J. Phys. Chem. B 2009, 113, 4379–4383
4379
Mutarotation in D-Fructose Melt Monitored by Dielectric Spectroscopy P. Wlodarczyk,* K. Kaminski, M. Paluch, and J. Ziolo Institute of Physics, Silesian UniVersity, ul. Uniwersytecka 4, 40-007 Katowice, Poland ReceiVed: August 12, 2008; ReVised Manuscript ReceiVed: December 17, 2008
Broadband dielectric spectroscopy was employed to investigate the kinetics of the mutarotation of D-fructose. In this article, D-fructose tautomerization behavior in pure melt is revealed. By monitoring structural relaxation time shift, we were able to determine speed in temperature range 303-333 K and the activation energy for this phenomenon. It is equal to 107 kJ/mol, which is very high in comparison to the solvated system. We have also shown the influence of this process to the γ-relaxation (secondary relaxation process). Dielectric spectroscopy derived complex information about mutarotational dynamics. Introduction Sugars are one of the most important classes of organic compounds widespread in nature. They play key role in many biological processes and biochemical reactions.1,2 Fructose is widely used in food industry as a sweetener. It is main ingredient of high-fructose corn syrup and the monosugar unit of sucrose. It has relatively high sweetness index. Many scientists focused their attention on understanding both chemistry and physics of carbohydrates. Consequently many various experiments have been performed.3-10 D-Fructose has been extensively studied by Tombari’s group by means of dynamic specific heat calorimetry and normal calorimetry11,12 Although sugars are one of the most investigated groups of compounds, there is still a number of fundamental issues that need to be clarified. Herein we address one of the most known phenomena occurring in sugars, that is, mutarotation. One can find information about this process even in chemistry text books for students.13Mutarotation of the carbohydrates has been mainly studied in solution. Sugar concentration in such solution is usually very small, and sugar-solvent interactions dominate over sugar-sugar ones. In this article, we report for the first time detailed experimental studies of mutarotation in pure, anhydrous, supercooled fructose. The mutarotation was studied by means of broadband dielectric spectroscopy, which turns out to be a useful tool for monitoring this process. One can be aware that polarimetry is a routine technique for monitoring mutarotation; however dielectric spectroscopy can be an alternative or supplementary method. It is sensitive to change of viscosity (this will be explained later) instead of optical activity, thus one can obtain additional informations about this phenomenon. One of the advantages of this method is that it can be easily adapted to the high pressure experiments. In general, monosugars can be found in four different cyclic forms: R-pyranose, β-pyranose, R-furanose, β-furanose. Pyranoses form six-membered rings, while furanoses form fivemembered ones. R and β tautomers differs by position (up or down) of the hydroxyl group attached to the first carbon in the ring. Mutarotation can be regarded as a process of transformation from one tautomer to another with the open ring (chain) stage.14 Anhydrous fructose occurs in β-pyranose form in crystalline state. By melting crystals, mutarotation is induced and β-pyra* To whom correspondence should be addressed.
nose begins transformation to R-pyranose, β-furanose, and R-furanose forms. A percentage of each form in equilibrium depends on temperature and solvent (environment). Each form has a different energy, so the Boltzmann distribution describes populations of tautomers. The solvent also has a large influence on the stability of the particular tautomer.15 Polarity of the solvent (which can be described by dielectric constant r) has the biggest influence on the cyclic forms population. For example, R-fructose in pyridine (εr ) 1.5) consists of 7% R-pyranose and 5% chain (keto) form, while in water (εr ) 80) D-pyranose and keto tautomers are undetectable. Dielectric spectroscopy is commonly employed to measure relaxation processes observed in glasses and liquids. In supercooled liquid phase, the structural relaxation of molecules is observed. It is the called R-process, and it determines liquid-glass transition. Structural relaxation time of system is directly related to its viscosity. When supercooled liquid is approaching glassy state on cooling, viscosity of the system increases and R-process is shifting toward lower frequencies. When relaxation time of structural relaxation approaches 100 s, sample undergoes glass transition. Every tautomer has its unique physical properties, and if so, mutarotational dynamics leads to viscosity fluctuations in the system. In consequence, one can expect change of R-relaxation time at constant temperature during equilibrating of the sample. Thus it should be possible to monitor mutarotation by means of the dielectric spectroscopy and describe kinetics of the whole process by performing isothermal time measurements. In literature, there are reports that dielectric spectroscopy could be sensitive to mutarotation;16 however there is no dielectric measurements describing the whole process. Experimental Anhydrous D-fructose (98% purity) was purchased from sigma-Aldrich and used as received. To avoid caramelization of the D-fructose, the melting process was performed very quickly (less than 0.5 min). Isothermal dielectric measurements on D-fructose were carried out using a Novo-Control GMBH Alpha dielectric spectrometer (10-2-107 Hz). The temperature was controlled by the Quattro system, employing a nitrogen gas cryostat, and temperature stability of the sample was achieved better than 0.1 K · s-1. Results and discussion Herein, we performed time dependent measurements of in several temperatures. As a result it was found that
D-fructose
10.1021/jp8095902 CCC: $40.75 2009 American Chemical Society Published on Web 03/02/2009
4380 J. Phys. Chem. B, Vol. 113, No. 13, 2009
Wlodarczyk et al.
Figure 1. Dieletric loss spectra time evolution at two different temperatures. The black line represents initial state and the red one final state.
dielectric spectra evolve with time in accordance with our expectation. Time evolution of structural relaxation times has been traced by the broadband dielectric spectroscopy. Rates of R-relaxation time changes correspond to the rates of mutarotation. From the data, we were able to construct kinetic curves that exhibit unexpected, sigmoidal shape. Moreover, we determined speed and activation energy of this process in the anhydrous environment. Figure 1 shows the dielectric loss spectra evolving in time at two different temperatures (only initial and final states are plotted). These measurements in two different temperatures have been performed independently from each other (two different samples). One can see that at 313 K the maximum of loss of the R-relaxation peak shifts toward lower frequencies more than 2 decades and that is a significant change. At 303 K, the peak shifts only 0.4 decade. One can conclude that the viscosity is increasing while mutarotation proceeds. We are certain that newly formed furanoses have material impact on system viscosity. Anomers transformation cannot be monitored by means of dielectric spectroscopy, because anomeric forms should have similar viscosity. Five-membered rings in equilibrium mixture are responsible for increased viscosity and consequently higher glass transition temperature (Tg). We studied this process at five different temperatures in the temperature range 303-333 K. Both speed of the process at all temperatures and the activation energy have been determined. By ploting relaxation time of the R process τR versus time t one can construct a kinetic curve. We have constructed curves for every temperature. We were able to compare those curves on one graph for all five temperatures by linear rescaling relaxation times to the 0-1 scale. It has been achieved by applying following procedure
R)
τ - τp τk - τp
(1)
For every temperature, τp is the first measured relaxation time (12 ( 3 min after melting), and τk is the last measured relaxation time in the equilibrium state. R-Parameter can be treated as a degree of conversion of the β-D-fructopyranose. Rescaled curves for all studied temperatures are displayed in Figure 2. At first sight, the curve at 313 K has clearly a sigmoidal shape. Until
Figure 2. Kinetic curves representing relaxation times dependence on time. Relaxation times are rescaled to the 0-1 scale.
now there has been no evidence of such behavior in previous studies of mutarotation. The sigmoidal shape of kinetic curve may imply that process has autocatalytic character. It is worth noting that in diluted solution, tautomerization is very fast, and it is impossible to monitor the beginning stage of this process. This process has tremendous speed in water solution, while in our pure system the transformation is relatively slow (see Table 1 for comparison). When the process is so slow, and it needs few days to be terminated, sigmoidal shape is fairly easy to be detected. There is also another possible explanation – such behavior occurs only for anhydrous, supercooled carbohydrates near glassy state. Sugars’ kinetic curves, representing the time dependence of change of the concentration, are usually fitted to the first-order kinetics (eq 2) equation. In our case, structural relaxation times are connected to the tautomers’ concentration fluctuations. Thus, the experimentally observed relaxation time of the R-process can be considered as an equivalent of concentration in this equation. Data from our measurements for all studied temperatures follow satisfactory the exponential dependence
τk - τ ) exp(-kt)
(2)
However, the above dependence is not valid at the early stage of the process where sigmoidal shape is manifested. Thus, at the lowest temperature (303 K) the fitting procedure by means of eq 1 was performed in a limited range of time. Recently, ultrasonic studies of fructose water solution were carried out by R. Behrends and U. Kaatze.17 They deduced that the ultrasonic method probes only one pathway of tautomer equilibration (6-membered pyranose to 5-membered furanose ring), while rotary polarization study probes all possible pathways. Thus the important question arises immediately: which pathway is monitored by dielectric spectroscopy? In order to answer this question, it should be noted that only pyranose to furanose transformation is in fact related to the viscosity change (i.e., pyranose and furanose have different glass transition temperature). Thus only furanose concentration fluctuation is manifested by change in R-relaxation time. Since pyranosefuranose transformation covers the change of β-pyranose, both to β-furanose and R-furanose, k can be viewed as a sum18 of these two tautomerizations. First-order fits and the temperature dependence of rate constant logarithm are presented in Figure 3. The energy
Mutarotation in D-Fructose Melt
J. Phys. Chem. B, Vol. 113, No. 13, 2009 4381
TABLE 1: Kinetics Data from First-Order Fitsa temperature (K) 298 303 313 318 323 328 333 a
reference rate constant (ref 19) from mixture water–ethanol 1:9 (s-1) 4.33e-4 1.03e-3 1.47e-3
measured rate constant (s-1)
fitted rate constant (s-1)
half-life time (min)
time of 99.2% reaction progress (hrs)
1.20e-5 4.57e-5 1.12e-4 2.10e-4 2.66e-4 5.72e-4
7.03e-6 1.39e-5 5.05e-5 9.41e-5 1.70e-4 3.03e-4 5.32e-4
1643 831.1 228.8 122.8 67.96 38.13 21.72
191.7 97.0 26.7 14.3 7.9 4.4 2.5
Time of 99.2% reaction progress (7× half-life time). Data for 298 K is from extrapolation.
Figure 3. First-order kinetics fits for different temperatures. Inset diagram represents temperature dependence of rate constant logarithm.
Figure 4. Fits to the Vogel–Fulcher–Tamman equation, describing two carbohydrate systems. β-fructopyranose and pyranose-furanose mixture. Angell plot is presented in the inset.
activation derived from this graph (slope index) is 107.0 ( 5.7 kJ/mol. Rate constants for different temperatures are summarized in Table 1. For the first-order kinetics, half-life of the process is given by the simple relation and depends only on rate constant k.
t1/2 )
ln(2) k
(3)
Figure 5. Influence of mutarotation on the γ-relaxation.
In Table 1, half-life times derived from first-order kinetics analysis are gathered. In addition we also included the time of 99.2% tautomerization progress. The mutarotation mechanism is not fully understood, but as far as we know the main stage of the reaction is proton transfer process correlated with ring opening. A review of the mechanisms is presented in recent theoretical work.19 In the presence of water or other polar molecules, the proton transfer is intermolecular, so these polar media catalyze reaction. Energy activation barrier derived from theoretical studies for glucose in the water environment is about 100 kJ/mol. In the absence of water, energy is two times higher. Calculations show that ring flattening occurs when transformation follows the noncatalyzed mechanism, which is the reason of very high energy barrier (>200 kJ/mol). Intramolecular origin of the process could occur for example in the gas phase, where there is no possibility for intermolecular proton transfer. Fructose has been already studied in the aqueous and ethanolic solutions.19 There were experimentally derived rate constants and energy barriers from such systems. Mutarotation activation energy, which is equal to Ea ) 53 ( 3 kJ/mol, was found. From our experiment, we determined that the activation energy barrier of mutarotation in anhydrous fructose to be equal to Ea ) 107.0 ( 5.7 kJ/mol. Thus the energy is two times higher than in the aqueous or ethanolic environment. In addition, rate constants differ significantly in comparison to our system (see Table 1). The process in water and ethanol-water mixture is much more rapid (in water, this process is five times faster than in water-ethanol 1:9 mixture). It can be concluded that the origin of the mechanism in molten fructose is not the same as in the polar liquid solutions. Furanoses and pyranoses have different physical properties, such as glass transition temperature. In Figure 4 we show two
4382 J. Phys. Chem. B, Vol. 113, No. 13, 2009
Wlodarczyk et al.
temperature dependences of τR. The curve labeled A refers to the sample quickly quenched after melting, and the second curve (labeled B) refers to the sample equilibrated at ambient temperature for 8 days. Both dependencies were described by means of Vogel-Fulcher-Tamman equation (eq 4)
log(τ) ) log(τ0) +
DTk T - Tk
(4)
The glass transition temperature was calculated using following definition: Tg ) T (τR ) 100 s). There is 10 K of difference in the value of Tg between two studied samples. The same change in Tg has been already reported from the differential scanning calorimetry experiment.20 In addition, we determined the fragility defined by the steepness index, m, as
m)
d[log(τR)] Tg T)Tg d T
()
(5)
Steepness index describes the deviation of temperature dependence of the relaxation time τR from Arrhenius behavior. Substances are classified as strong when this dependence is close to the Arrhenius behavior. Therefore their steepness index is relatively small. Compounds revealing a large deviation from the Arrhenius equation are classified as fragile, and their steepness index is high. The steepness index for curve A is m ) 94, and for curve B it is m ) 98. These values are almost equal, thus there is no change in fragility during mutarotation. We have already presented how this phenomenon affects the structural relaxation, and how the kinetics of the process can be obtained. However, structural relaxation (primary) is not the only process that can be monitor by dielectric spectroscopy. By using broadband dielectric spectroscopy, one can derive information about secondary relaxations. Therefore, we have studied also influence of mutarotation on the secondary processes. Secondary relaxations can be classified as Johari-Goldstein (JG) type or non-JG. It is stated that Johari-Goldstein secondary mode is a process that is universal for any glass-forming liquid and it mimics structural relaxation. This is in fact a precursor of R-relaxation. This phenomenon is always pressure sensitive. Non-JG type of processes is usually pressure independent. The origin of these relaxations is local. Monosugars have two secondary relaxations.13 The β-process at ambient pressure is hidden under the structural relaxation and is visible as the socalled excess wing. This process has been classified as JG. The faster γ-relaxation is well separated. The origin of this γ-relaxtaion in sugars has not been fully revealed, but it is classified as a local pressure-independent process (non-JG). The energy activation of this process is quite high (∼44 kJ/mol in fructose), thus it is very likely that this process is connected with the sugar ring conformational transformation. This transformation is known as chair-boat for pyranose or envelope-twist for furanose ring. One can be certain that the furanose ring is more flexible and it has more internal degrees of freedom than the pyranose one. This is the reason why the change of the γ-relaxation character is expected during mutarotation. To study this possible effect, we have performed two measurements of this process at one temperature in the glassy state for the quenched and equilibrated sample. The first measurement has been performed instantly and the second one has been done after two days of equilibrating sample at 303 K. The comparison of spectra has
TABLE 2: Havriliak-Negami Fit Parameters for Spectra at 243 K and Activation Energy for the γ-Process for Quickly Quenched and Equilibrated Sample
R β ∆ε τhn τmax (s) Ea (kJ/mol)
sample 1 ∼ pyranose
sample 2 (pyranose-furanose equilibrium mixture)
0.516 ( 0.009 0.325 ( 0.016 0.856 ( 0.009 (3.68 ( 0.27) e-4 7.42 e-6 44.7 ( 0.1
0.510 ( 0.007 0.329 ( 0.013 1.308 ( 0.012 (2.89 ( 0.18) e-4 5.80 e-6 44.3 ( 0.2
been done at 243 K. These measurements have shown that relaxation time τ is slightly shorter, while the intensity of the peak has increased (Figure 5). γ-Relaxation in sugars is asymmetric, thus spectra from Figure 5 have been fitted to the Havriliak-Negami function. Fit parameters have been collected in the Table 2. One can notice, that mutarotation has caused significant increase of γ-relaxation’s dielectric strength. It rises from 0.86 to 1.31 (by 52%). Activation energy, along with the relaxation times and shape of curve, are almost the same. One can find interesting that γ-process is also visible in di- and polysaccharides, which probably has the same origin. Disaccharides, which have in their structure one furanose ring, that is, lactulose, sucrose, have larger dielectric strength of γ-process, than disaccharides, which are build of two pyranose units. This is evidence that in disaccharides as well as in monosugars that the presence of a flexible furanose ring in the structure is responsible for higher dielectric strength of this relaxation process. The next interesting aspect is the correlation between R-relaxation and γ-relaxation. As one can see, the fast secondary process is not sensitive to the R-relaxation shift, caused by the mutarotation. The γ-relaxation is evidently not correlated with the structural relaxation. This is the next clear evidence of γ-relaxation locality. Conclusions In summary, we were able to determine the speed of mutarotation in the range of temperatures 303-333 K and activation energy of this phenomenon as well. Activation energy equal to Ea ) (107.0 ( 5.7) kJ/mol has been obtained. This value is two times higher than for the fructose in polar solution (Ea-polar ) 53.0 ( 3.0 kJ/mol). Speed is characterized by the rate constant k. In aqueous solution, mutarotation is very rapid, while in anhydrous fructose this process is several dozen times slower. This is evidence of different mutarotation character of tautomerization in melt than in the polar solvent sugars’ solutions. Moreover, dielectric studies revealed unexpected, sigmoidal shape of kinetic curve, suggesting that this process may be autocatalytic. Tg of furanose formed in the mutarotation process proved to be much higher than pyranose (equilibrium mixture has 10 K higher Tg than pyranose). This can be the reason of different Tg’s reported for sugars. Moreover, we have noticed that dielectric strength of D-fructose γ-relaxation rises greatly, while shape parameters and energy activation stay the same. Dielectric spectroscopy proved to be an excellent tool for monitoring mutarotation in anhydrous carbohydrates. Along with the kinetics data derived from the structural relaxation evolution in time, we get information about dynamics in the glass. Acknowledgment. The authors deeply appreciate the financial support of the Foundation for Polish Science within the
Mutarotation in D-Fructose Melt framework of the TEAM programme received from the European Union Structural Funds in Poland within the framework of the Innovative Economy Operational Programme. K.K. acknowledges financial assistance from FNP (2008). References and Notes (1) Sears, P.; Wong, C. H. Angew. Chem., Int. Ed. 1999, 38, 2300. (2) Gabius, H. J. Naturwissenschaften 2000, 87, 108. (3) Kirschner, K. N.; Woods, R. J. Proc. Natl. Acad. Sci. U.S.A. 2001, 98, 10541. (4) Chalikian, T. V. J. Phys. Chem. 1998, 102, 6921. (5) Molinero, V.; Goddard, W. A. Phys. ReV. Lett. 2005, 95, 045701. (6) Weinga¨rtner, H.; Knocks, A.; Boresch, S.; Ho¨chtl, P.; Steinhauser, O. J. Chem. Phys. 2001, 115, 1463. (7) Tran, V. H.; Brady, J. W. Biopolymers 2004, 29, 977. (8) Tyrell, P. M.; Prestegard, J. H. J. Am. Chem. Soc. 1986, 108 (14), 3990–3995. (9) Freedberg, D. I. J. Am. Chem. Soc. 202, 124 (10), 2358. (10) Kaminski, K.; Kaminska, E.; Paluch, M.; Ziolo, J.; Ngai, K. L. J. Phys. Chem. B 2006, 110, 25045.
J. Phys. Chem. B, Vol. 113, No. 13, 2009 4383 (11) Tombari, E.; Salvetti, G.; Ferrari, C.; Johari, G. P. J. Phys. Chem. B 2004, 106, 16877. (12) Tombari, E.; Ferrari, C.; Johari, G. P. J. Chem. Phys. 2008, 129, 054501. (13) Morrison, R. T.; Boyd, R. N.; Boyd, R. K.; Organic Chemistry, 6th ed.; Benjamin Cummings: San Francisco, CA, 1992. (14) Fan, J.; Angell, C. A. Thermochim. Acta 1995, 266, 9. (15) Shallenberger, R. S. Pure Appl. Chem. 1978, 50, 1409–1420. (16) Tombari, E.; Cardelli, C.; Salvetti, G.; Johari, G. P. J. Mol. Struct. 2001, 559, 245–254. (17) Behrends, R.; Kaatze, U. Biophys. Chem. 2004, 111, 89–94. (18) Flood, A. E.; Johns, M. R.; White, E. T. Carbohydr. Res. 1996, 288. (19) Silva, A. M.; da Silva, E. C.; da Silva, C. O. Carbohydr. Res. 2006, 341, 1029–1040. (20) Truong, V.; Bhandari, B. R.; Howes, T.; Adhikari, B. Int. J. Food Sci. Technol. 2004, 39, 56922. (21) Kaminski, K.; Kaminska, E.; Wlodarczyk, P.; Pawlus, S.; Kimla, D.; Kasprzycka, A.; Paluch, M.; Ziolo, J.; Szeja, W.; Ngai, K. L. J. Phys. Chem. B 2008, 112 (40), 12816–12823.
JP8095902
