Mode coupling behavior in glass-forming liquid crystalline isopentylcyanobiphenyl

Aleksandra Drozd-Rzoska,1 Sylwester J. Rzoska,1 Marian Paluch,1 Sebastian Pawlus,1 Jerzy Zioło,1 Patrick G. Santangelo,2

C. Michael Roland,2 Krzysztof Czupryński,3 and Roman Dąbrowski31Institute of Physics, Silesian University, ul. Uniwersytecka 4, 40-007 Katowice, Poland

2Naval Research Laboratory, Chemistry Division, Code 6120, Washington, D.C. 20375-5342, USA3Military University of Technology, ul. Kaliskiego 2, 00-908 Warsaw, Poland

sReceived 18 November 2003; revised manuscript received 10 September 2004; published 21 January 2005d

Linear and nonlinear dielectric measurements of liquid crystalline chiral isopentylcyanobiphenyls5*CBdand n-pentylcyanobiphenyls5CBd, combined with viscosityhsTd data, are presented. The 5*CB compoundglassifies on cooling in the cholesteric phase whereas 5CB crystallizes in the nematic phase. In both com-pounds the temperature evolution of dielectric relaxation times, the dc conductivity, and the viscosity are welldescribed by the “critical-like” description from mode coupling theorysMCTd. However, for 5*CB a uniquecoincidence of the MCT “critical” temperature and extrapolated temperature of the hypothetical continuousisotropic-cholestericsT* d phase transition was found. The temperature dependence of the strong electric-field-induced changes of the dielectric permittivity exhibits a strong anomaly in the direction of negative values onapproachingT*, not observed up to now. The anomaly is described by the susceptibility-related criticalexponentg=1. The divergence of the “nonlinear” dielectric relaxation follows a power dependence describedby the exponenty=1. This paper recalls the recent discussions on the glassy dynamics of a “hard-ellipsoid”liquid and the possible relationship between the glass transition, critical phenomena, and isotropic-nematictransition.

DOI: 10.1103/PhysRevE.71.011508 PACS numberssd: 64.70.Pf, 64.70.Ja, 64.70.Md

INTRODUCTION

Many studies of the dynamics of glass-forming liquidshave shifted their focus to temperatures considerably higherthan the glass temperatureTglassf1–30g. This challenging em-phasis has been stimulated by mode coupling theorysMCTdf1,11g. One of its important predictions is the existence of adynamic crossover between the ergodic and nonergodic re-gimes at the “critical” or crossover temperatureTMCT=s1.3±0.1dTglassf1–24g:

ta−1,h,d ~ sT − TMCTd−g, s1d

whereta is the structuralsad dielectric relaxation time,h theviscosity, andd the diffusion coefficient. In practice, the ex-ponent 2,g,4, in fair agreement with MCT.

MCT defines the singular temperatureTMCT as the point atwhich the dynamics of a liquid changes from the fluidlikeregime to the solidlike one, via a qualitative change in thecaging phenomenonf1,2,11,13,16–18g. This may be sup-ported by recent studies in benzophenone, where the corre-spondence ofTMCT with the onset of possible nucleation andgrowth of a new phase was notedf24g. It is noteworthy thatthe critical-like relations1d describes the experimental dataonly for T.TMCT+s15–20 Kd f1,2,5,7,8,11,15,17g. How-ever, in experiment the crossover between dynamic regimeson passingTMCT is always gradual. In the full range of tem-peratures thetasTd, hsTd, anddsTd dependences can be por-trayed by the VFT relationf1–23g

ts,h,d ~ expS DT0

T − T0D , s2d

whereT0 is the extrapolated ideal glass temperature andD isthe parameter related to the fragility of the glass former.

The extended MCT took this experimental behavior intoaccount by including the hopping diffusion to restore ergo-cidity even belowTMCT f11,23g. The significance of theMCT “critical temperature” supports the recent suggestionf17g that the value of the relaxation timetC=tasTMCTd<10−7 s may be “magic”—i.e., approximately the same forall glass-forming liquids. However, Casaliniet al. f18,19greasoned that the value of the relaxation timetCsPd=tasP,TMCTd is constant and independent from the pressureandTMCT value, but only for a given class of glass formers.

The original MCT considers spheres and hence focuses ontranslational degrees of freedomf1,11,25g. Recently, the mo-lecular MCT for a model system of hard ellipsoids, aiming atreaching insight into the interplay between translational andorientational degrees of freedom, was proposed by Schillingand co-workersf25–28g. It was found that for moleculeswith an aspect ratio greater than 2.5, an additional crossovertemperatureTMCT

upper=TMCTsorientd appears. It corresponds tothe freezing of the orientational freedom while the transla-tional degree of freedom remains still liquid like. This phe-nomenon was expected to appear near the nematic instabilityboundary—i.e., near the isotropic-nematicsI-Nd transition ina real system of a rodlike liquid crystalline compoundf26–28g. Very recently, the existence ofTMCT

lower andTMCTupperwas

shown in the isotropic phase of few rodlike nematogens withthe I-N transition, includingn-octylcyanobiphenyls8CBdfrom n-cyanobiphenylssnCBd homologous series, based onoptical heterodyne Kerr effectsOHD-OKEd studiesf29,30g.The temperatureTMCT

upper was estimated slightly below theclearing temperature while the extrapolated value ofTMCT

lower

was located already in the solid phase, well below theI-Nand isotropic-solidsI-Sd phase transitions. For 8CB the ob-tained value ofTMCT

lower was in agreement with the earlier esti-

PHYSICAL REVIEW E 71, 011508s2005d

1539-3755/2005/71s1d/011508s7d/$23.00 ©2005 The American Physical Society011508-1

mation based on a multifrequency electric conductivityanalysisf31g. Dielectric studies of isotropicnCB gave alsoclear evidence for canonical features for the “glassy dynam-ics” such as non-Arrhenius evolution relaxation times or thebroadening of dielectric loss curves on coolingf31,32g. Thedielectric relaxation timestad is a magnitude related to thetwo-point susceptibility fx2stdg. Recent simulations in amodel supercooled liquid showed that the four-point suscep-tibility x4sTd and the related relaxation timet4 also exhibit ananomaly on approachingTMCT f33–37g:

x4sTd ~ sT − TMCTd−g4, s3ad

t4sTd ~ sT − TMCTd−y4, s3bd

where numerical analysis gaveg4=0.5–1.2 andy4<1–1.4.These results are particularly noteworthy due to the fact

that x4sTd is a magnitude directly coupled to mesoscale het-erogeneitiesf34–37g. Theoretical analysis in Refs.f33–37genabled the identification of a growing dynamical lengthscale on approachingTMCT, corresponding to fluctuations ofthe order parameterf34g, as a source of anomalies ofx4sTdandt4sTd. The existence of spatially heterogeneous mesocaledynamics is well established experimentally close toTglassf4,11–15,20–23g. There is also an experimental evidence forthe existence of long-lived mesoscale heterogeneities athigher temperaturesf6,11,12g. However, no experimentalstudies aiming at testing relationss3ad and s3bd are beingconducted so far, to the best of the authors knowledge. Theproperties of the isotropic rodlike nematogens are verystrongly influenced by well-defined prenematic fluctuationsin the broad surrounding of the isotropic-nematic clearingtemperaturef29–32,38–43g. Hence their studies may be sig-nificant for a vivid discussion of the role of heterogeneitiesfor supercooled liquids.

Here we discuss the validity of the MCT behavior for aliquid crystalline compound which glassifies on cooling. Re-sults are presented based on a unique combination of “linear”and “nonlinear” dielectric studies together with shear viscos-ity measurements. Studies in 5*CB are compared with simi-lar tests onn-pentylcyanobiphenyls5CBd, a “classical” nem-atogen showing the isotropic-nematic-crystal phasesequence.

EXPERIMENT

Both tested compounds are rodlike mesogens, with a per-manent dipole moment parallel to the longsmaind axis of themolecule. 5*CB remains isotropic down toTI-Ch<250 K, atwhich the phase transition to the chiral nematicfcholesteric:N*, Chg phase occurs. On further cooling, the glassy state isattained at Tg<180 K f38g. 5*CB is the isomer ofn-pentylcyanobiphenyl. The latter forms on cooling the nem-atic phase atTI-N<308 K and next crystallizes atTN-S<280 K f32,39g. Structures of both compounds are shown inthe inset in Fig. 1. A comprehensive discussion of broadbanddielectric relaxation in 5*CB, both as a function of tempera-ture and pressure, is given in Ref.f38g. A similar discussionfor isotropic 5CB is given in Refs.f32,40g. For 5CB, this

paper is based on data from Ref.f32g. For 5*CB novel BDSmeasurements were conducted, extending the range of testedtemperatures in the isotropic phase up toT<TI-Ch+150 K, incomparison with Ref.f38g. Dielectric measurements werecarried out using the Novocontrol BDS 80 spectrometer withQuattro temperature control unit, the same as applied inRefs. f32,38g. Steady-state and dynamic shear viscositieswere measured using a Bohlin VOR rheometer with theparallel-plate geometry. Sample radii ranged from3 to 20 mm with gaps ca. 1.5 mm. Viscosity data for the fivehighest temperatures in isotropic 5*CB were obtained usingthe Ubbelohde-type viscometer with the capillary of diam-eter of 0.6 mm and 120 mm longf41g. The temperature sta-bility was at least ±0.1 K. For all measured temperatures, thesteady-state and dynamic shear viscosities were equivalent.Nonlinear dielectric measurements are related to the shift ofdielectric permittivity induced by a strong electric fieldf32,42g:

«E = « + D«EE2 + ¯ , s4d

where« and «E are respective permittivities in a weak andstrong electric fieldE. The measure of the nonlinearity,known as the nonlinear dielectric effectsNDEd, is defined asENDE=D«E/E2.

In experiments the changes of electric capacitanceDCE,related toD«E, are registered:DCE=1–10 fF for asample ofC=100 pF. This has to be detected with three-digit resolu-tion during the action of a pulse of a strong electric fieldlasting DtD=1–4 ms. Adetailed description of the appliedNDE spectrometer is given in Ref.f42g. It is based on thedual-field, single-generator, aperiodic principle withfrequency-time direct measurements supported by the modu-lation domain analyzer HP 53 310A. For NDE measurementssamples were placed in the flat-parallel capacitor made onlyfrom Invar and quartzsgap 0.5 mm, diameter 16 mmd. Thetested compounds were synthesized by Dąbrowski and

FIG. 1. The temperature evolution of shear viscosityhsTd in5*CB. The dashed and solid curves are portrayed by relation VFTs2d and MCT dependences5d, respectively. The inset shows struc-tures of 5CB and 5*CB. Fitted parameters for the MCT dependenceare given in the figure.

DROZD-RZOSKAet al. PHYSICAL REVIEW E 71, 011508s2005d

011508-2

Czupryński at the Military Technical University, Warsaw, Po-land. Samples were carefully degassed immediately prior tomeasurements.

RESULTS AND DISCUSSION

Figures 1–3 present Arrhenius plots of the shear viscosity,dielectric relaxation times, and dc conductivity in 5*CB, re-spectively. In each case the VFT dependence, with approxi-mately the same values ofT0 andD parameters, can portrayexperimental data:

hVFT = 355 ± 10smPd, D = 5 ± 0.2,

T0 = 186 ± 5 sKd for hsTd,

tVFT = 17 ± 1 spsd, D = 4.7 ± 0.2,

T0 = 186 ± 3 sKd for tasTd,

sVFT = 84.6 ± 0.3skS m−1d, D = 5.1 ± 0.2,

T0 = 187 ± 3 sKd for ssTd.

For tasTd the relaxation is the cholesteric phase is clearlymanifested by two processes, described bytVFT

uppersTd andtVFT

lowersTd relaxation times, which can be portrayed by VFTrelations with the following parameters:

tVFTupper= 18.5 ± 1 ps, D = 5.5 ± 1, T0 = 182 ± 5 K,

tVFTlower = 20 ± 1 ps, D = 4.8 ± 1, T0 = 186 ± 5 K.

The behavior of the dielectric relaxation time is in agreementwith results discussed in Ref.f38g. The small differencesmay be explained by the broader range of temperature andsignificant increase of the number of tested temperatures inthese studies. It is noteworthy that forhsTd, tasTd, andssTdexperimental dependences passing of theI-Ch transition—i.e., the change in the symmetry of the system—seems tohave no influence onT0 andD values. The analysis presentedin Figs. 1–3 shows that all the properties mentioned abovecan also be portrayed by the MCT “critical” relations1d:namely,

s−1sTd,tasTd,hsTd ~ sT − TMCTd−g, T . TMCT + 15 K.

s5d

Values of the fitted parameters are given in Figs. 1–3. To thebest of the authors’s knowledge this is the first evidenceshowing the validity of the MCT “critical relation also forthe dc conductivity. Although the single VFT dependenceseems to deliver a fair description of experimental data, thedistortion-sensitive analysis of experimental data proposedby Stickel et al. f43g confirms this only for the dc conduc-tivity as shown in Figs. 4 and 5. The analysis presented inFig. 4 reveals two VFT processes fortasTd: starting fromTB<TI-Ch+110 K=360 K, the fragility coefficient changesby ca. 20%. Moreover, a strong pretransitional distortion inthe immediate vicinity of theI-N transition occurs. The in-sets in Figs. 4 and 5 present a log-log-plot-based analysis forisotropic 5*CB, to test further the validity of the MCTcritical-like descriptionfrelation s5dg. In the opinion of theauthors, such a plot, which is a standard for critical phenom-enaf44g, is much more sensitive for possible distortions thanthe stad1/gsTd analysis, most often used in the physics ofglassy liquidsf2,11g. It is visible that fortasTd relations5d isvalid from T1<TI-Ch+20 K=270 K to T2<TI-Ch+110 K=360 K. ForssTd the value ofT1 is the same, but there is noupper limit sT2d. Noteworthy is the fact thatT2<TB for thetasTd dependence.

FIG. 2. The temperature evolution of dielectric relaxation timetasTd in 5*CB. The dashed and solid curves are portrayed by rela-tion VFT s2d and MCT dependences5d, respectively. The insetshows the validity of the MCT dependences2d for tasTd in isotropic5CB, applying the log-log scale. Fitted parameters for the MCTdependence are given in the figure.

FIG. 3. The temperature evolution of dc conductivityssTd in5*CB. The dashed and solid curves are portrayed by relation VFTs2d and MCT dependences5d, respectively. The inset shows thevalidity of the MCT dependences2d for ssTd in isotropic 5CB,applying the log-log scale. Fitted parameters for the MCT depen-dence are given in the figure.

MODE COUPLING BEHAVIOR IN GLASS-FORMING… PHYSICAL REVIEW E 71, 011508s2005d

011508-3

When discussing the analysis based on the MCT relations5d, noteworthy is a clear coincidence betweenTMCT andT*,describing the extrapolated temperature of the hypotheticalisotropic-mesophase continuous phase transition. The lattercan be estimated from the analysis pretransitional anomaliesof the static dielectric permittivity«8sTd f=1 kHz or maxima ofloss curves«9sTd1 MHz,fp,400 MHz: see Refs.f32,38g for 5CB

and 5*CB, respectively. One of the most classical methodsof the estimation ofT*, and hence the discontinuityDT=TI-N−T* of the I-N transition, employs the strong pretran-sitional anomalies of the Kerr effect, the Cotton-Mouton ef-fect, or light scatteringf39,44g. In the last decade it wasshown that the same behavior shows the low-frequency,static NDEf32,40,45–47g. However, for the NDE the mea-surement field can be shifted over a range characteristic ofthe decay time of prenematic fluctuationsf32,42g, which isnot possible for other methods mentioned above employinglight. Figure 6 shows results of NDE measurements for a fewfrequencies in the isotropic 5*CB. For the lowest measure-ment frequency the NDE strongly increases in the directionof positive values on approaching theI-N clearing point. Forhigher measurement frequencies the negative-sign pretransi-tional anomaly in the isotropic phase occurs. This finding isin contradistinction to the existing claims that fluctuationsalways have to introduce a positive-sign contribution to theNDE f48–52g. In isotropic 5CB the NDE pretransitionalanomaly in the isotropic phase remains always positive, al-though strong distortions from the linear dependence ofENDE

−1 sTd, not predicted by a simple mean-field analysis basedon the Landau–de Gennes model, occurf32g. To parameter-ize the pretransitional anomaly in 5*CB one may apply theapparent amplitude analysis, successfully applied in the iso-tropic 5CB f32g:

ENDE =ANDE

T − T*~ xTsTdkDM0lkDMfl = x0

* kuDM0ulkuDMfulsT − T * dg ,

s6d

ANDEsTdsT − T * d = ANDELF +

ANDEHF

1+v2t2 , s7d

where the indices 0 andf are for the static limit, the givenfrequency valuesv=2pf, kDM0l, andkDMfl are deviationsof the local order parameter, andg=1 is the critical exponentof susceptibilityscompressibilityd: xT

* =x0* /T−T*. The index

FIG. 4. The derivative analysis of the temperature dependenceof dielectric relaxation times, showing whether the description viathe VFT relation is possible in 5*CB. The inset presents the log-logplot of the MCT power relation for the same data. In both casesstraight lines show regions of the validity of the VFT or MCTdependences, respectively. Characteristic temperatures appearingdue to the applied analysis are given in the figure.

FIG. 5. The derivative analysis of the temperature dependenceof dc conductivity showing whether the description via the VFTrelation is possible in 5*CB. The inset presents the log-log plot ofthe MCT power relation for the same data. In both cases straightlines show regions of the validity of the VFT or MCT dependences,respectively. Characteristic temperatures appearing due to the ap-plied analysis are given in the figure.

FIG. 6. Nonlinear changes of dielectric permittivitysNDEd inthe isotropic phase of 5*CB, for selected measurement frequencies.The inset shows results of the apparent amplitude analysis.

DROZD-RZOSKAet al. PHYSICAL REVIEW E 71, 011508s2005d

011508-4

“LF” refers to “low frequency”—i.e., thev−1@t limit—and“HF” to “high frequency”—v−1!t. The relaxation timet=t0/ sT−T* dy, y=1. Regarding the amplitudes,ANDE

LF

~ kDM0l2~ sD«0d2 andANDEHF ~ kDM0lkDMfl~D«0D« f, where

D«0 andD« f denote anisotropies in the dielectric permittivityof a perfectly ordered sample.

The inset in Fig. 6 clearly confirms the validity of theassumption that the unusual and unique behavior observed inisotropic 5*CB is indeed associated with the amplitudeANDE. Distortions from the linear dependence ofENDE

−1 sTd canbe associated with the coincidence between two relevanttime scales. The first one is related to the reciprocal of themeasurement frequencysv−1=1/2pfd, and the second one isassociated with the relaxation time of premesomorphic fluc-tuations std. For higher frequencies, on approaching theclearing temperature,v−1,t, and thenkDMfl becomes re-lated to the dielectric anisotropy of a single, prenematic fluc-tuation. For lower frequencies or remote fromT*, the condi-tion v−1@t is fulfilled. In this domain the NDE measuringfield can detect only an averaged response from a set offluctuations andkDMfl is reduced to the dielectric anisotropyof a single molecule which is always positive. One of basicfeatures of the nematic phase is the equivalence ofnW and −nWvectors, known as “directors,” showing the preferable orien-tational ordering direction. This causes, within the nematicphase, including prenematic fluctuations, projections of thepermanent dipole moments on the director direction to becanceled. For 5CB this results in an almost perfect cancella-tion of the permanent dipole moments within the fluctuation,due to the rodlike structure of molecules. Hence, for 5CB,one may expectANDE

HF ~ kDMflkDM0l~D«sm=0dD«0<5.1311.2 andANDE

LF ~ kDM0l~ sD«0d<s11.2d2, where D«0 de-notes the dielectric anisotropy for a perfectly ordered me-sophase in a zero-frequency limit. This reasoning gave a fairagreement between the experimental and expected ratios ofthe amplitudes,ANDE

LF /ANDEHF , for 5CB f32g. However, for

5*CB the long molecular axis within premesomorphic fluc-tuations is less parallel to the director than in the case of5CB. This is associated with different molecular structures of5*CB and 5CB as shown in Fig. 1. For 5*CB there remainsa nonvanishing component of the permanent dipole perpen-dicular to nW, practically absent for 5CB. Hence, for 5*CB,one may expectkDMfl=kMi

f −M'f l,0 and henceANDE,0

for higher measurement frequencies. On moving away fromT* or for very low measurement frequencies,kDMfl=«i

−«'.0 andANDE.0. It is noteworthy that both in 5CBf32gand 5*CB the same values of the susceptibility-related ex-ponent and the exponent related to the decay of premesomor-phic fluctuations,g=1 andy=1, have been obtained. As sug-gested in Ref.f32g, the latter can be associated with thedynamical exponent for the nonconserved order parameterz=2.

Concluding, Canget al. f29,30g experimentally showedthe existence of the “lower” and “upper” MCT “critical”temperatures in the isotropic phase of classical rodlike nem-atogens, with theI-N-S phase sequence. The temperatureTMCT

uppercan be associated with the onset of orientational freez-ing and the temperatureTMCT

lower with the onset of the possibleloss of the translational degree of freedom. These results are

in agreement with basic predictions of molecular MCT de-veloped recently for a liquid consisting of hard-ellipsoidshape moleculesf25–28g. As shown above, the MCT“critical-like” relation s5d perfectly describes the experimen-tal data in the isotropic 5CB, forTI-N+50 K.T.TI-N, with-out any distortions. This can be explained by the fact that for5CB the estimated valueTMCT

lower=TI-N−35 K is located alreadyin the solid phase, whereas in glassy liquids relations5d isvalid for T.TMCT+20 K f30g. Following Refs.f25–30g onemay expect thatTMCT

upper is close to the clearing temperature;hence,TMCT

upper andTMCTlower are more than 35 K apart.

For glass-forming 5*CB, with the isotropic–chiral-nematic–glass phase sequence, a small distortion from therodlike structure of 5CB caused thatTMCT

upper=TMCTsorientd<TMCTstransd=TMCT

lower and these MCT “critical” temperaturescoincide with the thermodynamic phase transition atT=T*.It is noteworthy that for 5CB the extrapolated relaxation timetasTMCT

lowerd<ta(TMCTstransd)<2310−7 s is in fair agreementwith the “magic, universal” relaxation timetasTMCTd,10−7 s proposed by Novikov and Sokolovf17g. Regardingthe upper relaxation time in isotropic 5CB,tasTMCT

upperd=ta(TMCTsorientd)<3310−9 s, assuming thatTMCT

upper=T*=TI-N−DT and DT<1.1 K f32g. For 5*CB,ta(TMCTsorientd)<ta(TMCTstransd)<2310−5 s. In 5*CBthis relaxation time drops with rising pressure, reaching thevalue ta<3310−9 s, characteristic for 5CB at the atmo-spheric pressure, forP<140 MPa, as can be concluded fromthe results given in Ref.f38g. One of the still puzzling phe-nomena for supercooled liquids is the appearance of struc-tural heterogeneities and their relationship to the observedexperimental properties. The behavior of the isotropic liquidcrystalline compounds is naturally dominated by well-defined heterogeneities—i.e., the premesomorphic fluctua-tions f28–32,45–47g. In the case of 5CB and 5*CB there arewell-defined bond-ordering heterogeneities—i.e., ordered inan orientational way, “swarms” of rodlike molecules—in theliquidlike surrounding. They induce strong pretransitionalanomalies of the NDE, associated with exponentsg=1 forsusceptibility andy=1 for the relaxation time, which may beassociated with the expected evolution ofx4sTd and t4sTd,mentioned in the Introduction.

It is noteworthy that when analyzing existing experimen-tal data similar singular behavior associated with researchmethods directly coupled to mesoscale heterogeneities canbe found also for “normal” glass-forming liquids such asp-terphenylsPTPd f53g ftransient grating optical Kerr effectsTG-OKEdg 2-biphenylmethenolsBPMd foptical-heterodyne-detected optical Kerr effectsOHD-OKEdg f29g and epoxyresin EPON 5sNDEd f53,54g. When discussing the role ofheterogeneities it is worth recalling the recent simulations ona polymer meltf34g which showed that density fluctuationsremain short range whereas spatial correlations betweenmonomers become long range on approachingTMCT. In thisway a growing dynamical correlation length and the bond-ordering order parameter were identified as the onset of apossible connection between critical phenomena and glass-forming liquids f34g. This behavior resembles the experi-mental findings of this paper and a recent proposal by Tanakaf55g, who showed that the existence of both density and

MODE COUPLING BEHAVIOR IN GLASS-FORMING… PHYSICAL REVIEW E 71, 011508s2005d

011508-5

bond-ordering fluctuations may result in a general model forliquid-liquid transitions, covering supercooled liquids, liq-uids near the critical point, and the liquid-liquid transitions ina one-component liquid. In isotropic 5CB and 5*CB dy-namic and static properties are strongly influenced by or-dered in a premesomorphic way as fluctuations and hetero-geneities, even 100 K above the clearing point. There isstrong evidence for the glassy the fluidlike, tricritical, andpseudospinodal description of the isotropic phase resemblingthe one observed for the liquid-liquid demixing in binarysolutionsf32,45,46,56g. A clear relationship between criticalphenomena and the glassy behavior manifested via dynami-cal regimes and long-lived “glassy” heterogeneities, orderedin a positional and/or orientational way in a fluidlike sur-rounding, was also obtained in recent theoretical models byBerthier and Garrahanf57,58g and Bakai and Fischerf59g.For both tested compounds, 5*CB and 5CB, the temperatureevolution of dielectric relaxation times, the dc conductivity,and the viscosity are well described by the “critical-like”description from mode coupling theory. However, for 5*CBa unique coincidence of the MCT “critical” temperature andextrapolated temperature of the hypothetical continuousisotropic-cholestericsT* d phase transition was found. Thetemperature dependence of the strong electric-field-induced

changes of the dielectric permittivity exhibits a stronganomaly, described by the exponentsg=1 for susceptibilityandy=1 for heterogeneity-related relaxation time. It is note-worthy that for the homogeneous phase of near-critical mix-tures approximately the same critical exponent has been ob-tained f60–62g. Results presented above show a uniquesignificance of studies of mesogenic compounds for a betterinsight into the dynamics of supercooled, glass-forming liq-uids. They also recall questions of the proper description ofthe isotropic phase and theI-N transition which in basicmonographs on the physics of liquid crystals is most oftenpresented as the best example of a simple mean-field descrip-tion f39,44,62–65g.

ACKNOWLEDGMENTS

A.D.R. was supported by a grant of the Committee forScientific ResearchsKBN, Polandd sgrant respectively forJ.Z.d. S.J.R. was supported by a grant of the Committee forScientific ResearchsKBN, Polandd sRef. f2g, Grant No.PO3B 034 25d. The work at NRL was supported by the Of-fice of Naval Research. S.P. holds a scholarship from theFoundation for Polish SciencesGrant No. FNP 2004d.

f1g W. Goetze and L. Sjoegren, Rep. Prog. Phys.55, 241 s1992d.f2g A. Schoenhals, F. Kremer, A. Hofmann, and E. W. Fischer,

Phys. Rev. Lett.70, 3459s1993d.f3g A. P. Sokolv and W. Steffen, Phys. Rev. E52, 5105s1995d.f4g P. Lunkenheimer, A. Pimenov, M. Dressel, Yu. G. Goncharov,

R. Boehmer, and A. Loidl, Phys. Rev. Lett.77, 318 s1996d.f5g W. Knob, M. Naurothandt, and H. C. Andersen, inProceed-

ings of the 7th Yukawa International Seminar, KyotofSuppl.Prog. Theor. Phys.126, 35 s1997dg.

f6g B. Schiener, R. V. Chamberlin, G. Diezemann, and R. G. Bo-hmer, J. Chem. Phys.107, 7746s1997d.

f7g H. Z. Cummins, J. Phys.: Condens. Matter11, 95 s1999d.f8g S. J. Rzoska, A. Drozd-Rzoska, M. Paluch, and J. Zioło, J.

Phys.: Condens. Matter11, L451 s1999d.f9g G. Hinze, D. D. Brace, S. D. Gottke, and M. D. Fayer, J.

Chem. Phys.113, 3723s2000d.f10g P. G. Debenedetti and F. H. Stillinger, NaturesLondond 410,

259 s2001d.f11g E. Donth,The Glass Transition: Relaxation Dynamics in Liq-

uids and Disordered MaterialssSpringer Verlag, Berlin, 2001d.f12g A. Patkowski, H. Gläser, T. Kanaya, and E. W. Fischer, Phys.

Rev. E 64, 031503s2001d.f13g M. Dzugutov, S. I. Simdyankin, and F. H. M. Zettèrling, Phys.

Rev. Lett. 89, 195701s2002d.f14g A. Mukherjee, S. Bhattacharya, and B. Bagchi, J. Chem. Phys.

116, 4577s2002d.f15g E. W. Fischer, A. Bakai, A. Patkowski, W. Steffen, and L.

Reinhardt, J. Non-Cryst. Solids307, 584 s2002d.f16g K. L. Ngai and C. M. Roland, Polymer43, 567 s2002d.f17g V. N. Novikov and A. P. Sokolov, Phys. Rev. E67, 031507

s2003d.

f18g R. Casalini, M. Paluch, J. J. Fontanella, and C. M. Roland, J.Chem. Phys.117, 4901s2002d.

f19g R. Casalini, M. Paluch, and C. M. Roland, J. Chem. Phys.118,5701 s2003d.

f20g J. Kim, W.-X. Li, and T. Keyes, Phys. Rev. E67, 021506s2003d.

f21g V. Zoelmer, K. Raetzke, and F. Faupel, Phys. Rev. Lett.90,195502s2003d.

f22g R. Chelli, G. Cardini, P. Procacci, R. Righini, and S. Califano,J. Chem. Phys.119, 357 s2003d.

f23g G. Foffi, W. Goetze, F. Sciortino, and P. Tartaglia, Phys. Rev. E69, 011505s2004d.

f24g N. A. Davydova, V. I. Melnik, J. Baran, and M. Drozd, J. Mol.Struct. 651–653, 171 s2003d.

f25g M. Letz and R. Schilling, Philos. Mag. B79, 1815s1999d.f26g A. Winkler, A. Latz, R. Schilling, and C. Theis, Phys. Rev. E

62, 8004s2000d.f27g M. Letz, R. Schilling, and A. Latz, Phys. Rev. E62, 5173

s2000d.f28g T. Theenhaus, M. P. Allen, M. Letz, A. Latz, and R. Schilling,

Eur. Phys. J. E8, 269 s2002d.f29g H. Cang, J. Li, V. N. Novikov, and M. D. Fayer, J. Chem.

Phys. 118, 9303s2003d.f30g H. Cang, J. Li, V. N. Novikov, and M. D. Fayer, J. Chem.

Phys. 119, 10 421s2003d.f31g S. J. Rzoska, M. Paluch, A. Drozd-Rzoska, J. Zioło, P. Janik,

and K. Czupryński, Eur. Phys. J. E7, 387 s2002d.f32g A. Drozd-Rzoska and S. J. Rzoska, Phys. Rev. E65, 041701

s2002d.f33g S. Franz, C. Donati, G. Parisi, and S. C. Glotzer, Philos. Mag.

B 79, 1827s1999d.

DROZD-RZOSKAet al. PHYSICAL REVIEW E 71, 011508s2005d

011508-6

f34g C. Bennemann, C. Donati, J. Baschnagel, and S. Glotzer, Na-ture sLondond 399, 246 s1999d.

f35g S. Franz and G. Parisi, J. Phys.: Condens. Matter12, 6335s2000d.

f36g S. C. Glotzer, V. N. Novikov, and T. B. Schroeder, J. Chem.Phys. 112, 509 s2000d.

f37g N. Lacevic, F. W. Starr, T. B. Schroeder, V. N. Novikov, and S.C. Glotzer, Phys. Rev. E66, 030101s2002d.

f38g S. J. Rzoska, M. Paluch, S. Pawlus, A. Drozd-Rzoska, K.Czupryński, and R. Dąbrowski, Phys. Rev. E68, 031705s2003d.

f39g Handbook of Liquid Crystals, edited by D. Demus, J. Goodby,G. W. Gray, H. W. Spiess, and V. VillsSpringer, Berlin, 1998d,Vol. 1.

f40g S. J. Rzoska and A. Drozd-Rzoska,Liquids Under NegativePressures, NATO Science Series II Vol. 84, edited by A. R.Imre, H. J. Maris, and P. R. WilliamssKluwer, Dordrecht,2002d, pp. 117–126.

f41g A. Drozd-Rzoska, Phys. Rev. E62, 8071s2000d.f42g M. Górny and S. J. Rzoska, inNonlinear Dielectric Phenom-

ena in Complex Liquids, NATO Science Series II Vol. 157,edited by S. J. Rzoska and V. P. ZheleznysKluwer, Dordrecht,2004d, pp. 45–54.

f43g F. Stickel, R. W. Fischer, and R. Richert, J. Chem. Phys.104,2043 s1996d.

f44g M. A. Anisimov, Critical Phenomena in Liquids and in LiquidCrystalssGordon and Breach, Reading, MA, 1994d.

f45g A. Drozd-Rzoska, Liq. Cryst.24, 835 s1998d.f46g A. Drozd-Rzoska, Phys. Rev. E59, 5556s1999d.f47g A. Drozd-Rzoska, inNonlinear Dielectric Phenomena in Com-

plex Liquidsf42g, pp. 201–210.f48g A. Chełkowski, Dielectric PhysicssPWN-Elsevier, Warsaw,

1980d; Fizyka DielektrykówsPWN, Warsaw, 1994d sthe up-to-date edition in Polishd.

f49g C. J. F. Boettcher,Theory of Electric PolarizationsElsevier,Amsterdam, 1993d.

f50g J. Małecki, J. Mol. Struct.67, 23 s1995d.f51g J. Małecki, J. Mol. Struct.437, 595 s1997d.f52g P. KJdziora, J. Jadzyn, and L. Hellemans, Phys. Rev. E66,

031702s2002d.f53g R. Torre, Chem. Phys. Lett.212, 90 s1993d.f54g J. Zioło, S. J. Rzoska, and A. Drozd-Rzoska, inNonlinear

Dielectric Phenomena in Complex Liquidsf42g, pp. 269–274.f55g H. Tanaka, Phys. Rev. E62, 6968s2000d.f56g P. K. Mukherjee, J. Phys.: Condens. Matter10, 9191s1998d.f57g L. Berthier and J. P. Garrahan, Phys. Rev. E68, 041201

s2003d.f58g S. Whitelam, L. Berthier, and J. Garrahan, Phys. Rev. Lett.92,

185705s2004d.f59g A. S. Bakai and E. W. Fischer, J. Chem. Phys.120, 5235

s2004d.f60g S. J. Rzoska, Phys. Rev. E48, 1136s1993d.f61g A. Drozd-Rzoska, inNonlinear Dielectric Phenomena in Com-

plex Liquidsf42g, pp. 201–210.f62g G. Vertogen and W. H. de Jeu,Thermotropic Liquid Crystals:

Fundamentals, Springer Series in Chemical Physics Vol. 45sSpringer, Berlin, 1988d.

f63g P. G. de Gennes and J. Prost,The Physics of Liquid CrystalssOxford University Press, Oxford, 1994d.

f64g S. Chandrasekhar,Liquid Crystals sCambridge UniversityPress, Cambridge, England, 1994d.

f65g S. Urban and A. Wuerflinger, inAdvances in Chemical Phys-ics, edited by I. Prigogine and S. A. RicesWiley, New York,1997d, Vol. XCVIII, p. 143.

MODE COUPLING BEHAVIOR IN GLASS-FORMING… PHYSICAL REVIEW E 71, 011508s2005d

011508-7