Phase Transitions of Naphthalene and Its Derivatives Confined in

ARTICLE pubs.acs.org/JPCC

Phase Transitions of Naphthalene and Its Derivatives Confined in Mesoporous Silicas Janice A. Lee,† Harald R€osner,‡ John F. Corrigan,† and Yining Huang*,† † ‡

Department of Chemistry, The University of Western Ontario, London Institut f€ur Materialphysik, Westf€alische Wilhelms-Universit€at M€unster, Wilhelm-Klemm-Str. 10, D-48149 M€unster, Germany

bS Supporting Information ABSTRACT: The phase transitions of naphthalene and its derivatives including 2-methylnaphthalene, 2-methyoxynaphthalene, and 2-chloronaphthalene confined in mesoporous silicas of spherical and cylindrical pore geometry were investigated by differential scanning calorimetry (DSC), and a depression of the phase transition temperature was observed. It was found that depression and freeze-melt hysteresis are generally greater in spherical pores. Using powder X-ray diffraction, it was observed that the crystal structure of confined naphthalene is the same as the bulk. Using DSC and Raman spectroscopy, it was shown on both a thermodynamic and molecular level that there is a nonfreezing layer on the pore walls for each confined compound and its thickness depends on the pore diameter.

’ INTRODUCTION It is well-known that the properties of materials confined in pores differ from those of their bulk counterparts due to geometrical confinement and interaction with the surface of the pore wall.1 Most commonly observed are the hysteresis between freezing and melting cycles and the shift in phase transition temperatures.2,3 The topic of phase transitions of confined materials, or materials within pores, is of considerable importance in commercial areas such as catalysis, interfacial adhesion, and separation science,4,5 and also serves practical interest in such areas as frost heave, weathering, and oil recovery.6 In the last two decades, there has been an incredible expansion in research devoted to the controlled preparation of nanocrystals using porous materials as nanoreactors.7 However, crystallization in confining media is still not completely understood.8 It has been well-established that materials in confinement exhibit a transition temperature depression ΔT. The general form of the Gibbs-Thomson equation9,10 is typically used to interpret these measurements9,11,12 and relates the temperature shift ΔT of crystallization to the pore size of the confining material according to ΔT γ vs ¼ - 2 sl T0 rp ΔHsl

ð1Þ

where T0 is the bulk transition temperature, γsl the solid-liquid surface energy, vs the molar volume of the solid phase, rp the pore radius, and ΔHsl the latent heat of the solid-liquid transition. According to this equation, the shift of the transition temperature of a confined liquid is inversely proportional to the radius of the pore in which it is confined. However, this is not the case, as it is known that not all of the solvent takes part in the transition; a significant portion remains adsorbed on the pore surface. As a result, eq 1 should be reformulated to include the r 2011 American Chemical Society

thickness t of this adsorbed layer:4 ΔT γsl vs ¼ -2 ð2Þ T0 ðrp - tÞ ΔHsl Much of the previous research on phase transitions of confined materials has been conducted in porous systems such as Vycor glasses, controlled pore glasses (CPG), or porous silica. These materials exhibit broad pore size distributions, inducing disorder which complicates the interpretation of results. For this reason, more recent research has turned its focus to mesoporous media with simple geometry instead. Specifically, mesoporous silicas are typically well-ordered and characterized by narrower pore size distributions, lending an advantage over CPG and Vycor glasses. While there is considerable interest in the behavior of molecules confined within simple geometries, the number of studies conducted on slit and cylindrical pores13-18 far outweighs those regarding spherical pores.19-21 To this effect, we examined in the present work the effect of confinement on the phase behavior of naphthalene and several derivatives in spherical pores using differential scanning calorimetry (DSC) as the primary experimental method in combination with other techniques such as powder X-ray diffraction (XRD) and Raman spectroscopy. For the purpose of comparison, the phase transition behavior of the abovementioned materials in cylindrical pores was also investigated.

’ EXPERIMENTAL SECTION Spherical Mesoporous Silicas. Spherical mesopores with diameters ranging from 6.6 to 28.6 nm were prepared using three different synthesis procedures in order to access such sizes. Received: December 1, 2010 Revised: February 3, 2011 Published: February 28, 2011 4738

dx.doi.org/10.1021/jp111432j | J. Phys. Chem. C 2011, 115, 4738–4748

The Journal of Physical Chemistry C

ARTICLE

Table 1. Summary of Spherical (S) and Cylindrical (C) Mesopores

a

pore volume (cm3 g-1)

SSAc

type

9.5

1.14

812

MCF

8.6

0.87

413

MTS

9.0

4.6

0.58

780

SBA-16

S4

6.6

N/A

0.34

559

SBA-16

C1

8.0

7.3

1.03

674

SBA-15

C2

6.5

6.2

0.82

562

SBA-15

C3

5.5

4.8

0.51

419

SBA-15

sample

DBdB-FHHa (ads) (nm)

S1

28.5

S2

18.3

S3

DBdB-FHHb (des) (nm)

Body diameter. b Window diameter. c Specific surface area.

Mesocellular foams (MCFs) have diameters ranging from 24 to 42 nm and were prepared according to Lettow22 and Zhao et al.23 Following this method, we synthesized a 28.5 nm pore. Micelletemplated silicas (MTSs) have been previously reported to provide pore diameters ranging from 2 to 15 nm,24 and we were able to synthesize an 18.3 nm MTS following the procedure described by Ottaviani et al.24 SBA-16 materials, which possess pore diameters ranging from 4.7 to 7.2 nm, were prepared according to Kleitz et al.25 We synthesized 6.6 and 9.0 nm SBA-16 in this way, and details of the above preparations can be found in the Supporting Information. Cylindrical Mesoporous Silicas. SBA-15 materials ranging from 5.5 to 8.0 nm in diameter were prepared according to Zhao et al.23 For synthetic details, see the Supporting Information. Characterization of Mesopores. The template-free mesoporous silica materials were characterized by nitrogen adsorption analysis using a Micromeritics ASAP2020 analyzer at 77 K. Before acquiring the adsorption-desorption isotherms, the samples were degassed at 473 K for 8 h. The Micromeritics software uses the Brumauer-Emmett-Teller (BET) method to calculate specific surface areas and the Barrett-Joyner-Halanda (BJH) method to calculate pore volumes. However, these methods were found to underestimate both values; instead, the Broekhoff-de Boer-Frenkel-Halsey-Hill (BdB-FHH) method was used to calculate the pore size distributions (PSDs) and pore volumes.26 The adsorption and desorption branches of the isotherms were used to derive the pore and window diameters, respectively, for the spherical pores. For cylindrical pores, both branches of the isotherm give PSDs with similar values for the pore diameter. For all pores, the diameters are reported according to the adsorption branches. Transmission electron microscopy (TEM) micrographs of the spherical mesopores were recorded using a Philips CM10 Transmission Electron Microscope (LaB6 filament, 100 kV, 0.2 nm resolution) and on a FEI Tecnai F20 G2 (field-emission gun, 200 kV, 0.23 nm point resolution) microscope. Method of Loading Organics into Mesopores. All samples for DSC experiments were prepared directly in the DSC crucibles. In a typical loading, a sample of the silica material was degassed in a glass tube on a vacuum line at 673 K for 2 h. A portion of the sample was placed into the crucible and weighed, and the total pore volume was obtained to calculate the correct mass of the organic required achieving a 60% overload. The compound was added to the crucible and weighed to obtain the exact mass of the organic present. The crucible was hermetically sealed and heated at 403 K for several hours to allow the organic to equilibrate inside the pores. For Raman spectroscopic studies, organic compound was added to the silica sample inside a glass tube which was then

quickly sealed by flame. This was allowed to equilibrate at 403 K for several hours and then cooled slowly to room temperature before recording spectra. Measurements of Confined Materials. DSC measurements were performed using a TA Instruments DSC Q20 equipped with the Liquid Nitrogen Cooling Accessory and Universal Analysis 2000 software. Instrument calibration was performed using 1 mg of certified high purity In sealed in a standard crimped Al pan from 183 to 673 K at a rate of 10 K/min. All samples were overloaded in order to detect both the bulk and confined transition temperatures and thus obtain precise determinations of temperature depression. All scans were carried out at heating or cooling rates of 10 K/min after a 10 min isothermal stage at the starting temperature of the experiment. All exotherms are given as positive peaks, in the direction of heat flow. Raman spectra were recorded on a Bruker RFS 100/S FTRaman Spectrometer equipped with a Nd3þ/YAG laser operating at 1064.1 nm and a liquid-nitrogen-cooled Ge detector. The laser power was 120 mW at the sample, and the resolution was 2 cm-1. Powder XRD patterns were recorded on a Rigaku diffractometer using Co KR radiation (λ = 1.7902 Å). Samples were scanned from 5 e 2θ e 65 at a scan rate of 10/min with a step-size of 0.02.

’ RESULTS AND DISCUSSION Mesoporous Silica Host Materials. In order to obtain a range of pore sizes for study, several types of spherical and cylindrical mesopores were synthesized. Sorption isotherms and pore size distributions are given in Figures S1 and S2, respectively, in the Supporting Information. Values for the BdB-FHH pore size distributions calculated from the adsorption branches are given in Table 1. The properties of the selected pores are also summarized in Table 1. The selected TEM images (Figure S3 of the Supporting Information) of typical spherical and cylindrical mesopores show differences in pore morphology. Cylindrical pores are arranged in a regular hexagonal array with well-defined pore walls, whereas spherical pores form a less orderly network. Naphthalene. At room temperature, pure C10H8 exists in the solid phase. It has a rigid lattice structure and shows no plastic crystalline phase.27 The DSC curve of pure naphthalene is shown in Figure S4a of the Supporting Information, with melting and freezing occurring at 79.4 and 66.1 K, respectively. This thermal hysteresis is expected and is due to the small delay in the heterogeneous freezing nucleation of the pure liquid;11 the melting temperature is then identified as the true transition temperature and taken as the reference transition temperature T0. The measured enthalpy values ΔHsl upon heating and cooling were found to be 147.18 and 145.80 J g-1. Note that 4739

dx.doi.org/10.1021/jp111432j |J. Phys. Chem. C 2011, 115, 4738–4748


ARTICLE

Figure 2. ΔTm plotted as a function of r-1 for spherical (9) and p cylindrical (2) pores.

the liquid phase, and ΔHsl is the bulk latent heat of melting. For sufficiently large pores (>6 nm diameter), the shift in freezing temperature ΔTf can be related to the pore width D on the basis of the Gibbs-Thomson equation that is obtained by equating the free energies of the confined liquid and solid.

Figure 1. DSC curves of bulk C10H8 (outside the pore) and C10H8 confined in spherical pores upon heating and cooling. Confined phase transitions are indicated by arrows, and the transition temperatures are given.

Table 2. DSC Values of the Transition Temperatures, Transition Temperature Shifts (ΔT), and Hysteresis Values of C10H8 Confined in Spherical Pores sample

a

rp

t

T0

Tm

Tf

ΔTm

ΔTf

(nm)

(nm)

(K)

(K)

(K)

(K)

(K)

hysteresis

bulk

¥

S1 S2

14.3 9.2

2.4 1.1a

352.6 338.9 334.9 351.4 331.9 316.0

13.7 19.5

17.7 35.4

4.0 15.9

S3

4.5

1.3

352.3 303.3 213.7

49.0

138.6

89.6

S4

3.3

0.9

351.8 286.7 200.6

65.1

151.2

86.1

352.5 339.2

13.3

Extrapolated from experimental data.

the complete freezing event usually takes place over a broader temperature range due to crystal rearrangement, so the tail end of the peak extends such that differentiation from the baseline is difficult.28 It follows that the integrated peak area usually gives rise to a smaller enthalpic value than that given by the melting peak. The enthalpy calculated from the melting curve is therefore taken as the true value. The properties of pure C10H8 are given in Table S1 in the Supporting Information and agree closely with literature values.29 DSC Studies of C10H8 Confined in Spherical Pores. By overloading the pores, we were able to obtain the DSC melting and freezing curves of both the bulk and confined C10H8 for comparison, as shown in Figure 1. The peaks corresponding to phase transitions of the confined C10H8 are denoted by arrows, and it can be seen that both melting and freezing transition temperatures decrease with pore diameter. A summary of these values can be found in Table 2. It is evident from Figure 1 and Table 2 that there is a depression in both the confined melting and freezing transitions of C10H8 relative to those of the bulk transitions. Furthermore, this depression increases as the pore sizes become smaller. Evans et al.30 explained the freezing temperature shift of confined liquids using eq 3 shown below, where γws and γwl are the wallsolid and wall-liquid surface tensions, v is the molar volume of

Tf , pore - T0 ΔTf ðγ - γwl Þv ¼ ¼ 2 ws ð3Þ DΔHsl T0 T0 The sign of the shift in freezing temperature is given by the difference of the surface tensions γws-γwl; thus, the equation predicts that the freezing temperature will, as in our case, decrease compared with the bulk value if the pore wall prefers the liquid phase over the solid phase. Conversely, if the solid phase is preferred, then there will be an increase in the freezing temperature.30 This fluid-wall interaction affects nucleation, and therefore hysteresis, in the confined phase change.31 As we observe a decrease in the freezing temperature of confined C10H8 in the present work, this suggests that our silica-surface pore walls prefer the liquid phase of naphthalene over the solid. However, because eq 3 uses derivations of macroscopic concepts such as surface tension and does not account for the strong inhomogeneity of the confined phase, it fails for smaller pores.32 According to eq 1, the shift of the transition temperature of a confined liquid is inversely proportional to the pore radius. Among the assumptions made in this relationship are that the values γsl, ΔHsl, and vs are independent of crystal size. Furthermore, considerations that have not been incorporated into this equation are the confined fluid-fluid and fluid-wall interactions,14 though it has been found that the strength of the fluid-fluid interaction relative to the fluid-wall interaction plays an important role in determining the sign of the shift of the freezing point.33 Nevertheless, many previous studies6,13,14 have largely confirmed the melting point depression expressed from eq 1 for nanoparticles as well as materials in confinement, where deviations are observed only for the smallest pore sizes.6 In plotting ΔTm as a function of r-1 p as prescribed by eq 1, our results show a reasonably linear trend (Figure 2). At r-1 p = 0, rp = ¥, so the pore size is infinitely large; this situation corresponds with unconfined bulk naphthalene and so it is expected that there is no transition temperature depression. However, this trend line does not exactly pass through the origin. It has previously been noted34-38 that a so-called “contact layer” exists on the pore walls when a material is introduced into the pore interior. This layer remains in the liquid state below the melting temperature of the substance, and many authors have reported experimental evidence of such a layer for numerous substances.34,37,39-42 The presence of such a nonfreezing layer thereby reduces the effective pore radius in the application of eq 1 such that the effective pore radius is r = rp - t, where rp is the 4740



ARTICLE

Figure 5. t plotted as a function of rp for C10H8 in spherical pores. Figure 3. Some typical DSC curves recorded for various flushing times for spherical pore S2 overloaded with C10H8.

Figure 4. Evolution of Hc (O) and Hb (0) as a function of the mass of C10H8 present in spherical pore S2.

radius of the pore and t is the thickness of the contact layer. We therefore use eq 2 to treat our data, as it allows for the evaluation of ΔT with respect to the effective pore radius rp - t. The value of t can be determined by the traditional calibration procedure using materials of various pore sizes, where the transition temperature depression is related to the pore radius through a value of t. From experimental data, t can be determined by calculating the limit of the pore radius as ΔTf approaches ¥ (no confined freezing).43 However, there is an underlying assumption that t does not vary with pore size.8 Instead, we determined t according to a previously reported procedure8,44 which does not assume that t is constant. Briefly, in the present study, a known mass of porous sample S2 is placed in a DSC pan with an excess of C10H8. Several holes are drilled in the crucible lid, and C10H8 is progressively evaporated at 80 K by N2 gas flushing between thermal cycles. We were able to calculate the amount of C10H8 remaining after each flush by weighing the pan; from there, we were able to obtain semiquantitative information for C10H8 (both free and confined) undergoing the solid-liquid transition. Figure 3 presents some typical DSC curves recorded upon evaporation of C10H8 by flushing. The labels b and c denote the bulk and confined transition peaks, respectively. It can be seen that, as the flushing time increases, peak b decreases and eventually cannot be observed. Following this, peak c also decreases until no transition can be observed. A plot of the heats corresponding to peaks b and c as a function of the quantity of naphthalene present is shown in Figure 4. The point at which Hc differs from zero corresponds to the mass of the contact layer mt. In the portion where Hc remains constant, the pores are completely filled and we can measure Hc,max. From the graph, we extrapolated a mass of 2.2 mg required to fill the pores, in close agreement with the 2.4 mg calculated from N2

sorption. At Hc,max, the enthalpy corresponds to the mass of the C10H8 present mvp minus the mass participating in the contact layer, namely m = mvp - mt. We can then deduce the enthalpy of melting per gram for confined C10H8 (ΔHc = 101 J g-1). Using the mass of the contact layer mt determined from the plot, we were able to calculate the thickness of the layer according to8 mt ð4Þ t ¼ FðSSAÞðmSiO2 Þ where SSA is the specific surface area of the silica sample given in Table 1. Knowing the pore radius rp and t for sample S2, we calculated the surface tension γsl = 27.6 mJ m-2 from eq 2. There are a number of γsl values for naphthalene that can be found in the literature. Jackson and McKenna4 found values of 8.2 and 6.1 mJ m-2 from experimental data and report a value of 31.7 mJ m-2 from empirical calculations based on previous work by Turnbull45 and Dunning.46 Jones47 reported a value of 61 ( 11 mJ m-2. Our value is generally comparable to the empirical value within the large variability of such measurements reported in the literature. As such, we chose to use our γsl value for further calculations. Furthermore, use of the lower and higher reported values presents unreasonable results (i.e., negative values of t). Jackson and McKenna4 suggested one reason for such great variation in the reported values of γsl for C10H8 may be that there are changes in the crystallographic form when C10H8 crystallizes in confined geometry. However, we have found that this is not the case and our results are discussed below. If the value of t is constant for varying pore sizes, then t will be 75% of the pore radius for the smallest pore S4. This is clearly unreasonable, and we conclude that, for our system, t does indeed vary with pore size. Using the γsl value derived, we determined the thickness t of the contact layer in each pore size. It was found that t generally decreases with pore size. This trend agrees with results reported by Meziane et al.8 for carbon tetrachloride confined in mesoporous silica gels. Our results are presented in Table 2, and a plot of t as a function of r is shown in Figure 5. One way to verify the accuracy of the evaporation experiment is to use the measured mvp to deduce the porous volume Vp according to8 mvp Vp ¼ ð5Þ FmSiO2 where F is the solid density of naphthalene. We extrapolated the value 0.832 g mol-1 at 80 K from previous work by Gryzll et al.48 The value of 0.85 cm3 g-1 found for Vp is comparable to the value of 0.87 cm3 g-1 obtained by N2 sorption (VN2); such close agreement indirectly confirms the reliability of the measured t value. It is clear from the DSC curves in Figure 1 that the breadths of confined transitions broaden as the pore size decreases, suggesting a progressive layer-by-layer melting in confinement where the outer 4741



ARTICLE

Table 3. DSC Values of the Transition Temperatures, Transition Temperature Shifts (ΔT), and Hysteresis Values of C10H8 Confined in Cylindrical Pores rp sample (nm)

t

T0

Tm

Tf

ΔTm

ΔTf

(nm)

(K)

(K)

(K)

(K)

(K)

bulk

a

352.5 339.2

hysteresis 13.3

C1

4.0

0.59a

352.0 309.5 290.4

42.5

61.6

19.1

C2

3.3

0.48

352.0 304.6 283.3

47.4

68.7

21.3

C3

2.8

0.32

351.2 287.4 258.4

63.8

92.8

29.0


portions of the confined material melt first and those in the center melt last.4,49 In the case of freezing, the liquid outside the pore freezes first and freezing inside the pores occurs through the slow penetration of a freezing front. This penetration is delayed in the smallest pore apertures, broadening the confined transition peak.11 DSC Studies of C10H8 Confined in Cylindrical Pores. Using the same method of pore loading for spherical pores, we loaded our cylindrical pores with naphthalene such that the DSC curves provide information regarding the phase transitions of both the bulk and confined C10H8. Curves can be found in Figure S5 of the Supporting Information, and values are given in Table 3. Results following the Gibbs-Thomson equation are shown in Figure 2, where ΔTm is plotted against the inverse of the pore radius without taking into account the thickness t of the contact layer. Again, the trend is reasonably linear. However, as we have determined that the contact layer does indeed exist, the calculation of ΔH will be underestimated if this layer is not accounted for. We therefore performed the evaporation experiment described above in order to determine t. Since the quantities of naphthalene used were small, we performed the experiment three times using samples of C1 (8.0 nm diameter) loaded with C10H8 to ensure that the result is reasonably accurate. Some typical DSC curves recorded upon evaporation of C10H8 from trial 1 are presented in Figure S6 of the Supporting Information, and the plots of the heats as a function of the quantity of naphthalene present are shown in Figure S7 of the Supporting Information. The t values from the three sets of parallel experiments are reasonably consistent. The calculated volumes Vp correspond well with the pore volume VN2 determined by nitrogen adsorption porosimetry and again provide verification of the described evaporation experiments. These results are summarized in Table S2 of the Supporting Information and used in the calculations of t for the other pores in this system. As in the previous section, we calculated the value of the surface tension γsl using eq 2. For C10H8 in cylindrical pores, we find that it has a value of 19.0 mJ m-2. We note a small discrepancy between the γsl values found for naphthalene in our spherical pores (27.6 mJ m-2) and our cylindrical pores. Although both numbers are within the range of those values reported in the literature (see above), the value calculated from the cylindrical system is lower than the value calculated from the spherical system. According to Jackson and McKenna,4 a possible reason for this is a reduced value of the solid density F in a small pore, and thus increased value of the molar volume vs, resulting from poor packing in confinement which causes a lower value of γsl. Using the value of γsl = 19.0 mJ m-2, we calculated t for the other cylindrical pores in this system. Again, we find that there is a decrease in t with decreasing pore size. The values are found in Table 3, and the trend is shown in Figure 6a.

Figure 6. (a) t plotted as a function of rp for C10H8 in cylindrical pores and (b) ΔTf versus ΔTm for C10H8 confined in spherical and cylindrical pores.

Comparison of Cylindrical and Spherical Pores. The DSC curves indicate a thermal hysteresis of the confined C10H8 which increases with decreasing pore radius. Because the transition mechanisms of a confined liquid are different from those of the bulk, the physical cause of such a feature differs. Materials confined in small pores exhibit transition temperature depressions due to their high acquired surface area to volume (S/V) ratios; the presence of this surface introduces excess energy for the solid phase and shifts the solid-liquid equilibrium toward the liquid state. Petrov and Furo12 attribute this hysteresis to a free-energy barrier between metastable and stable states of porefilling material. They show that the freezing temperature depression is defined by the surface to volume ratio S/V, whereas the melting temperature depression is defined by the mean curvature κ of the pore surface such that

ΔTm ¼ ΔTf

2kV S

ð6Þ

It turns out that for any pore 2κ < (S/V). As a result, freeze-melt hysteresis always occurs with ΔTm < ΔTf.12 In particular, for a spherical pore 2κV/S = 2/3 and for a cylindrical pore 2κV/S = 1/2. It follows that ΔTm = 2/3ΔTf for a spherical pore and ΔTm = 1 /2ΔTf for a cylindrical one. This implies that, in plotting ΔTm as a function of ΔTf for both geometries, we should find that the slope of the linear relationship for the spherical system should be 1.3 times greater than that of the cylindrical systems. Plotting our data in Figure 6b, we find that the slope of this relationship for our spherical pores is 1.4 times greater than that of our cylindrical ones. This trend is consistent with Petrov and Fur o’s model.12,50 To further describe the relationship between the cylindrical and spherical pore systems, we determined the ratio of the change in the freeze-melt hysteresis of C10H8 in the spherical system versus the cylindrical system. Figure 7a shows a plot of hysteresis values as a function of effective pore radius rp - t for both cylindrical and spherical pores. Data from the spherical pores indicates that, with increasing effective pore radii, the hysteresis value approaches zero. It is unlikely for hysteresis to take negative values, as this would imply that ΔTf is less than ΔTm, but as explained by Petrov and Furo’s hysteresis model,12 this situation does not occur. Thus, it appears that with increasing pore size hysteresis approaches, but does not reach, zero. As 4742



Figure 7. Hysteresis plotted as a function of effective pore radius rp - t for the spherical and cylindrical pore systems.

Scheme 1. 2-Methylnaphthalene, 2-Methoxynaphthalene, and 2-Chloronaphthalene

hysteresis exists even in the bulk (i.e., no confinement), this trend is intuitively reasonable. Because the cylindrical pores used in this study were relatively small, we do not see this approach to zero in the data. However, it is clear that hysteresis cannot decrease linearly with increasing pore size, as it would eventually take on negative values. We suggest that, with larger cylindrical pore radii, hysteresis would also approach zero as with the spherical pores but with a faster rate. Naphthalene Derivatives. To gain an understanding of the effects of molecular properties on confined phase behavior, we examine a system of C10H8 derivatives confined in cylindrical pores. 2-Methylnaphthalene (2-C10H7CH3), 2-methoxynaphthalene (2C10H7OCH3), and 2-chloronaphthalene (2-C10H7Cl) (Scheme 1) were used. The three derivatives, substituted at the same position on the parent C10H8 moiety, differ in their electronic properties. For example, the methoxy group is strongly electron donating while the choloro group is electron withdrawing. These properties affect the aromatic π system of the naphthalene ring, thereby affecting the molecule’s ability to interact with the pore surface through π-type hydrogen bonding.51,52 The DSC curves of each pure compound and their properties are shown in Figure S4 and Table S1 of the Supporting Information, respectively. According to Chanh et al.,53 2-methylnaphthalene undergoes a solid-state transition at 291 K upon melting (253 K upon freezing), though our DSC curves do not indicate such an event for either pure or confined C10H7CH3. It should be noted that confinement has been found to depress the solid-solid phase transition in n-alkanes.54 It has also been shown that some 2(R)-substituted naphthalene compounds (R = F, Cl, Br) exhibit crystalline order-disorder transitions.53,55,56 In the inset of Figure S4d of the Supporting Information, the transition between these two forms for 2-C10H7Cl is indicated by arrows; the solid-solid transition from form II (semiordered) to form I (disordered) upon heating occurs at 315.0 K, comparable to findings by Chanh et al.,53 and the form I to II transition upon cooling at 297.6 K. The enthalpy, ΔH, of the solid-liquid transition for 2-C10H7OCH3 is almost 2 times greater than that of the other derivatives. This is because 2-C10H7OCH3 does not

ARTICLE

have a disordered phase, whereas the high temperature solid phases of 2-C10H7Cl and 2-C10H7CH3 are both disordered.53,56 DSC curves of the confined compounds are shown in Figure 8. Pores were overloaded to obtain the melting and freezing curves of both the bulk and confined organic for comparison. The peaks corresponding to solid T liquid phase transitions are denoted by arrows, and it can be seen that both melting and freezing transition temperatures decrease with pore diameter. Summaries of these values can be found in Table 4. The DSC curves of 2-C10H7Cl in Figure 8c show that there are two confined transition peaks for freezing. The two peaks are due to the confined solid-solid and liquid-to-solid transitions, respectively. Plotting ΔTm as a function of r-1 p according to eq 1 gives a reasonably linear trend (Figure S8 of the Supporting Information). However, since a contact layer does exist, we therefore determined the thicknesses of the contact layers using the evaporation experiment described earlier. Some typical curves recorded upon evaporation of C10H7CH3, C10H7OCH3, and C10H7Cl, as well as their corresponding heat plots, are shown in Figures S9-11 of the Supporting Information, respectively. Values determined from these plots and calculated using eqs 2, 4, and 5 are summarized in Table 5. Values of the solid densities F used for calculations were 1.0058 g mL-1 (2-C10H7CH3),29 1.072 g mL-1 (2-C10H7OCH3),29 and 1.327 g mL-1 (2C10H7Cl).57 We found that ΔHc for 2-C10H7OCH3 is approximately twice the values for 2-C10H7CH3 and 2-C10H7Cl, maintaining the same general relationship displayed by their bulk values. This indicates that the disordered and semiordered solid forms of 2-C10H7CH3 and 2-C10H7Cl do exist even in confinement. The pore volumes Vp calculated from the experiments with 2-C10H7CH3 and 2-C10H7OCH3 are in good agreement with the volume VN2 for C1 determined by nitrogen adsorption porosimetry, confirming the reliability of the measurements. The value obtained from 2-C10H7Cl, however, differs from the nitrogen adsorption value. This is a result of the solid-solid transition seen in 2-C10H7Cl. In the DSC curves, this peak overlaps with that of the confined solid-liquid transition under study. As such, the heat corresponding to the confined peak is overestimated when there is bulk 2-C10H7Cl present. The mass of the contact layer and hence thickness are also overestimated as a result. Comparison of Naphthalene and Its Derivatives. The surface of confining pores may have attractive or repulsive properties which affect the confined material,4 and the SBA-15 materials used in this study contain the usual surface hydroxyl groups58 which can interact with delocalized π systems through π-type hydrogen bonding.51,52 This type of interaction has also been found to occur between the hydrogens of water and the πelectronic system of aromatic rings.59-61 The degree of the πtype hydrogen bonding is dependent on the specific π system. For example, the π-type hydrogen bonding in naphthalene differs from that of benzene due to a reduction of the negative charge density of the naphthalene π system compared to benzene.62 To this effect, it is possible that electronic effects from substituent groups on C10H8's delocalized π system will affect confined phase behavior. As C10H8 interacts with the pore surface through its aromatic π system, the electron-withdrawing and -donating properties of each substituent affect this interaction through alteration of the electron density. The strength of the fluid-fluid interaction relative to the fluid-wall interaction plays an important role in determining the sign of the freezing point shift,33 and fluid-wall interactions appear to play a crucial role in the melting point depression.4,63,64 4743



ARTICLE

Figure 8. DSC curves of bulk (outside the pore) and confined (a) 2-C10H7CH3, (b) 2-C10H7OCH3, and (c) 2-C10H7Cl confined in cylindrical pores C1, C2, and C3 (from left to right). Confined phase transitions are indicated by arrows, and the transition temperatures are given. The asterisks (*) and triangles (3) indicate solid-solid transitions of the bulk and confined materials, respectively.

Table 4. DSC Values of the Transition Temperatures, Transition Temperature Shifts (ΔT), and Hysteresis Values 2-C10H7CH3, 2-C10H7OCH3, and 2-C10H7Cl Confined in Cylindrical Pores rp t sample (nm) (nm)

T0 (K)

Tm (K)

Tf (K)

ΔTm (K)

ΔTf (K)

hysteresis (K)

2-methylnaphthalene bulk

307.8 300.6

7.2

C1

4.0

1.2a

305.9 272.5 252.2

33.4

53.7

20.3

C2

3.3

0.8

305.8 268.9 245.0

36.9

60.8

23.9

C3

2.8

0.7

305.2 260.8 228.3

44.4

76.9

32.5

67.9

23.7 25.3

2-methoxynaphthalene bulk C1

4.0

1.5a

346.5 322.8 344.6 302.0 276.7

42.6

C2

3.3

1.1

344.8 297.0 267.5

47.8

77.3

29.5

C3

2.8

1.0

344.5 285.3 237.8

59.2

106.7

47.5

2-chloronaphthalene bulk

a

331.0 326.3

4.7

C1

4.0

2.1a

330.8 297.8 278.9

33.0

51.9

18.9

C2

3.3

1.5

330.1 294.2 270.7

35.9

59.4

23.5

C3

2.8

1.3

330.1 287.0 250.4

43.1

79.7

36.6


Takei and co-workers52 have shown that, for a given pore size, increasing the number of surface hydroxyl groups in the same type of porous silica samples caused a greater melting point depression for benzene but not for n-hexane. n-Hexane was determined by IR spectroscopy to have negligible surface interactions with the pore. They proposed that the interaction between benzene π electrons with the surface hydroxyl groups causes a change in the liquid structure of benzene, such that benzene molecules in the vicinity of the silica surface are oriented parallel to the surface. Melting point depression occurs as a result, and the degree of orientation depends on the concentration of surface hydroxy groups. On the basis of these findings, we expect that 2-C10H7OCH3 will show the greatest depression, followed by 2-C10H7CH3, C10H8, and 2-C10H7Cl. While our results show that 2-C10H7OCH3 is indeed further depressed than 2-C10H7CH3 and 2-C10H7Cl, the overall trend is not in accordance with what we expected initially. The methyl group in 2-C10H7CH3 is slightly electron donating.65 The electron density of the fused-ring portion of 2-C10H7CH3 is thus increased slightly. In comparison, -OCH3 is heavily electron donating due to the oxygen lone electron pair,65 significantly increasing the electron density of the fused-ring portion of 2-C10H7OCH3. The increased electron densities in 2-C10H7CH3 and 2-C10H7OCH3 in turn improve the molecules’ abilities to form hydrogen bonds with surface hydroxyl groups. The greater electron donating capacity of -OCH3 allows 2-C10H7OCH3 to form stronger π4744



ARTICLE

Table 5. Mass, Thickness t, Porous Volumes, and Transition Enthalpies Measured by DSC and N2 Sorption for 2-C10H7CH3, 2-C10H7OCH3, and 2-C10H7Cl in Sample C1 (Radius 4.0 nm) mvp (mg)

t (nm)

γsl (mJ m-2)

Vp (cm3 g-1)

VN2 (cm3 g-1)

ΔHc (J g-1)

2-C10H7CH3

0.15

1.2

12.4

0.8

1.03

31

2-C10H7OCH3 2-C10H7Cl

0.12 0.17

1.5 2.1

25.3 10.8

1.0 1.4

1.03 1.03

70 35

derivative

Figure 9. Comparison of ΔTm as a function of r-1 for C10H8, p 2-C10H7CH3, 2-C10H7OCH3, and 2-C10H7Cl.

type hydrogen bonds with the surface. It follows that the melting point depression of 2-C10H7OCH3 is greater than that of 2-C10H7CH3, as can be seen in Figure 9. The chlorine atom is electron withdrawing due to its electronegativity.65 Thus, the electron density of 2-C10H7Cl’s π system is decreased, causing a decrease in the molecule’s ability to form π-type hydrogen bonds with the surface hydroxyl groups. Of the four compounds, 2-C10H7Cl generally has the least depressed confined melting point. From this reasoning, it is expected that ΔTm for 2-C10H7Cl differs more significantly from ΔTm for 2-C10H7CH3, and that ΔTm for unsubstituted C10H8 should be somewhere between 2-C10H7CH3 and 2-C10H7Cl. However, it is likely that other factors are involved which contribute to ΔTm. For instance, the effects of size and intermolecular interactions have not been accounted for. A 29Si MAS NMR study66 on p-dichlorobenzene and p-xylene adsorbed in the pores of ZSM-5 showed that shape and size characteristics of organic molecules are the main factors involved in interactions between the sorbate and the confining walls, and that the bond polarity is less important. Our data shows that 2-C10H7CH3 and 2-C10H7Cl have very similar ΔTm trends, which can be explained by the similar size and shape of the two molecules. However, this reasoning cannot be extended to the similarity of ΔTm between C10H8 and 2-C10H7OCH3 and we propose that intermolecular interactions may require consideration. In a study of the steric inhibition of such π-π stacking by Moorthy et al.,67 it was found that sterically hindered aryl rings did not undergo close π-π stacking, leading to solid-state properties that parallel those in the solution state. Unsubstituted C10H8 does not experience steric hindrance of π-π stacking, so intermolecular interactions are undisturbed. Pore-wall interactions cause molecules near the silica surface to align in an orderly arrangement, increasing the depression of ΔTm.52 In the case of 2-C10H7CH3 and 2-C10H7Cl, steric hindrance prevents the ordered arrangement of the molecules near the surface, resulting in a less depressed ΔTm. Although the methoxy group is a relatively bulky group, its electron donating properties increase the electron density of the aromatic π system of the parent naphthalene moiety, improving the molecules’ ability to form πtype hydrogen bonds with the surface as well as intermolecular π-π stacking interactions. Furthermore, the oxygen from the methoxy group may also interact favorably with hydrogens on

Figure 10. Raman spectra of pure solid C10H8 (PS), liquid C10H8 (PL), C10H8 confined in pores (80% loading) (PL), and C10H8 coating the pore walls (CL).

the pore surface, increasing the overall ability of 2-C10H7OCH3 to interact with the surface. Raman Spectroscopic Studies. The mass of the contact layer is very small, and it may be questionable whether a phase transition was simply undetectable in the DSC curves. To verify the nature of this layer in C10H8 and the three derivatives, we obtained the Raman spectrum of each participating in only the contact layer and compared it to that of the pure solid and solution. All pure solid spectra were acquired at 173 K, and the solution spectra were acquired at room temperature by dissolving in CCl4. Pore sample S2 was used for C10H8 and C1 for the derivatives. A portion of each prepared contact layer sample was first used to obtain a DSC curve (Figure S12 in the Supporting Information). No thermal event has occurred, indicating that all of the organic present is within the contact layer. The contact layer spectrum of 2-C10H7CH3 in pore C1 was acquired at 253 K, C10H8 and 2-C10H7OCH3 at room temperature, and 2-C10H7Cl at 273 K; these temperatures are all below their respective confined melting points. Raman spectra for C10H8 in selected regions are shown in Figure 10. The spectrum of C10H8 filling the pore (FP) was also obtained for comparison, and it shows that, below the confined melting point, the spectrum corresponds to that of the pure solid. Several Raman bands of naphthalene are sensitive to phase transition, as can be seen in Figure 10. For example, the spectrum 4745


The Journal of Physical Chemistry C of pure liquid C10H8 in the C-H stretching region looks distinctly different from that of pure solid. In CCl4 solution, C10H8 exhibits two bands at 3063 and 3055 cm-1, but in the spectrum of pure solid, there is one peak at 3055 cm-1 in this region. At room temperature, the spectrum of the sample containing only C10H8 as a contact layer in this region is quite similar to that of the solution, indicating that the contact layer is not frozen. For the sample where naphthalene is occupying the entire pore, its spectrum is identical to that of the pure solid, confirming that the confined C10H8 in the pore has crystallized. The v6 mode (CCH in-plane bending vibration) of C10H8 appears at 1460 cm-1 in the liquid spectrum but shifts to a higher energy at 1464 cm-1 upon transformation to the solid state. The frequencies of this mode in the spectra of C10H8 only in the contact layer and the completely filled pore are 1460 and 1464 cm-1, respectively. Similarly, the C-C stretching vibration mode v3 of C10H8 shows up at 1025 cm-1 in both pure liquid and the contact layer. The peak position of this mode in both pure solid and the silica with pore filled completely is at 1020 cm-1. The v1 (CCC in-plane bending) and v13 (CCC out-of-plane bending) modes both appear as singlets in the spectra of liquid and contact layer C10H8 but split into doublets in the pure and confined solid. These results show unambiguously that the spectrum of the contact layer is identical to that of solution but differs from that of solids, indicating that the contact layer is nonfreezing. Figure S13 of the Supporting Information presents the spectrum 2-C10H7CH3 in selected regions from the pure solid, liquid, and the contact layer. The differences in the solid and solution spectra are subtle compared to the differences in C10H8 spectra. Nevertheless, there are several indicators that contact layer 2-C10H7CH3 is in the liquid state, as the spectrum resembles that of the solution but not the pure solid. The strongest C-H stretching band of the CH3 group68 appears at 2915 cm-1 in pure solid, but the same peak significantly shifted to 2923 and 2926 cm-1 in pure liquid and the contact layer, respectively. The C-C stretching peak69 seen in the pure solid at 1384 cm-1 has an fwhh of 5 cm-1, which is shifted to 1381 cm-1 and broadened to 9 cm-1 in the pure liquid. The shift and line broadening are even more pronounced in the contact layer, appearing at 1378 cm-1 with an fwhh of 12 cm-1. There is also a low-frequency shift of the ring torsion vibration69 from 452 cm-1 in the pure solid to 450 and 449 cm-1 in the solution and contact layer spectra, respectively. The Raman spectra of 2-C10H7OCH3 and 2-C10H7Cl in pure solids, solution, and contact layer are show in Figures S14 and S15 of the Supporting Information. Overall, the appearance of the spectra for both compounds in the contact layer is more similar to that of the CCl4 solution rather than that in pure solids. In summary, the Raman data suggest that the contact layer of naphthalene and its derivatives in mesoporous silica hosts does exist and is nonfreezing. Powder XRD Studies. The possibility that growing a crystal inside a small pore may result in a crystal structure different from that of the bulk has been previously mentioned by Jackson and McKenna.4 It has also been noted that confinement may favor the formation of one solid form over another in polycrystalline materials.70,71 If this were the case, all of the thermodynamic parameters defined in the Gibbs-Thomson and modified Gibbs-Thomson equations would change. To this effect, powder X-ray diffraction patterns of naphthalene existing inside and outside of the spherical pore of sample S2 were obtained in order

ARTICLE

Figure 11. (a) DSC curve of C10H8 existing outside pores. (b) DSC curve of C10H8 existing only inside pores, exhibiting the confined melting transition.

Figure 12. Powder XRD pattern of C10H8 existing outside and inside pore S2. Reflections are expanded below.

to determine whether or not confinement causes a change in the crystal structure. In a typical experiment for the former, the total pore volume of the mesoporous sample was calculated after degassing and the amount of C10H8 required to fill the pores was introduced. The sample tube was flame-sealed, heated at 408 K for several hours, and then allowed to cool on the bench. To prepare a comparable sample in which C10H8 is outside of the pores, the amount of C10H8 required to fill the pores was ground into a powder and physically mixed with the mesopore sample without heating. To ensure the samples were correctly prepared, we first acquired the DSC curves shown in Figure 11. We expected that, for the sample in which C10H8 exists outside the pores, the DSC curve will show no confined melting transition of the solid. For the sample in which C10H8 resides only inside the pores, we expected to see only the confined phase transitions. Indeed, the DSC curve corresponding to the sample containing C10H8 outside of the pores shows no confined melting transition. For the DSC curve corresponding to the sample containing C10H8 inside the pores, the confined melting transition is clearly visible. Figure 12 illustrates the powder XRD patterns of C10H8 inside and outside of pore S2 at room temperature. The powder patterns were comparable to the pattern simulated on the basis of work by Brock and Dunitz.72 It is clear from Figure 12 that crystallization in the pore does not cause a change in the crystal structure. There is, however, a shift in the 2θ values to lower angles in the pattern for confined naphthalene. This implies a larger d-spacing in confinement, or lattice expansion, which is the 4746


The Journal of Physical Chemistry C result of the interfacial energy and mutual surface tension attraction of the confined material.73 It is well-known that, compared to the bulk, there is a broadening in the fwhh of XRD reflections of nanoparticles. In comparing the fwhh values of our powder patterns, we have found that there is indeed line broadening in confined C10H8. The most significant broadening can be seen in the (200), (211), and (210) reflections. The fwhh of these lines are greater than their corresponding bulk lines by 0.10-0.18. The broadening seen throughout the confined pattern further verifies that the shift in the 2θ values is a result of confinement.

’ CONCLUSIONS Materials in confinement undergo a shift in phase transition temperatures. Confined naphthalene, 2-methylnaphthalene, 2-methyoxynaphthalene, and 2-chloronaphthalene exhibit a phase transition temperature depression dependent on pore size and geometry, where a greater degree of confinement results in greater depression. Although confinement does not change the crystal structure of C10H8, our XRD results show that it does cause lattice expansion. The modified Gibbs-Thomson equation accounts for the thickness t of the nonfreezing contact layer, which can be experimentally deduced using DSC for a particular pore. Although t has sometimes been assumed constant, we have found that it generally decreases with decreasing pore size. For both the spherical and cylindrical pore systems, we have found that the freeze-melt hysteresis is generally greater in smaller pores; that is, ΔTf is greater than ΔTm. There appears a nonlinear change in hysteresis as a function of pore size, approaching zero for large pores. For a given pore diameter, the hysteresis is larger in a spherical pore. Electronic effects, relative molecular sizes and shapes, and steric hindrance must all be considered in the analysis of differences in ΔTm within this series of compounds in this study. Ultimately, such factors affect the ability of molecules to interact with the pore surface and the propensity of these interactions to determine the degree of depression of ΔTm. ’ ASSOCIATED CONTENT

bS

Supporting Information. Additional experimental results (15 figures and 2 tables). This material is available free of charge via the Internet at http://pubs.acs.org.

’ ACKNOWLEDGMENT Y.H. and J.F.C. thank the Natural Science and Engineering Research Council of Canada for research grants. Funding from the Canada Research Chair program is also gratefully acknowledged. We also thank Drs. Baines and Ragogna for using their DSC instruments and Mr. D. Li for technical help. We thank Dr. G. D. Stucky for providing the pore analysis macro. ’ REFERENCES (1) Smirnov, P.; Yamaguchi, T.; Kittaka, S.; Kuroda, Y. J. Phys. Chem. B 2000, 104, 5498–5504. (2) Morishige, K.; Kawano, K. J. Chem. Phys. 2000, 112, 11023–11029. (3) Mu, R.; Zue, Y.; Henderson, D. O.; Frazier, D. O. Phys. Rev. B 1996, 53, 6041–6047. (4) Jackson, C. L.; McKenna, G. B. J. Chem. Phys. 1990, 93, 9002–9011.

ARTICLE

(5) Hemelsoet, K.; Nollet, A.; Vandichel, M.; Lesthaeghe, D.; Van Speybroek, V.; Waroquier, M. ChemCatChem 2009, 1, 373–378. (6) Christenson, H. K. J. Phys.: Condens. Matter 2001, 13, R95–R133. (7) Pileni, M. P. J. Phys. Chem. B 2001, 105, 3358–3371. (8) Meziane, A.; Grolier, J.-P. E.; Baba, M.; Nedelec, J. M. Faraday Discuss. 2007, 136, 384–394. (9) Gibbs, J. Collected Works; Yale University Press: New Haven, CT, 1928. (10) Thomson, S. W. Philos. Mag. B 1871, 42, 448. (11) Faivre, C.; Bellet, D.; Dolino, G. Eur. Phys. J. B 1999, 7, 19–36. (12) Petrov, O.; Furo, I. Phys. Rev. E 2006, 73, 011608-1–011608-7. (13) Alba-Simionesco, C.; Coasne, B.; Dosseh, G.; Dudziak, G.; Gubbins, K. E.; Radhakrishnan, R.; Sliwinska-Bartkowiak, M. J. Phys.: Condens. Matter 2006, 18, R15–R68. (14) Alcoutlabi, M.; McKenna, G. B. J. Phys.: Condens. Matter 2005, 17, R461–R524. (15) Anandan, S.; Okazaki, M. Microporous Mesoporous Mater. 2005, 87, 77–92. (16) Zhou, Y.; Stell, G. Mol. Phys. 1989, 66, 791–796. (17) Evans, R. J. Phys.: Condens. Matter 1990, 2, 8989–9007. (18) Macpherson, A. K.; Carignan, Y. P.; Vladimiroff, T. J. Chem. Phys. 1987, 86, 4228–4232. (19) Macpherson, A. K.; Carignan, Y. P.; Vladimiroff, T. J. Chem. Phys. 1987, 87, 1768–1770. (20) Rowlinson, J. S. J. Chem. Soc., Faraday Trans. 2 1986, 82, 1801–1815. (21) Zhou, Y.; Stell, G. Mol. Phys. 1989, 66, 767–789. (22) Lettow, J. S.; Han, Y. J.; Schmidt-Winkel, P.; Yang, P.; Zhao, D.; Stucky, G. D.; Ying, J. Y. Langmuir 2000, 16, 8291–8295. (23) Zhao, D.; Feng, J.; Huo, Q.; Melosh, N.; Fredrickson, G. H.; Chmelka, B. F.; Stucky, G. D. Science 1998, 279, 548–552. (24) Ottaviani, M. F.; Moscatelli, A.; Desplantier-Giscard, D.; Renzo, F. D.; Kooyman, P. J.; Alonso, B.; Galarneau, A. J. Phys. Chem. B 2004, 108, 12123–12129. (25) Kleitz, F.; T.-W. Kim, T.-W.; Ryoo, R. Langmuir 2006, 22, 440–445. (26) Lukens, W. W., Jr.; Schmidt-Winkel, P.; Zhao, D.; Feng, J.; Stucky, G. D. Langmuir 1999, 15, 5403–5409. (27) Mitchell, J.; Strange, J. H. Mol. Phys. 2004, 102, 1997–2005. (28) Hatakeyma, T.; Kasuga, H.; Tanaka, M.; Hatakeyama, H. Thermochim. Acta 2007, 465, 59–66. (29) Lide, D. R. Enthalpy of Fusion in CRC Handbook of Chemistry and Physics, 88th ed.; CRC Press/Taylor and Francis: Boca Raton, FL, 2007. (30) Evans, R.; Marconi, U. M. B. J. Chem. Phys. 1987, 86, 7138–7148. (31) Men, Y.; Yan, Q.; Jiang, G.; Zhang, X.; Wang, W. Phys. Rev. E 2009, 79, 051602-1–051602-12. (32) Sliwinksa-Bartkowiak, M.; Gras, J.; Sikorski, R.; Radhakrishnan, R.; Gelb, L. D.; Gubbins, K. E. Langmuir 1999, 15, 6060–6069. (33) Radhakrishnan, R.; Gubbins, K. E.; Sliwinska-Bartkowiak, M. J. Chem. Phys. 2000, 112, 11048–11057. (34) Booth, H. F.; Strange, J. H. Mol. Phys. 1999, 15, 5403–5409. (35) Morishige, K.; Nobuoka, K. J. Chem. Phys. 1997, 107, 6965–6969. (36) Schreiber, A.; Ketelsen, I.; Findenegg, G. H. Phys. Chem. Chem. Phys. 2001, 3, 1185–1195. (37) Sliwinska-Bartkowiak, M.; Dudziak, G.; Sikorski, R.; Gras, R.; Radhakrishnan, R.; Gubbins, K. E. J. Chem. Phys. 2001, 114, 950–962. (38) Miyahara, M.; Gubbins, K. E. J. Chem. Phys. 1997, 106, 2865–2880. (39) Shao, Y.; Hoang, G.; Zerda, T. W. J. Non-Cryst. Solids 1995, 182, 309–314. (40) Tang, X. P.; Wang, J. C.; Cary, L. W.; Kleinhammes, A.; Wu, Y. J. Am. Chem. Soc. 2005, 127, 9255–9259. (41) Prouzet, E.; Brubach, J.; Roy, P. J. Phys. Chem. B 2010, 114, 8081–8088. (42) Takahara, S.; Nakano, M.; Kittaka, S. J. Phys. Chem. B 1999, 103, 5814–5819. 4747



ARTICLE

(43) Billamboz, N.; Baba, M.; Grivet, M.; Nedelec, J. M. J. Phys. Chem. B 2004, 108, 12032–12037. (44) Baba, M.; Nedelec, J. M.; Lacoste, J. J. Phys. Chem. B 2003, 107, 12884–12890. (45) Turnbull, D. J. Appl. Phys. 1950, 21, 1022–1028. (46) Dunning, W. J. Physics and Chemistry of the Organic Solid State; Wiley-Interscience: 1963. (47) Jones, D. R. H. J. Mater. Sci. 1974, 9, 1–17. (48) Grzyll, L. R.; Ramos, C.; Black, D. D. J. Chem. Eng. Data 1996, 41, 446–450. (49) Morineau, D.; Dosseh, G.; Alba-Simionesco, C.; Llewellyn, P. Philos. Mag. B 1999, 79, 1847–1855. (50) Petrov, O.; Furo, I. Prog. Nucl. Magn. Reson. Spectrosc. 2009, 54, 97–122. (51) Svishchev, I. M.; Plugatyr, A.; Nahtigal, I. G. J. Chem. Phys. 2008, 128, 124514-1–124514-7. (52) Takei, T.; Konishi, T.; Fuji, M.; Watanabe, T.; Chikazawa, M. Thermochim. Acta 1995, 267, 159–167. (53) Chanh, N. B.; Haget, Y.; Bonpunt, L.; Meresse, A.; Housty, J. Anal. Calorim. 1977, 4, 233–240. (54) Jiang, K.; Xie, B.; Fu, D.; Luo, F.; Liu, G.; Su, Y.; Wang, D. J. Phys. Chem. B 2010, 114, 1388–1392. (55) Chanh, N. B.; Haget, Y.; Leroy, F.; Hannoteaux, F. Acta Crystallogr. 1973, 29, 1469–1473. (56) Chanh, N. B.; Haget, Y.; Meresse, A.; Housty, J. Mol. Cryst. Liq. Cryst. 1978, 45, 307–312. (57) Meresse, A.; Haget, Y.; Chanh, N. B. Cryst. Struct. Commun. 1980, 9, 699–703. (58) Moreno, J.; Sherrington, D. C. Chem. Mater. 2008, 20, 4468–4474. (59) Linse, P. J. Am. Chem. Soc. 1990, 112, 1744–1750. (60) Ravishanker, G.; Mehrotra, P. K.; Mezei, M.; Beveridge, D. L. J. Am. Chem. Soc. 1948, 106, 4102–4108. (61) Schravendijk, P.; van der Vegt, N. F. A. J. Chem. Theory Comput. 2005, 1, 643–652. (62) Tsuzuki, S.; Honda, K.; Uchimaru, T.; Mikami, M. J. Chem. Phys. 2004, 120, 647–659. (63) Dosseh, G.; Xia, Y.; Alba-Simionesco, C. J. Phys. Chem. B 2003, 107, 6445–6453. (64) Mu, R.; Malhotra, V. M. Phys. Rev. B 1991, 44, 4296–4303. (65) Clayden, J.; Greeves, N.; Warren, S.; Wothers, P. Organic Chemistry; Oxford University Press Inc.: New York, 2001. (66) Fyfe, C. A.; Strobl, H.; Gies, H.; Kokotailo, G. T. Can. J. Chem. 1988, 66, 1942–1947. (67) Moorthy, J. N.; Natarajan, p.; Venkatakrishnan, P.; Huang, D. F.; Chow, T. J. Org. Lett. 2007, 9, 5215–5218. (68) Gussoni, A.; Zerbi, G.; van Es, J. J. G. S.; Biemans, H. A. M.; Meijer, E. W. Synth. Met. 1996, 80, 201–204. (69) Librando, V.; Alparone, A. Polycyclic Aromat. Compd. 2007, 27, 65–94. (70) Berwanger, R.; Henschel, A.; Knorr, K.; Huber, P.; Pelster, R. Phys. Rev. B 2009, 79, 125442-1–125442-12. (71) Sliwinksa-Bartkowiak, M.; Jazdzewska, M.; Huang, L. L.; Gubbins, K. E. Phys. Chem. Chem. Phys. 2008, 10, 4909–4919. (72) Brock, C. P.; Dunitz, J. D. Acta Crystallogr., Sect. B: Struct. Crystallogr. Cryst. Chem. 1982, 38, 2218–2228. (73) Wei, Z.; Xia, T.; Ma, J.; Feng, W.; Dai, J.; Wang, Q.; Yan, P. Mat. Char. 2007, 58, 1019–1024.

4748


Phase Transitions of Naphthalene and Its Derivatives Confined in

Recommend Documents