Synergistic effects in competitive adsorption of carbohydrates on an ion-exchange resin

https://doi.org/10.1016/j.chroma.2007.07.022Get rights and content

Abstract

Adsorption of the three carbohydrates sucrose, glucose and fructose from aqueous solutions was investigated on an ion-exchange resin. The adsorption equilibrium of single components, binary and ternary mixtures was quantified by frontal analysis and the adsorption–desorption method. The experiments covered a concentration range up to 600 g/L at 60 °C and 80 °C. Within this range the adsorption isotherms of carbohydrates exhibited anti-Langmuirian behavior. Data of mixture adsorption revealed reversed competitive (synergistic or cooperative) effects, i.e., an increase of the concentration of one component of the mixture enhanced the adsorption of others. To model such an adsorption behavior the anti-Langmuir model has been used. The isotherm parameters determined for single components were used to simulate the competitive adsorption equilibria through the IAS (ideal adsorbed solution) theory. Finally, dynamic concentration profiles of multicomponent mixtures have been recorded. The shapes of adsorption and desorption curves confirmed the observed competitive effects found in the equilibrium studies. The breakthrough curves measured were simulated using the equilibrium theory as well as a numerical solution of the equilibrium dispersive model.

Introduction

Separation of different sugars belong to the industrial chromatographic processes conducted on very large scale [1], [2], [3]. The separations are preferably performed on ion-exchange resins [4], [5], [6], [7], [8], typically on a polystyrene-based cation exchange resin in calcium form using warm water as eluent. Temperature is often above 60 °C as a good compromise between reduction of viscosity and preservation of carbohydrates. Separation of sugars occurs due to specific interactions between calcium ions held by the resin and the hydroxyl groups of the different sugars. Hereby, oligosaccharides form a complex with the calcium ions; fructose is strongly bound to the surface and retained in the column, while glucose and other oligosaccharides are more easily eluting [9]. Adsorption equilibria of sugars has been investigated in a number of studies. To acquire adsorption isotherm data some authors applied frontal analysis [9], [10], [11], [13], [16], or used other methods [12], [13], [14], [15]. In the case of ion-exchange resins, many authors assumed linear isotherms. Ching et al. [17] pointed out that the frequently made assumption of linearity of glucose and fructose isotherms on cation exchange resins might not be valid at higher concentrations. In that work, isotherms were determined for solutions of concentrations up to 260 g/L. The glucose isotherm was found to be strictly linear, but the isotherm of fructose possessed a significant curvature in the higher concentration range. The non-linearity of sugar isotherms up to 400 g/L has been recently reported by Vente et al. [13], [14].

In this work the adsorption equilibria of the three sugars sucrose, fructose and glucose have been measured on a Ca2+ ion-exchange resin within a concentration range up to 600 g/L. The data of adsorption equilibrium of single, binary and ternary mixtures were acquired by frontal analysis and the adsorption–desorption method [12].

A considerable number of adsorption isotherm models of single component have been suggested in the literature [18], [19], [20]. In studies of liquid–solid equilibria most frequently Langmuir-type isotherms were used. The standard Langmuir model can describe concave isotherms reaching a saturation behavior. This model assumes homogenous surfaces and neglects adsorbate interactions. Analogously, the anti-Langmuir model can be used for convex isotherms. Adsorption isotherms often possess more complicated shapes. They may have inflection points, e.g., isotherm being concave at low concentration and convex at high concentration [21], [22].

The extension of the Langmuir model to competitive adsorption (competitive Langmuir model) is based on the same hypotheses as that for single one. It violates the Gibbs's equation and is thermodynamically inconsistent when the saturation capacities for all components are different. As it was reported in Ref. [23], this fact is also true in the case of anti-Langmuir and mixed Langmuir isotherms (i.e., when one of the component follows the Langmuir while the other the anti-Langmuir model). Moreover, even when saturation capacities of components are identical, mixed Langmuir isotherms are not consistent with the Gibbs isotherm. Thermodynamic consistency is guaranteed in the case of competitive isotherms by the IAS (ideal adsorbed solution) framework [24].

In this work to model the single component adsorption behavior of sugars the anti-Langmuir model was used. The isotherm parameters determined for single components were used to calculate the competitive adsorption equilibria through the IAS theory. The isotherm models were employed to simulate the individual band profiles using the equilibrium theory [25], [26], [27] as well as the numerical solutions of a more detailed dynamic model. To determine the individual band profiles of a binary mixture the analytical solution of the model of column dynamics can be used provided that the adsorption behavior of the components follows the Langmuir-type competitive model. However, an analytical solution cannot be derived for the competitive IAS model capable to incorporate arbitrary single component isotherm models.

Section snippets

Single component adsorption isotherms

The generalized Langmuir model for a component i is given by the following equation [28]:qi*=qs,iKiCi1+piKiCi,p=±1where Ci is the concentration of the component i in the mobile phase, qi* the loading (concentration in the solid phase) in equilibrium with Ci, qs,i the monolayer saturation capacity (loading capacity) and Ki is the adsorption equilibrium constant. The parameter pi can take the values ±1 and define the sign of the corresponding term in the denominator. Langmuirian (favorable)

Chemicals and materials

For the experiments performed four columns marked were available (250 mm × 8 mm). The columns were filled with sugar grade gel-type resins MDS 1368 (Bayer-Lewatit, Leverkusen, Germany) in Na+ form, subsequently changed into Ca2+ form. The bead size (90%) was 0.35 mm (±0.05). In each case, solutions were prepared with quality Milli-Q aseptic water (obtained by passing doubly distilled water through a Millipore Milli-Q purification system (Millipore, Billerica, MA, USA)). Column porosity was

Single component isotherms

Fig. 3 shows a comparison between experimental equilibrium loadings for the individual sugars obtained by FA and AD experiments as well as the isotherms calculated with the anti-Langmuir model (parameters in Table 1). The sugar isotherms obtained are slightly convex, which is noticeable in particular for glucose and sucrose. The experiments presented in this work confirm earlier reports of non-linearity [13], [14].

The comparison between experimental results of the FA and the AD method (Fig. 3)

Conclusions

The adsorption equilibrium data of single-component systems acquired by frontal analysis confirmed that investigated sugar isotherms have convex shapes within the range of concentrations (0–600 g/L). In order to predict this behavior the anti-Langmuir model has been employed. The competitive adsorption behavior of the binary and ternary mixtures of sucrose, glucose and fructose revealed the reversed competitive (synergistic) effect, i.e., the presence of additional sugar (or sugars) enhanced

Nomenclature

    Notation

    C

    concentration in mobile phase (g/L)

    C¯

    concentration of all the mixture components in mobile phase (g/L)

    C˜i0

    fictitious concentration (g/L)

    Da

    apparent diffusion coefficient (m2/s)

    F

    phase ratio, (1  ɛt)/ɛt

    K

    equilibrium constant (L/g)

    L

    column length (m)

    m

    mass (g)

    n

    number of components

    N

    column efficiency

    p

    parameter in the generalized Langmuir isotherm

    q

    concentration in the solid phase (g/L)

    qs

    saturation capacity (g/L)

    r1, r2

    roots of Eq. (12)

    S

    sum of squared differences in Eq. (15) (g/L)

    t

    time coordinate (s)

    t0

    dead

References (41)

  • C.B. Ching et al.

    Chem. Eng. Sci.

    (1985)
  • C.B. Ching et al.

    Chem. Eng. Sci.

    (1985)
  • P. Jandera et al.

    J. Chromatogr.

    (1974)
  • R.W. Goulding

    J. Chromatogr.

    (1975)
  • M. Saska et al.

    J. Chromatogr.

    (1992)
  • L. Matijasevic et al.

    Biochem. J.

    (2000)
  • A. Jupke et al.

    J. Chromatogr. A

    (2002)
  • R. Lewandowski et al.

    Chem. Eng. Process.

    (1992)
  • J.A. Vente et al.

    J. Chromatogr. A

    (2005)
  • Y.A. Beste et al.

    J. Chromatogr. A

    (2000)
  • C.B. Ching et al.

    Chem. Eng. Sci.

    (1987)
  • A. Seidel-Morgenstern et al.

    Chem. Eng. Sci.

    (1993)
  • W. Piątkowski et al.

    J. Chromatogr. A

    (2003)
  • M. Mazzotti

    J. Chromatogr. A

    (2006)
  • A. Seidel et al.

    Chem. Eng. Sci.

    (1988)
  • J.M. Jacobson et al.

    J. Chromatogr.

    (1990)
  • A. Seidel-Morgenstern

    J. Chromatogr. A

    (2004)
  • O. Lisec et al.

    J. Chromatogr. A

    (2001)
  • F. Gritti et al.

    J. Chromatogr. A

    (2005)
  • K. Hashimoto et al.

    J. Chem. Eng. Jpn.

    (1983)
  • Cited by (0)

    View full text