2229
J. Sep. Sci. 2015, 38, 2229–2237
Dengjun He1,2 Jiachun Zhou1 Quanming Xia1 Lihua Jiang1 Yongjun Qiu1,2 Liming Zhao1,2 1 State
Key Laboratory of Bioreactor Engineering, R&D Center of Separation and Extraction Technology in Fermentation Industry, East China University of Science and Technology, Shanghai, China 2 Shanghai Collaborative Innovation Center for Biomanufacturing Technology (SCIBT), Shanghai, China Received January 7, 2015 Revised April 3, 2015 Accepted April 4, 2015
Research Article
Kinetics and equilibria of the chromatographic separation of maltose and trehalose Trehalose, a nonreducing disaccharide, has been extensively applied to food, cosmetics, and pharmaceutical goods. The resultant solution of trehalose prepared by enzymatic methods includes high amounts of maltose. However, it is quite difficult to separate maltose and trehalose on an industrial scale because of their similar properties. In this paper, a highperformance resin was selected as a stationary phase to separate trehalose and maltose, and the resolution of these sugars was 0.59. The potential of a cation exchange resin was investigated as the stationary phase in separating trehalose and maltose using deionized water as the mobile phase. Based on the equilibrium dispersive model, the axial dispersion coefficients and overall mass transfer coefficients of maltose and trehalose were determined by moment analysis at two different temperatures, 50 and 70⬚C. Other parameters, including the column void and the adsorption isotherms, were also determined and applied to simulate the elution curves of trehalose and maltose. The simulated results matched the experimental data, validating the parameters. The optimized parameters are critical to the chromatographic separation of trehalose and maltose on an industrial scale. Keywords: Adsorption isotherm / Axial dispersion / Mass transfer / Maltose / Trehalose DOI 10.1002/jssc.201500005
1 Introduction Trehalose (␣-D-glucopyranosyl-(1,1)-␣-D-glucopyranoside), a nonreducing disaccharide, is present in many organisms, such as bacteria, fungi, seeds, plants, and fungal spores. It preserves the life of organisms during freezing, heating, dehydration, and exposure to toxic chemicals [1, 2]. As a unique sugar, trehalose has been extensively applied to food, cosmetics, and pharmaceutical goods as, for example, preservatives in unstable foods, humectants in cosmetics, and antifreeze in medicine [3, 4]. Its usefulness has been recognized for decades, but trehalose was not produced on a large scale until 1994 [2]. A number of methods for the enzymatic preparation of trehalose have been reported [5–8]. Reaction of maltooligosyl trehalose synthase and maltooligosyl trehalose trehalohydrolase with starch is a large-scale procedure for the commercial production of trehalose. The reaction of converting starch to trehalose is not completely efficient, and the resulting product Correspondence: Professor Liming Zhao, State Key Laboratory of Bioreactor Engineering, R&D Center of Separation and Extraction Technology in Fermentation Industry, East China University of Science and Technology, Shanghai 200237, China E-mail: [email protected] Fax: +86-021-64250829
Abbreviations: HETP, height equivalent to a theoretical plate; ODE, ordinary differential equation; SMB, simulated moving bed C 2015 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
solution is complex, including both trehalose and maltose. It is technically difficult to utilize the byproduct maltose after extracting a major portion of the trehalose from the mother liquor by crystallization [6, 9–11]. Therefore, the resulting solution containing a large amount of maltose is not useful, economical, or environmentally friendly because of the complex mixture of chemicals and high-energy consumption. In recent years, chromatographic techniques such as simulated moving bed (SMB) chromatography have been successfully applied in the food industry, particularly for sugar mixtures [12–14]. Al-Eid [15] used a chromatographic column filled with cation exchange resin to separate fructose and glucose from date syrup. Lee [16] used a two-section SMB to separate glucose and fructose from an aqueous mixture at a high concentration of 500 mg/mL. Underivatized samples of trehalose and maltose have been successfully separated by many methods, such as thin-layer chromatography [17], high performance thin-layer chromatography [18], HILIC [19], and high-performance anion-exchange chromatography [20]. However, separation of these sugars on a semipreparative column or SMB has not been previously reported. The resins loaded with cations such as Ca2+ , Na+ , or K+ are commercially available, which can form complexes with sugars in aqueous solution [21], and the contribution of sugar-calcium complex formation appears to be directly correlated with the elution of trehalose and maltose in ligand-exchange chromatography [22]. Specifically, the formation of sugar-calcium complexes between the cations immobilized on the ion exchange resin and the sugar www.jss-journal.com
2230
J. Sep. Sci. 2015, 38, 2229–2237
D. He et al.
hydroxyl groups drives separation [22, 23]. According to the above studies, the trehalose-calcium complex is stronger than that of the maltose-calcium complex, and the trehalose is therefore strongly attracted to the calcium on the resin and spends more time in the column. Conversely, maltose is less attracted and spends less time in the column. The kinetics and equilibria of these two commercially valuable products on column are critical to the chromatographic separation of trehalose and maltose on a large scale. The present study investigated the potential of an ion-exchange resin as the stationary phase in the separation of trehalose and maltose on a semipreparative scale. The column packed with this stationary phase was characterized by the bed voidage and total porosity. The axial dispersion coefficients and overall mass transfer coefficients for the chromatographic separation were determined by moment analysis. The adsorption isotherms and equilibrium constants were evaluated by frontal analysis. These optimized parameters were then used to simulate the band profiles, and the simulation and experimental results were compared.
2 Materials and methods
2.1.1 Chromatographic rate model A mathematical model can analyze and explain data from a chromatographic experiment. A series of theories and models have been reported to describe the phase equilibrium and mass transfer kinetics in LC [24–26]. In this paper, the equilibrium dispersive model was used to interpret the chromatographic process. Hence, the mass balance equation for each component in the mobile phase can be expressed as [24]:
1 − gT gT
q i∗ = K i c i
where Ki is the dimensionless equilibrium constant. In this approach, it was assumed in Eq. (2) that the difference between qi * and qi was the driving force of the mass transfer, and the mass transfer rate was in proportion to the driving force. In this linear driving force expression, it was assumed that the dominating resistance in the particle diffusion step could be represented by an overall mass transfer coefficient. 2.1.2 Moment analysis Moment analysis is an effective method to determine the axial dispersion coefficient and mass transfer coefficient from pulse experiments in a chromatographic column [26–29]. The first and second moments can be expressed as follows:
2L 2 = v
(1)
Here, ci and qi denote the average values of the mobile phase and stationary phase concentrations, for respectively, for component i; v represents the velocity of the interstitial mobile phase; DL is the axial dispersion coefficient; g T is the total porosity of column; and t and z are the time and the space coordinates, respectively. The following assumptions were considered in the derivation of Eq. (1), including homogeneous packing, a constant volumetric flow rate, a constant dispersion coefficient, isothermal conditions, and no radial gradients. The sugar concentration on the stationary phase was in proportion to the mass transfer coefficient on the basis of the linear driving force approximation model [25]: ∂q i = ki (q i∗ − q i ) ∂t C 2015 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
(2)
(4)
2 DL 1 − gT 1 − gT K 1+ K + v2 gT gT km
(5)
In the above equations, 1 and 2 are the first and second moments; L is the column length; km is the overall mass transfer coefficient; and K is the equilibrium constant. Thus, the expression of the height equivalent to a theoretical plate (HETP) can be derived: HETP =
∂q i ∂ 2ci ∂c i +v = DL 2 ∂t ∂z ∂z
(3)
1 − gT L K 1+ 1 = v gT
2.1 Theoretical procedures
∂c i + ∂t
where ki is the mass transfer coefficient of component i; and qi * is the equilibrium concentration of the stationary phase with a solution of concentration ci . qi * was related to these mobile phase concentrations through the equilibrium isotherm. In this case, the linear equilibrium isotherm was given as:
2 L 2DL L = 2 = N v 1 −2 1 − gT 1 − gT K 1+ K +2v gT km gT
tR 2 N = 5.545 w1/2
(6)
(7)
where N is the theoretical plate number; w1/2 is the peak width at half height; and tR is the retention time. According to different expressions of the axial dispersion coefficient DL and the mass transfer coefficient km , HETP can be expressed in various ways. 2.1.3 Axial dispersion On the basis of Eq. (6), the axial dispersion and mass transfer coefficients cannot be evaluated by a simple linear regression of their nonlinear relationships with HETP. However, www.jss-journal.com
Liquid Chromatography
J. Sep. Sci. 2015, 38, 2229–2237
2231
Figure 1. Typical chromatogram of maltose and trehalose. Experimental conditions: binary solution, 25 mg/mL; flow rate, 1.0 mL/min; mobile phase, deionized water; temperature, 70⬚C.
Figure 2. HPLC–RID chromatogram of a standard mixture of maltose and trehalose of 20 mg/mL. Experimental conditions: column, Waters SugarPak I (300 mm × 6.5 mm id); mobile phase, distilled water; flow rate, 0.5 mL/min; temperature, 80⬚C; injection volume, 20 L.
the axial dispersion primarily consisted of two mechanisms, molecular diffusion and eddy diffusion. Therefore, the axial dispersion can be expressed as [30]: DL = Dm + v
(8)
where, is the tortuosity factor for a packed column; is a constant dependent on flow geometry; and Dm is the molecular diffusivity. In liquid systems, the contribution of molecular diffusivity to axial dispersion was too small to be mentioned, as previously reported [26, 30] even at low Reynolds numbers [28]. Diffusivities of organic molecules with low molecular weights in water range from 6 × 10 −5 to 9 × 10 −4 cm2 min−1 , but can be smaller with increasing C 2015 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
molecular weight [31]. The molecular diffusivity can thus be ignored, and Eq. (8) can be written as: DL = v
(9)
Using Eq. (9) in Eq. (6), we obtain: HETP = 2 + 2v
1 − gT gT
−2 K 1 − gT 1+ K km gT
(10)
As a result, by calculating the first and second moments of a series of elution curves and plotting HETP versus fluid velocity of the mobile phase, the axial dispersion DL and the overall mass transfer coefficient km can be calculated from the intercept and the slope of the straight line, respectively. www.jss-journal.com
2232
J. Sep. Sci. 2015, 38, 2229–2237
D. He et al.
Figure 3. Plot of retention times of (A) acetone and (B) blue dextran versus the inverse superficial velocity of the mobile phase.
2.2 Materials D-(+)-Maltose monohydrate (purity > 99%), D-(+)-trehalose (purity > 99%), and acetone (purity > 99%) were purchased from Sinopharm Chemical Reagent (Shanghai, China). Blue dextran (purity > 99%) was purchased from Sigma–Aldrich (St. Louis, MO, USA). Deionized water was used as the eluent and solvent in all experiments. Purolite CGC 100 × 8 Ca ion exchange gel-type resin was used as the stationary phase. The resin was a spherical bead, and its polymer structure combined with gel polystyrene, which was cross-linked by divinylbenzene. Sulfonic acid and Ca2+ were the functional group and ionic form, respectively. The resin with a typical mean size of 100 m was used to pack the column.
2.3 Semipreparative chromatography The system used to determine the hydrodynamics and thermodynamic model parameters, i.e. bed porosity, adsorption isotherms, and axial dispersion coefficient, was composed of a single glass column (40 cm × 1.6 cm id) equipped with a jacket, pump (BT100, Huxi, Shanghai, China), recirculation water bath (MP-5H, Yiheng, Shanghai, China), and fraction collector (BSZ-100, Huxi, Shanghai, China).
2.4 Sample analysis Samples were collected at the exit of the column and analyzed by three methods. The mono sugar solutions (samples) of trehalose and maltose at greater than 10 mg/mL were analyzed by a Pocket refractometer (PAL-LOOP, Japan). If samples were below this concentration, the fractions were analyzed on C 2015 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
an analytical HPLC system equipped with a refractive index detector (Series 1200, Agilent, USA) and a Waters Sugar-Pak I column (300 mm × 6.5 mm id). The binary solutions of trehalose and maltose were also analyzed on the HPLC system. The blue dextran and acetone solutions were analyzed by a UV/Vis detector (UV-2000, Unico, Shanghai, China) set at the wavelengths of 630 and 310 nm, respectively.
2.5 Experimental procedures 2.5.1 Bed porosities Moment analysis was applied to measure the bed porosities. The residence time distribution curves were obtained from tracer experiments [32]. The tracers blue dextran (3 mg/mL) and acetone (20% v/v) assessed external (g) and total (g T ) bed porosity, respectively. Pulses of 2.0 mL of tracer were injected into the column under 30⬚C at a flow rate of 1.0–2.0 mL/min. 2.5.2 Adsorption isotherms The frontal analysis method [32, 33] was used to measure the single component adsorption isotherms of trehalose and maltose. The concentrations of trehalose and maltose were both between 25 and 150 mg/mL. The mobile phase flow rate was approximately 1.5 mL/min for both trehalose and maltose. A step series method was used to pump the feed solutions into the column. A known concentration of a single sugar was pumped into the column, and the breakthrough curve was monitored at the exit of the column. When the column became saturated, deionized water was used as a wash until reequilibration was reached. For each maltose (or trehalose) liquid-phase concentration, three experiments were performed, and the average value of the equilibrium www.jss-journal.com
Liquid Chromatography
J. Sep. Sci. 2015, 38, 2229–2237
2233
Figure 4. Adsorption isotherms of maltose and trehalose at 50⬚C (A) and 70⬚C (B).
concentration in the liquid and solid phases was applied to a single compound adsorption isotherm. The experiments were carried out at 50 and 70⬚C. The concentration of the adsorbed phase in equilibrium with the liquid phase concentration C0 was calculated by mass balance as expressed in Eq. (11): q i∗ =
Ai Q − gVc C0 (1 − g)Vc
Ai =
∞
[C0 − Ci (t)]dt − C0 tdead
(11)
(12)
0
In Eqs. (11) and (12), Ai is the mass of solute accumulated inside of the column; Q is the fluid flow rate; Vc is the bed
C 2015 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
volume; and tdead denotes the dead time. tdead consists of the time spent in the extra column volumes of the experimental setup, which includes tubing, connections, and pump heads.
2.5.3 Axial dispersion and mass transfer coefficients In the given semipreparative chromatographic system, the experiments were performed at five different mobile phase flow rates from 0.7 to 2.0 mL/min and at two different temperatures, 50 and 70⬚C. The band profiles were recorded to measure the axial dispersion and the overall mass transfer coefficients.
www.jss-journal.com
2234
J. Sep. Sci. 2015, 38, 2229–2237
D. He et al.
Figure 5. Plot of the HETP of maltose and trehalose versus the interstitial velocity of the mobile phase at 50⬚C (A) and 70⬚C (B).
3 Results and discussion 3.1 Separation of trehalose and maltose Square pulses of solutions of maltose and trehalose mixtures were injected into the chromatographic column at 1.0 mL/min. A typical chromatogram of these mixtures is illustrated schematically in Fig. 1. Trehalose was better retained by this stationary phase, and maltose was less retained. The resolution [25] between trehalose and maltose can be calculated by Eq. (13): R=
tR,T − tR,M (wT + wM )/2
resolution of the mixture was not good enough to completely separate the two sugars, and SMB should therefore be used to separate maltose and trehalose on an industrial scale in future studies. The fractions collected upon exit of the chromatographic column were analyzed by HPLC. Figure 2 shows a typical chromatogram of trehalose and maltose separated on a Waters Sugar-Pak I, in which complete separation of trehaose and maltose was achieved. The retention times were 9.3 and 7.5 min for trehalose and maltose, respectively, representing an elution order identical to that on the Purolite CGC 100 × 8 Ca ion exchange resin.
(13) 3.2 Determination of porosity
where R is the resolution; tR,T and tR,M are the retention times of trehalose and maltose, respectively; and wT and wM are the peak widths of trehalose and maltose, respectively. According to Eq. (13) and Fig. 1, the resolution of trehalose and maltose was 0.59. However, as seen in the chromatogram, the C 2015 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
According to Eq. (14), the external porosity g and the total porosity g T can be evaluated from the first moments of blue dextran and acetone, respectively, because these compounds were not retained on this resin. Blue dextran is a polymer that www.jss-journal.com
Liquid Chromatography
J. Sep. Sci. 2015, 38, 2229–2237 Table 1. Adsorption isotherm parameters determined by frontal analysis
Table 2. Axial dispersion and mass transfer coefficients
T (⬚C) T (⬚C)
Maltose
50 70
R2
Adsorption isotherm
R2
y = 0.203x y = 0.220x
0.999 0.998
y = 0.335x y = 0.341x
0.996 0.993
is too large to pass through the particle pores, but it can move through the interparticle voids of the column. However, acetone is so small that it can penetrate not only the interparticle voids of the column but also the particle pores. = 0 ∞ 0
tc(t)dt c(t)dt
=
gL gVc = Q u
(14)
where represents the residence time; c(t) is the concentration of blue dextran or acetone at the exit of the column; u is the superficial fluid velocity; and Vc and L are the volume and length of the column, respectively. According to Suzuki [31] and Ruthven [34], the bed voidage can be calculated by Eq. (15): g T = 0.45 + 0.55g
Trehalose
gT − g 1−g
(16)
Plotting the retention times of blue dextran and acetone in the column versus L/u, as shown in Fig. 3, shows that the total void fraction and the external porosity can be calculated from the slopes of the straight lines using Eq. (14). The total void fraction g T was 0.84, and the external porosity g was 0.46. The correlation coefficients for acetone and blue dextran were R2 = 0.995 and 0.996, respectively. The internal porosity of the resin g p was 0.70, which was calculated by Eq. (16).
3.3 Determination of the adsorption isotherms The experimental adsorption isotherms of maltose and trehalose, with concentrations varying from 25 mg/mL to 150 mg/mL at a flow rate of 1.5 mL/min are shown in Fig. 4. These isotherms illustrate the relationships between the concentrations of maltose and trehalose in solution and on the stationary phase at 50 and 70⬚C. A linear function was applied to describe this experimental data. The correlation coefficients were all greater than 0.990. According to Fig. 4 and the correlation coefficients, the calculated and experimental data were in good agreement. The parameters of the adsorption isotherms including the linear equations and R-squared values are summarized in Table 1. The K values of trehalose C 2015 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
50 70
DL (cm2 min–1 )
km (min–1 )
DL (cm2 min–1 )
km (min–1 )
0.0237v 0.0262v
1.10 1.84
0.0203v 0.0176v
1.14 1.29
were clearly larger than that of maltose, and their values increased with increasing temperature. Sugar adsorption was promoted by the rise of temperature. Moreover, the resin had a higher affinity for trehalose than for maltose. According to Ramos [35], the higher affinity for trehalose can be calculated by the following equation: ␣=
∗ q T∗ /q M c T0 /c 0M
(17)
Here, ␣ represents the selectivity of trehalose against maltose. The selectivities were 1.65 and 1.55 at temperatures of 50 and 70⬚C, respectively. Selectivity decreased with increasing temperature, which may useful in purifying trehalose and maltose on an industrial scale.
(15)
Hence, the internal porosity of the adsorbent can be calculated by Eq. (16): gp =
Maltose
Trehalose
Adsorption isotherm
∞
2235
3.4 Determination of axial dispersion and mass transfer coefficients Based on Eq. (10), the values of HETP were slightly related to the axial dispersion and overall mass transfer coefficients. The plots of HETP versus the interstitial velocity of the mobile phase for maltose and trehalose in Fig. 5 show that HETP increased linearly with velocity. The axial dispersion coefficients can be calculated from the intercepts of the straight lines and are listed in Table 2. The axial dispersion coefficients for maltose and trehalose were of the same order of magnitude. The overall mass transfer coefficients can be determined from the slopes of the straight lines, and their values are also presented in Table 2. The overall mass transfer coefficients were 1.10 and 1.14 min−1 at 50⬚C and 1.84 and 1.29 min−1 at 70⬚C for maltose and trehalose, respectively. The mass transfer rate on this column was slow according to the magnitudes of the km values. The mass transfer rates of each sugar increased with increasing temperature from 50 to 70⬚C because the liquid phase viscosity decreased with increasing temperature.
3.5 Simulation of the elution profile with optimized parameters The optimized parameters, including the total porosity of the column g T , equilibrium constants K, axial dispersion coefficient DL , and overall mass transfer coefficients km , were used to simulate the elution profiles and verify the effectivewww.jss-journal.com
2236
J. Sep. Sci. 2015, 38, 2229–2237
D. He et al.
Figure 6. Comparison of the simulated and experimental band profiles of maltose and trehalose. Experimental conditions: binary solution, 200 mg/mL; flow rate, 1.0 mL/min; mobile phase, deionized water; temperature, 70⬚C.
ness of these values. The corresponding initial conditions for Eqs. (1) and (2) are given by: c i (t = 0, z) = q (t = 0, z) = 0
(18)
The boundary condition at entrance of the column can be represented as a square pulse: c i (t, z = 0) = c i0 , 0 < t ≤ t p
(19)
c i (t, z = 0) = 0, t > t p
(20)
Here, the isotherm was linear and can be expressed by Eq. (3). The boundary condition at exit of the column can be calculated as follows: ∂c i (t, z = L ) = 0 ∂z
(21)
The method of lines was used to obtain the calculated elution profiles by solving Eqs. (1) and (2) together with the initial and boundary conditions. In this method, the finite difference method discretized the partial differential equations Eqs. (1) and (2) into a set of ordinary differential equations (ODEs). A MATLAB program solved the resultant ODEs of the initial value problems. Figure 6 gives the comparison of the simulated results and experimental data, which matched reasonably well. This validated the simulation parameters. The methods used to measure these parameters and the model itself were also valid and can be used for future simulation and design of operating conditions for preparative chromatography, such as SMB.
C 2015 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
4 Conclusions The chromatographic separation of maltose and trehalose was investigated on a semipreparative chromatographic column packed with an ion exchange resin. The bed voidage, axial dispersion coefficient, and overall mass transfer coefficients were determined through moment analysis. The adsorption isotherms of maltose and trehalose were determined by frontal analysis. The stationary phase exhibited a greater affinity for trehalose over maltose. The equilibrium constants of trehalose at 50 and 70⬚C were larger than those of maltose. The magnitude of the overall mass transfer coefficient indicated that mass transfer on the column was slow. In addition, the optimized parameters were used to simulate elution profiles, and the simulated results echoed the experimental data. These reported results will be used to optimize simulated moving bed chromatography on an industrial scale in future work.
Nomenclature ci ci 0 DL Dm Dp HETP kf km K L N qi
concentration of component i in mobile phase (mg/mL) concentration of inlet (mg/mL) axial dispersion coefficient (cm2 min−1 ) molecular diffusion coefficient (cm2 min−1 ) pore diffusion coefficient (cm2 min−1 ) height equivalent to a theoretical plate (cm) external film mass transfer coefficient (cm min−1 ) overall mass transfer coefficient (min−1 ) equilibrium constant column length (cm) theoretical plate number concentration of component i on stationary phase (mg/mL)
www.jss-journal.com
J. Sep. Sci. 2015, 38, 2229–2237
qi * Q R t t0 tR tp u v Vc w1/2 wi z
equilibrium concentration of component i on stationary phase (mg/mL) flow rate of mobile phase (mL/min) resolution time coordinate retention time of unretained compound (min) retention time of sugar (min) duration time of injection (min) superficial fluid velocity (cm min−1 ) interstitial fluid velocity (cm min−1 ) column volume (cm3 ) peak width at half height (min) peak width component i (min) space coordinate
Liquid Chromatography
2237
Shigoyuki, A., Kurimoto, M., Food. Chem. Toxicol. 2002, 40, 871–898. [10] Birch, G. G., Adv. Carbohydr. Chem. 1963, 18, 201–225. [11] Stewart, L. C., Richtmyer, N. K., Hudson, C. S., J. Am. Chem. Soc. 1950, 329, 8–10. [12] Beste, Y. A., Lisso, M., Wozny, G., Arlt, W., J. Chromatogr. A 2000, 868, 169–188. [13] Bubnik, Z., Pour, V., Gruberova, A., Starhova, H., Hinkova, A., Kadlec, P., J. Food. Eng. 2004, 61, 509–513. [14] Mostafazadeh, A. K., Sarshar, M., Javadian, S., Zarefard, M., Haghighi, Z. A., Sep. Purif. Technol. 2011, 79, 72–78. [15] Al-Eid, S. M., Int. J. Food. Sci. Nutr. 2006, 57, 83–96. [16] Lee, K. N., Korean. J. Chem. Eng. 2003, 20, 532–537. [17] Hansen, S. A., J. Chromatogr. 1975, 107, 224–226.
Greek symbols
[18] Ranganathan, T. V., Kulkarni, P. R., Food. Chem. 2002, 77, 263–265.
␣ T 1 2
[19] Ikegami, T., Horie, K., Saad, N., Hosoya, K., Fiehn, O., Tanaka, N., Anal. Bioanal. Chem. 2008, 391, 2533– 2542.
selectivity of trehalose against maltose, dimensionless bed voidage total column porosity tortuosity factor flow geometry-dependent constant first moment second moment
This work is financially supported by the National Natural Science Foundation of China (NO. 31371725) and National High Technology Research and Development Program of China (SS2014AA021202).
The authors have declared no conflict of interest.
5 References [1] Chen, C., Desmet, T., VanDer Borght, J., Lin, S. K. C., Soetaert, W., Sep. Purif. Technol. 2012, 96, 161–167. [2] Higashiyama, T., Pure. Appl. Chem. 2002, 74, 1263– 1269. [3] Wu, T. T., Lin, S. C., Shaw, J. F., Process. Biochem. 2011, 46, 1481–1485. [4] Zhang, Y., Zhang, T., Chi, Z., Wang, J. M., Liu, G. L., Chi, Z. M., Carbohyd. Polym. 2010, 80, 13–18.
[20] Ouchemoukh, S., Schweitzer, P., Bachir Bey, M., DjoudadKadji, H., Louaileche, H., Food. Chem. 2010, 121, 561– 568. [21] Pedruzzi, I., Borges da Silva, A. E., Rodrigues, E.A., Sep. Purif. Technol. 2008, 63, 600–611. [22] Caruel, H., Rigal, L., Gaset, A., J. Chromatogr. 1991, 558, 89–104. [23] Landers, J. P., BioEssays 1991, 13, 253–258. [24] Wang, X., Liu, Y., Ching, C. B., Sep. Purif. Technol. 2006, 50, 204–211. [25] Foley, J. P., Analyst 1991, 116, 1275–1279. [26] Bao, Z., Su, B., Ren, Q., J. Sep. Sci. 2008, 31, 16–22. [27] Guiochon, G., Felinger, A., Shirazi, D. G., Katti, A. M., Fundamentals of Preparative and Nonlinear Chromatography, Academic Press, New York 1994. [28] Duan, G., Ching, C., Swarup, S., Chem. Eng. J. 1998, 69, 111–117. [29] Goncalves, C., Carpes, M. J., Correa, C. R., Santana, C., Chem. Eng. J. 2007, 133, 151–156. [30] Yu, H. W., Ching, C. B., Chromatographia 2001, 54, 697– 702. [31] Suzuki, M., Adsorption Engineering, Elsevier, Amsterdam 1990.
[5] Cho, Y. J., Park, O. J., Shin, H. J., Enzyme. Microb. Tech. 2006, 39, 108–113.
[32] Luz, D., Rodrigues, A. K., Silva, F. R., Torres, A. E., Jr, C. L. C., Brito, E., Azevedo, D. C., Bioresour. Technol. 2008, 99, 2455–2465.
[6] Yoshida, M., Nakamura, N., Horikoshi, K., Enzyme. Microb. Tech. 1998, 22, 71–75.
[33] Wu, J., Peng, Q., Arlt, W., Minceva, M., J. Chromatogr. A 2009, 1216, 8793–8805.
[7] Kato, M., J. Mol. Catal. B-Enzym. 1999, 6, 223–233.
[34] Ruthven, D. M., Principles of Adsorption and Adsorption Processes, John Wiley, New York 1984.
[8] Zhou, Y., Yuan, Q., Gao, H. L., Ma, R. Y., J. Mol. Catal. B-Enzym. 2006, 43, 137–141. [9] Richards, A. B., Krakowka, S., Dexter, L. B., Schmid, H., Wolterbeek, A. P. M., Waalkens-Berendsen, D. H.,
C 2015 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
[35] Ramos, J. E., Duarte, T. C., Rodrigues, A. K., Silva Jr., I. J., Cavalcante, Jr., C. L., Azevedo, D. C., J. Food. Eng. 2011, 102, 355–360.
www.jss-journal.com
J. Sep. Sci. 2015, 38, 2229–2237
Dengjun He1,2 Jiachun Zhou1 Quanming Xia1 Lihua Jiang1 Yongjun Qiu1,2 Liming Zhao1,2 1 State
Key Laboratory of Bioreactor Engineering, R&D Center of Separation and Extraction Technology in Fermentation Industry, East China University of Science and Technology, Shanghai, China 2 Shanghai Collaborative Innovation Center for Biomanufacturing Technology (SCIBT), Shanghai, China Received January 7, 2015 Revised April 3, 2015 Accepted April 4, 2015
Research Article
Kinetics and equilibria of the chromatographic separation of maltose and trehalose Trehalose, a nonreducing disaccharide, has been extensively applied to food, cosmetics, and pharmaceutical goods. The resultant solution of trehalose prepared by enzymatic methods includes high amounts of maltose. However, it is quite difficult to separate maltose and trehalose on an industrial scale because of their similar properties. In this paper, a highperformance resin was selected as a stationary phase to separate trehalose and maltose, and the resolution of these sugars was 0.59. The potential of a cation exchange resin was investigated as the stationary phase in separating trehalose and maltose using deionized water as the mobile phase. Based on the equilibrium dispersive model, the axial dispersion coefficients and overall mass transfer coefficients of maltose and trehalose were determined by moment analysis at two different temperatures, 50 and 70⬚C. Other parameters, including the column void and the adsorption isotherms, were also determined and applied to simulate the elution curves of trehalose and maltose. The simulated results matched the experimental data, validating the parameters. The optimized parameters are critical to the chromatographic separation of trehalose and maltose on an industrial scale. Keywords: Adsorption isotherm / Axial dispersion / Mass transfer / Maltose / Trehalose DOI 10.1002/jssc.201500005
1 Introduction Trehalose (␣-D-glucopyranosyl-(1,1)-␣-D-glucopyranoside), a nonreducing disaccharide, is present in many organisms, such as bacteria, fungi, seeds, plants, and fungal spores. It preserves the life of organisms during freezing, heating, dehydration, and exposure to toxic chemicals [1, 2]. As a unique sugar, trehalose has been extensively applied to food, cosmetics, and pharmaceutical goods as, for example, preservatives in unstable foods, humectants in cosmetics, and antifreeze in medicine [3, 4]. Its usefulness has been recognized for decades, but trehalose was not produced on a large scale until 1994 [2]. A number of methods for the enzymatic preparation of trehalose have been reported [5–8]. Reaction of maltooligosyl trehalose synthase and maltooligosyl trehalose trehalohydrolase with starch is a large-scale procedure for the commercial production of trehalose. The reaction of converting starch to trehalose is not completely efficient, and the resulting product Correspondence: Professor Liming Zhao, State Key Laboratory of Bioreactor Engineering, R&D Center of Separation and Extraction Technology in Fermentation Industry, East China University of Science and Technology, Shanghai 200237, China E-mail: [email protected] Fax: +86-021-64250829
Abbreviations: HETP, height equivalent to a theoretical plate; ODE, ordinary differential equation; SMB, simulated moving bed C 2015 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
solution is complex, including both trehalose and maltose. It is technically difficult to utilize the byproduct maltose after extracting a major portion of the trehalose from the mother liquor by crystallization [6, 9–11]. Therefore, the resulting solution containing a large amount of maltose is not useful, economical, or environmentally friendly because of the complex mixture of chemicals and high-energy consumption. In recent years, chromatographic techniques such as simulated moving bed (SMB) chromatography have been successfully applied in the food industry, particularly for sugar mixtures [12–14]. Al-Eid [15] used a chromatographic column filled with cation exchange resin to separate fructose and glucose from date syrup. Lee [16] used a two-section SMB to separate glucose and fructose from an aqueous mixture at a high concentration of 500 mg/mL. Underivatized samples of trehalose and maltose have been successfully separated by many methods, such as thin-layer chromatography [17], high performance thin-layer chromatography [18], HILIC [19], and high-performance anion-exchange chromatography [20]. However, separation of these sugars on a semipreparative column or SMB has not been previously reported. The resins loaded with cations such as Ca2+ , Na+ , or K+ are commercially available, which can form complexes with sugars in aqueous solution [21], and the contribution of sugar-calcium complex formation appears to be directly correlated with the elution of trehalose and maltose in ligand-exchange chromatography [22]. Specifically, the formation of sugar-calcium complexes between the cations immobilized on the ion exchange resin and the sugar www.jss-journal.com
2230
J. Sep. Sci. 2015, 38, 2229–2237
D. He et al.
hydroxyl groups drives separation [22, 23]. According to the above studies, the trehalose-calcium complex is stronger than that of the maltose-calcium complex, and the trehalose is therefore strongly attracted to the calcium on the resin and spends more time in the column. Conversely, maltose is less attracted and spends less time in the column. The kinetics and equilibria of these two commercially valuable products on column are critical to the chromatographic separation of trehalose and maltose on a large scale. The present study investigated the potential of an ion-exchange resin as the stationary phase in the separation of trehalose and maltose on a semipreparative scale. The column packed with this stationary phase was characterized by the bed voidage and total porosity. The axial dispersion coefficients and overall mass transfer coefficients for the chromatographic separation were determined by moment analysis. The adsorption isotherms and equilibrium constants were evaluated by frontal analysis. These optimized parameters were then used to simulate the band profiles, and the simulation and experimental results were compared.
2 Materials and methods
2.1.1 Chromatographic rate model A mathematical model can analyze and explain data from a chromatographic experiment. A series of theories and models have been reported to describe the phase equilibrium and mass transfer kinetics in LC [24–26]. In this paper, the equilibrium dispersive model was used to interpret the chromatographic process. Hence, the mass balance equation for each component in the mobile phase can be expressed as [24]:
1 − gT gT
q i∗ = K i c i
where Ki is the dimensionless equilibrium constant. In this approach, it was assumed in Eq. (2) that the difference between qi * and qi was the driving force of the mass transfer, and the mass transfer rate was in proportion to the driving force. In this linear driving force expression, it was assumed that the dominating resistance in the particle diffusion step could be represented by an overall mass transfer coefficient. 2.1.2 Moment analysis Moment analysis is an effective method to determine the axial dispersion coefficient and mass transfer coefficient from pulse experiments in a chromatographic column [26–29]. The first and second moments can be expressed as follows:
2L 2 = v
(1)
Here, ci and qi denote the average values of the mobile phase and stationary phase concentrations, for respectively, for component i; v represents the velocity of the interstitial mobile phase; DL is the axial dispersion coefficient; g T is the total porosity of column; and t and z are the time and the space coordinates, respectively. The following assumptions were considered in the derivation of Eq. (1), including homogeneous packing, a constant volumetric flow rate, a constant dispersion coefficient, isothermal conditions, and no radial gradients. The sugar concentration on the stationary phase was in proportion to the mass transfer coefficient on the basis of the linear driving force approximation model [25]: ∂q i = ki (q i∗ − q i ) ∂t C 2015 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
(2)
(4)
2 DL 1 − gT 1 − gT K 1+ K + v2 gT gT km
(5)
In the above equations, 1 and 2 are the first and second moments; L is the column length; km is the overall mass transfer coefficient; and K is the equilibrium constant. Thus, the expression of the height equivalent to a theoretical plate (HETP) can be derived: HETP =
∂q i ∂ 2ci ∂c i +v = DL 2 ∂t ∂z ∂z
(3)
1 − gT L K 1+ 1 = v gT
2.1 Theoretical procedures
∂c i + ∂t
where ki is the mass transfer coefficient of component i; and qi * is the equilibrium concentration of the stationary phase with a solution of concentration ci . qi * was related to these mobile phase concentrations through the equilibrium isotherm. In this case, the linear equilibrium isotherm was given as:
2 L 2DL L = 2 = N v 1 −2 1 − gT 1 − gT K 1+ K +2v gT km gT
tR 2 N = 5.545 w1/2
(6)
(7)
where N is the theoretical plate number; w1/2 is the peak width at half height; and tR is the retention time. According to different expressions of the axial dispersion coefficient DL and the mass transfer coefficient km , HETP can be expressed in various ways. 2.1.3 Axial dispersion On the basis of Eq. (6), the axial dispersion and mass transfer coefficients cannot be evaluated by a simple linear regression of their nonlinear relationships with HETP. However, www.jss-journal.com
Liquid Chromatography
J. Sep. Sci. 2015, 38, 2229–2237
2231
Figure 1. Typical chromatogram of maltose and trehalose. Experimental conditions: binary solution, 25 mg/mL; flow rate, 1.0 mL/min; mobile phase, deionized water; temperature, 70⬚C.
Figure 2. HPLC–RID chromatogram of a standard mixture of maltose and trehalose of 20 mg/mL. Experimental conditions: column, Waters SugarPak I (300 mm × 6.5 mm id); mobile phase, distilled water; flow rate, 0.5 mL/min; temperature, 80⬚C; injection volume, 20 L.
the axial dispersion primarily consisted of two mechanisms, molecular diffusion and eddy diffusion. Therefore, the axial dispersion can be expressed as [30]: DL = Dm + v
(8)
where, is the tortuosity factor for a packed column; is a constant dependent on flow geometry; and Dm is the molecular diffusivity. In liquid systems, the contribution of molecular diffusivity to axial dispersion was too small to be mentioned, as previously reported [26, 30] even at low Reynolds numbers [28]. Diffusivities of organic molecules with low molecular weights in water range from 6 × 10 −5 to 9 × 10 −4 cm2 min−1 , but can be smaller with increasing C 2015 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
molecular weight [31]. The molecular diffusivity can thus be ignored, and Eq. (8) can be written as: DL = v
(9)
Using Eq. (9) in Eq. (6), we obtain: HETP = 2 + 2v
1 − gT gT
−2 K 1 − gT 1+ K km gT
(10)
As a result, by calculating the first and second moments of a series of elution curves and plotting HETP versus fluid velocity of the mobile phase, the axial dispersion DL and the overall mass transfer coefficient km can be calculated from the intercept and the slope of the straight line, respectively. www.jss-journal.com
2232
J. Sep. Sci. 2015, 38, 2229–2237
D. He et al.
Figure 3. Plot of retention times of (A) acetone and (B) blue dextran versus the inverse superficial velocity of the mobile phase.
2.2 Materials D-(+)-Maltose monohydrate (purity > 99%), D-(+)-trehalose (purity > 99%), and acetone (purity > 99%) were purchased from Sinopharm Chemical Reagent (Shanghai, China). Blue dextran (purity > 99%) was purchased from Sigma–Aldrich (St. Louis, MO, USA). Deionized water was used as the eluent and solvent in all experiments. Purolite CGC 100 × 8 Ca ion exchange gel-type resin was used as the stationary phase. The resin was a spherical bead, and its polymer structure combined with gel polystyrene, which was cross-linked by divinylbenzene. Sulfonic acid and Ca2+ were the functional group and ionic form, respectively. The resin with a typical mean size of 100 m was used to pack the column.
2.3 Semipreparative chromatography The system used to determine the hydrodynamics and thermodynamic model parameters, i.e. bed porosity, adsorption isotherms, and axial dispersion coefficient, was composed of a single glass column (40 cm × 1.6 cm id) equipped with a jacket, pump (BT100, Huxi, Shanghai, China), recirculation water bath (MP-5H, Yiheng, Shanghai, China), and fraction collector (BSZ-100, Huxi, Shanghai, China).
2.4 Sample analysis Samples were collected at the exit of the column and analyzed by three methods. The mono sugar solutions (samples) of trehalose and maltose at greater than 10 mg/mL were analyzed by a Pocket refractometer (PAL-LOOP, Japan). If samples were below this concentration, the fractions were analyzed on C 2015 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
an analytical HPLC system equipped with a refractive index detector (Series 1200, Agilent, USA) and a Waters Sugar-Pak I column (300 mm × 6.5 mm id). The binary solutions of trehalose and maltose were also analyzed on the HPLC system. The blue dextran and acetone solutions were analyzed by a UV/Vis detector (UV-2000, Unico, Shanghai, China) set at the wavelengths of 630 and 310 nm, respectively.
2.5 Experimental procedures 2.5.1 Bed porosities Moment analysis was applied to measure the bed porosities. The residence time distribution curves were obtained from tracer experiments [32]. The tracers blue dextran (3 mg/mL) and acetone (20% v/v) assessed external (g) and total (g T ) bed porosity, respectively. Pulses of 2.0 mL of tracer were injected into the column under 30⬚C at a flow rate of 1.0–2.0 mL/min. 2.5.2 Adsorption isotherms The frontal analysis method [32, 33] was used to measure the single component adsorption isotherms of trehalose and maltose. The concentrations of trehalose and maltose were both between 25 and 150 mg/mL. The mobile phase flow rate was approximately 1.5 mL/min for both trehalose and maltose. A step series method was used to pump the feed solutions into the column. A known concentration of a single sugar was pumped into the column, and the breakthrough curve was monitored at the exit of the column. When the column became saturated, deionized water was used as a wash until reequilibration was reached. For each maltose (or trehalose) liquid-phase concentration, three experiments were performed, and the average value of the equilibrium www.jss-journal.com
Liquid Chromatography
J. Sep. Sci. 2015, 38, 2229–2237
2233
Figure 4. Adsorption isotherms of maltose and trehalose at 50⬚C (A) and 70⬚C (B).
concentration in the liquid and solid phases was applied to a single compound adsorption isotherm. The experiments were carried out at 50 and 70⬚C. The concentration of the adsorbed phase in equilibrium with the liquid phase concentration C0 was calculated by mass balance as expressed in Eq. (11): q i∗ =
Ai Q − gVc C0 (1 − g)Vc
Ai =
∞
[C0 − Ci (t)]dt − C0 tdead
(11)
(12)
0
In Eqs. (11) and (12), Ai is the mass of solute accumulated inside of the column; Q is the fluid flow rate; Vc is the bed
C 2015 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
volume; and tdead denotes the dead time. tdead consists of the time spent in the extra column volumes of the experimental setup, which includes tubing, connections, and pump heads.
2.5.3 Axial dispersion and mass transfer coefficients In the given semipreparative chromatographic system, the experiments were performed at five different mobile phase flow rates from 0.7 to 2.0 mL/min and at two different temperatures, 50 and 70⬚C. The band profiles were recorded to measure the axial dispersion and the overall mass transfer coefficients.
www.jss-journal.com
2234
J. Sep. Sci. 2015, 38, 2229–2237
D. He et al.
Figure 5. Plot of the HETP of maltose and trehalose versus the interstitial velocity of the mobile phase at 50⬚C (A) and 70⬚C (B).
3 Results and discussion 3.1 Separation of trehalose and maltose Square pulses of solutions of maltose and trehalose mixtures were injected into the chromatographic column at 1.0 mL/min. A typical chromatogram of these mixtures is illustrated schematically in Fig. 1. Trehalose was better retained by this stationary phase, and maltose was less retained. The resolution [25] between trehalose and maltose can be calculated by Eq. (13): R=
tR,T − tR,M (wT + wM )/2
resolution of the mixture was not good enough to completely separate the two sugars, and SMB should therefore be used to separate maltose and trehalose on an industrial scale in future studies. The fractions collected upon exit of the chromatographic column were analyzed by HPLC. Figure 2 shows a typical chromatogram of trehalose and maltose separated on a Waters Sugar-Pak I, in which complete separation of trehaose and maltose was achieved. The retention times were 9.3 and 7.5 min for trehalose and maltose, respectively, representing an elution order identical to that on the Purolite CGC 100 × 8 Ca ion exchange resin.
(13) 3.2 Determination of porosity
where R is the resolution; tR,T and tR,M are the retention times of trehalose and maltose, respectively; and wT and wM are the peak widths of trehalose and maltose, respectively. According to Eq. (13) and Fig. 1, the resolution of trehalose and maltose was 0.59. However, as seen in the chromatogram, the C 2015 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
According to Eq. (14), the external porosity g and the total porosity g T can be evaluated from the first moments of blue dextran and acetone, respectively, because these compounds were not retained on this resin. Blue dextran is a polymer that www.jss-journal.com
Liquid Chromatography
J. Sep. Sci. 2015, 38, 2229–2237 Table 1. Adsorption isotherm parameters determined by frontal analysis
Table 2. Axial dispersion and mass transfer coefficients
T (⬚C) T (⬚C)
Maltose
50 70
R2
Adsorption isotherm
R2
y = 0.203x y = 0.220x
0.999 0.998
y = 0.335x y = 0.341x
0.996 0.993
is too large to pass through the particle pores, but it can move through the interparticle voids of the column. However, acetone is so small that it can penetrate not only the interparticle voids of the column but also the particle pores. = 0 ∞ 0
tc(t)dt c(t)dt
=
gL gVc = Q u
(14)
where represents the residence time; c(t) is the concentration of blue dextran or acetone at the exit of the column; u is the superficial fluid velocity; and Vc and L are the volume and length of the column, respectively. According to Suzuki [31] and Ruthven [34], the bed voidage can be calculated by Eq. (15): g T = 0.45 + 0.55g
Trehalose
gT − g 1−g
(16)
Plotting the retention times of blue dextran and acetone in the column versus L/u, as shown in Fig. 3, shows that the total void fraction and the external porosity can be calculated from the slopes of the straight lines using Eq. (14). The total void fraction g T was 0.84, and the external porosity g was 0.46. The correlation coefficients for acetone and blue dextran were R2 = 0.995 and 0.996, respectively. The internal porosity of the resin g p was 0.70, which was calculated by Eq. (16).
3.3 Determination of the adsorption isotherms The experimental adsorption isotherms of maltose and trehalose, with concentrations varying from 25 mg/mL to 150 mg/mL at a flow rate of 1.5 mL/min are shown in Fig. 4. These isotherms illustrate the relationships between the concentrations of maltose and trehalose in solution and on the stationary phase at 50 and 70⬚C. A linear function was applied to describe this experimental data. The correlation coefficients were all greater than 0.990. According to Fig. 4 and the correlation coefficients, the calculated and experimental data were in good agreement. The parameters of the adsorption isotherms including the linear equations and R-squared values are summarized in Table 1. The K values of trehalose C 2015 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
50 70
DL (cm2 min–1 )
km (min–1 )
DL (cm2 min–1 )
km (min–1 )
0.0237v 0.0262v
1.10 1.84
0.0203v 0.0176v
1.14 1.29
were clearly larger than that of maltose, and their values increased with increasing temperature. Sugar adsorption was promoted by the rise of temperature. Moreover, the resin had a higher affinity for trehalose than for maltose. According to Ramos [35], the higher affinity for trehalose can be calculated by the following equation: ␣=
∗ q T∗ /q M c T0 /c 0M
(17)
Here, ␣ represents the selectivity of trehalose against maltose. The selectivities were 1.65 and 1.55 at temperatures of 50 and 70⬚C, respectively. Selectivity decreased with increasing temperature, which may useful in purifying trehalose and maltose on an industrial scale.
(15)
Hence, the internal porosity of the adsorbent can be calculated by Eq. (16): gp =
Maltose
Trehalose
Adsorption isotherm
∞
2235
3.4 Determination of axial dispersion and mass transfer coefficients Based on Eq. (10), the values of HETP were slightly related to the axial dispersion and overall mass transfer coefficients. The plots of HETP versus the interstitial velocity of the mobile phase for maltose and trehalose in Fig. 5 show that HETP increased linearly with velocity. The axial dispersion coefficients can be calculated from the intercepts of the straight lines and are listed in Table 2. The axial dispersion coefficients for maltose and trehalose were of the same order of magnitude. The overall mass transfer coefficients can be determined from the slopes of the straight lines, and their values are also presented in Table 2. The overall mass transfer coefficients were 1.10 and 1.14 min−1 at 50⬚C and 1.84 and 1.29 min−1 at 70⬚C for maltose and trehalose, respectively. The mass transfer rate on this column was slow according to the magnitudes of the km values. The mass transfer rates of each sugar increased with increasing temperature from 50 to 70⬚C because the liquid phase viscosity decreased with increasing temperature.
3.5 Simulation of the elution profile with optimized parameters The optimized parameters, including the total porosity of the column g T , equilibrium constants K, axial dispersion coefficient DL , and overall mass transfer coefficients km , were used to simulate the elution profiles and verify the effectivewww.jss-journal.com
2236
J. Sep. Sci. 2015, 38, 2229–2237
D. He et al.
Figure 6. Comparison of the simulated and experimental band profiles of maltose and trehalose. Experimental conditions: binary solution, 200 mg/mL; flow rate, 1.0 mL/min; mobile phase, deionized water; temperature, 70⬚C.
ness of these values. The corresponding initial conditions for Eqs. (1) and (2) are given by: c i (t = 0, z) = q (t = 0, z) = 0
(18)
The boundary condition at entrance of the column can be represented as a square pulse: c i (t, z = 0) = c i0 , 0 < t ≤ t p
(19)
c i (t, z = 0) = 0, t > t p
(20)
Here, the isotherm was linear and can be expressed by Eq. (3). The boundary condition at exit of the column can be calculated as follows: ∂c i (t, z = L ) = 0 ∂z
(21)
The method of lines was used to obtain the calculated elution profiles by solving Eqs. (1) and (2) together with the initial and boundary conditions. In this method, the finite difference method discretized the partial differential equations Eqs. (1) and (2) into a set of ordinary differential equations (ODEs). A MATLAB program solved the resultant ODEs of the initial value problems. Figure 6 gives the comparison of the simulated results and experimental data, which matched reasonably well. This validated the simulation parameters. The methods used to measure these parameters and the model itself were also valid and can be used for future simulation and design of operating conditions for preparative chromatography, such as SMB.
C 2015 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
4 Conclusions The chromatographic separation of maltose and trehalose was investigated on a semipreparative chromatographic column packed with an ion exchange resin. The bed voidage, axial dispersion coefficient, and overall mass transfer coefficients were determined through moment analysis. The adsorption isotherms of maltose and trehalose were determined by frontal analysis. The stationary phase exhibited a greater affinity for trehalose over maltose. The equilibrium constants of trehalose at 50 and 70⬚C were larger than those of maltose. The magnitude of the overall mass transfer coefficient indicated that mass transfer on the column was slow. In addition, the optimized parameters were used to simulate elution profiles, and the simulated results echoed the experimental data. These reported results will be used to optimize simulated moving bed chromatography on an industrial scale in future work.
Nomenclature ci ci 0 DL Dm Dp HETP kf km K L N qi
concentration of component i in mobile phase (mg/mL) concentration of inlet (mg/mL) axial dispersion coefficient (cm2 min−1 ) molecular diffusion coefficient (cm2 min−1 ) pore diffusion coefficient (cm2 min−1 ) height equivalent to a theoretical plate (cm) external film mass transfer coefficient (cm min−1 ) overall mass transfer coefficient (min−1 ) equilibrium constant column length (cm) theoretical plate number concentration of component i on stationary phase (mg/mL)
www.jss-journal.com
J. Sep. Sci. 2015, 38, 2229–2237
qi * Q R t t0 tR tp u v Vc w1/2 wi z
equilibrium concentration of component i on stationary phase (mg/mL) flow rate of mobile phase (mL/min) resolution time coordinate retention time of unretained compound (min) retention time of sugar (min) duration time of injection (min) superficial fluid velocity (cm min−1 ) interstitial fluid velocity (cm min−1 ) column volume (cm3 ) peak width at half height (min) peak width component i (min) space coordinate
Liquid Chromatography
2237
Shigoyuki, A., Kurimoto, M., Food. Chem. Toxicol. 2002, 40, 871–898. [10] Birch, G. G., Adv. Carbohydr. Chem. 1963, 18, 201–225. [11] Stewart, L. C., Richtmyer, N. K., Hudson, C. S., J. Am. Chem. Soc. 1950, 329, 8–10. [12] Beste, Y. A., Lisso, M., Wozny, G., Arlt, W., J. Chromatogr. A 2000, 868, 169–188. [13] Bubnik, Z., Pour, V., Gruberova, A., Starhova, H., Hinkova, A., Kadlec, P., J. Food. Eng. 2004, 61, 509–513. [14] Mostafazadeh, A. K., Sarshar, M., Javadian, S., Zarefard, M., Haghighi, Z. A., Sep. Purif. Technol. 2011, 79, 72–78. [15] Al-Eid, S. M., Int. J. Food. Sci. Nutr. 2006, 57, 83–96. [16] Lee, K. N., Korean. J. Chem. Eng. 2003, 20, 532–537. [17] Hansen, S. A., J. Chromatogr. 1975, 107, 224–226.
Greek symbols
[18] Ranganathan, T. V., Kulkarni, P. R., Food. Chem. 2002, 77, 263–265.
␣ T 1 2
[19] Ikegami, T., Horie, K., Saad, N., Hosoya, K., Fiehn, O., Tanaka, N., Anal. Bioanal. Chem. 2008, 391, 2533– 2542.
selectivity of trehalose against maltose, dimensionless bed voidage total column porosity tortuosity factor flow geometry-dependent constant first moment second moment
This work is financially supported by the National Natural Science Foundation of China (NO. 31371725) and National High Technology Research and Development Program of China (SS2014AA021202).
The authors have declared no conflict of interest.
5 References [1] Chen, C., Desmet, T., VanDer Borght, J., Lin, S. K. C., Soetaert, W., Sep. Purif. Technol. 2012, 96, 161–167. [2] Higashiyama, T., Pure. Appl. Chem. 2002, 74, 1263– 1269. [3] Wu, T. T., Lin, S. C., Shaw, J. F., Process. Biochem. 2011, 46, 1481–1485. [4] Zhang, Y., Zhang, T., Chi, Z., Wang, J. M., Liu, G. L., Chi, Z. M., Carbohyd. Polym. 2010, 80, 13–18.
[20] Ouchemoukh, S., Schweitzer, P., Bachir Bey, M., DjoudadKadji, H., Louaileche, H., Food. Chem. 2010, 121, 561– 568. [21] Pedruzzi, I., Borges da Silva, A. E., Rodrigues, E.A., Sep. Purif. Technol. 2008, 63, 600–611. [22] Caruel, H., Rigal, L., Gaset, A., J. Chromatogr. 1991, 558, 89–104. [23] Landers, J. P., BioEssays 1991, 13, 253–258. [24] Wang, X., Liu, Y., Ching, C. B., Sep. Purif. Technol. 2006, 50, 204–211. [25] Foley, J. P., Analyst 1991, 116, 1275–1279. [26] Bao, Z., Su, B., Ren, Q., J. Sep. Sci. 2008, 31, 16–22. [27] Guiochon, G., Felinger, A., Shirazi, D. G., Katti, A. M., Fundamentals of Preparative and Nonlinear Chromatography, Academic Press, New York 1994. [28] Duan, G., Ching, C., Swarup, S., Chem. Eng. J. 1998, 69, 111–117. [29] Goncalves, C., Carpes, M. J., Correa, C. R., Santana, C., Chem. Eng. J. 2007, 133, 151–156. [30] Yu, H. W., Ching, C. B., Chromatographia 2001, 54, 697– 702. [31] Suzuki, M., Adsorption Engineering, Elsevier, Amsterdam 1990.
[5] Cho, Y. J., Park, O. J., Shin, H. J., Enzyme. Microb. Tech. 2006, 39, 108–113.
[32] Luz, D., Rodrigues, A. K., Silva, F. R., Torres, A. E., Jr, C. L. C., Brito, E., Azevedo, D. C., Bioresour. Technol. 2008, 99, 2455–2465.
[6] Yoshida, M., Nakamura, N., Horikoshi, K., Enzyme. Microb. Tech. 1998, 22, 71–75.
[33] Wu, J., Peng, Q., Arlt, W., Minceva, M., J. Chromatogr. A 2009, 1216, 8793–8805.
[7] Kato, M., J. Mol. Catal. B-Enzym. 1999, 6, 223–233.
[34] Ruthven, D. M., Principles of Adsorption and Adsorption Processes, John Wiley, New York 1984.
[8] Zhou, Y., Yuan, Q., Gao, H. L., Ma, R. Y., J. Mol. Catal. B-Enzym. 2006, 43, 137–141. [9] Richards, A. B., Krakowka, S., Dexter, L. B., Schmid, H., Wolterbeek, A. P. M., Waalkens-Berendsen, D. H.,
C 2015 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
[35] Ramos, J. E., Duarte, T. C., Rodrigues, A. K., Silva Jr., I. J., Cavalcante, Jr., C. L., Azevedo, D. C., J. Food. Eng. 2011, 102, 355–360.
www.jss-journal.com
