the solutions in each bottle wasot controlled during the adsorption process (in order noto introduce any additional ions into the solution). After8 h, samples were taken from the bottles for the determi-ation of the final Cu concentrations in the solutions. Cuoncentrations in the samples were analyzed using an Induc-ively Coupled Plasma-Optical Emission Spectrometer (ICP-ES, Perkin Elmer Optima 3000 DV). Both crosslinked andon-crosslinked chitosan–cellulose beads were studied. Themount of adsorption, q (mg/g), was calculated from the fol-owing equation:
= (C0 − Ce)V
m(2)
here C0 (mg/l) and Ce (mg/l) are the initial and final Cuoncentrations of the solution in each bottle, respectively, Vl) is the volume of the Cu solution in each bottle, andm(g) ishe weight of chitosan–cellulose hydrogel beads added intoach bottle.
240 N. Li, R. Bai / Separation and Purification Technology 42 (2005) 237–247
2.5.2. Adsorption equilibrium studyThe experiments were conducted with solutions of Cu con-
centrations ranging from 5 to 100 mg/l at pH 6. About a 0.1 gamount of the chitosan–cellulose beads was added to a 5 mlsolution containing a different amount of Cu in 25 ml plasticbottles. The mixtures in the bottles were stirred in an orbitalshaker operated at 200 rpm and at room temperature for 7days (the pH of the solutions was again not adjusted duringthe adsorption process). The initial and final Cu concentra-tions in solution were determined using an ICP-OES. Similarexperiments were carried out for both the chitosan–cellulosehydrogel beads and the crosslinked chitosan–cellulose hy-drogel beads. The adsorption capacities were also calculatedusing Eq. (2).
2.5.3. Kinetic adsorption experimentsKinetic study was conducted at an initial pH of 6 and an
initial Cu concentration of 15 mg/l. A 12.5 g amount of thechitosan–cellulose beads (non-crosslinked or crosslinked)was added into 500 ml of the Cu solution in a one-liter flask.The mixture in the flask was shaken in an orbit shaker oper-ated at 200 rpm and room temperature for a period of up to24 h or until adsorption equilibrium was established. Sampleswere taken and analyzed for Cu periodically. The adsorbedamounts of Cu per unit weight of the hydrogel beads at time ti ,qa
q
wttci
2e
cctt(helt
2
ct1
KBr in an agate mortar and pressed into a tablet. FTIRspectra were obtained for the tablet from a FTS3500 FTIRSpectrometer.
2.8. X-ray photoelectron spectroscopy (XPS)
XPS analyses of the chitosan–cellulose and thecrosslinked chitosan–cellulose beads before and after Cu ad-sorption were made on a VGESCALAB MKII spectrometerwith an Al K� X-ray source (1486.6 eV of photons), follow-ing a similar procedure described by Bai et al. [14]. The XPSspectra peaks were decomposed into subcomponents by fix-ing the 0% Lorentzian–Gaussian curve-fitting program witha linear background to the spectra through an XPSpeak 4.1software package. The full width half maximum was main-tained at 1.4. The calibration of the binding energy (BE) ofthe spectra was performed with the C 1s peak of the aliphaticcarbons at 284.6 eV.
3. Results and discussion
3.1. Surface morphology
Macroscopically, the synthesized chitosan–cellulose hy-dchotctwhcgsbbm(gbssv
3
eTwsFc
(ti) (mg/g), was calculated from the mass balance equations:
(ti) =∑n
i=1(Cti−1 − Cti )Vti−1
m(3)
here Ct0 (=C0) and Cti (mg/l) are the initial Cu concentra-ion and the Cu concentrations at time ti , respectively, Vti ishe volume of the solution at time ti (samples taken for Cuoncentration analysis were not returned to the flask), and ms the weight of the hydrogel beads added into the flask.
.6. Surface morphology observation with scanninglectron microscope (SEM)
The chitosan–cellulose and the crosslinkedhitosan–cellulose beads were first placed into a desic-ator at 10 mTorr vacuum and room temperature for 72 ho remove excess moisture. The surface morphology washen observed using a JEOL scanning electron microscopeJSM-5600LV) with a high resolution of 3.5 nm underigh vacuum. Samples were platinum-coated by a vacuumlectric sputter coater (JEOLJFC-1300) to a thickness of ateast 500 A before glue-mounted onto the sample stud forhe SEM scan.
The dried chitosan–cellulose and crosslinkedhitosan–cellulose beads with or without Cu adsorp-ion were ground into powder. For each type of the powder,mg of the powder was blended with 100 mg of IR-grade
rogel beads were generally spherical (Fig. 1). Both thehitosan–cellulose and the crosslinked chitosan–celluloseydrogel beads had a milk-white color with a mean diameterf about 3.1 mm. The SEM images in Fig. 2 clearly show thathe addition of cellulose into chitosan made the surface of thehitosan–cellulose hydrogel beads much denser than that ofhe chitosan hydrogel beads, and the crosslinking reactionith EGDE also made the surface of the chitosan–celluloseydrogel beads very porous. It has been reported that therosslinking agent EGDE was prone to react with amineroups instead of hydroxyl groups [5]. Therefore, the poroustructure of the crosslinked chitosan–cellulose hydrogeleads may be formed due to the bridging connection of EGDEetween different amine groups in chitosan through intra-olecular and/or inter-molecular crosslinking. FTIR analysis
spectra are not shown here) of the chitosan–cellulose hydro-el beads and the crosslinked chitosan–cellulose hydrogeleads showed that the peaks representing the amine groupshifted and became weaker after the crosslinking reaction,uggesting that crosslinking reaction with EGDE indeed in-olved the amine groups in chitosan.
.2. Hydration and solubility properties
It has been found that the addition of cellulose in chitosannhanced the mechanical strength of the hydrogel beads.his can be partly supported by the SEM images in Fig. 2here the surface of the chitosan–cellulose beads was ob-
erved to be much denser than that of the chitosan beads.rom the results of hydration rate study given in Table 1, thehitosan–cellulose hydrogel beads are found to have much
N. Li, R. Bai / Separation and Purification Technology 42 (2005) 237–247 241
Fig. 2. SEM images showing the surface morphologies of the chitosan,chitosan–cellulose and crosslinked chitosan–cellulose hydrogel beads: (a)chitosan, (b) chitosan–cellulose and (c) crosslinked chitosan–cellulose.
greater mass ratio than the chitosan hydrogel beads, suggest-ing that the addition of cellulose in chitosan reduced the watercontent ratio in the hydrogel beads and thus made the beadsmaterially denser and hence mechanically stronger.
Table 1Hydration rate of various chitosan-based hydrogel beads
a The weight of chitosan in the solution to make the hydrogel beads is 2%.b The weight of chitosan and cellulose in the solution to make the hydrogel beads
Fig. 3. Zeta potentials of the chitosan–cellulose and the crosslinkedchitosan–cellulose beads at different solution pH values.
The results of the dissolution tests are presented in Table 2.It can be found that the addition of cellulose in chitosan didnot improve the acidic resistance of the hydrogel beads, butthe crosslinking reaction significantly increased the chemi-cal stability of the hydrogel beads in acidic media. Chitosanhydrogel beads have been known to be unstable or to start todissolve in solutions of pH < 4 [3]. The crosslinking reactionin this study however extended the chemical stability of thehydrogel beads to solutions of pH < 1 and did not affect thechemical stability of the hydrogel beads in neutral and basicpH solutions. It is generally known that the high hydrophilic-ity of chitosan or chitosan hydrogel beads is attributed to theprimary amine groups in chitosan. The fact that the crosslink-ing reaction improved the acid resistance of the hydrogelbeads therefore provides the evidence that some amine groupsin chitosan were consumed or shielded by the crosslink-ing reactions. In the following sections, the results for boththe chitosan–cellulose hydrogel beads and the crosslinkedchitosan–cellulose hydrogel beads will be presented anddiscussed.
3.3. Zeta potentials
The zeta potentials of the chitosan–cellulose and thecrosslinked chitosan–cellulose beads under different solu-tcsa
ight (g) Mass percentage of the hydrogels Hydration rate
2.71% 97.29%3.14% 96.86%5.78% 94.22 %6.51% 93.49%
is 2% for chitosan and 2% for cellulose, respectively.
ion pH values are shown in Fig. 3. It is observed thathitosan–cellulose beads had positive zeta potentials in acidicolutions and negative zeta potentials in basic solutions, withpoint of zero zeta potential at about pH 6.7. The pKa value of
242 N. Li, R. Bai / Separation and Purification Technology 42 (2005) 237–247
Table 2Solubility behavior of various chitosan-based hydrogel beads
Types of beads Dissolution condition
0.1 M HAc 0.1 M HCl 0.1 M H2SO4 DI water 0.1 M NaOH
the amine group in chitosan molecule has been known to bedependent on the extent of deacetylation (α). It was reportedthat the pKa was 6.1 when α < 0.72 and 6.7 when α > 0.72[15]. The α value of chitosan used in this study was given as0.85 (specified by Sigma Co.), and, hence, the point of zerozeta potential at about pH 6.7 for the chitosan–cellulose beadswas essentially the same as the pKa value of the amino groupin chitosan [15,16]. From the electrostatic interaction pointof view, the positive zeta potentials of the chitosan–cellulosebeads under acidic solution conditions would favor the ad-sorption of negatively charged species and the negative zetapotentials of the beads under basic solution conditions mayenhance the adsorption of positively charged species. Thecrosslinking reaction is observed to slightly reduce the posi-tive zeta potentials of the beads at pH < 6.7. Since the positivezeta potentials of the beads may partially result from the pro-tonation of the amine groups in chitosan, the reduction of thepositive zeta potentials of the crosslinked chitosan–cellulosebeads may again indicates that some of the amine groups onthe chitosan–cellulose beads were consumed or blocked bythe cross-linking reactions. (The negative zeta potentials ofthe beads in high solution pH conditions may be from theadsorption of the OH− to the amine groups, due to the addedNaOH for the adjustment of the solution pH.)
3
a
FtC
initial pH values from 3 to 10 (Cu mainly existed as Cu2+
in this pH range). In general, the adsorption capacity in-creased with the increase of the solution pH values for boththe chitosan–cellulose and the crosslinked chitosan–cellulosebeads. The maximum adsorption capacity reached about14–16 mg/g. It is also observed that the chitosan–cellulosebeads always had greater adsorption capacities than thecrosslinked chitosan–cellulose beads in the pH range stud-ied.
As mentioned earlier, the zeta potentials of both thechitosan–cellulose and the crosslinked chitosan–cellulosebeads were positive at pH < 6.7 and negative at pH > 6.7.The electrostatic interaction between the two types of beadsand the Cu would be repulsive at pH < 6.7 but becomes attrac-tive at pH > 6.7. The repulsive electrostatic interaction at pH< 6.7 may hinder Cu in the bulk solution from approachingthe surface of the beads for adsorption to take place. However,the electrostatic repulsion decreased with the increase of thesolution pH. This probably explains the observed increase ofthe adsorption capacities with solution pH from 3 to 6.7. AtpH above 6.7, the attractive electrostatic interaction betweenthe beads and the Cu would favor the adsorption to take place,but the adsorption capacity did not increase further with thesolution pH at pH > 6.7. The change in the adsorption char-acteristics with solution pH in Fig. 4 may be more clearly ex-pbt
–
–
–
Et(afpif–b
.4. Effect of pH on Cu adsorption
Fig. 4 shows the Cu adsorption on the chitosan–cellulosend the crosslinked chitosan–cellulose beads in solutions of
ig. 4. Effect of initial solution pH values on Cu adsorption capacities onhe chitosan–cellulose and the crosslinked chitosan–cellulose beads (initialu concentration in the solution: 30 mg/l).
lained by the acid–base properties of the beads. Eqs. (4)–(7)elow may be used to show the major characteristic reactionshat can take place at the solid–solution interface:
NH2 + H+ ↔ –NH3+ (4)
NH2 + Cu2+ → –NH2Cu2+ (5)
NH2 + OH− ↔ –NH2OH− (6)
–NH2OH− + Cu2+(or CuOH+)
↔ –NH2OH− · · · Cu2+(or –NH2OH− · · · CuOH+) (7)
q. (4) indicates the protonation and deprotonation reac-ions of the amine groups of chitosan in the solution, Eq.5) shows the formation of surface complexes of Cu with themine groups, and Eq. (6) describes the adsorption of OH−rom the solution through hydrogen bond at high solutionH conditions [4]. At lower solution pH values, the reactionn Eq. (4) favored the protonation of the amine groups toorm –NH3
+. Since more –NH2 groups were converted toNH3
+, there were only fewer –NH2 sites available on theeads’ surface for Cu adsorption through Eq. (5). In addi-
N. Li, R. Bai / Separation and Purification Technology 42 (2005) 237–247 243
tion, the formation of more –NH3+ sites on the surface in-
creased the electrostatic repulsion between the Cu and thesurfaces of the beads. All these effects would result in thereduction of Cu adsorption on the beads with decreasing so-lution pH values. In comparison with the adsorption capaci-ties of the chitosan–cellulose beads, the significant reductionin Cu adsorption on the crosslinked chitosan–cellulose beadsat pH below about 7 (Fig. 4) can be attributed to the factthat the total number of –NH2 and –NH3
+ groups on thecrosslinked chitosan–cellulose beads was much smaller (dueto the crosslinking reaction which consumed some of thesegroups) and the reaction in Eq. (4) had dramatic impact onthe number of –NH2 available for Cu adsorption through Eq.(5). On the other hand, with the increase of solution pH, thereaction in Eq. (4) proceeds to the left, resulting in an increaseof the number of –NH2 sites on the surface of the beads forCu adsorption through Eq. (5), thus increasing the adsorptioncapacity. At higher solution pH, the reaction in Eq. (6) mayproceed. This reaction on one hand can reduce the adsorp-tion of Cu through surface complexation in Eq. (5), but on theother hand may increase the adsorption of Cu on the beadsthrough electrostatic attraction (non-specific or physical ad-sorption) as indicated in Eq. (7). It appears that, at pH val-ues greater than 8 or 9, the formation of –NH2OH− throughEq. (6) in fact reduced the adsorption of Cu on the beads,iet
3
auCcc
Fc(
tion, the adsorption capacity of both the chitosan–cellulosebeads and the crosslinked chitosan–cellulose beads increasedsignificantly. At lower initial Cu concentrations, the adsorp-tion capacities increased linearly with the initial Cu concen-trations, suggesting that the adsorption sites on the beads weresufficient and the amount of adsorption in these cases weredependent on the number of Cu that were transported from thebulk solution to the surfaces of the beads. At higher initial Cuconcentrations, however, the adsorption capacities no longerincreased proportionally with the initial Cu concentrations,indicating that the number of adsorption sites on the surfacesof the beads actually limited the adsorption capacities.
The adsorption isotherm data were analyzed by the Lang-muir isotherm model in the linearlized form
Ce
qe= Ks
qmax+ Ce
qmax(8)
where qe is the equilibrium Cu adsorption amounts on thebeads (mg/g), Ce is the equilibrium Cu concentrations inthe solution (mg/l), qmax represents the maximum amountof Cu that could be adsorbed on the beads (mg/g), and Ksis a constant of the Langmuir model (mg/l). The plot of theexperimental Ce/qe against Ce for the experimental data inFig. 5 is shown in Fig. 6a. It is observed that Cu adsorptionon the chitosan–cellulose beads can be fitted to the LangmuiriomttLoeFsq
ae
l
w(iflbtua2icic
ndicating that surface complexation had a more importantffect than the electrostatic interaction on Cu adsorption onhe beads in this pH range.
.5. Adsorption isotherms
Adsorption isotherms describe how adsorbates inter-ct with adsorbents and are important in optimizing these of adsorbents. The experimental adsorption data ofu on the chitosan–cellulose beads and the crosslinkedhitosan–cellulose beads are shown in Fig. 5. The resultslearly indicate that with an increase in initial Cu concentra-
ig. 5. Adsorption capacities of Cu on the chitosan–cellulose and therosslinked chitosan–cellulose beads at various initial Cu concentrationsinitial solution pH = 6).
sotherm model very well. From the slope and the interceptf the straight line, the values of qmax and Ks can be esti-ated to be 53.2 mg/g and 2.29 mg/l, respectively. However,
he Langmuir isotherm model cannot describe the adsorp-ion of Cu on the crosslinked chitosan–cellulose beads. Theangmuir isotherm model assumes that the adsorbed layer isne molecule in thickness and that all adsorption sites havequal energies and enthalpies of adsorption [2]. As shown inig. 2, the crosslinked chitosan–cellulose beads had poroustructures. Hence, the homogeneous surface conditions re-uired by the Langmuir isotherm model may not be met.
Alternatively, the experimental data in Fig. 5 may be an-lyzed with the Freundlich isotherm model which in its lin-arized form is:
og qe = 1
nlog Ce + log P (9)
here P is a constant representing the adsorption capacitymg/g)(L/mg)n, and n is a constant depicting the adsorptionntensity (dimensionless). The plot of log qe versus logCeor the experimental data in Fig. 5 is shown in Fig. 6b. It il-ustrates that Cu adsorption on both the chitosan–celluloseeads and the crosslinked chitosan–cellulose beads obeyshe Freundlich isotherm model reasonably well. The val-es of the Freundlich model constants P and n are 13.167nd 2.285 for the chitosan–cellulose beads and 5.563 and.259 for the crosslinked chitosan–cellulose beads. The sim-lar values of n for the chitosan–cellulose beads and therosslinked chitosan–cellulose beads indicate that the driv-ng force for Cu adsorption on the two types of beads isomparable. However, the significantly greater value of P
244 N. Li, R. Bai / Separation and Purification Technology 42 (2005) 237–247
Fig. 6. Illustration of the experimental adsorption isotherm data presented interms of the linearized Langmuir and Freundlich models, respectively, for Cuadsorption on the chitosan–cellulose and the crosslinked chitosan–cellulosebeads (pH = 6): (a) plot according to the Langmuir model in Eq. (8), (b) plotaccording to the Frendlich model in Eq. (9).
for the chitosan–cellulose beads (13.167) than that for thecrosslinked ones (5.563) indicates that the affinity of Cu tothe chitosan–cellulose beads is much greater than that to thecrosslinked chitosan–cellulose beads. The Freundlich modelis based on an assumption of adsorption on heterogeneroussurfaces and also possibly in multi-layer adsorption pattern.Cu adsorption is better described by the Freundlich modelfor the crosslinked chitosan–cellulose beads than that forthe chitosan–cellulose beads. The analysis hence suggests,at least partly, that the crosslinked chitosan–cellulose beadswere more heterogenerous in the surface properties than thechitosan–cellulose beads.
3.6. Adsorption kinetics
The kinetics of adsorption is to establish the time course ofCu uptake on the beads. It is also desirable to examine whetherthe behavior of Cu adsorption on the beads can be describedby a theoretical model that is predictive [17,18]. The typi-cal experimental results of adsorbed Cu on the two types ofbeads versus time are shown in Fig. 7. The adsorption kineticsof the crosslinked chitosan–cellulose beads was faster than
Fig. 7. Typical kinetic adsorption results of Cu on the two types of hydrogelbeads (initial solution pH = 6, initial Cu concentration = 15 mg/l, amount ofadsorbent = 12.5 g, and volume of solution = 500 ml).
the chitosan–cellulose beads. The adsorption equilibrium wasreached at about 7 h for the crosslinked chitosan–cellulosebeads but took about 15 h for the chitosan–cellulose beads,possibly due to their higher adsorption capacity. The fast ad-sorption rate of the crosslinked chitosan–cellulose beads mayprobably be due to their three dimensional structures andhence possibly large specific surface areas.
The adsorption of Cu on the beads may be considered toconsist of two processes: (a) the transport of Cu from thebulk solution to the surfaces of the beads (including externaland/or intraparticular diffusions), and (b) the attachment ofCu to the active adsorption sites on the surfaces of the beads.In the initial stage, the surfaces of the beads were relativelyfree of Cu and the Cu that arrived at the beads’ surfaces mayattach instantly to the surface sites. Hence, the adsorptionrate may be dominated by the number of Cu diffused fromthe bulk solution to the surfaces of the beads in this case.Based on the Fickian diffusion law, McKay and Poots [19]showed that the amount of adsorption by diffusion-controlleddynamics as a function of time can be given as:
q(t) = 2C0S√
Dt/π = kdt0.5 (10)
where q(t) represents the amount of Cu adsorbed per unitweight of the beads at time t (mg/g), C0 is the initialconcentration of Cu in the bulk solution, D is the dif-ftbtrk
riooob
usion coefficient, and S is the specific surface area ofhe chitosan–cellulose or the crosslinked chitosan–celluloseeads. Eq. (10) indicates that under diffusion-controlledransport mechanism, q(t) versus t0.5 would follow a linearelationship, with kd = 2C0S
√D/π depicting the intrinsic
inetic rate constant for diffusion-controlled adsorption.Fig. 8 shows a plot of q(t) versus t0.5 for the experimental
esults in Fig. 7. A linear relationship of q(t) against t0.5 isndeed observed in the initial step of adsorption for each typef the beads. The results confirm the existence and importancef diffusion-controlled transport mechanism in Cu adsorptionn the chitosan–cellulose or crosslinked chitosan–celluloseeads. From the correlation analysis, the value of kd is found
N. Li, R. Bai / Separation and Purification Technology 42 (2005) 237–247 245
Fig. 8. The fitting of diffusion-controlled kinetic model, Eq. (10), to thedynamic adsorption amounts of Cu for the experimental results in Fig. 7.
to be 0.577 and 0.386 for the chitosan–cellulose beads andthe crosslinked chitosan–cellulose beads, respectively.
In the later stage of Cu adsorption, the experimental datain Fig. 7 no longer obey the model in Eq. (10). This indi-cates that other factors started to play an important role incontrolling the adsorption. Since most of the adsorption siteson the beads’ surfaces were occupied by previously adsorbedCu, the Cu that were subsequently transported to the surfaceof the beads had to find available adsorption sites before theattachment can occur. The adsorption of Cu on the beadsin the later stage hence slowed down and would probablytransit from the initial diffusion-controlled process to a finalattachment-controlled process.
3.7. Adsorption mechanisms
To identify the possible sites of Cu bonding to thetwo types of beads, FTIR spectra were obtained for thechitosan–cellulose and the crosslinked chitosan–cellulosebeads before and after Cu adsorption, as shown inFig. 9.
In general, it is observed that the major changes in theFTIR spectra after Cu adsorption are quite similar for boththe chitosan–cellulose and the crosslinked chitosan–cellulosebeads, indicating that the mechanisms of Cu adsorption ontc1wtbcbiatstwl
Fig. 9. FTIR spectra for the two types of hydrogel beads before andafter Cu adsorption: (a) chitosan–cellulose beads and (b) crosslinkedchitosan–cellulose beads.
ited lower adsorption capacities than the chitosan–cellulosebeads, as shown in Fig. 4.
In addition, the changes in the FTIR spectra at wavenum-bers of 3434.9 and 1076.8 cm−1 may also be assigned to the–OH stretching vibration in the alcohol group and the –C Ostretching vibrations, respectively (transmittance was signif-icantly reduced; see Fig. 9). These changes may suggest thepossibility that the oxygen atoms in the hydroxyl groups inchitosan and cellulose were involved in Cu adsorption as well.
To further verify the findings from the FTIR spectra,XPS were employed. XPS spectra have widely been usedto identify the existence of a particular element and to distin-guish the different forms of the same element in a material[4,14,16,20,21]. Fig. 10 shows the typical XPS wide scanspectra for the crosslinked chitosan–cellulose beads beforeand after Cu adsorption. It is clear that a new peak at the BEof about 933 eV appeared after Cu adsorption. The presenceof a satellite band nearby is representative of the oxidationstate +2 for the Cu 2p3/2 orbital. Therefore, the peak at BE of933 eV provides evidence of Cu (i.e., Cu2+) being adsorbedon the surface of the beads.
he two types of beads are essentially the same. Significanthanges in the FTIR spectra are found at the wavenumbers of435.7 and 1327.6 cm−1 after Cu adsorption. The peaks at theavenumbers of 1435.7 and 1327.6 cm−1 can be assigned to
he –N H deformation vibration and –C N– stretching vi-ration in chitosan molecules, respectively. Another majorhange in the FTIR spectra can be observed at the wavenum-er of 2365.8 cm−1 which can be assigned to the –NH stretch-ng vibration. Therefore, Cu adsorption on the beads affectedll the chemical bonds with N atoms. It is, thus, reasonableo assume that the nitrogen atoms should be the main ad-orption sites for Cu attachment on the beads. Since some ofhe nitrogen atoms were involved in the crosslinking reactionith EGDE and became unavailable for Cu adsorption, it is
ikely that the crosslinked chitosan–cellulose beads exhib-
246 N. Li, R. Bai / Separation and Purification Technology 42 (2005) 237–247
Fig. 10. Typical wide scan XPS spectra for the crosslinkedchitosan–cellulose beads before and after Cu adsorption: (a) beforeCu adsorption and (b) after Cu adsorption.
In Fig. 11, the typical N 1s XPS spectra ofthe chitosan–cellulose beads and the crosslinkedchitosan–cellulose beads before and after Cu adsorp-tion are presented. Before Cu adsorption, there is onlyone peak at 399.1 eV for the chitosan–cellulose beads orat 399.3 eV for the crosslinked chitosan–cellulose beads.This is attributed to the N atom in the –NH2 and/or the–NH– groups on the surfaces of the chitosan–celluloseor crosslinked chitosan–cellulose beads [16]. After Cuadsorption, however, a new peak at BE of 400.9 eV forthe chitosan–cellulose beads or at BE of 401.4 eV for thecrosslinked chitosan–cellulose beads is observed. Thisindicates that some N atoms existed in a more oxidizedstate on the beads’ surfaces due to Cu adsorption. This phe-nomenon can be attributed to the formation of R–NH2Cu2+
complexes, in which a lone pair of electrons in the nitrogenatom was donated to the shared bond between the N andCu2+, and, as a consequence, the electron cloud density ofthe nitrogen atom was reduced, resulting in a higher BE peakobserved. Therefore, the XPS spectra provide evidence ofCu binding to nitrogen atoms, in agreement with the FTIRfindings. The O 1s XPS spectra however did not clearlyshow significant changes of the O 1s BEs before and afterCu adsorption (less than 0.5 eV and the results not shown).Since both FTIR and XPS spectra could not provide cleareaaC
Fig. 11. Typical N 1s XPS spectra for the two types of hydrogel beads before (lcrosslinked chitosan–cellulose beads.
vidence that the chemical bonds associated with the oxygentoms on both beads were significantly changed after Cudsorption, it may be speculated that the contribution ofu-oxygen interaction to Cu adsorption on the beads was
eft) and after (right) Cu adsorption: (a) chitosan–cellulose beads and (b)
N. Li, R. Bai / Separation and Purification Technology 42 (2005) 237–247 247
mainly through a non-specific interaction (physical adsorp-tion, electrostatic attraction, etc.) or a very weak chemicalinteraction.
4. Conclusion
Chitosan–cellulose hydrogel beads can be used as an ef-fective adsorbent for Cu removal from aqueous solutions.The addition of cellulose to chitosan made the hydrogelbeads materially denser and hence mechanically stronger.The crosslinking reaction using EGDE improved the chem-ical stability of the chitosan–cellulose beads in acid solu-tions with pH down to 1. Both the chitosan–cellulose andthe crosslinked chitosan–cellulose beads had high adsorp-tion capacities for Cu adsorption, although the adsorptionwas pH-dependent (with maximum adsorption at a neu-tral pH) and the crosslinked beads had slightly lower ad-sorption capacities. The Langmuir isotherm model can de-scribe the adsorption isotherm of the chitosan–cellulosebeads very well, but the Freundlich isotherm model has tobe used for the crosslinked chitosan–cellulose beads. Cu ad-sorption on both the chitosan–cellulose and the crosslinkedcctttia
Acknowledgement
The financial support of the Academic Research Funds,National University of Singapore, is acknowledged.
References
[1] R. Schmuhl, H.M. Krieg, K. Keizer, Water SA 27 (2001) 1.[2] J.C.Y. Ng, W.H. Cheung, G. McKay, J. Colloid Interface Sci. 255
(2002) 64.[3] G. Bayrammoglu, A. Denizli, S. Sektas, M.Y. Arica, Microchem. J.
72 (2002) 63.[4] L. Jin, R.B. Bai, Langmuir 18 (2002) 9765.[5] M. Jansson-Charrier, E. Guibal, J. Roussy, B. Delanghe, P. Le
Cloirec, Water Res. 30 (1996) 465.[6] W.S. Wan Ngah, C.S. Endud, R. Mayanar, React. Funct. Polym. 50
(2002) 181.[7] J.R. Evans, W.G. David, J.D. MacRae, A. Amirbahman, Water Res.
[10] E. Piron, A. Domard, Int. J. Biol. Macromol. 22 (1998) 33.[11] C. Milot, J. McBrien, S. Allen, E. Guibal, J. Appl. Polym. Sci. 68
(1998) 571.[12] L. Yang, W.W. Hsiao, P. Chen, J. Membrane Sci. 5084 (2001) 1.[13] R.B. Bai, C. Tien, J. Colloid Interface Sci. 218 (1999) 488.[14] R.B. Bai, X. Zhang, J. Colloid Interface Sci. 243 (2001) 52.[15] G.A.F. Roberts (Ed.), Chitin Chemistry, Macmillan, London, 1992
[[[[
[[
hitosan–cellulose beads is found to obey a transport-ontrolled adsorption kinetics in the initial stage but mayransit to an attachment-controlled adsorption kinetics inhe later stage. FTIR and XPS analyses clearly revealhat the nitrogen atoms in chitosan were the main bind-ng sites for Cu to form surface complexes during thedsorption.
(Chapter 5).16] X. Zhang, R.B. Bai, J. Colloid Interface Sci. 264 (2003) 30.17] Y.A. Ho, G. McKay, Process Biochem. 34 (1999) 451.18] Y. Sag, Y. Aktay, Biochem. Eng. J. 12 (2002) 143.19] G. McKay, R.E. Poots, J. Chem. Technol. Biotechnol. 30 (1980)
