The True Johari-Goldstein â-Relaxation of Monosaccharides

K. Kaminski, † E. Kaminska,† M. Paluch,*,†,‡ J. Ziolo,† and K. L. Ngai§

Institute of Physics, Silesian UniVersity, ul. Uniwersytecka 4, 40-007 Katowice, Poland, Department of PolymerScience, The UniVersity of Akron, Akron, Ohio 44325-3909, and NaVal Research Laboratory, Washington, DC20375-5320

ReceiVed: July 24, 2006; In Final Form: September 27, 2006

Broadband isothermal dielectric relaxation measurements of anhydrous fructose, glucose, galactose, sorbose,and ribose were made at ambient pressure in their liquidus and glassy states. We found a new secondaryrelaxation in fructose and glucose that is slower than those seen before by others. This new secondary relaxationalso appears in the dielectric spectra of galactose, sorbose, and ribose, and hence it is a general feature of therelaxation dynamics of the monosaccharides. Dielectric measurements at elevated pressure of fructose andribose show that the new secondary relaxation shifts to lower frequencies with applied pressures, mimickingthe behavior of theR-relaxation. In contrast, the faster secondary relaxation remains stationary on applyingpressure. These results together with other inferences indicate that the slower secondary relaxation bearsrelations to theR-relaxation, and hence, it is the true Johari-Goldstein secondary relaxation of themonosaccharides.

1. Introduction

Glucose, fructose, galactose, sorbose, and ribose, membersof the sugar family, are hydrogen-bonded glass-forming organicsubstance. They have important applications in food science,medicine, and biology. For this reason, the molecular relaxationdynamics of some of these sugars, namely, glucose and fructose,have been investigated in their liquidus and glassy states byvarious techniques, including dielectric relaxation1-6 and en-thalpy relaxation.7 Two relaxation mechanisms were observedin glucose and fructose.1-4 The slower one is the usual structuralR-relaxation, and the much faster one is a secondary relaxation.Recent advances in research of the relaxation dynamics of glass-forming substances have shown that a certain class of secondaryrelaxations is universal and these secondary relaxations play afundamental role in glass transition.8-11 They are called theJohari-Goldstein (JG)â-relaxations,12 which have propertiesmimicking theR-relaxation, and not all of these properties areshared by other secondary relaxations of lesser or no importanceto glass transition. A list of distinctive properties that character-ize the JGâ-relaxation has been given before.8 Here we citesome of these properties that are relevant to the present study.The relaxation time of JGâ-relaxation,τJG, is sensitive topressure and is approximately the same as the primitiverelaxation time,τ0, of the coupling model. If more than onesecondary relaxation is observed, then the faster one is not theJGâ-relaxation. This is because the JGâ-relaxation, the ultimateprecursor of theR-relaxation, must involve motion of essentiallythe entire molecule. Hence, it must be the slowest secondaryrelaxation. In view of the fundamental importance of the JGâ-relaxation and the claim of its universal appearance in allglass-formers, it is high time to find out if the previously

observed secondary relaxations in glucose and fructose are JGâ-relaxations or not. Previous dielectric relaxation studies ofanhydrous sugars are limited to glucose and fructose. Themeasurements carried out over limited frequency ranges atambient pressure are not sensitive and detailed enough foranyone to make any conclusion or to detect additional secondaryrelaxation that is hidden. For this reason, we have performedbroadband dielectric measurements over the wide frequencyrange, 10-2 e ν e 109 Hz, on anhydrous glucose and fructoseas well as galactose, sorbose, and ribose in the liquidus andglassy states at ambient pressure. In addition, dielectric relax-ation measurements under applied pressures have been madeon fructose and ribose to investigate the pressure dependenceof the relaxation processes. As shown by the data to bepresented, our detailed experimental data indicate the presenceof a new secondary relaxation slower than the previouslyobserved secondary relaxation in all five sugars. The discoveryof the slower secondary relaxation is sufficient for us to inferthat the faster secondary relaxation is not the JGâ-relaxationand that possibly this slower secondary relaxation is theuniversal JGâ-relaxation of the sugars. These inferences areconfirmed by the significant shift of the slower secondaryrelaxation in fructose and ribose to lower frequencies byapplying pressure. On the other hand, the faster secondaryrelaxation remains stationary, indicating that it is not the JGâ-relaxation. The relations between the true JGâ-relaxation andtheR-relaxation of the sugars are discussed and compared withthe general pattern established before in many other glass-formers. These results are discussed in the context of thecoupling model.

2. Experimental Section

Isothermal dielectric measurements at ambient pressure werecarried out using a Novo-Control GMBH Alpha dielectricspectrometer (10-2-107 Hz) in combination with an Agilentnetwork analyzer (107-109 Hz). The temperature was controlled

* Corresponding author. Office: (+48) 32 3591484. Fax: (+48) 322588431. E-mail: mpaluch@us.edu.pl.

† Silesian University.‡ The University of Akron.§ Naval Research Laboratory.

25045J. Phys. Chem. B2006,110,25045-25049

10.1021/jp064710o CCC: $33.50 © 2006 American Chemical SocietyPublished on Web 11/21/2006

by the Quatro system, employing a nitrogen-gas cryostat, andthe temperature stability achieved for the sample was better than0.1 K. For measurements at high pressures, we used a techniquethat was described in detail in ref 13. The capacitor, filled withtested material, was placed in the high-pressure chamber.Pressure was exerted on the chamber from a pressure generatorthrough silicone oil. The sample capacitor was sealed andmounted inside a Teflon ring to separate the sample from thesilicon oil. Pressure was measured by a Nova Swiss tensometricpressure meter with an accuracy of 0.1 MPa. Temperature wascontrolled to within 0.5 K by means of liquid flow from athermostatic bath. The anhydrous sugar samples of over 98%purity were purchased from Aldrich Chemicals. These com-pounds were crystals and friable powders and were measuredimmediately after receiving them. During the melting pro-cess we heated the samples very slowly to avoid carmelization.The chemical structures of the sugars studied are shown inFigure 1.

3. Results and Discussions

(A) Dielectric Data at Ambient Pressure. The dielectricrelaxation of each sugar was measured at ambient pressure overthe broad frequency range of 10-2 e ν e 109 Hz. Starting at atemperature well above the glass transition temperature,Tg,measurements were made isothermally down to temperaturesdeep in to the glassy state of each sugar. The loss data are shownin Figure 2 for fructose, Figure 3 for glucose, Figure 4 forgalactose, Figure 5 for sorbose, and Figure 6 for ribose. As faras we know, the loss data of galactose, sorbose, and ribose havenot been obtained before. In the equilibrium liquid state, theloss data of fructose and glucose are essentially the same asobtained before by others over much narrower frequencyranges.1,3 Besides the loss from ionic conductivity at the lowestfrequencies, the spectra show two relaxation peaks contributedby the R-relaxation and the secondary relaxation seen beforeby others. However, when temperature is lowered sufficientlybelowTg and theR-loss peak is moved out of the experimentalfrequency window, a shoulder appears in the loss spectra. Thisnew feature indicates the presence of a new secondary relaxation(labeledâ in Figures 2 and 3) located at frequencies intermediatebetween theR-relaxation and the previously found secondaryrelaxation (labeledγ).

The newâ-relaxation is also found in galactose (Figure 4)and sorbose (Figure 5). In galactose and sorbose, as thetemperature is further decreased, the shoulder is transformedto a loss peak, which provides the best evidence of the presenceof theâ-relaxation in the sugars. Their most probableâ-relax-ation times are determined and are shown as a function ofreciprocal temperature in Figure 7 together with theR- and theγ-relaxation times of galactose and sorbose, as well as that offructose, glucose, and ribose. By inspection of Figures 2-5,the loss from theâ-relaxation increases with increasing tem-perature, and this behavior is typical of secondary relaxationsin glass-formers.

Our findings of two secondary relaxations in all five sugarsraise the following question: Which one is the true Johari-Goldstein secondary relaxation according to the criteria givenin ref 8? Since the presently discoveredâ-relaxation is theslower of the two secondary relaxations, it is likely the JGâ-relaxation. However, this criterion may not be enough.We would like to see which secondary relaxation bears acloser relation in properties to theR-relaxation. One propertyused as a criterion for JGâ-relaxation is the pressure sensi-tivity of its relaxation time. For this reason, we carried out

dielectric relaxation measurements under applied pressure atconstant temperature. The results are reported in the followingsubsection.

(B) Dielectric Spectra at Elevated Pressures.We measuredthe dielectric relaxation spectra of fructose at the constanttemperature of 253 K under applied pressures increasing from1000 to 4800 bar. This choice of the temperature and pressureranges makes it possible to see both theâ- and theγ-relaxationsand how they change within the spectral range of our instru-ments. The loss data presented in Figure 8 clearly show thatthe â-relaxation shifts to lower frequencies with increasingapplied pressure from 1000 to 4800 bar. On the other hand, theγ-relaxation time is rather insensitive to pressure.

For ribose, we bring out the difference of the pressuredependence between theâ- and theγ-relaxations in a differentmanner. Measurements are made at 5000 bar for severaltemperature ranges and the loss data (symbols) are shown inFigure 9. At lower temperatures, a shoulder appears in the

Figure 1. The chemical structures of studied sugars. Both glasstransition temperature,Tg, and fragility parameter,m (steepness index),are indicated.

25046 J. Phys. Chem. B, Vol. 110, No. 49, 2006 Kaminski et al.

spectrum, indicating the presence of theâ-relaxation in ribose.The â-relaxation does not show up clearly in the isothermalloss data of ribose taken at ambient pressure (Figure 6). Shownalso in Figure 9 as thick solid lines are the ambient pressure (1bar) data of ribose taken at four temperatures, 265, 267, 269,and 271 K. When comparing the data taken atT ) 269 K andP ) 1 bar with that taken at almost the same temperature butat 5000 bar (∆) in Figure 9, one can see that theâ-relaxationtime is highly sensitive to pressure, while theγ-relaxation isnot.

From the loss data under an applied pressure of the tworepresentatives of the five sugars studied, it is clear that thenewly discoveredâ-relaxation has its relaxation time stronglydependent on pressure mimicking the same property of theR-relaxation. Hence,â-relaxation is the JG secondary relaxationof the sugars. In contrast, the lack of any significant pressuredependence of theγ-relaxation indicates that it has little or noconnection with theR-relaxation and the mechanism leadingto glass transition of the sugars. Figure 9 shows that while therelaxation strengths of theR- and â-relaxation are almostunchanged by elevating pressure and temperature, the relaxation

strength of theγ-relaxation is significantly modified. Sincehydrogen bonds are prevalent in the sugars and the scheme ofhydrogen bondings is sensitive to elevated pressure and tem-perature, the data in Figure 9 suggest that theγ-relaxation mayhave some connection with the hydrogen bonds. This suggestionis further supported by theγ-relaxations of fructose, glucose,galactose, and sorbose being nearly the same (see Figure 7)over a wide temperature range spanning the liquid and glassystates of these substances. Furthermore, the relaxation strengthof the γ-relaxation of glucose is substantially modified onaddition of water,5,6 which is likely due to the modification ofthe hydrogen-bonding schemes of the anhydrous sugars. Attemperatures belowTg, theγ-relaxation time of glucose has theArrhenius dependence with an activation enthalpyEγ equal to42.20 kJ/mol, which is smaller than the 60 kJ/mol reported inref 1. The dielectric glass transition temperature defined byτR-(Tg) )100 s is 309 K. The ratioEγ/RTg is equal to 16 and issmaller than the ratio for Johari-Goldstein secondary relaxationsin other glass-formers.14

Although theR-relaxation is the dominant feature in the lossspectra, the loss contributed by theγ-relaxation is relativelyhigh and increases with increasing temperature. As can be seen

Figure 2. The dielectric loss spectra of fructose at ambient pressure.The dc-conductivity has been subtracted. The spectra were measuredfrom T ) 333 to 225 K with steps of 4 K and fromT ) 218 to 178 Kwith steps of 10 K.

Figure 3. The dielectric loss spectra of glucose at ambient pressure.The dc-conductivity has been subtracted. The spectra were measuredfrom T ) 363 to 328 K with steps of 5 K, fromT ) 325 to 303 K withsteps of 2 K, fromT ) 298 to 273 K with steps of 5 K, and, finally,from T ) 263 to 163 K with steps of 10 K.

Figure 4. The dielectric loss spectra of galactose at ambient pressure.The dc-conductivity has been subtracted. The spectra were measuredfrom T ) 353 to 305 K with steps of 4 K and fromT ) 299 to 175 Kwith steps of 4 K.

Figure 5. The dielectric loss spectra of sorbose at ambient pressure.The dc-conductivity has been subtracted. The spectra were measuredfrom T ) 343 to 195 K with steps of 4 K and fromT ) 188 to 173 Kwith steps of 5 K.

Johari-Goldsteinâ-Relaxation of Monosaccharides J. Phys. Chem. B, Vol. 110, No. 49, 200625047

from any of the spectra of fructose, glucose, galactose, andsorbose in their liquidus states, the height of theγ-loss peakbecomes comparable to that of theR-loss peak and they mergeinto a single peak at higher temperatures. The effect is abroadening of theR-loss peak. Even though theâ-relaxation is

weaker than theγ-relaxation, it also is effective in broadeningthe R-loss peak because theâ-relaxation is closer to theR-relaxation. As best seen from the spectra of galactose (Figure4) and sorbose (Figure 5), theâ-loss maximum increasesmonotonically with temperature while theR-loss maximumremains practically constant. Thus, in the liquidus state, we canexpect theâ-relaxation to contribute part of the loss seen onthe high-frequency flank of theR-loss peak. There is also alarge conductivity contribution on the low-frequency side ofthe R-loss peak. These complications preempt any attempt toextract reliably the loss peak contributed exclusively by theR-relaxation, and hence the true Kohlrausch exponent of theR-relaxation or the coupling parameter of the coupling model.8,9

Furthermore, the lack of distinctive features in the spectraobtained in the liquidus state makes it impossible to locate theâ-relaxation and determine its relaxation time at temperaturesaboveTg. For these reasons, we make no effort here to calculatethe primitive relaxation time of the coupling model and compareit with the experimentalâ-relaxation time.

4. ConclusionIn all past experimental studies of glucose and fructose, only

one secondary relaxation has been found and it is well separatedfrom the R-relaxation at the glass transition temperature. Animportant advance of the present study is the discovery of thepresence of a slower secondary relaxation in these twomonosaccharides as well as in galactose, sorbose, and ribose.Naturally, we call this new secondary relaxation theâ-relaxationand the previously known secondary relaxation theγ-relaxation.Dielectric relaxation measurements of fructose and ribose showthat theâ-relaxation shifts to lower frequency when pressure iselevated, mimicking the pressure dependence of theR-relax-ation. On the contrary, the relaxation time of theγ-relaxationis practically pressure independent. The experimental factstogether with other ancillary considerations lead us to concludethat the â-relaxation is the fundamental Johari-Goldsteinsecondary relaxation of the monosaccharides. Theγ-relaxationpossibly originates from the hydrogen-bonding scheme in thesesubstances, but the current data are not sufficient to make anydefinite conclusion.

Acknowledgment. Financial support of the Committee forScientific Research, Poland KBN, Grant No 1P03B 075 28, isgratefully acknowledged. The work at the Naval ResearchLaboratory was supported by the Office of Naval Research.

Figure 6. The dielectric loss spectra of ribose at ambient pressure.The spectra were measured fromT ) 323 to 251 K with steps of 4 K,from T ) 245 to 173 K with steps of 6 K, and fromT ) 163 to 153K with steps of 10 K.

Figure 7. Temperature dependences of theR-, â- (the true JG), andγ-relaxation processes of studied sugars.

Figure 8. Dielectric γ-relaxation peak for fructose measured atT )253 K and different pressures.

Figure 9. The evolution of the dielectric loss spectra of ribose underhigh pressure (p ) 5000 bar) at different temperatures. Solid curvesare the dielectric spectra of ribose measured at ambient pressure.

25048 J. Phys. Chem. B, Vol. 110, No. 49, 2006 Kaminski et al.

References and Notes

(1) Chan, R. K.; Pathmanathan, K.; Johari, G. P.J. Phys. Chem.1986,90, 6358

(2) Noel, T. R.; Ring, S. G.; Whittam, M. A.J. Phys. Chem.1976, 96,5667.

(3) Tombari, E.; Cardellia, C.; Salvettia, G.; Johari, G. P.J. Mol. Struct.2001, 559 245.

(4) Gangasharan; Murthy, S. S. N.J. Chem. Phys.1993, 99, 9865.(5) Tyagi, M.; Murthy, S. S. N.Carbohydr. Res.2006, 341, 650.(6) Seo, J.-A.; Oh, J.; Kwon, H.-J.; Kim, H. K.; Hwang, Y.-H. InAIP

Conference Proceedings;Tokuyama, M., Maruyama, S., Eds.; 2006; Vol.832, p 37.

(7) Wungtanagorn, R.; Schmidt, S.J. Thermochim. Acta2001, 36995.

(8) Ngai, K. L.; Paluch, M.J. Chem. Phys.2004, 120, 857.(9) Ngai, K. L. J. Phys.: Condens. Matter2003, 15, S1107.

(10) Paluch, M.; Roland, C. M.; Pawlus, S.; Ziolo, J.; Ngai, K. L.Phys.ReV. Lett. 2003, 91, 115701.

(11) Hensel-Bielowka, S.; Paluch, M.; Ngai, K. L.J. Chem. Phys.2005,123, 014502.

(12) Johari, G. P.; Goldstein, M.J. Chem. Phys. 1970, 53, 2372.(13) Urbanowicz, P.; Rzoska, S. J.; Paluch, M.; Sawicki, B.; Szulc, A.;

Ziolo, J. Chem. Phys.1995, 201, 575.(14) Ngai, K. L.; Capaccioli, S.Phys. ReV. E. 2004, 69, 031501.

Johari-Goldsteinâ-Relaxation of Monosaccharides J. Phys. Chem. B, Vol. 110, No. 49, 200625049