Removal of phenylamine and catechol by adsorptive

Colloids and Surfaces A: Physicochem. Eng. Aspects 279 (2006) 28–33

Removal of phenylamine and catechol by adsorptive micellar flocculation Thiago D’Orsi Almeida a,b,1 , Federico I. Talens-Alesson a,∗ a

Scheme, University of Nottingham, UK b Mott MacDonald, Cambridge, UK

Received 23 July 2005; received in revised form 15 November 2005; accepted 16 December 2005 Available online 23 January 2006

Abstract The surfactant/water partition ratios for phenylamine in the flocculates formed during adsorptive micellar flocculation (AMF) are found to be around four to five times higher than the ratios reported in the literature for micelle enhanced ultra filtration. Binding ratios phenylamine/surfactant may be as high a 0.25. This suggests that highly soluble organic compounds present in acidic form in aqueous solution are strong candidates for removal by adsorptive micellar flocculation. At concentrations of phenylamine 0.0128 M or higher, micelles may not aggregate into large easily filtered flocs if the concentration of Al3+ is less than 0.024 M, but into small colloidal aggregates in the size range of hundreds of nanometers. It is still possible to remove significant amounts of phenylamine and surfactant by a sequence of filtration of coarse solids and filtration through 200 nm pore membranes. The high efficiency of the smaller, less hydrophobic particles in capturing phenylamine, with binding ratios phenylamine/SDS of 0.35 and higher, strongly supports the view that hydrophobicity cannot be taken as the main reason why micellar flocculates capture contaminants. Catechol shows binding ratios up to 0.15, well above binding ratios up to 0.1 for phenol. That, being other properties similar, strong complexant catechol adsorbs much better than phenol, reinforces the view that superficial complexation with Al3+ is a key mechanism in adsorptive micellar flocculation. © 2005 Elsevier B.V. All rights reserved. Keywords: Adsorptive micellar flocculation; Phenylamine; Catechol; Phenol; Laurylsulphate; Water treatment

1. Introduction For the last 25 years research on the potential of surfactant-mediated separation methods for water treatment has taken place. Topics investigated have included hemimicellar/admicellar systems (adsolubilisation) and micellar solutions, the latter in combination with membrane filtration apparatuses (micelle enhanced ultra filtration, MEUF) and phase change devices like changes in temperature to change the hydrophylliclipophillic balance of surfactants (cloud point separation). One such technique is adsorptive micellar flocculation. This technique is based on the flocculation of micelles of some anionic surfactants (laurylsulphate and ␣-olefinsulphonate) in the presence of Al3+ [1,2]. The flocculation process allows the adsorption onto the flocculate of complexes between the flocculant cation and organic species present in solution. This gives the ∗

Correspondence to: Talenco Consulting, Psge Canti 8, 2-2, 08005 Barcelona, Spain. E-mail address: [email protected] (F.I. Talens-Alesson). 1 Present address: Mott MacDonald, Dememter House, Station Road, Cambridge CB1 2RS, UK. 0927-7757/$ – see front matter © 2005 Elsevier B.V. All rights reserved. doi:10.1016/j.colsurfa.2005.12.029

technique some similarities with precipitation–coagulation techniques, in which insoluble compounds between organic species and either Al3+ or Fe3+ form and their particles are subsequently aggregated by additional amounts of coagulant [3,4], or such compounds are deposited on the surface of solid particles already available [5] or the organic compounds are adsorbed on nascent aluminum hydroxide [6]. 1.1. Interactions organic compound—flocculant cation Because the solution is moderately acidic due to the presence of Al3+ with pH usually between 3 and 3.5, most compounds investigated so far may be present in anionic form like pesticide 2,4-D (pKa 2.85) or in neutral form like phenol (pKa 9.89). However, even in the case of compounds present in the solution as non-dissociated like phenol complexation may take place [7,8]. High local concentration of cations must be expected around the micelles during the flocculation: molar ratios bound Al3+ to micellar surfactant may be as high as 0.3 [1,9,10]. The results of complex formation under pH conditions similar to those in AMF indicate that at high Al3+ concentrations the stoichiometries are aberrant (e.g. Benzoic:Al3+ 1:9) [3,7]. A

T.D. Almeida, F.I. Talens-Alesson / Colloids and Surfaces A: Physicochem. Eng. Aspects 279 (2006) 28–33

29

suitable explanation is that these are stoichiometries of aquacomplexes including water molecules and hydroxyl groups. There is evidence of the adsorption of aqua-complexes of Al3+ [11] and Fe3+ [12] on micellar flocculates. A mechanism of adsorption of organic pollutants onto flocculates due to complex formation is consistent with a number of observations on the removal of 2,4-dichlorophenoxyacetic acid by AMF: it follows a Temkin-Frumkin isotherm and if the pollutant is mixed with the surfactant before the flocculant is added the proportion of surfactant flocculated is lower [13]. Association of dissociated organic acids and anionic micelles must increase the negative charge of the micelle, as they tend to associated to micelles without incorporating into the hydrophobic core, as described in the following section. This increases the available negative charge, which must be neutralised, and causes a less effective flocculation.

bound, but also mechanically bound within the hydrodynamic boundary layer around the micelles, as an explanation for the discrepancies between their predictions of micelle binding ratios for Na+ and binding ratios derived by Rathman and Scamehorn from selective electrode measures. In this paper we discuss the adsorption of two model organic pollutants, phenylamine and catechol, on micellar flocculates. Phenylamine has been chosen to represent compounds which are present themselves in a cationic form under the conditions of adsorptive micellar flocculation. Phenylamine is also known to form at best highly unstable complexes with Al3+ , although it is reported that such complexes play a role on mineral surface adsorption, for example on gibbsite [19]. Catechol is a well know chelating agent, although its abilities are low in the range of pH below 3.

1.2. Interactions organic compound—SDS micelle

2. Experimental procedures

Organic compounds associate to SDS micelles in a variety of forms. Non-polar compounds like toluene dissolving within the core of the micelle, whereas compounds like phenol remain close to the aqueous phase [14]. They may be present as phenoxide, completely submersed in the hydrophobic phase but close to its surface, or as phenol, with the polar head into the aqueous phase. There is a moderate change in the UV absorption spectrum for phenol in water or in SDS solutions, which confirms that the hydrophilicity of the environment in which phenol is present differs to some extent from water [15]. Other compounds, like benzoic acid have even smaller changes in their UV absorption spectrum, which suggests that the environment around the acid is mostly water [15]. Phenylammonium behaves similarly to phenol, solubilising with the aromatic ring within the organic part of the micelle, but in this case the ammonium fragment is completely surrounded by water [16]. Phenylammonium has the advantage of electrostatic attraction with the polar heads of the anionic surfactant.

2.1. Adsorptive micellar flocculation

1.3. Interactions anionic micelle—Al3+ The surface of an anionic micelle generates an electrical field, which segregates the ions present in the solution around it. This creates three regions around the micelle: a Stern layer, where anionic species are forbidden, and which is usually considered the region where “bound” cations are; a diffuse layer across which the concentration of cations drops from that at the Stern layer to the concentration in the bulk phase, and the concentration of anions raises from zero to the concentration in the bulk phase; and the bulk phase, with an homogeneous concentration. Assuming micelle-bound Al3+ to be within a distance 10–20% of the micellar radius of a 10 nm micelle, [Al3+ ] may well be in the order of magnitude of 10 M [1,8,9]. This is consistent with calculations using triple-planar models for electrostatic binding of Na+ to SDS micelles [17], which suggest concentrations within the Stern layer at least 8 M (against predictions of 16 M using earlier double-planar models [18]) and is sufficiently high to allow for abnormal complexation ratios. Chen and Jafvert [17] also suggest that “bound” may not only mean electrostatically

The procedure is relatively similar to a jar test, although constant agitation is not provided. Micellar flocculation has been found to follow the Schulze Hardy rule in some cases [20], and have a clearly defined boundaries for the concentration region where slow flocculation of small particles leading to turbid suspension begins, and the concentration region where fast flocculation occurs. The later leads to the separation from the solution of large, well formed flocs [1]. Both concentrations are very close for the case of SDS and AOS. Stock solutions included Al2 (SO4 )3 , SDS, the model pollutants and, in some cases, NaOH and H2 SO4 for the experiments with adjusted pH. These stock solutions are kept at 25 ◦ C, mixed, shaken vigorously, left to settle for five minutes and filtered through qualitative Whatman paper Nr 1 (800 ␮m pore size). In this work catechol, phenol and phenylamine from Fischer Chemicals, SR were used as model pollutants. Aluminum sulphate was also from Fisher, general laboratory quality, as were NaOH and H2 SO4 . SDS was technical grade, kindly provided by Unilever Research Vlaardingen. In some experiments AMF was investigated on samples where the interaction between phenylamine, SDS and Al3+ led to turbid suspension of small size particles. These samples were first filtered and the content in surfactant and pollutant analysed in the way described above, and then filtered again through 0.2 ␮m PURADISC 25 PP polypropylene filtration disks. The clarified liquid was again analysed. 2.2. Analysis of substances The surfactant content is determined by a double phase titration using benzethonium chloride (Aldrich) as standard and a mixture of dimidium bromide and disulfine blue (Fisher Chemicals) as indicator. The method conforms to ISO-2271-1972. The concentration of organic compound (all model pollutants are aromatic) is measured in a Jenway 6405 spectrophotometer at 280 nm. A diluted solution (dilution factor 1:20) formed from an aliquot of the sample is alkalinised with solid NaOH to pH at

30


least 11 for phenylamine and 14 for catechol, to disrupt possible Al-pollutant complexes. The process also destroys any small colloidal particles, making the solutions perfectly clear. Phenylamine is known not to have stable complexes, but interactions may still occur. This method allows to compare absorbance readings in a sample with Al3+ with a calibration line for the pollutant in the absence of Al3+ , at the same pH.

rium phenylamine-phenylammonium by removing phenylamine from the solution more efficient than less hydrophobic, stable micelles do in MEUF. However, in the second case it must be taken into account that the flocculate is actually a large, single bulk normally floating on top of the solution, with a much smaller contact surface than the micelles or the nascent flocculate during aggregation. This should impair the adsorption of phenylamine onto the flocculate.

3. Results and discussion 3.1. Ratios phenylamine/SDS in AMF and MEUF Phenylamine with a pKa 4.63 is found as Phenylammonium cation at the low pH values found during micellar flocculation, between 3 and 3.5. Fig. 1 shows the partition coefficient for phenylamine between flocs and water versus the fraction of phenylamine in the floc. The results range between 2000 to 6000, with the solute present as phenylammonium. The partition coefficient is calculated as: K=

([PhNH2 ]initial − [PhNH2 ]final )/([SDS]initial − [SDS]final ) [PhNH2 ]final /55

with 55 being the concentration of water (55 mole per litre). In this way, the partition coefficient is represented as molar ratio phenylamine/SDS over molar ratio phenylamine/water. For MEUF, Dougherty and Berg give values between 400 and 1400, using the same definition of partition coefficient, with the solute present as phenylamine [21]. The ionic form adsorbs much better on the floc than the neutral form on the micelle. Phenylammonium salts are not soluble in non-polar solvents like ether, whereas their solubility in water is much higher than that of phenylamine. On the other hand, micellar flocculates are more hydrophobic than micelles, as they cannot remain colloidal in water. One explanation for enhanced removal of phenylammonium by adsorptive micellar flocculation may be that flocculation of micelles by Al3+ and removal of organic compounds are simultaneous but not directly related. Another is that the more hydrophobic flocculate displaces the equilib-

Fig. 1. Partition coefficient (molar fraction in micelle over molar fraction in solution) for phenylamine in micellar SDS flocculates. [SDS]initial range from 0.017 to 0.0765 M; [Al2 (SO4 )3 ]initial = 0.015 M; [Ph-NH2 ]initial = 0.0025–0.0086 M.

3.2. Binding of phenylammonium onto imperfectly formed micellar flocculates Some experiments were carried with solutions containing 0.072 M SDS and 0.01286 M phenylamine. [Al2 (SO4 )3 ] ranged from 0.005 to 0.0125 M. These experiments allowed to study the binding of phenylamine to micellar aggregates obtained under different flocculation conditions. For [Al2 (SO4 )3 ] = 0.005 M, no large flocs formed, although small aggregates formed that could be rejected with a 200 nm PURADISC membrane. For an aluminum sulphate concentration of 0.0125 M, flocculation of micelles was complete and no small aggregates were collected through the 200 nm membrane. In between, different proportions of flocculated surfactant were presented as large flocs or small aggregates. Fig. 2 shows the concentration of SDS and phenylamine remaining in solution after flocculation. Initially depletion of both SDS and phenylamine is proportional, but afterwards the removal of phenylamine in the large flocs slows down, whereas the removal of phenylamine in the small aggregates is enhanced. In fact, at aluminum sulphate 0.01 M, the combined effect of large flocs and small colloids leads to a residual phenylamine concentration of less than 0.002 M, whereas at aluminum sulphate 0.0125 M, in the presence of only large flocs, a more hydrophobic substrate, the residual phenylamine concentration is higher, about 0.003 M. Fig. 3 shows that small aggregates actually show higher binding ratios than large flocs occurring at the same time. Large flocs have constant binding ratios over the range investigated, whereas small colloids have increasing binding ratios. This suggests that hydrophobicity may not be the leading factor in removal of phenylamine. Adsorption of phenylamine on small aggregates is favoured by their larger specific surface.

Fig. 2. Remaining concentrations of phenylamine and SDS after flocculation. MF refers to residual concentrations after filtration through a 0.2 mm membrane, and “filter” refers to residual concentrations after filtration through a Nr 1 Whatman qualitative filer. [SDS]initial = 0.072 M, [Ph-NH2 ]initial = 0.0129 M.


Fig. 3. Binding ratios on large flocs and small particles, same conditions as in Fig. 2. The pollutant load on the fraction of small colloidal particles is higher than in the large flocs.

3.3. Comparison between pollutant/surfactant binding ratios in floc for different organic compounds Solubility in solvent data and octanol/water partition coefficients are taken from the literature [22,23]. Fig. 4 shows binding ratios pollutant/SDS versus free concentration of pollutant for 2,4-D, phenol, catechol and phenylamine. Fig. 5 shows the same information in a semi-logarithmic plot for clearer observation of the behaviour at low pollutant concentration. Phenylamine shows a shape like a Temkin isotherm with phase change [24]. This refers to an occupied surface, which can no longer be homogeneous. Such behaviour would be consistent with the fact that phenylamine has a strong effect in modifying the susceptibility of micelles to flocculation. Catechol exhibits higher binding ratios than phenol [8], although its octanol water partition coefficient of 6.3 and its partition coefficient in SDS micelles (as ratio of molar fractions in micelle and solution) of 6.3 are lower than the octanol water partition coefficient of 28.8 and micelle partition coefficient of 10 for phenol. The solubility of catechol in benzene is 1% (w/w), against 8.2% (w/w) of phenol. Finally, 2,4-dichlorphenoxyacetic acid, which is significantly less sol-

Fig. 4. Binding ratios pollutant surfactant for phenylamine, catechol, phenol and 2,4-D [12]. The dotted line represents binding ratios calculated from the equilibrium constant for phenol-SDS-water by Scamehorn and Christian [22]. The solid line represents binding ratios calculated from the equilibrium constant by Sabate et al. [23]. Initial [SDS] = 0.02 M for phenol, catechol and phenylamine. [Al2 (SO4 )3 ] = 0.015 M for phenol, and 0.025 M for catechol and phenylamine. Data for 2,4-D are from a previous work, with initial [SDS] = 0.05 M and [Al2 (SO4 )3 ] either 0.023 M or 0.035 M.

31

Fig. 5. Same as Fig. 4, but with linear scale for better detail of the catechol and phenylamine series.

uble in water, but has an octanol/water partition coefficient of 6.2, is sparingly soluble in organic compounds, and has a pKa of 2.8 shows high ratios pollutant/SDS in the floc at very low free pollutant concentrations, compared with pollutant-micelle binding values for phenol. The first dissociation constant of catechol (9.29) is less than one unit lower than the dissociation constant for phenol of 9.89. 3.4. Effect of the availability of Al3+ on pollutant binding to micellar flocculates Fig. 6 shows binding ratios phenylamine/SDS for a range of aluminum sulphate concentrations. There is substantial scatter although most of the points follow roughly a trend. In Fig. 7, however, there is a clear effect of the increase in the availability of Al3+ , with binding ratios catechol/SDS falling significantly. Comparing with the data shown in Figs. 4 and 5, the binding ratios for catechol in the presence of 0.0125 M aluminium sulphate are in the range of the binding ratios for phenol. The results suggest that hydrophobicity is not the main reason for pollutant binding to flocs in AMF, although in the case of moderately hydrophobic compounds with minor complexing abilities like phenol, hydrophobic interactions between flocculate and pollutant may play an important role.

Fig. 6. The results presented in the figure have been obtained in a variety of conditions, with initial [Phenylamine] from 2 × 10−3 to 12.9 × 10−3 M, and with initial [SDS] from 0.018 to 0.072 M. Binding ratios may be up to 0.25.

32


lammonium and there is less aluminum present as Al3+ . The sharp increase in the solubility of phenylamine is connected with the increase in the concentration of SDS not flocculated above pH 4. Although this means that the colloids are more hydrophilic, this does not prevent an increase in the binding ratio phenylamine/SDS. Such increase in binding ratios is possibly a consequence of the increased concentration of phenylamine in solution displacing the equilibrium. Fig. 9 shows a similar set of plots for catechol. The concentration of catechol in solution remains almost constant, with minor variations in binding ratio catechol/SDS. This is to be expected because over most of the range of pH the predominant form of catechol is the same. Fig. 7. Initial [catechol] = 0.02 M. Initial [SDS] = 0.02 M. At the higher aluminum sulphate concentration the binding ratios drop dramatically.

4. Conclusions

Fig. 8. Initial [phenylamine] = 0.02 M. Initial [SDS] = 0.02 M. Full triangles (residual SDS) and open rhombi (residual phenylamine) must be read against the right-hand Y-axis. Full rhombi (binding ratio phenylamine/SDS) must be read against the left-hand Y-axis.

Removal of phenylamine by AMF seems based on direct adsorption of phenylammonium. The binding ratios of phenylammonium to flocs are higher than the binding ratios of phenylamine to micelles. There does not seem to be a significant influence of the concentration of Al3+ on the removal of phenylamine. The effect of Al3+ seems to be to provide a more favourable pH. This would suggest that Fe3+ , which is less efficient in the removal of donors of protons like phthalic acid or phenol precisely because of the acidity of its solutions [12], might be an alternative to Al3+ in the case of amines. However, there is evidence [25] that micelle bound Fe3+ reacts with phenylamine in a redox process. Therefore, if recovery of phenylamine is the target, Fe3+ is not a viable alternative. The high binding ratios of catechol, compared with phenol, and the effect of bulk phase concentrations of Al3+ on the binding of catechol to micellar flocculates suggests that complexes may play a role in its adsorption onto micellar flocculates. An increase in pH from pH 3 to pH 4.5 causes, without affecting the flocculation of the surfactant, a decrease in the ratio catechol/SDS. An increase in pH reduces the availability of Al3+ , although within limits this would not affect surfactant flocculation. However, this may explain why the efficiency of the flocculate as adsorbent decreases. The results suggest that direct adsorption into the micellar palisade or adsorption of complexes are the more likely mechanisms for the capture of pollutants by adsorptive micellar flocculation. Some scarcely dissociated molecules, with higher hydrophobicity, may be attracted by the flocculate as a hydrophobic surface. However, well-formed and highly hydrophobic micellar flocculates are single massive aggregates and the loss in surface during the process of aggregation is very important.

3.5. Effect of pH on removal of catechol and phenylamine The removal of phenylamine by AMF is influenced by the pH of the system, with higher pH resulting in worse separation (Fig. 8). At higher pH less phenylamine is present as pheny-

Acknowledgement EPSRC funded this research through grant GR/R 44393/01. Fig. 9. Initial [catechol] = 0.02 M. Initial [SDS] = 0.02 M. Open symbols must be read against the right-hand Y-axis. Rhombi are residual catechol concentrations, squares residual SDS concentrations. Full symbols (squares) must be read against the left-hand Y-axis. They represent catechol/SDS binding ratios in flocs.

References [1] F. Talens, P. Paton, S. Gaya, Langmuir 14 (18) (1998) 5046–5050. [2] P. Paton, F.I. Talens, J. Surfactants Deterg. 1 (3) (1998) 399–402.

T.D. Almeida, F.I. Talens-Alesson / Colloids and Surfaces A: Physicochem. Eng. Aspects 279 (2006) 28–33 [3] G. Cathalifaud, J. Ayele, M. Mazet, Water Res. 31 (4) (1997) 689–698. [4] E. Lefebvre, B. Legube, Water Res. 7 (3) (1993) 433–447. [5] V. Lahoussine-Turcaud, M. Wiesner, J.Y. Bottero, J. Mallevialle, Water Res. 26 (5) (1992) 695–702. [6] F. Julien, B. Gueroux, M. Mazet, Water Res. 28 (12) (1994) 2567–2574. [7] F.I. Talens-Alesson, S. Anthony, M. Bryce, Water Res. 38 (2004) 1477–1483. [8] M. Bryce, Colloids Surf. A, 2005. doi: 10.1016/j.colsurfa.2005.08.023. [9] P. Paton, F.I. Talens-Alesson, Langmuir 17 (20) (2001) 6059–6064. [10] P. Paton, F.I. Talens-Alesson, Langmuir 18 (22) (2002) 8295–8301. [11] P. Paton, F.I. Talens-Alesson, Colloid Polym. Sci. 279 (2000) 196–199. [12] F.I. Talens-Alesson, S.T. Hall, N.P. Hankins, B.J. Azzopardi, Colloids Surf. A 204 (2001) 85–91. [13] M. Porras, F.I. Talens-Alesson, Env. Sci. Technol. 33 (18) (1999) 3206–3209. [14] V. Suratkar, S. Mahapatra, J. Colloid Interf. Sci. 225 (2000) 32.

[15] [16] [17] [18] [19] [20] [21] [22] [23] [24]

[25]

33

X-N. Yang, M.A. Matthews, J. Colloid Interf. Sci. 229 (2000) 53. B.J. Kim, S.S. Im, S.G. Oh, Langmuir 17 (2001) 565–566. C.T. Jafvert, C.-C. Lin, Langmuir 16 (2000) 2450. J.F. Rathman, J.F. Scamehorn, Langmuir 88 (1984) 5807–5816. L. Yane, J.B. Fein, Geochimica et Cosmochimica Acta 62 (12) (1998) 2077–2085. F.I. Talens-Alesson, J. Dispersion Sci. Technol. 20 (7) (1999) 1861–1871. S.J. Dougherty, J.C. Berg, J. Colloid Interf. Sci. 48 (1) (1974) 110–120. Environmental Contaminant Reference Databook, 1998. John Wiley & Sons. J.A. Dew (Ed.), Lange’s Handbook of Chemistry, 15th ed., McGrawHill, 2001. H.H. Kohler, Thermodynamics of adsorption from solution, in: B. Dobias (Ed.), Coagulation and Flocculation, Marcel Dekker, New York, 1993, pp. 14–20. F.I. Talens-Alesson, Chem. Eng. Technol. 6 (2003) 684–687.