- Email: [email protected]
Behaviors and mechanisms of copper adsorption on hydrolyzed polyacrylonitrile fibers
Journal of Colloid and Interface Science 260 (2003) 265–272 www.elsevier.com/locate/jcis
Behaviors and mechanisms of copper adsorption on hydrolyzed polyacrylonitrile fibers Shubo Deng, Renbi Bai,∗ and J.P. Chen Department of Chemical and Environmental Engineering, National University of Singapore, 10 Kent Ridge Crescent, 119260 Singapore Received 10 August 2002; accepted 20 December 2002
Abstract Polyacrylonitrile fiber (PANF) was hydrolyzed in a solution of sodium hydroxide and the hydrolyzed polyacrylonitrile fiber (HPANF) was used as an adsorbent to remove copper ions from aqueous solution. Scanning electron microscopy (SEM) showed that the hydrolysis process made the surface of HPANF rougher than that of PANF. Fourier transform infrared (FTIR) spectroscopy revealed that the HPANF contained conjugated imine (–C=N–) sequences. Batch adsorption results indicated that the HPANF was very effective in adsorbing copper, and the adsorption equilibrium could be reached within 10–20 min. Atomic force microscopy (AFM) showed that some aggregates formed on the surface of the HPANF after copper ion adsorption and the average surface roughness (Ra ) value of the HPANF changed from 0.363 to 3.763 nm due to copper adsorption. FTIR analysis indicated that copper adsorption caused a decrease of the light adsorption intensity of the imine (–C=N–) groups at 1573 and 1406 cm−1 wavenumbers, and X-ray photoelectron spectroscopy (XPS) showed that the binding energy (BE) of some of the nitrogen atoms in the HPANF increased to a greater value due to copper adsorption. The FTIR and XPS results suggest that the adsorption of copper ions to the HPANF is attributed to the imine groups on the surface of the HPANF. 2003 Elsevier Science (USA). All rights reserved. Keywords: Hydrolyzed polyacrylonitrile fiber (HPANF); Copper adsorption; FTIR and XPS spectra; SEM and AFM images
1. Introduction Fibers, which have large specific surface areas and high adsorption rates, have attracted great attention in heavy metal removal from wastewater in recent years. There are many types of fibers or modified fibers, including activated carbon fiber, cotton fiber, glass fiber, and polymer fiber (polyester, polypropylene, polyethylene terephthalate, polystyrene), that have been used to adsorb metal ions or to enrich trace amount of metals from aqueous solutions [1–7]. The adsorption of metal ions from aqueous solutions to adsorbents is usually controlled by the properties of the surface functional groups of the adsorbents. Particularly, modified fibers that contain carboxyl, tetrazine, amidoxime– hydroxam, imidazoline, amino, and phosphoric groups have been found to be effective in metal ion adsorption [1,7–9]. An example is polyacrylonitrile fiber (PANF), a common and cheap commercial product, that has been modified to introduce various types of chelating groups and been used * Corresponding author.
E-mail address: [email protected] (R. Bai).
as an adsorbent for metal ions removal [8]. The modification of PANF may be done through hydrolysis in an acid or a base [10,11]. In general, the hydrolysis of PANF can be a complex process and different types of hydrolyzed products with different capabilities for metal adsorption may be obtained. From alkaline hydrolysis of PANF, the products may include those with conjugated sequences of –C=N–, acrylamide, sodium acrylate, and amidine from different stages of the hydrolysis [11–13]. Although many adsorbents were used to adsorb metal ions from aqueous solutions, the mechanisms of metal ion adsorption are not always fully understood. X-ray photoelectron spectroscopy (XPS) has found increasing applications in studying the interactions between an adsorbate and the active groups on the surface of an adsorbent in the adsorption process [14–17]. This is due to the fact that the creation of chemical bonds during adsorption would change the distribution of the electrons around the corresponding atoms, and, as a consequence, the binding energy of the electrons could be shifted to a lower or higher level, which would provide information on the adsorption mechanisms. Alternatively, Fourier transform infrared (FTIR) spectroscopy may
0021-9797/03/$ – see front matter 2003 Elsevier Science (USA). All rights reserved. doi:10.1016/S0021-9797(02)00243-6
266
S. Deng et al. / Journal of Colloid and Interface Science 260 (2003) 265–272
be used in the study of adsorption mechanisms, through examining the changes of adsorption to the light energy when a material is exposed to radiation in the infrared portion of the electromagnetic spectrum at certain frequencies since each absorption band in an FTIR spectrum is caused by different chemical bonds [4,17]. In addition, scanning electron microscopy (SEM) and atomic force microscopy (AFM) have often been used to visually show the morphological changes of the adsorbents after adsorption with an adsorbate [14–19]. One of the major advantages of AFM as compared to SEM is that the images can be shown in three dimensions and at nanometer scales. In this study, PANF was hydrolyzed in a sodium hydroxide solution and the performance of the hydrolyzed polyacrylonitrile fiber (HPANF) in copper ion adsorption was studied. The mechanisms of surface interactions were investigated through FTIR and XPS analyses, and the surface morphologies of the HPANF before and after adsorption were examined through SEM and AFM images.
2.4. FTIR spectroscopy The samples of PANF, HPANF, and copper-adsorbed HPANF were cut into about 1-mm pieces, blended with KBr, and pressed into discs. The spectra were recorded on a BioRad FTIR Model FTS135 spectrophotometer under ambient conditions. 2.5. X-ray photoelectron spectroscopy
2. Experimental
XPS analysis of the HPANF before and after the adsorption of copper ions was carried out on an AXIS HIS spectrometer (Kratos Analytical Ltd., England) with an AlKα X-ray source (1486.71 eV of photons) to determine the elements such as C, N, and O atoms present. The X-ray source was run at a reduced power of 150 W, and the pressure in the analysis chamber was maintained at less than 10−8 Torr during each measurement. All binding energies were referenced to the neutral C(1s) peak at 284.71 eV. Surface elemental stoichiometries were determined from the sensitivity-factorcorrected peak-area ratios, and the software XPSpeak 4.1 was used to fit the XPS spectra peaks [14].
2.1. Materials
2.6. Adsorption and desorption experiments
Polyacrylonitrile fiber (PANF), a commercial product, was kindly provided by Daqing Polyacrylonitrile Corp., PR China. The PANF contained 90% polyacrylonitrile and 10% vinyl acetate and had a diameter of about 20–50 µm. The other chemicals used in the study were of reagent grade. 2.2. Hydrolysis of PANF The hydrolysis of PANF was carried out in a solution of sodium hydroxide, with the conditions following: 10 g of PANF was added to a 500-ml flask with 300 ml of ethanol/water (mole ratio of 0.22) solution containing 10% (w/w) NaOH; the solution was stirred at 75 ◦ C in a shaking water bath for 20 min; then the HPANF was collected, rinsed with deionized (DI) water until neutral, and then dried in an oven at 50 ◦ C to constant weight and stored in a desiccator prior to use in the study. 2.3. AFM and SEM analysis A scanning electron microscope (SEM, JEOL JSM-6400) at 10–20 kV was used to examine the surface morphologies of PANF before and after hydrolysis. The morphologies of HPANF surfaces before and after copper adsorption were studied with a Nanoscope IIIa atomic force microscope (AFM) from Digital Instruments Inc. using the tapping mode at a scan size of 500 nm. The drive frequency was 330 kHz, and the voltage was between 3.0 and 4.0 V. The drive amplitude and scan rate were 300 mV and 1.0 Hz, respectively.
Adsorption experiments were conducted as follows. To a 100-ml flask, 50 ml of Cu(NO3 )2 ·3H2 O solution (pH 4.5) with a specific copper concentration was added. Then 0.1 g of HPANF was added to the solution. The flask was shaken at 120 rpm in a water bath with a temperature of 25 ◦ C. Copper concentrations in the solution were determined at desired time intervals until adsorption equilibrium was established. For the desorption study, 0.1 g of HPANF was added to 50 ml of 500 mg/l Cu(NO3 )2 ·3H2 O solution with other conditions the same as in the adsorption experiments. After the adsorption equilibrium was reached, the fibers were separated by filtration and rinsed with DI water to remove residual solution trapped among the fibers. Then the HPANF with adsorbed copper was transferred to a flask with 50 ml of 1 mol/l HCl solution or DI water (pH 5.0). The flask was shaken at 150 rpm in a rotary shaker at 25 ◦ C and the copper ion concentrations in the solution were analyzed against time. An inductively coupled plasma emission spectroscope (ICP-ES) (Perkin–Elmer Optima 3000, USA) was used to determine the copper concentrations in the samples.
3. Results and discussion 3.1. SEM images As shown in Fig. 1, the surface of PANF before the hydrolysis was relatively smooth and uniform. After the hydrolysis, the fiber surface was corroded and became rough, indicating that the hydrolysis reaction occurred on the surface of the PANF.
S. Deng et al. / Journal of Colloid and Interface Science 260 (2003) 265–272
267
(a) Fig. 2. FTIR spectra of (a) PANF and (b) HPANF.
Scheme 1. Hydrolysis of PANF to HPANF.
(b) Fig. 1. SEM images showing the surface morphologies of (a) PANF and (b) HPANF before and after hydrolysis.
3.2. FTIR spectroscope analysis The spectra of PANF before and after hydrolysis is shown in Fig. 2. It is seen that the peaks of the PANF (Fig. 2a) can be assigned as follows: 3632 cm−1 (γ OH), 2946 and 2874 cm−1 (γ C–H asymmetric and symmetric in CH, CH2 , and CH3 groups), 2246 cm−1 (γ C≡N), 1739 cm−1 (γ C≡O), 1457 cm−1 (δ CH3 and δS CH2 ), 1363 cm−1 (δ CH3 symmetric in CCH3 ), 1172 cm−1 (γ C–N), 1074 cm−1 (δ C–N), and 538 cm−1 (δτ C=O), where γ represents a stretching vibration, δ a blending vibration, δS a scissor vibration, and δτ a twisting vibration. After the hydrolysis with sodium hydroxide, the spectrum of the HPANF (Fig. 2b) changed greatly. The broad adsorption band at 3632 cm−1 shifted to 3423 cm−1 , with the intensity of the peak increased significantly. This band corresponds to the stretching vibration of the OH group and indicates the introduction of the OH groups onto the HPANF. The new peaks at 1573, 1406, and 1224 cm−1 wavenumbers indicate the existence of imine (–C=N–) conjugated sequences in the modified fiber [13,20]. Since the peak at 2246 cm−1 for –C≡N did not change significantly, only part
Scheme 2. Reaction of the nitrile group in PANF during hydrolysis.
of the –C≡N groups in the PANF were therefore converted to the –C=N– groups in the HPANF. The intensity of the peak at 2946 cm−1 (γ CH2 ) was reduced greatly, attributed to the hydrolysis of the ester group in the PANF. The peaks at 1739 cm−1 (γ C=O) and 538 cm−1 (δτ C=O) became weaker after the hydrolysis, suggesting acetate ester being removed from the surface of PANF. Based on the FTIR results and other references [11–13], the reaction taking place during the hydrolysis may be proposed to be as in Scheme 1. The cyclization reaction of the nitrile groups in the hydrolysis may be given as in Scheme 2. 3.3. Adsorption behaviors 3.3.1. Effect of pH The effect of pH (pH 6) on the adsorption amount (0.1 g of HPANF in 50 ml of 100 mg/l Cu(NO3 )2 ·3H2 O) is shown in Fig. 3. The amount of copper adsorbed onto the HPANF at adsorption equilibrium increased with increasing solution pH values. At pH less than 3, the adsorption amount was very small (less than 2.8 mg/g), but the adsorption amount increased sharply for pH from 3.0 to 4.0 and reached
268
S. Deng et al. / Journal of Colloid and Interface Science 260 (2003) 265–272
Fig. 3. Effect of pH on the adsorption of copper ions onto HPANF.
Fig. 5. Adsorption isotherm of copper ions on PANF and HPANF (pH 4.5).
3.3.2. Effect of Cu2+ concentration As shown in Fig. 5, the amount of copper adsorbed at adsorption equilibrium on the HPANF was significantly more than that on the PANF. When the copper ion concentration was 132 mg/l, the copper amount adsorbed by the HPANF reached 27.95 mg/g, as compared to only 1.3 mg/g by the PANF. The adsorption amount (qe ) of copper ion by HPANF can conform to the Langmuir equation as (R 2 = 0.973) qe =
Fig. 4. Zeta potentials of HPANF at different solution pH values.
up to 12.4 mg/g. However, no further increase of adsorption was observed for pH increasing from 4 to 6. The zeta potentials of the HPANF at different pH values are shown in Fig. 4. It can be seen that the point of zero ζ potential of the fiber is at pH about 5.3. When the solution pH values are below 5.3, the zeta potentials of the HPANF are positive and therefore the interaction between copper ions and the surface of the HPANF is electrostatically repulsive. Decreasing solution pH increases the repulsive force between copper ions and the surface of the HPANF and thus hinders or inhibits the transport of copper ions from the bulk solution to the surface of the HPANF for adsorption to take place. This would explain the decrease of the adsorption capacities with decreasing solution pH at pH below 5, as observed in Fig. 3. On the other hand, the adsorption of copper ions onto the surface of the HPANF will depend on the number of sites available for copper attachment on the surface of the HPANF, possibly through complexation with the nitrogen atoms (see the discussion in the section of adsorption mechanisms). The adsorption capacity will not increase further if all the adsorption sites have been taken up by copper ions. This would explain the observed maximum and constant adsorption capacity for the copper ions at pH 4–6; see Fig. 3.
qm Ce , 1/b + Ce
(1)
where qm is the maximum amount of adsorption (mg/l), b is the adsorption equilibrium constant (l/mg), and Ce is the equilibrium concentration of the copper ions in the solution (mg/l). The solid lines in Fig. 5 show the predictions based on Eq. (1). The maximum adsorption amount (qm ) of copper ions by the HPANF is predicted as 29.64 mg/g in this case. However, the Freundlich equation was found not to be able to fit the experimental adsorption data well (results not shown). 3.4. Adsorption kinetics Adsorption kinetics was studied by adding 0.1 g of the HPANF to 50 ml of Cu(NO3 )2 ·3H2 O solution with a copper concentration of 132 mg/l (pH 4.5) in a flask. As shown in Fig. 6, the adsorption of copper ions onto the HPANF was very fast, and the adsorption equilibrium could be reached in about 10–20 min. This phenomenon may be attributed to the large surface areas of and possibly the high density of adsorption sites on the HPANF. Therefore, copper adsorption can be controlled mainly by the diffusion of copper ions from the bulk solution onto the surface of the HPANF. Adsorption of metal ions from aqueous solution onto many conventional granular adsorbents, such as activated carbon and resin, has been reported to be a long process (adsorption equilibrium reached in several hours or more) because the kinetics of metal ion adsorption onto those porous adsorbents is mainly controlled by an innersurface diffusion process which is generally much slower than the bulk diffusion of metal ions in solutions [21]. Other
S. Deng et al. / Journal of Colloid and Interface Science 260 (2003) 265–272
269
Fig. 6. Kinetics of copper ion adsorption onto HPANF (pH 4.5). (a)
Fig. 7. Rate of Cu desorption from HPANF in DI water (pH 5.0) and 1 mol/l HCl solution. (b)
researchers also reported that modified polyacrylonitrile and polyethylene terephthalate fibers were faster adsorbents for metal ions from aqueous solutions [1,7]. 3.5. Desorption of copper from HPANF using DI water and hydrochloric acid solution Desorption of the copper ions from the surface of HPANF in DI water and in 1 mol/l HCl solution, respectively, was investigated. As shown in Fig. 7, 3.6% of the copper ions may be desorbed in DI water in the first 10 min, but the desorption did not increase further after 10 min. In the acid solution, up to 89% of the adsorbed copper ions may be desorbed within 2 min, and the percentage of desorption can reach 98% after 30 min. Hydrochloric acid desorption therefore is much more efficient than DI water. 3.6. Surface morphology of the HPANF with adsorbed copper ions The surface morphologies of HPANF before and after copper adsorption were studied through AFM. The tridimensional AFM images of the HPANF with or without copper ion adsorption are shown in Fig. 8. The average roughness (Ra ) of the surfaces can be calculated from the roughness
Fig. 8. Tridimensional AFM images of (a) HPANF and (b) HPANF with copper adsorption.
profile determined from the AFM images. For the HPANF, the value of Ra is found to be 0.363 nm. After the adsorption of copper ions onto the HPANF, the value of Ra becomes 3.763 nm, which indicates that the surfaces of the HPANF became much rougher after copper adsorption. The increase of roughness may be attributed to the formation of copper ion complexes on the surfaces. 3.7. FTIR results The HPANF with adsorbed copper ions at an adsorbed content of 29.32 mg/g was dried at 50 ◦ C overnight and then analyzed by FTIR. The spectra in Fig. 9 show that, after the HPANF was coordinated with Cu2+ , the peak at 1573 cm−1 was slightly shifted to 1572 cm−1 and the band intensities at 1573 and 1406 cm−1 were greatly decreased. These changes can be attributed to the adsorption of copper ions onto the imine (–C=N–) groups on the surface of the HPANF. Since the C≡N bonds at the wavenumber 2246 cm−1 were not changed, the nitrile groups in the HPANF were therefore not involved in the adsorption of copper ions.
270
S. Deng et al. / Journal of Colloid and Interface Science 260 (2003) 265–272
(a)
Fig. 9. FTIR spectra of (A) HPANF and (B) HPANF with copper adsorption.
(b) Fig. 11. N(1s) core-level spectra of HPANF: (a) before copper adsorption and (b) after copper adsorption (pH 4.5).
(a)
(b) Fig. 10. Wide scan spectra of the HPANF (a) before copper adsorption and (b) after copper adsorption (pH 4.5).
3.8. XPS study XPS has increasingly been used in studying the interactions between metal compounds and membranes [22], as well as between adsorbates and adsorbents [14–17]. Figure 10a shows a wide-scan XPS spectrum of the HPANF surface, and Fig. 10b, a typical spectrum of the HPANF with copper adsorption. The peaks in Fig. 10a at binding energies
of 283.8, 398.6, and 530.6 eV are for the C1s, N1s and O1s, respectively. After copper adsorption, a new peak at binding energy 932.8 eV for Cu2p/3/2 is observed in Fig. 10b. Since the nitrogen-containing groups are suggested to be responsible for the adsorption of copper ions from the FTIR results, the detailed N1s core-level spectra from XPS analysis are examined in Fig. 11. In consideration of the fact that the binding energies of imine (398.4–399.8 eV) and nitrile (398.7–399.4 eV) groups are in the same region [23], a single peak centered at 398.6 eV is employed to fit the curve for the HPANF in Fig. 11a. A new peak at a binding energy 400 eV is, however, observed for the HPANF with adsorbed copper ions in Fig. 11b. The results indicate that copper ions were incorporated with the nitrogen atoms onto the surface of the HPANF and drew electrons from the imine sites toward the copper ions, which increased the binding energy of the N1s. Due to copper nitrate being used as absorbate in this study, the nitrate ion may be adsorbed onto the surface of the HPANF and affect the N1s spectrum. XPS analysis, however, did not show the peak of N1s at around 407 eV which is characteristic of the N1s of nitrate anion [24]. It is found that the oxidation state of copper ions was altered after Cu(II) was adsorbed on the surface of the HPANF. Fig. 12 shows the Cu(2p 3/2) core-level spectrum of the HPANF with adsorbed copper ions at pH 4.5. The binding energy at 932.8 eV is attributed to Cu(I) + Cu(0) [18], and no binding energy of Cu(II) (935.1 eV) is observed in
S. Deng et al. / Journal of Colloid and Interface Science 260 (2003) 265–272
271
hydrogen may be released and a series of pyridine rings could be generated [23]. When polyacrylonitrile is hydrolyzed by sodium hydroxide, many different complex products can be produced at different stages of the hydrolysis [11–13]. As the hydrolysis reaction in our study was restricted to 20 min and the PANF changed to a red-brown color (due to cyclization), the production of conjugated sequences (–C=N–) was expected [13].
4. Conclusions Fig. 12. Cu(2p3/2 ) core-level spectrum of HPANF after adsorption of copper ions at pH 4.5.
Scheme 3. Adsorption of copper ions onto HPANF.
the spectrum. This result indicates that some chemical bonds were formed between copper ions and the imine groups on the surface of the HPANF and the electron cloud density of the copper ions was increased, resulting in the observed lower binding energy for the adsorbed copper ions. 3.9. Adsorption mechanisms The uptakes of metal ionic species on an adsorbent can be governed by electrostatic interaction, and specific interactions. During the adsorption processes, the following aspects need to be considered: (1) the characteristics of the adsorbent which has many functional groups attributable to adsorption, and (2) the speciation diagram of the adsorbate. Both aspects are related to pH. Therefore, it is necessary to know about the speciation diagram of the adsorbate and the surface nature of the adsorbent at different pH values in order to understand the adsorption mechanism. The hydrolysis of Cu(II) to form Cu2 (OH)2 2+ and mononuclear species Cu(OH)+ , Cu(OH)2 , and Cu(OH)3− occurs in dilute solution with increasing pH from 8 to 12, with Cu(OH)4 2− formed in the more alkaline solutions. In this study, the adsorption tests were conducted at pH below 5.0, and thus copper existed mainly in the form of Cu2+ . Formation of chemical bonds during the adsorption of copper ions onto the HPANF is suggested by the FTIR and XPS results, and this adsorption mechanism may be expressed as in Scheme 3 [24,25]. In the literature, the adsorption of transition metal ions onto activated carbon derived from polyacrylonitrile has been attributed to the nitrogen in the pyridine functional group [24]. During pyrolysis of polyacrylonitrile, cyclization by polymerization of the cyano group took place and conjugated imine was formed. With further heat treatment,
The surface of polyacrylonitrile fiber (PANF) was modified through a simple hydrolysis process in a sodium hydroxide solution. The hydrolyzed polyacrylonitrile fiber (HPANF) possessed conjugated sequences of –C=N– on the surface and was much more effective in removing copper ions from aqueous solution than the PANF that contained the –C≡N groups. The amount of copper ion adsorption onto the HPANF increased with solution pH values in the pH range of 2–6 studied. This adsorption behavior could be attributed to the effect of electrostatic interaction between the copper ions and the HPANF, as ζ potential measurements indicated that the electrostatic interaction became less repulsive or more attractive at a higher solution pH value. FTIR and XPS analyses showed that copper ion adsorption onto the HPANF changed the spectra of the nitrogen atoms in the imine groups (–C=N–) of the HPANF, suggesting that the removal of copper ions from the solution by the HPANF was at least partly effected through the formation of chemical bonds with the nitrogen atoms in the –C=N– groups on the surface of the HPANF. It is seen that surface modification is a useful tool for improving the performance of an adsorbent, and HPANF has potential applications in wastewater treatment for heavy metal ion removal.
Acknowledgment A postdoctoral fellowship from A*Star of Singapore to Dr. S.B. Deng is appreciated.
References [1] [2] [3] [4] [5] [6] [7]
R.X. Liu, B.W. Zhang, H.X. Tang, J. Appl. Polym. Sci. 70 (1998) 7. R.V.S. Alfaya, Y. Gushikem, J. Colloid Interface Sci. 213 (1999) 438. S.K. Karin, S. Strnad, V. Ribitsch, Polym. Eng. Sci. 39 (1999) 1412. R.X. Liu, B.W. Zhang, H.X. Tang, React. Funct. Polym. 39 (1999) 71. P. Li, G.H. Xiu, L. Jiang, Sep. Sci. Technol. 36 (2001) 2147. M. Celik, M. Sacak, J. Appl. Polym. Sci. 59 (1996) 609. R. Coskun, M. Yigitoglu, M. Sacak, J. Appl. Polym. Sci. 75 (2000) 766. [8] B.W. Zhang, K. Fischer, D. Bieniek, A. Kettrup, React. Polym. 24 (1994) 49. [9] F. Vernon, T. Shah, React. Polym. 1 (1983) 301. [10] B.K. Liya, V.K. Yaroslav, I.R. Alexandr, D.L. Arkady, A.P. Nicolai, Macromolecules 34 (2001) 5607.
272
S. Deng et al. / Journal of Colloid and Interface Science 260 (2003) 265–272
[11] A.D. Litmanovich, N.A. Plate, Macromol. Chem. Phys. 201 (2000) 2176. [12] I.V. Ermakov, A.I. Rebrov, A.D. Litmanovich, Macromol. Chem. Phys. 201 (2000) 1415. [13] G.P. Karpacheva, L.M. Zemtsov, G.N. Bondarenko, Polym. Sci. Ser. A 42 (2000) 620. [14] R.B. Bai, X. Zhang, J. Colloid Interface Sci. 243 (2001) 52. [15] X. Zhang, R.B. Bai, Langmuir 18 (2002) 3459. [16] X. Zhang, R.B. Bai, J. Mater. Chem. 12 (2002) 2733. [17] J. Li, R.B. Bai, Langmuir 18 (2002) 9765. [18] L. Dambies, C. Guimon, S. Yiacoumi, E. Guibal, Colloids Surf. A 177 (2001) 203.
[19] J.R. Orlando, E. Marie, D.N. Ronald, M.C. Per, Langmuir 18 (2002) 1604. [20] Y. Ikezawa, T. Sawatari, H. Terashima, Electrochim. Acta 46 (2001) 1333. [21] J.M. Sanchez, M. Hidalgo, V. Salvado, React. Funct. Polym. 46 (2001) 283. [22] D.R. Lide (Ed.), CRC Handbook of Chemistry and Physics, CRC Press, Boca Raton, FL, 1998. [23] J.R. Pels, F. Kapteijn, J.A. Moulijn, Q. Zhu, K.M. Thomas, Carbon 33 (1995) 1641. [24] M.H. Liu, G. Xu, Y.H. Liu, Q. Chen, Langmuir 17 (2001) 427. [25] Y.F. Jia, B. Xiao, K.M. Thomas, Langmuir 18 (2002) 470.
Behaviors and mechanisms of copper adsorption on hydrolyzed polyacrylonitrile fibers Shubo Deng, Renbi Bai,∗ and J.P. Chen Department of Chemical and Environmental Engineering, National University of Singapore, 10 Kent Ridge Crescent, 119260 Singapore Received 10 August 2002; accepted 20 December 2002
Abstract Polyacrylonitrile fiber (PANF) was hydrolyzed in a solution of sodium hydroxide and the hydrolyzed polyacrylonitrile fiber (HPANF) was used as an adsorbent to remove copper ions from aqueous solution. Scanning electron microscopy (SEM) showed that the hydrolysis process made the surface of HPANF rougher than that of PANF. Fourier transform infrared (FTIR) spectroscopy revealed that the HPANF contained conjugated imine (–C=N–) sequences. Batch adsorption results indicated that the HPANF was very effective in adsorbing copper, and the adsorption equilibrium could be reached within 10–20 min. Atomic force microscopy (AFM) showed that some aggregates formed on the surface of the HPANF after copper ion adsorption and the average surface roughness (Ra ) value of the HPANF changed from 0.363 to 3.763 nm due to copper adsorption. FTIR analysis indicated that copper adsorption caused a decrease of the light adsorption intensity of the imine (–C=N–) groups at 1573 and 1406 cm−1 wavenumbers, and X-ray photoelectron spectroscopy (XPS) showed that the binding energy (BE) of some of the nitrogen atoms in the HPANF increased to a greater value due to copper adsorption. The FTIR and XPS results suggest that the adsorption of copper ions to the HPANF is attributed to the imine groups on the surface of the HPANF. 2003 Elsevier Science (USA). All rights reserved. Keywords: Hydrolyzed polyacrylonitrile fiber (HPANF); Copper adsorption; FTIR and XPS spectra; SEM and AFM images
1. Introduction Fibers, which have large specific surface areas and high adsorption rates, have attracted great attention in heavy metal removal from wastewater in recent years. There are many types of fibers or modified fibers, including activated carbon fiber, cotton fiber, glass fiber, and polymer fiber (polyester, polypropylene, polyethylene terephthalate, polystyrene), that have been used to adsorb metal ions or to enrich trace amount of metals from aqueous solutions [1–7]. The adsorption of metal ions from aqueous solutions to adsorbents is usually controlled by the properties of the surface functional groups of the adsorbents. Particularly, modified fibers that contain carboxyl, tetrazine, amidoxime– hydroxam, imidazoline, amino, and phosphoric groups have been found to be effective in metal ion adsorption [1,7–9]. An example is polyacrylonitrile fiber (PANF), a common and cheap commercial product, that has been modified to introduce various types of chelating groups and been used * Corresponding author.
E-mail address: [email protected] (R. Bai).
as an adsorbent for metal ions removal [8]. The modification of PANF may be done through hydrolysis in an acid or a base [10,11]. In general, the hydrolysis of PANF can be a complex process and different types of hydrolyzed products with different capabilities for metal adsorption may be obtained. From alkaline hydrolysis of PANF, the products may include those with conjugated sequences of –C=N–, acrylamide, sodium acrylate, and amidine from different stages of the hydrolysis [11–13]. Although many adsorbents were used to adsorb metal ions from aqueous solutions, the mechanisms of metal ion adsorption are not always fully understood. X-ray photoelectron spectroscopy (XPS) has found increasing applications in studying the interactions between an adsorbate and the active groups on the surface of an adsorbent in the adsorption process [14–17]. This is due to the fact that the creation of chemical bonds during adsorption would change the distribution of the electrons around the corresponding atoms, and, as a consequence, the binding energy of the electrons could be shifted to a lower or higher level, which would provide information on the adsorption mechanisms. Alternatively, Fourier transform infrared (FTIR) spectroscopy may
0021-9797/03/$ – see front matter 2003 Elsevier Science (USA). All rights reserved. doi:10.1016/S0021-9797(02)00243-6
266
S. Deng et al. / Journal of Colloid and Interface Science 260 (2003) 265–272
be used in the study of adsorption mechanisms, through examining the changes of adsorption to the light energy when a material is exposed to radiation in the infrared portion of the electromagnetic spectrum at certain frequencies since each absorption band in an FTIR spectrum is caused by different chemical bonds [4,17]. In addition, scanning electron microscopy (SEM) and atomic force microscopy (AFM) have often been used to visually show the morphological changes of the adsorbents after adsorption with an adsorbate [14–19]. One of the major advantages of AFM as compared to SEM is that the images can be shown in three dimensions and at nanometer scales. In this study, PANF was hydrolyzed in a sodium hydroxide solution and the performance of the hydrolyzed polyacrylonitrile fiber (HPANF) in copper ion adsorption was studied. The mechanisms of surface interactions were investigated through FTIR and XPS analyses, and the surface morphologies of the HPANF before and after adsorption were examined through SEM and AFM images.
2.4. FTIR spectroscopy The samples of PANF, HPANF, and copper-adsorbed HPANF were cut into about 1-mm pieces, blended with KBr, and pressed into discs. The spectra were recorded on a BioRad FTIR Model FTS135 spectrophotometer under ambient conditions. 2.5. X-ray photoelectron spectroscopy
2. Experimental
XPS analysis of the HPANF before and after the adsorption of copper ions was carried out on an AXIS HIS spectrometer (Kratos Analytical Ltd., England) with an AlKα X-ray source (1486.71 eV of photons) to determine the elements such as C, N, and O atoms present. The X-ray source was run at a reduced power of 150 W, and the pressure in the analysis chamber was maintained at less than 10−8 Torr during each measurement. All binding energies were referenced to the neutral C(1s) peak at 284.71 eV. Surface elemental stoichiometries were determined from the sensitivity-factorcorrected peak-area ratios, and the software XPSpeak 4.1 was used to fit the XPS spectra peaks [14].
2.1. Materials
2.6. Adsorption and desorption experiments
Polyacrylonitrile fiber (PANF), a commercial product, was kindly provided by Daqing Polyacrylonitrile Corp., PR China. The PANF contained 90% polyacrylonitrile and 10% vinyl acetate and had a diameter of about 20–50 µm. The other chemicals used in the study were of reagent grade. 2.2. Hydrolysis of PANF The hydrolysis of PANF was carried out in a solution of sodium hydroxide, with the conditions following: 10 g of PANF was added to a 500-ml flask with 300 ml of ethanol/water (mole ratio of 0.22) solution containing 10% (w/w) NaOH; the solution was stirred at 75 ◦ C in a shaking water bath for 20 min; then the HPANF was collected, rinsed with deionized (DI) water until neutral, and then dried in an oven at 50 ◦ C to constant weight and stored in a desiccator prior to use in the study. 2.3. AFM and SEM analysis A scanning electron microscope (SEM, JEOL JSM-6400) at 10–20 kV was used to examine the surface morphologies of PANF before and after hydrolysis. The morphologies of HPANF surfaces before and after copper adsorption were studied with a Nanoscope IIIa atomic force microscope (AFM) from Digital Instruments Inc. using the tapping mode at a scan size of 500 nm. The drive frequency was 330 kHz, and the voltage was between 3.0 and 4.0 V. The drive amplitude and scan rate were 300 mV and 1.0 Hz, respectively.
Adsorption experiments were conducted as follows. To a 100-ml flask, 50 ml of Cu(NO3 )2 ·3H2 O solution (pH 4.5) with a specific copper concentration was added. Then 0.1 g of HPANF was added to the solution. The flask was shaken at 120 rpm in a water bath with a temperature of 25 ◦ C. Copper concentrations in the solution were determined at desired time intervals until adsorption equilibrium was established. For the desorption study, 0.1 g of HPANF was added to 50 ml of 500 mg/l Cu(NO3 )2 ·3H2 O solution with other conditions the same as in the adsorption experiments. After the adsorption equilibrium was reached, the fibers were separated by filtration and rinsed with DI water to remove residual solution trapped among the fibers. Then the HPANF with adsorbed copper was transferred to a flask with 50 ml of 1 mol/l HCl solution or DI water (pH 5.0). The flask was shaken at 150 rpm in a rotary shaker at 25 ◦ C and the copper ion concentrations in the solution were analyzed against time. An inductively coupled plasma emission spectroscope (ICP-ES) (Perkin–Elmer Optima 3000, USA) was used to determine the copper concentrations in the samples.
3. Results and discussion 3.1. SEM images As shown in Fig. 1, the surface of PANF before the hydrolysis was relatively smooth and uniform. After the hydrolysis, the fiber surface was corroded and became rough, indicating that the hydrolysis reaction occurred on the surface of the PANF.
S. Deng et al. / Journal of Colloid and Interface Science 260 (2003) 265–272
267
(a) Fig. 2. FTIR spectra of (a) PANF and (b) HPANF.
Scheme 1. Hydrolysis of PANF to HPANF.
(b) Fig. 1. SEM images showing the surface morphologies of (a) PANF and (b) HPANF before and after hydrolysis.
3.2. FTIR spectroscope analysis The spectra of PANF before and after hydrolysis is shown in Fig. 2. It is seen that the peaks of the PANF (Fig. 2a) can be assigned as follows: 3632 cm−1 (γ OH), 2946 and 2874 cm−1 (γ C–H asymmetric and symmetric in CH, CH2 , and CH3 groups), 2246 cm−1 (γ C≡N), 1739 cm−1 (γ C≡O), 1457 cm−1 (δ CH3 and δS CH2 ), 1363 cm−1 (δ CH3 symmetric in CCH3 ), 1172 cm−1 (γ C–N), 1074 cm−1 (δ C–N), and 538 cm−1 (δτ C=O), where γ represents a stretching vibration, δ a blending vibration, δS a scissor vibration, and δτ a twisting vibration. After the hydrolysis with sodium hydroxide, the spectrum of the HPANF (Fig. 2b) changed greatly. The broad adsorption band at 3632 cm−1 shifted to 3423 cm−1 , with the intensity of the peak increased significantly. This band corresponds to the stretching vibration of the OH group and indicates the introduction of the OH groups onto the HPANF. The new peaks at 1573, 1406, and 1224 cm−1 wavenumbers indicate the existence of imine (–C=N–) conjugated sequences in the modified fiber [13,20]. Since the peak at 2246 cm−1 for –C≡N did not change significantly, only part
Scheme 2. Reaction of the nitrile group in PANF during hydrolysis.
of the –C≡N groups in the PANF were therefore converted to the –C=N– groups in the HPANF. The intensity of the peak at 2946 cm−1 (γ CH2 ) was reduced greatly, attributed to the hydrolysis of the ester group in the PANF. The peaks at 1739 cm−1 (γ C=O) and 538 cm−1 (δτ C=O) became weaker after the hydrolysis, suggesting acetate ester being removed from the surface of PANF. Based on the FTIR results and other references [11–13], the reaction taking place during the hydrolysis may be proposed to be as in Scheme 1. The cyclization reaction of the nitrile groups in the hydrolysis may be given as in Scheme 2. 3.3. Adsorption behaviors 3.3.1. Effect of pH The effect of pH (pH 6) on the adsorption amount (0.1 g of HPANF in 50 ml of 100 mg/l Cu(NO3 )2 ·3H2 O) is shown in Fig. 3. The amount of copper adsorbed onto the HPANF at adsorption equilibrium increased with increasing solution pH values. At pH less than 3, the adsorption amount was very small (less than 2.8 mg/g), but the adsorption amount increased sharply for pH from 3.0 to 4.0 and reached
268
S. Deng et al. / Journal of Colloid and Interface Science 260 (2003) 265–272
Fig. 3. Effect of pH on the adsorption of copper ions onto HPANF.
Fig. 5. Adsorption isotherm of copper ions on PANF and HPANF (pH 4.5).
3.3.2. Effect of Cu2+ concentration As shown in Fig. 5, the amount of copper adsorbed at adsorption equilibrium on the HPANF was significantly more than that on the PANF. When the copper ion concentration was 132 mg/l, the copper amount adsorbed by the HPANF reached 27.95 mg/g, as compared to only 1.3 mg/g by the PANF. The adsorption amount (qe ) of copper ion by HPANF can conform to the Langmuir equation as (R 2 = 0.973) qe =
Fig. 4. Zeta potentials of HPANF at different solution pH values.
up to 12.4 mg/g. However, no further increase of adsorption was observed for pH increasing from 4 to 6. The zeta potentials of the HPANF at different pH values are shown in Fig. 4. It can be seen that the point of zero ζ potential of the fiber is at pH about 5.3. When the solution pH values are below 5.3, the zeta potentials of the HPANF are positive and therefore the interaction between copper ions and the surface of the HPANF is electrostatically repulsive. Decreasing solution pH increases the repulsive force between copper ions and the surface of the HPANF and thus hinders or inhibits the transport of copper ions from the bulk solution to the surface of the HPANF for adsorption to take place. This would explain the decrease of the adsorption capacities with decreasing solution pH at pH below 5, as observed in Fig. 3. On the other hand, the adsorption of copper ions onto the surface of the HPANF will depend on the number of sites available for copper attachment on the surface of the HPANF, possibly through complexation with the nitrogen atoms (see the discussion in the section of adsorption mechanisms). The adsorption capacity will not increase further if all the adsorption sites have been taken up by copper ions. This would explain the observed maximum and constant adsorption capacity for the copper ions at pH 4–6; see Fig. 3.
qm Ce , 1/b + Ce
(1)
where qm is the maximum amount of adsorption (mg/l), b is the adsorption equilibrium constant (l/mg), and Ce is the equilibrium concentration of the copper ions in the solution (mg/l). The solid lines in Fig. 5 show the predictions based on Eq. (1). The maximum adsorption amount (qm ) of copper ions by the HPANF is predicted as 29.64 mg/g in this case. However, the Freundlich equation was found not to be able to fit the experimental adsorption data well (results not shown). 3.4. Adsorption kinetics Adsorption kinetics was studied by adding 0.1 g of the HPANF to 50 ml of Cu(NO3 )2 ·3H2 O solution with a copper concentration of 132 mg/l (pH 4.5) in a flask. As shown in Fig. 6, the adsorption of copper ions onto the HPANF was very fast, and the adsorption equilibrium could be reached in about 10–20 min. This phenomenon may be attributed to the large surface areas of and possibly the high density of adsorption sites on the HPANF. Therefore, copper adsorption can be controlled mainly by the diffusion of copper ions from the bulk solution onto the surface of the HPANF. Adsorption of metal ions from aqueous solution onto many conventional granular adsorbents, such as activated carbon and resin, has been reported to be a long process (adsorption equilibrium reached in several hours or more) because the kinetics of metal ion adsorption onto those porous adsorbents is mainly controlled by an innersurface diffusion process which is generally much slower than the bulk diffusion of metal ions in solutions [21]. Other
S. Deng et al. / Journal of Colloid and Interface Science 260 (2003) 265–272
269
Fig. 6. Kinetics of copper ion adsorption onto HPANF (pH 4.5). (a)
Fig. 7. Rate of Cu desorption from HPANF in DI water (pH 5.0) and 1 mol/l HCl solution. (b)
researchers also reported that modified polyacrylonitrile and polyethylene terephthalate fibers were faster adsorbents for metal ions from aqueous solutions [1,7]. 3.5. Desorption of copper from HPANF using DI water and hydrochloric acid solution Desorption of the copper ions from the surface of HPANF in DI water and in 1 mol/l HCl solution, respectively, was investigated. As shown in Fig. 7, 3.6% of the copper ions may be desorbed in DI water in the first 10 min, but the desorption did not increase further after 10 min. In the acid solution, up to 89% of the adsorbed copper ions may be desorbed within 2 min, and the percentage of desorption can reach 98% after 30 min. Hydrochloric acid desorption therefore is much more efficient than DI water. 3.6. Surface morphology of the HPANF with adsorbed copper ions The surface morphologies of HPANF before and after copper adsorption were studied through AFM. The tridimensional AFM images of the HPANF with or without copper ion adsorption are shown in Fig. 8. The average roughness (Ra ) of the surfaces can be calculated from the roughness
Fig. 8. Tridimensional AFM images of (a) HPANF and (b) HPANF with copper adsorption.
profile determined from the AFM images. For the HPANF, the value of Ra is found to be 0.363 nm. After the adsorption of copper ions onto the HPANF, the value of Ra becomes 3.763 nm, which indicates that the surfaces of the HPANF became much rougher after copper adsorption. The increase of roughness may be attributed to the formation of copper ion complexes on the surfaces. 3.7. FTIR results The HPANF with adsorbed copper ions at an adsorbed content of 29.32 mg/g was dried at 50 ◦ C overnight and then analyzed by FTIR. The spectra in Fig. 9 show that, after the HPANF was coordinated with Cu2+ , the peak at 1573 cm−1 was slightly shifted to 1572 cm−1 and the band intensities at 1573 and 1406 cm−1 were greatly decreased. These changes can be attributed to the adsorption of copper ions onto the imine (–C=N–) groups on the surface of the HPANF. Since the C≡N bonds at the wavenumber 2246 cm−1 were not changed, the nitrile groups in the HPANF were therefore not involved in the adsorption of copper ions.
270
S. Deng et al. / Journal of Colloid and Interface Science 260 (2003) 265–272
(a)
Fig. 9. FTIR spectra of (A) HPANF and (B) HPANF with copper adsorption.
(b) Fig. 11. N(1s) core-level spectra of HPANF: (a) before copper adsorption and (b) after copper adsorption (pH 4.5).
(a)
(b) Fig. 10. Wide scan spectra of the HPANF (a) before copper adsorption and (b) after copper adsorption (pH 4.5).
3.8. XPS study XPS has increasingly been used in studying the interactions between metal compounds and membranes [22], as well as between adsorbates and adsorbents [14–17]. Figure 10a shows a wide-scan XPS spectrum of the HPANF surface, and Fig. 10b, a typical spectrum of the HPANF with copper adsorption. The peaks in Fig. 10a at binding energies
of 283.8, 398.6, and 530.6 eV are for the C1s, N1s and O1s, respectively. After copper adsorption, a new peak at binding energy 932.8 eV for Cu2p/3/2 is observed in Fig. 10b. Since the nitrogen-containing groups are suggested to be responsible for the adsorption of copper ions from the FTIR results, the detailed N1s core-level spectra from XPS analysis are examined in Fig. 11. In consideration of the fact that the binding energies of imine (398.4–399.8 eV) and nitrile (398.7–399.4 eV) groups are in the same region [23], a single peak centered at 398.6 eV is employed to fit the curve for the HPANF in Fig. 11a. A new peak at a binding energy 400 eV is, however, observed for the HPANF with adsorbed copper ions in Fig. 11b. The results indicate that copper ions were incorporated with the nitrogen atoms onto the surface of the HPANF and drew electrons from the imine sites toward the copper ions, which increased the binding energy of the N1s. Due to copper nitrate being used as absorbate in this study, the nitrate ion may be adsorbed onto the surface of the HPANF and affect the N1s spectrum. XPS analysis, however, did not show the peak of N1s at around 407 eV which is characteristic of the N1s of nitrate anion [24]. It is found that the oxidation state of copper ions was altered after Cu(II) was adsorbed on the surface of the HPANF. Fig. 12 shows the Cu(2p 3/2) core-level spectrum of the HPANF with adsorbed copper ions at pH 4.5. The binding energy at 932.8 eV is attributed to Cu(I) + Cu(0) [18], and no binding energy of Cu(II) (935.1 eV) is observed in
S. Deng et al. / Journal of Colloid and Interface Science 260 (2003) 265–272
271
hydrogen may be released and a series of pyridine rings could be generated [23]. When polyacrylonitrile is hydrolyzed by sodium hydroxide, many different complex products can be produced at different stages of the hydrolysis [11–13]. As the hydrolysis reaction in our study was restricted to 20 min and the PANF changed to a red-brown color (due to cyclization), the production of conjugated sequences (–C=N–) was expected [13].
4. Conclusions Fig. 12. Cu(2p3/2 ) core-level spectrum of HPANF after adsorption of copper ions at pH 4.5.
Scheme 3. Adsorption of copper ions onto HPANF.
the spectrum. This result indicates that some chemical bonds were formed between copper ions and the imine groups on the surface of the HPANF and the electron cloud density of the copper ions was increased, resulting in the observed lower binding energy for the adsorbed copper ions. 3.9. Adsorption mechanisms The uptakes of metal ionic species on an adsorbent can be governed by electrostatic interaction, and specific interactions. During the adsorption processes, the following aspects need to be considered: (1) the characteristics of the adsorbent which has many functional groups attributable to adsorption, and (2) the speciation diagram of the adsorbate. Both aspects are related to pH. Therefore, it is necessary to know about the speciation diagram of the adsorbate and the surface nature of the adsorbent at different pH values in order to understand the adsorption mechanism. The hydrolysis of Cu(II) to form Cu2 (OH)2 2+ and mononuclear species Cu(OH)+ , Cu(OH)2 , and Cu(OH)3− occurs in dilute solution with increasing pH from 8 to 12, with Cu(OH)4 2− formed in the more alkaline solutions. In this study, the adsorption tests were conducted at pH below 5.0, and thus copper existed mainly in the form of Cu2+ . Formation of chemical bonds during the adsorption of copper ions onto the HPANF is suggested by the FTIR and XPS results, and this adsorption mechanism may be expressed as in Scheme 3 [24,25]. In the literature, the adsorption of transition metal ions onto activated carbon derived from polyacrylonitrile has been attributed to the nitrogen in the pyridine functional group [24]. During pyrolysis of polyacrylonitrile, cyclization by polymerization of the cyano group took place and conjugated imine was formed. With further heat treatment,
The surface of polyacrylonitrile fiber (PANF) was modified through a simple hydrolysis process in a sodium hydroxide solution. The hydrolyzed polyacrylonitrile fiber (HPANF) possessed conjugated sequences of –C=N– on the surface and was much more effective in removing copper ions from aqueous solution than the PANF that contained the –C≡N groups. The amount of copper ion adsorption onto the HPANF increased with solution pH values in the pH range of 2–6 studied. This adsorption behavior could be attributed to the effect of electrostatic interaction between the copper ions and the HPANF, as ζ potential measurements indicated that the electrostatic interaction became less repulsive or more attractive at a higher solution pH value. FTIR and XPS analyses showed that copper ion adsorption onto the HPANF changed the spectra of the nitrogen atoms in the imine groups (–C=N–) of the HPANF, suggesting that the removal of copper ions from the solution by the HPANF was at least partly effected through the formation of chemical bonds with the nitrogen atoms in the –C=N– groups on the surface of the HPANF. It is seen that surface modification is a useful tool for improving the performance of an adsorbent, and HPANF has potential applications in wastewater treatment for heavy metal ion removal.
Acknowledgment A postdoctoral fellowship from A*Star of Singapore to Dr. S.B. Deng is appreciated.
References [1] [2] [3] [4] [5] [6] [7]
R.X. Liu, B.W. Zhang, H.X. Tang, J. Appl. Polym. Sci. 70 (1998) 7. R.V.S. Alfaya, Y. Gushikem, J. Colloid Interface Sci. 213 (1999) 438. S.K. Karin, S. Strnad, V. Ribitsch, Polym. Eng. Sci. 39 (1999) 1412. R.X. Liu, B.W. Zhang, H.X. Tang, React. Funct. Polym. 39 (1999) 71. P. Li, G.H. Xiu, L. Jiang, Sep. Sci. Technol. 36 (2001) 2147. M. Celik, M. Sacak, J. Appl. Polym. Sci. 59 (1996) 609. R. Coskun, M. Yigitoglu, M. Sacak, J. Appl. Polym. Sci. 75 (2000) 766. [8] B.W. Zhang, K. Fischer, D. Bieniek, A. Kettrup, React. Polym. 24 (1994) 49. [9] F. Vernon, T. Shah, React. Polym. 1 (1983) 301. [10] B.K. Liya, V.K. Yaroslav, I.R. Alexandr, D.L. Arkady, A.P. Nicolai, Macromolecules 34 (2001) 5607.
272
S. Deng et al. / Journal of Colloid and Interface Science 260 (2003) 265–272
[11] A.D. Litmanovich, N.A. Plate, Macromol. Chem. Phys. 201 (2000) 2176. [12] I.V. Ermakov, A.I. Rebrov, A.D. Litmanovich, Macromol. Chem. Phys. 201 (2000) 1415. [13] G.P. Karpacheva, L.M. Zemtsov, G.N. Bondarenko, Polym. Sci. Ser. A 42 (2000) 620. [14] R.B. Bai, X. Zhang, J. Colloid Interface Sci. 243 (2001) 52. [15] X. Zhang, R.B. Bai, Langmuir 18 (2002) 3459. [16] X. Zhang, R.B. Bai, J. Mater. Chem. 12 (2002) 2733. [17] J. Li, R.B. Bai, Langmuir 18 (2002) 9765. [18] L. Dambies, C. Guimon, S. Yiacoumi, E. Guibal, Colloids Surf. A 177 (2001) 203.
[19] J.R. Orlando, E. Marie, D.N. Ronald, M.C. Per, Langmuir 18 (2002) 1604. [20] Y. Ikezawa, T. Sawatari, H. Terashima, Electrochim. Acta 46 (2001) 1333. [21] J.M. Sanchez, M. Hidalgo, V. Salvado, React. Funct. Polym. 46 (2001) 283. [22] D.R. Lide (Ed.), CRC Handbook of Chemistry and Physics, CRC Press, Boca Raton, FL, 1998. [23] J.R. Pels, F. Kapteijn, J.A. Moulijn, Q. Zhu, K.M. Thomas, Carbon 33 (1995) 1641. [24] M.H. Liu, G. Xu, Y.H. Liu, Q. Chen, Langmuir 17 (2001) 427. [25] Y.F. Jia, B. Xiao, K.M. Thomas, Langmuir 18 (2002) 470.