Edinburgh Research Explorer High-Pressure Structural Behavior of para-Xylene

,


INTRODUCTION
−3 Several organic solvents have been crystallized at room temperature upon compression.−7 Liquid benzene crystallizes at room temperature and about 0.07−0.15GPa; this phase is identified as equivalent to low-temperature orthorhombic phase I (Pbca).A new form, phase II (P2 1 /c), can be obtained at about 1 GPa.Other examples of pressure-induced crystallization include toluene (C 7 H 8 ), 8 cyclohexane (C 6 H 12 ), 9 and carbon tetrachloride (CCl 4 ). 10 The phenomenon of organic molecules crystallizing in different arrangements is known as polymorphism. 11The application of elevated pressures has been demonstrated to be an excellent means of exploring polymorphism in molecular materials. 12,13Polymorphism can be critically important in material manufacturing and processing, as different polymorphs may exhibit very different physicochemical properties, thereby affecting their performance and functionality. 14ara-Xylene (p-xylene), with the chemical formula C 8 H 10 , is an aromatic hydrocarbon with a normal freezing point (the freezing point at p = 1 atm) of 286 Ksee Figure 1a for the molecular structure.It is a raw material in the large-scale synthesis of many polymers, including polyethylene terephthalate, which is further used as polyester fiber, resin, and film. 15The annual production of p-xylene is industrially significant with more than 8 kg produced per capita each year in the United States 16 (i.e., nearly 3 million metric tons annually).The similar boiling points of various xylene isomers (p-xylene: 411 K, m-xylene: 412 K, and o-xylene: 417 K) make them difficult to separate by distillation.On an industrial scale, p-xylene is separated from a mixture of its aromatic isomers by a two-stage crystallization process at low temperatures. 17he first crystallographic study of p-xylene was reported in 1986 from a single-crystal X-ray diffraction (SCXRD) measurement at 180 K, and it was found to crystallize in the monoclinic space group P2 1 /n with a unit-cell volume per formula unit (V/Z) of 160.5 Å 3 . 18No structural phase transition was observed between 110 and 273 K.In a later study, the crystallographic structure was also determined by neutron powder diffraction. 19This study confirmed that there is no temperature-induced phase transition at ambient pressure down to a temperature of 5 K. Two other low-melting isomeric forms of xylenes meta-xylene (1,3-dimethyl benzene, m.p. 225 K) and ortho-xylene (1,2-dimethyl benzene, m.p. 248 K) crystallize at low temperatures into orthorhombic (Pbca) 19 and monoclinic (P2 1 /a) 20 crystal systems, respectively.
The molecular structure of p-xylene is given in Figure 1a and we use the same atomic numbering scheme as Ibberson et al. 19 The crystallographic structure comprises half a molecule in the asymmetric unit and the full molecule is generated through an inversion center.The p-xylene molecules are packed into chains along the b-axis (as shown in Figure 1b) by nearly equal van der Waals contacts between the six ring C atoms and one methyl H of a neighboring molecule.The chains are linked by three dominant H•••H contacts (as shown in Figure 1c).The rotational orientation of the methyl groups can be measured Pressure-induced crystallization in both o-and p-xylene has been reported previously. 21,22The crystallization behavior of pxylene under high pressures was investigated by optical microscopy, powder X-ray diffraction (PXRD), and vibrational spectroscopy. 21p-xylene is reported to crystallize at ∼0.1 GPa.In contrast to the low-temperature structure, the PXRD pattern of p-xylene at 0.84 GPa was indexed to a monoclinic unit cell with space group Cc.The authors were unable to solve the crystal structure, and so little is known about the bonding and intermolecular contacts in the new polymorphic structure.The study also reported that pressure-induced polymerization occurred around 13 GPa.This polymeric phase was characterized using the results from PXRD investigations at 17.2 GPa.The polymerized solid phase was quenchable to ambient conditions.Again, the structure could not be solved from the limited data obtained from the X-ray diffraction experiment inside the diamond anvil cell (DAC).
No such pressure-induced phase transitions are, however, observed upon compression of o-xylene.In 2017, Marciniak and Katrusiak conducted a high-pressure single-crystal diffraction study 22 on o-xylene and reported that the crystal structure of the high-pressure phase (freezing pressure 0.25 GPa) is consistent with the monoclinic structure of the perdeuterated o-xylene crystal at 2 K. 20 The apparent discrepancy between pressure effects on o-and p-xylene could be related to the rotation of methyl groups.Ab initio calculations of the isolated molecule show that the methyl groups of p-xylene and m-xylene rotate almost freely. 23The situation is different for o-xylene as two methyl groups in close proximity hinder free internal rotation.The apparent discrepancy between the pressure effects on p-and o-xylene is particularly intriguing.A new investigation into the pressure response of p-xylene with the aim of solving the proposed highpressure phase is important.
In order to determine the structure of the Cc phase of pxylene, we have performed neutron powder diffraction (NPD) and single-crystal X-ray diffraction studies at elevated pressure.The NPD technique proves to be particularly beneficial for polycrystalline organic materials containing atoms with low Z due to the comparable scattering power of C and D in neutrons.Therefore, the accurate location of hydrogen (D) positions can be obtained from neutron diffraction on the deuterated sample.Moreover, it has been shown that neutron techniques cause much less radiation damage for molecular materials. 24,25o further complement our experimental findings, we performed first-principles simulation as well.Density functional theory (DFT) has demonstrated a remarkable ability to predict the evolution of structures with pressure, particularly when weak dispersion forces are accounted for. 26The continued validation of increasingly sophisticated dispersion models has demonstrated that the DFT plus dispersion (DFT-D) model not only can reproduce experimentally determined hydrostatic compression behavior but can also be used to guide experimental endeavors. 25he objectives of this research effort were therefore as follows: (i) to perform the first high-pressure neutron diffraction study on perdeuterated p-xylene to solve the highpressure phase, (ii) to reproduce the pressure-induced crystallization on hydrogenous p-xylene inside a DAC and to compare the high-pressure single-crystal structure with the neutron powder data; (iii) to explore the ability of modern DFT models to predict the hydrostatic compression behavior, and (iv) to obtain accurate pressure−volume equations of state.
The extent of deuteration and isomeric purity of the sample were verified by 2 H (D) NMR and 13 C NMR, respectively (Figure S1, Supporting Information).
2.2.Neutron Powder Diffraction.High-pressure time-of-flight neutron-diffraction measurements were performed in two separate experiments on the PEARL instrument 27 at the ISIS Neutron and Muon Source, U.K. In the first experiment, the sample was loaded as a powdered solid at low temperatures.In the majority of cases, simply freezing the sample, from the liquid phase, promotes the growth of large crystallites resulting in significant preferred-orientation effects, and thus structure determination becomes challenging.For this experiment, crystalline powder samples were prepared by a lowtemperature (77 K, liquid nitrogen) hand grinding method using 5 mL of p-xylene-d 10 (normal m.p. 278 K). 28 The powdered sample was loaded together with a mixture of perdeuterated n-pentane/ isopentane as the pressure-transmitting medium (PTM) 29 to maintain hydrostatic pressure conditions.In the second experiment, the liquid sample was loaded without any PTM.In both experiments, a small Crystal Growth & Design piece of lead foil was included in the sample container to act as a pressure calibrant. 30Pressure was generated using a V3 variant Paris− Edinburgh press 31 with single-toroidal zirconia-toughened alumina anvils.A null-scattering-encapsulated TiZr gasket was used 32 and neutron-diffraction data were collected in approximately 2.5 tonne steps up to a maximum applied load of 60 tonnes (∼4.7 GPa).A beamline-developed correction for the wavelength and scatteringangle dependence of the neutron attenuation by the anvil and gasket materials was applied to the measured pattern.The data were normalized using the MANTID software. 33Rietveld refinements of the time-of-flight NPD patterns were performed in Topas Academic V5.2. 34A rigid-body model was used to limit the number of refined parameters.We used a rigid body described with a Z-matrix, and bond distance and angle restraints were informed by high-quality neutron structure ZZZITY02 (HRPD, 4.5 K). 19 All the carbon atoms in the molecule were constrained to be planar.The bond lengths C1−C2/ C3, C1−C4, and C−D were set to 1.39, 1.54, and 1.07 Å, respectively; the methyl group angles (D−C−D), the D2/D3 angles with the aromatic ring, and the half C2−C1−C3 angles were set to 110.6, 117, and 59°, respectively.We allowed a single parameter to refine describing a common D−C−D angle and refined a single parameter relating to the torsion angle made between a methyl deuterium atom and the main carbon ring.The other two deuterium atoms were constrained to be ±120 degrees of the latter parameter, which imposed the C 3v symmetry.
The liquid sample was loaded into a 250 μm diameter hole drilled in a pre-indented stainless steel gasket (thickness ≈ 150 μm).Small pieces of ruby were loaded in the DAC along with the sample as a pressure calibrant.The pressure within the gasket hole was determined by the ruby fluorescence method 36 using a Jobin-Yvon LabRam 300 spectrometer equipped with a 50 mW He−Ne laser of wavelength 632.8 nm.The DAC chamber was filled with p-xylene and sealed; p-xylene compressed at RT freezes into a polycrystalline mass at 0.1 GPa.A single crystal was obtained from this polycrystalline phase by slowly reducing the pressure to allow the crystallites to melt until only one remained in equilibrium with the liquid.Pressure was subsequently increased gradually until this crystallite grew to form a sufficiently large single crystal.
High-pressure SCXRD data were measured on a Bruker D8 Venture diffractometer with a Mo Kα X-ray source (λ = 0.71073 Å).
Diffraction data were processed using the APEX3 suite of programs. 37he data were integrated using dynamic masks (generated using the program ECLIPSE) 38 using SAINT 37 to mask shaded detector areas.Absorption and shading corrections were applied using the multi-scan procedure SADABS. 39Data sets were initially analyzed and space group of the crystal structure was determined using XPREP. 37The structure of p-xylene was solved by direct methods and refined with full-matrix least-squares against F 2 using ShelXS and ShelXL implemented in Olex2. 39,40.4.Computational Methods.Plane-wave DFT calculations were performed using CASTEP v17.21, 41 the initial coordinates for the structures were taken from the previously reported ambient pressure low-temperature structure of p-xylene. 18The Hamiltonian operator was approximated using the Perdew−Burke−Ernzerhof (PBE) 42 exchange−correlation functional augmented with the many-body dispersion correction. 43The molecular wave function description was provided by "on-the fly" pseudopotentials and a plane wave basis set operating at 650 eV, which gave the convergence of the energy to within 4 meV per atom.The electronic structure was sampled in the reciprocal space using a Monkhorst−Pack (MP) grid, with a spacing of 0.035 Å −1 . 44The potential energy surface was searched for energy minima by means of the Broyden−Fletcher− Goldfarb−Shanno (BFGS) algorithm. 45Structures were considered to be optimized when the total energy of the system, maximum force, maximum stress, and maximum atomic displacement converged to the values of 10 −8 eV, 10 −2 eV Å −1 , 10 −2 GPa, and 10 −3 Å, respectively.The structures were initially optimized without the application of an external applied stress.Once the first, ambient pressure, model was optimized, an external hydrostatic pressure of 0.5 GPa was applied and the system re-optimized; this process was repeated at 0.5 GPa steps until an external pressure of 5 GPa was reached.

RESULTS AND DISCUSSION
3.1.Crystal Structure of High-pressure Polymorphs from NPD.In our first experiment, a crystalline solid powder sample of p-xylene (prepared by rapid freezing using liquid nitrogen) was loaded, and NPD patterns were collected across a pressure range of 1.05(1)−3.90(4)GPa, (Figure S2a, Supporting Information).In the second compression experiment, liquid p-xylene was loaded and was allowed to crystallize upon compression.NPD patterns of the second experiment were obtained across a pressure range of 0.11(1)−4.72(2)GPa (Figure S2b, Supporting Information).The purpose of the Crystal Growth & Design liquid loading experiment was to check whether the material crystallizes differently compared with the solid-loading experiment.Powder diffraction patterns obtained from both the NPD experiments could be fitted to the known lowtemperature crystal structure (P2 1 /n).
Figure 2a shows the Rietveld refinement of the NPD pattern collected at 1.05(1) GPa (experiment 1, solid loaded sample), using the low-temperature P2 1 /n structure.The full structural details are provided in Table S1, Supporting Information.The quality of the Rietveld fit (R wp = 3.98%) is very good, and note that all peaks are accounted for.Figure 2b shows a Pawley fit of the same NPD pattern using the Cc cell of the reported HP phase; 21 the cell dimensions at 0.84 GPa were used as starting values.We have fixed all the lead/anvil parameters and allowed the remainder to be fitted by the Cc cell (a = 10.77Å, b = 9.16 Å, c = 13.06Å, and β = 136.78°).It is evident that a couple of peaks (particularly peaks at d-spacings of 2.35 and 2.73 Å) are clearly not fitted in the Pawley fit.We believe that the indexing of the PXRD from the previous report 21 is erroneous due to the possible inclusion of two peaks resulting from an impurity phase.A simulated PXRD pattern from the crystal structure obtained from our refinement of neutron powder data at 1.05(1) GPa is found to be very similar to the PXRD pattern obtained at 0.84 GPa in the previous study (see Figure S3, Supporting Information).We, therefore, suggest that the true structure actually has a P2 1 /n symmetry.
The quality of the Rietveld fit for the liquid sample is lower because of the preferred orientation resulting from the rapid freezing of the liquid upon compression.However, lattice parameters obtained from both the first and second compression experiments are in good agreement (Tables S3  and S4, Supporting Information).As we did not see the proposed phase transition with the neutron experiment, we wanted to rule out the possibility that there is some effect of deuteration that might cause the transition to be suppressed.
3.2.Pressure-Induced Crystallization inside a DAC to Further Investigate the Effect of Deuteration.To exclude the possibility of the deuteration effect, we conducted pressureinduced crystallization on hydrogenous p-xylene (C 8 H 10 ) inside a DAC.Small crystallites appeared to form at 0.1 GPa.We succeeded in growing a sufficiently large single crystal by careful pressure cycling (see the Experimental and Computational Methods section for details) as shown in Figure S4, Supporting Information.We performed a SCXRD at 1.0(1) GPa.The high-pressure crystal structure was solved in the P2 1 /n space group.Table 1 compares the lattice parameters of p-xylene at around 1 GPa from the NPD refinement and from the SCXRD structure solution.The lattice parameters obtained from the high-pressure SCXRD measurement are in good agreement with the results obtained from high-pressure neutron diffraction data (see Table 1).The torsion angles (τ1, τ2, and τ3) from SCXRD data are 12.1, 46.6, and 73.4°, respectively.The van der Waals contacts between H43 and the C atoms of the ring in a neighboring molecule are within 2.94−3.17Å.The three short H•••H contacts (H3•••H41, H3•••H42, and H3•••H43) are 2.41, 2.75, and 2.57 Å, respectively.Note that H positions could not be individually refined due to the poor quality data from DAC, and therefore these values may have larger errors.We will compare the torsion angles and intermolecular contacts with the values obtained from NPD data in Section 3.5.
3.3.Computational Structure at Ambient Pressure.Structural parameters were obtained from geometry optimization of the starting form (P2 1 /n, Z = 4) at ambient pressure.The unit cell determined from the calculation (a = 6.0432Å, b = 4.9447 Å, c = 12.4096 Å, β = 105.221°,and v = 357.814Å 3 ) is larger than the values obtained from the experimental reports, which suggests that this correction scheme overestimates the effects of dispersion for this crystal structure at ambient pressure.Hybrid DFT functionals can be considered, as these contain the formalism for exact exchange interactions, which pure DFT lacks; however, it has been shown that many of these hybrid functionals only work through the cancellation of errors. 46It is worth noting that this overestimation is predominantly the result of overestimating the crystallographic c-axis length.This axis is in the direction of the largest proportion of the void space, with weak dispersion interactions between the p-xylene molecules, and is therefore most responsive to liquid to solid crystallization effects. 26It also should be emphasized that the DFT-D calculations optimize a static crystal structure at essentially 0 K and thus the effects of thermal motion are neglected.Crystal Growth & Design angles (∼121°), while the substituted site (C1) is slightly deformed with a smaller C−C−C bond angle (∼117°).This result is consistent with previous experimental reports. 18,19econd, the rotational orientation of the methyl group from the DFT-optimized structure matches very well with a lowtemperature form; deviations of torsion angle values are < 2°f or NPD (5 K) and < 5°for SCXRD (180 K).Third, the van der Waals contacts between H43 and the C atoms of the ring in a neighboring molecule are in excellent agreement (deviation <0.1 Å).Fourth, the three dominant H  S2, Supporting Information, along with previous experimental reports.To summarize, our computational model overestimates the effects of dispersion for this crystal structure; however, the calculated molecular geometry and intermolecular contacts are in excellent agreement with experimental results.We, therefore, decided to use this model for highpressure computational investigation of p-xylene.

Compressibility and EoS.
In both the NPD measurements (solid and liquid loading), no significant changes are observed up to the highest pressure, other than the shifting of the Bragg reflections to the lower d-spacings.The Rietveld refinements were performed at each pressure point on diffraction patterns collected from a solid loaded sample (Figure S5, Supporting Information).This allowed us to determine the unit-cell parameters and atomic positions with pressure; details of all structural refinements along with figures-of-merit (R wp ) are available in Table S3, Supporting Information.For the liquid loaded sample, the quality of the patterns deteriorated due to non-hydrostatic conditions in the sample in the absence of a PTM.Moreover, we observed a significant preferred orientation and so it was difficult to perform Rietveld refinements; lattice parameters were obtained from Pawley fitting (Table S4, Supporting Information).In both experiments, all three lattice parameters decrease continuously and show a smooth trend (Figure 3).The compression is nearly isotropic; the crystallographic c-and aaxes are slightly more compressible than the b-axis.The overall compressibility over the range 0.1−4.7 GPa shows V/V 0 = 0.70, with a/a 0 = 0.911, b/b 0 = 0.924, and c/c 0 = 0.906.The smooth compression trend confirms that no first-order phase transitions take place between 0 and 4.7 GPa under hydrostatic conditions.We have determined the principal axes (a set of orthogonal axes) and their relation to the unit-cell axes along with their corresponding compressibilities. 47Directions of principal axes and their linear compressibilities are provided in Table S6, Supporting Information.
Computationally, the ambient pressure crystal structure (P2 1 /n, Z = 4) was compressed up to 5 GPa.The aim was to see any discontinuity in structural parameters with the variation of pressure, which can be indicative of an underlying phase transition. 48The effect of pressure on the lattice parameters is shown as relative unit-cell compression (%) in Figure S6, Supporting Information.
The variations in the unit cell parameters (a, b, c, β, and V) are tabulated in Table S5, Supporting Information.The overall smooth compression trend indicated that no phase transition is observed.With the increase of pressure, the c-axis is found to be the most compressible axis, followed by a-and b-axes.Note the compressibility of the c-axis is higher due to the overestimation of the c-axis in the ambient pressure DFT calculation.This resulting overestimation stems from the use of the PBE functional and is minimized through the use of a dispersion correction. 26igure 4 shows the variation in the experimental and calculated unit-cell volumes of p-xylene as a function of pressure.The hydrostatic compression trend predicted by the DFT-D method is in good agreement with the compression trend determined from the NPD experiment.The variation in the experimental unit-cell volumes as a function of pressure was fit to a semi-empirical third-order Birch−Murnaghan EoS; a second-order fit yielded an unreliable V 0 value, and a poor fit to the data.Given the limited number of data points in the low-pressure region for the solid-loading sample, we report our experimental bulk modulus coefficients from the liquid loading  Crystal Growth & Design experiment: bulk modulus B 0 = 3.5(4) GPa, bulk modulus derivative B′ = 14(1), and V 0 = 340(2) Å 3 .The values determined from the solid loading were determined to be the same within the error provided by the least-squares fit and the data points lie on the EoS determined from the liquid data set.A smaller B 0 and a larger B′ value of p-xylene indicate that the material is highly compressible at lower pressure, and the solid stiffens at higher pressures.The bulk modulus value of p-xylene is broadly consistent with other low-melting organic molecular liquids (e.g., B 0 = 1.05 GPa for benzene). 49.5.Effect on Molecular Geometry with Pressure.Figure 5a shows the experimental and computational variation of torsion angles with pressure.The changes of torsion angles determined from NPD refinements are broadly in agreement with the DFT-D obtained values.At low pressures, the torsion angles vary with increasing pressure.However, at higher pressures (>2 GPa), the torsion angles appear to remain invariant with pressure.It is interesting to note with increasing pressure, two torsion angles (τ1 and τ3) decrease and the third one (τ2) increases.We believe that the methyl D atoms move slightly in such a way that any unfavorable interactions can be avoided.The torsion angles obtained from our high-pressure single-crystal structure (see Section 3.2) follow the general trend, that is, two torsion angles (τ1, τ3) are smaller and the third one (τ2) is larger than the values obtained at ambient pressure.The values of torsion angles at 1 GPa differ between the two measurement techniques (SCXRD and NPD).The difference in refined torsion angles is almost certainly to do with how they have been refined.The refinements probably do not make a meaningful comparison between the two techniques due to poor data quality (high-pressure experiments).We expect that the torsion angles obtained from neutron powder diffraction data are more reliable, as the methyl group was constrained to have a C 3v symmetry and thus D positions can be reliably refined.
All the three dominant D•••D contacts and H43•••C (ring) decrease smoothly with increasing pressures; the corresponding values are provided in Table S7, Supporting Information.

CONCLUSIONS
This work reports the high-pressure crystal structure of pxylene using neutron powder diffraction, SCXRD, and DFT-D methods.We find no evidence for a new high-pressure phase with the Cc space group, as reported in a recent study.We believe that the phase transition reported in the earlier investigation is erroneous because of the inclusion of reflections from an impurity phase.Neutron powder diffraction patterns of perdeuterated p-xylene (C 8 D 10 ) obtained between 0.1 and 4.7 GPa can be indexed with the known lowtemperature P2 1 /n structure.Moreover, pressure-induced crystallization of hydrogenous p-xylene inside a DAC allowed us to solve the crystal structure from SCXRD.The highpressure single-crystal structure is consistent with the known low-temperature form and further supports our neutron diffraction results.Moreover, the calculated molecular geometry and intermolecular contacts from the DFT-D calculations agree well with the experimental results.The smooth compression trend produced by the two separate neutron powder diffraction studies (solid-loaded and liquidloaded sample) on perdeuterated p-xylene is consistent with the hydrostatic compression trend predicted by the DFT-D method.The overall good agreement between our neutron diffraction results and high-pressure single-crystal measurement with computational findings shows the importance of combining both experimental and computational tools for the study of structural evolution with pressure for molecular materials.

Figure 1 .
Figure 1.(a) Molecular structure and atomic numbering of p-xylene along with three torsion angles (τ); (b) molecular arrangement of monoclinic p-xylene and the six van der Waals contacts between H(43) and the C atoms of the ring in a neighboring molecule are shown in red dotted lines; and (c) along the b-axis, H•••H contacts form chains of molecules (shown as blue dotted lines).

Figure 2 .
Figure 2. (a) Rietveld refinement fit of p-xylene at 1.05(1) GPa using the P2 1 /n space group; (b) Pawley fit of the same NPD pattern using a Cc space group.In both (a,b), experimental (observed) data are shown as red dots, the solid black line shows the calculated profile from the refinements, and the bottom blue traces show the residual intensities I(obs) − I(calc).The simulated Bragg reflections for each phase are given as vertical tick marks: from top to bottom p-xylene, Pb (the pressure calibrant), Al 2 O 3 , and ZrO 2 (the anvil material).

Figure 3 .
Figure 3. Lattice parameters as a function of hydrostatic pressure of pxylene.Solid symbols: first NPD measurements (loaded as solid samples), open symbols: second NPD measurements (loaded as a liquid sample).a-axis is shown in blue, b-axis in red, c-axis in green, and β-angle in purple; note that error bars are smaller than the symbols plotted.

Figure 4 .
Figure 4. Unit-cell volumes as a function of pressure for p-xylene; open red squares: NPD data (solid-loading); open green circles: NPD data (liquid-loading); open black circles: DFT-D data; and green solid line: the EoS fit using liquid P−V data; error bars are shown but are smaller than the symbols.
Figure 5b shows the variation of the shortest D•••D contacts (D41•••D3) with pressure along with values obtained from DFT-D calculations and SCXRD measurements at 1 GPa.The experimental values from NPD data are consistent with the computational results.The shortest H•••H contact from the SCXRD data measured at 1 GPa is slightly larger than the corresponding D•••D distances from DFT-D and the neutron data set.The slight variation of distances could be related to the isotope effect hydrogen/deuterium as H/D distances are known to differ.In brief, both NPD results and computational calculations show a smooth decrease in intermolecular contacts with a slight variation of torsion angles in the low-pressure range (0−2 GPa).

Figure 5 .
Figure 5. (a) Variation of torsion angles of p-xylene with pressure and (b) variation of shortest D•••D contacts (D3•••D41) of p-xylene with pressure.Note that in both (a,b) DFT values are shown as open circles, NPD as solid circles, and the corresponding values from the SCXRD measurement at 1 GPa as triangle symbols.

Table 2
compares computational torsion angles, H•••H contacts, and the van der Waals distance (between H43•••C of the ring in a neighboring molecule) with previous experimental reports.Calculated molecular geometries and intermolecular contacts are in reasonable agreement with experimental results.First, in the DFT optimized structure, the unsubstituted carbon sites (C2 and C3) have nearly equal C−C−C bond

Table 1 .
Unit-Cell Parameters of p-Xylene at p ≈ 1 GPa; Note That Values Obtained from the Hydrogenous Sample (SCXRD) Are Compared against the Values Obtained from the Perdeuterated Sample (NPD Refinement)

Table 2 .
Torsion Angles and Intermolecular Contacts As Obtained from the DFT-D Geometry Optimization for p-Xylene, along with Data from Two Previous Experimental Reports (Refs 18 and 19)