16232
J. Phys. Chem. B 2006, 110, 16232-16238
19F
Single-Quantum and 19F-33S Heteronuclear Multiple-Quantum Coherence NMR of SF6 in Thermotropic Nematogens and in the Gas Phase Henri Tervonen, Jani Saunavaara, L. Petri Ingman,† and Jukka Jokisaari* NMR Research Group, Department of Physical Sciences, UniVersity of Oulu, P.O. Box 3000, FIN-90014 Oulu, Finland ReceiVed: April 13, 2006; In Final Form: June 20, 2006
19F
single-quantum (SQC) and 19F-33S heteronuclear multiple-quantum coherence (HMQC) NMR spectroscopy of sulfur hexafluoride (SF6) dissolved in thermotropic liquid crystals (TLCs) were used to investigate the properties of TLCs. On one hand, environmental effects on the NMR parameters of SF6, 19F nuclear shielding, 19F-33S spin-spin coupling, secondary isotope effects of sulfur on 19F shielding, and the self-diffusion coefficient in the direction of the external magnetic field were studied as well. The temperature dependence of the 19F shielding of SF6 in TLCs was modeled with a function that takes into account the properties of both TLC and SF6. It appears that the TLC environment deforms the electronic system of SF6 so that the 19F shielding tensor becomes slightly anisotropic, with the anisotropy being from -0.5 to -1.4 ppm, depending upon the TLC solvent. On the contrary, no sign of residual dipolar coupling between 19F and 33S was found, meaning that the so-called deformational effects, which arise from the interaction between vibrational and reorientational motions of the molecule, on the geometry of the molecule are insignificant. Diffusion activation energies, Ea, were determined from the temperature dependence of the self-diffusion coefficients. In each TLC, Ea increases when moving from an isotropic phase to a nematic phase. The spin-spin coupling constant, J(19F,33S), increases by ca. 10 Hz when moving from the gas phase to TLC solutions. The secondary isotope shifts of 19F shielding are practically independent of TLC solvent and temperature. For the first time, 19F-33S heteronuclear multiple-quantum NMR spectra were recorded for SF6 in the gas phase and in a liquid-crystalline solution.
Introduction NMR spectroscopy of noble gases (3He, 21Ne, 83Kr, 129Xe, and 131Xe) has been used for years to probe physical properties of liquid crystals.1 Particularly popular has been 129Xe NMR because of its exceptional sensitivity to local environmental effects. However, the low signal intensity of natural and even 129Xe enriched xenon often leads to a fairly long experiment duration which may be a limiting factor, for example, in 2D EXSY and diffusion experiments.2 On the other hand, 129Xe NMR sensitivity can be enhanced by several orders of magnitude by using optical pumping.3 Despite this, it is worth attempting to find alternative and supportive methods for studies focusing on the properties of thermotropic liquid crystals. One possibility is the 19F NMR of sulfur hexafluoride (SF6) which has recently been used, for example, in NMR imaging of rat lungs4 and porous materials5 and in studying cross-linking of rubbers.6 19F NMR has some experimental benefits compared to 129Xe NMR. First, the sensitivity of 19F is ca. 145 times that of 129Xe, and the six equivalent fluorine atoms further improve the situation. Second, the 19F spin-lattice relaxation time is significantly shorter than that of 129Xe for the gases dissolved in thermotropic liquid crystals (TLCs). This allows for fast signal accumulation and thus makes it feasible to carry out, for example, diffusion experiments, within a reasonable amount of time. * To whom correspondence should be addressed. E-mail:
[email protected]. Tel: +358 8 5531308. Mobile: +358 40 5956146. Fax: +358 8 5531287. † Present address: Instrument Centre, Department of Chemistry, University of Turku, FIN-20014 Turku, Finland.
In the present study, we have applied 19F NMR of SF6 dissolved in TLCs to study the properties of liquid crystals on one hand and the effect of TLC solvent on the NMR properties of SF6 on the other hand. For these purposes, we have determined the shielding tensor and diffusion coefficient through 19F NMR experiments at variable temperatures in four TLCs: ZLI 3125 and ZLI 2806 (the director of these LCs orients perpendicularly to the external magnetic field, B0, because their anisotropy of the diamagnetic susceptibility is negative), HAB and Phase 4 (the director orients itself parallel with B0, because their anisotropy of diamagnetic susceptibility is positive). The earlier developed theoretical model for describing the 129Xe shielding behavior in TLCs was applied to interpret the temperature dependence of 19F shielding.7 Furthermore, the SF6 self-diffusion coefficient in the direction of the external magnetic field, D| , was determined in the above-mentioned liquid crystal samples at variable temperatures, allowing for the determination of the activation energies Ea in the isotropic and liquidcrystalline phases. Apart from shielding and diffusion, the 19F-33S spin-spin coupling constant, J(19F,33S), and the secondary isotope effect caused by sulfur isotopes on 19F shielding, 1∆19F(m′/mS), were studied in the TLC solvents as a function of temperature. To better resolve the 33S satellites in the 19F spectra, 19F33S HMQC experiments were performed in the Phase 4 LC and in the gas phase as well. Earlier, Jackowski et al. carried out such experiments for a high pressure (ca. 20 atm) gas sample, but to our knowledge, this is the first time that such experiments have been reported for an LC solution.8
10.1021/jp062296m CCC: $33.50 © 2006 American Chemical Society Published on Web 07/29/2006
19F
SQ and
19F-33S
HMQC NMR of SF6
J. Phys. Chem. B, Vol. 110, No. 33, 2006 16233
TABLE 1: Compositions and Phase Transition Temperatures of the Used TLCs code name
composition
transition temperatures (K)a
HAB Phase 4 ZLI 2806 ZLI 3125
p,p′-di-n-heptylazoxybenzene eutectic mixture of p-methoxy-p′-n-butylazoxybenzenes mixture of alkylbicyclohexyl and alkyltercyclohexyl alkylbicyclohexyl
C-307-SmA-326.5-N-344.5-I C-293-N-347-I C-270-N-373-I C-266-N-336-I
a Transition temperatures are given for pure liquid crystals. Dissolved gas shifts them to some extent: C ) crystal phase, SmA ) smectic A phase, N ) nematic phase, and I ) isotropic phase.
Figure 1.
19
F-33S HMQC 2D pulse scheme.
Experimental Section SF6 (99.75%, from Sigma-Aldrich) gas was dissolved in four TLCs, ZLI 3125, ZLI 2806, HAB, and Phase 4. All the other liquid crystals, except HAB, were delivered by Merck (Darmstad, Germany). HAB was obtained from the University of Basel (Switzerland). The compositions and phase transition temperatures of the used TLCs are shown in Table 1. The TLCs were placed in 5 and 10 mm thick wall Wilmad glass tubes and carefully degassed prior to SF6 gas transfer and flame sealing. The 19F SQ experiments were performed on Bruker Avance DRX500 and DPX400 spectrometers equipped with 5 and 10 mm inverse broadband probes and integrated Z-gradient coils. The equilibrium pressure of the SF6 gas was ca. 2.0 atm in the 5 mm samples. In the 19F-33S HMQC experiments, carried out on the DRX500, a 10 mm broad band probe was used. The used pulse scheme is shown in Figure 1. The pressure of the 10 mm gas sample was 3 atm, and the pressure of the SF6 gas in the 10 mm Phase 4 sample was about 5 atm. The proton coils of all probes were tuned for fluorine. The fluorine resonance frequency was 470.62 and 376.50 MHz on 500 and 400 MHz instruments, respectively, and the sulfur resonated at 38.39 MHz on DRX500. A slight annoyance arose from the common amplifier (BLAXH40) and the preamplifier (HPPR X-BB-19F-2HP) of fluorine and sulfur on the DRX500 instrument. Because the Bruker standard phase-cycled HMQC pulse sequences (without decoupling) were used in HMQC experiments, they had to be edited to route 19F and 33S pulses into different outputs from the common amplifier. Tuning and 90° pulse determinations were able to be done separately with the common preamplifier, but in every two nuclei experiment, the cable of the nonobserved nuclei had to bypass the preamplifier. Its effect on the 90° pulse length was minimized by adjusting the power of the bypassing pulse by the oscilloscope to be the same as when the pulses were passing through the preamplifier. When the signal was observed in a short time, the pulses were further optimized with the actual spectrum. A (NH4)2SO4 sample was used to determine the 90° pulse length for 33S and to calibrate the chemical shift range. The 90° pulses of both 19F and 33S were ca. 40 µs at full power (PL ) -6). 2D HMQC was measured to locate the sulfur signal in the Phase 4 sample. Although the fluorine-decoupled 33S signal was observed in the gas sample, no signal was detected in the liquid
Figure 2. (a) Rf pulses and (b) pulse gradients of the PGSE experiment.9 (c) Rf pulses and (d) pulse gradients of the DSE experiment.10
crystal sample after 60 h of accumulation most probably due to low solubility. We used the information obtained from the gas sample, that is, resonance frequencies, coupling constant (about 250 Hz), and short repetition rate due to the fast relaxation of both nuclei. An overnight run with 128 transients in an indirect direction, 896 scans on each spectrum, and with a 0.83 s repetition rate gave a reasonable spectrum. After the shift of the sulfur signal was found in the two-dimensional (2D) spectrum, the coupling could be determined more precisely by 1D HMQC. A spectrum with a good signal-to-noise ratio could be observed in half an hour. (An even faster repetition rate could be used with the gas sample, and a 1D HMQC spectrum could be detected in a few minutes.) Although 2D HMQC was a helpful component in this study, it also proved that the method itself can be successfully used with more complex sulfur compounds. The temperature was calibrated with samples of ethylene glycol (80%) in DMSO-d6 (high temperatures) and methanol (4%) in methanol-d4 (low temperatures). The chemical shift of 19F was referenced to the corresponding signal in free gas (pressure ca. 2.0 atm). The self-diffusion measurements were carried out by utilizing double spin-echo (DSE) and the pulse gradient spin-echo (PGSE) sequences with ramped Z-gradient pulses.9,10 The DSE sequence was used at high temperatures to compensate for possible thermal convection flows, and the less time-consuming PGSE pulse sequence was used at low temperatures where convection may not occur. Thermal convection may start if the sample is heated from the bottom, as it usually is in most NMR spectrometers, when heating creates a large enough temperature gradient over the sample volume.11 The field gradient steps and the values for the parameters shown in Figure 2 were individually adjusted for each sample at each temperature to achieve the best possible results. To avoid possible problems induced
16234 J. Phys. Chem. B, Vol. 110, No. 33, 2006
Tervonen et al.
by gradient switching, such as eddy currents in the surrounding material, ramped gradient pulses were used. The parameters were Gz ) 0-0.485 T m-1, ∆ ) 50-330 ms, δ ) 0.8-7 ms, and d ) 0.09-0.8 ms. Measurements were started by first warming the sample to the isotropic state and then cooling it down using temperature intervals of 2-5 K.
[
]
χd ≈ 3M χd σo[1 - (R + β1)(T - T0)] - F0[1 - R(T - T0)] (3) 3M F0[1 - R(T - T0)] σ′0[1 - β1(T - T0)] -
and in the nematic phase
Results and Discussion A. 19F Shielding. In our earlier investigation on the behavior of noble gases in TLCs, it was shown that nuclear shielding depends on the density and the orientational order parameter of the TLC and the isotropic average and anisotropy of the shielding tensor are temperature dependent. For a description of the behavior of SF6 in TLCs, a similar approach is applied. Thus, the average shielding tensor element in the direction of the external magnetic field is represented in the form7
{
〈σzz〉 - σref ) F(T) σ′0[1 - β1(T - T0)] +
}
2 ∆σ′0[1 - β2(T - T0)]P2(cos θ)S(T) (1) 3 where σref is the shielding constant of a reference and σ′0 and ∆σ′0 are the shielding constant and shielding anisotropy divided by density, respectively, at the reference temperature T0. The coefficients β1 and β2 describe the assumed linear temperature dependence of the shielding constant and shielding anisotropy, respectively. P2(cos θ) ) (1/2)(3 cos2 θ - 1) is the second-order Legendre polynomial with θ being the angle between the LC director n and the external magnetic field B0. Unfortunately, the temperature dependence of the density of the present TLCs is not known. It is, however, a good approximation to assume a linear dependence:12 F(T) ) F0[1 R(T - T0)], where F0 is the density at the reference temperature and R is the isobaric thermal expansion coefficient. In the various phases (isotropic and nematic phases were analyzed in the present cases), R is not necessarily the same. Moreover, the TLC density may change discontinuously by a few tenths of a percent at the isotropic-nematic phase transition.12 In the present analysis, this factor has been omitted. The temperature dependence of the second rank orientational order parameter is described by the Haller function S(T) ) (1 - yT/TNI)z, where TNI is the isotropic-nematic phase transition temperature (chosen to be the reference temperature T0), and y and z are adjustable parameters.13 Because the 19F shielding was measured relative to an external low-pressure gas sample, a bulk susceptibility contribution, σb, has to be taken into account to obtain the pure medium effect on the shielding, that is, the gas-tosolution shift. For a long cylindrical sample with its axis parallel with the external magnetic field, the contribution may be calculated from7,14
σb ) -
〈σzz(T)〉 - σref ) σm + σb )
F(T) 1 2 χd + ∆χdS(T) 3 3 M
[
]
(2)
where both the isotropic average of the diamagnetic volume susceptibility tensor, χd, and the anisotropy of the tensor, ∆χd, are given in m3 mol-1 and M is the molar mass of the solvent. With a relatively good approximation, we can now write the equations applicable in the isotropic phase
〈σzz(T)〉 - σref ) σm + σa + σb )
{
F0[1 - R(T - T0)] σ′0[1 - β1(T - T0)] + 2 P (cos θ)∆σ′0[1 - β2(T - T0)]S(T) 3 2 1 2 χd + ∆χdS(T) 3M 3
]}
[
≈ σo[1 - (R + β1)(T - T0)] + 2 T z P2(cos θ)∆σ′0[1 - (R + β2)(T - T0)] 1 - y 3 T0 F0[1 - R(T - T0)] 2 T z χd + ∆χd 1 - y (4) 3M 3 T0
[
(
(
)
)]
Equations 3 and 4 are based on the first-order approximation, that is, terms of the form Rβ(T - T0)2 have been omitted. This is justified because both R and β are of the order of 10-4-10-3 and the temperature difference is not very large in any case. (In fact, both β values were set equal to zero. See below.) The shielding contributions to the total shielding are the following: σm is the gas-to-solution shift (equal to σ0 at T ) T0), σb is the bulk susceptibility effect on the shielding, and σa arises from the shielding tensor anisotropy in liquid-crystalline phases. In, principle, the least-squares fit applied to eqs 3 and 4 provides versatile information concerning the TLC properties (density, thermal expansion coefficients, diamagnetic susceptibility tensor, and order parameter) and about the 19F shielding tensor of SF6. The revelation of all this information is restricted by the fact that the experimental shielding is not very sensitive to all of these parameters. Therefore, some of them had to be fixed. On the basis of the electronic structure of SF6, one may conclude that the β factors are insignificant for fluorine, unlike for xenon, and consequently they were set equal to zero. The fitting procedure proceeded in two stages: first, the function (3) was fitted to the data in the isotropic phase, and second, the function (4) was fitted to the data collected in the nematic phase of the liquid crystals. The values of the most relevant adjusted and fixed parameters are listed in Table 2. The temperature and the phase dependence of the 19F shielding of SF6 are very similar to those earlier measured for the 129Xe shielding of xenon in various liquid crystals. The following changes are clearly visible in Figures 3-5: (i) There is an abrupt jump in shielding at the isotropic-nematic phase transition. This jump is toward lower shielding in liquid crystals with positive ∆χd (see, Figure 3) and toward higher shielding in liquid crystals with negative ∆χd (see, Figure 4). (ii) At the nematic-smectic A phase transition (see, Figure 5), no abrupt jump can be detected but the shielding starts to increase and the curvature changes. On the basis of 129Xe NMR experiments, this behavior was interpreted to arise from the redistribution of xenon atoms during the formation of the layer structure of the smectic A phase. The analysis of the data in the smectic A phase makes necessary the inclusion of two new order parameters as well.15
19F
SQ and
19F-33S
HMQC NMR of SF6
J. Phys. Chem. B, Vol. 110, No. 33, 2006 16235
TABLE 2: Parameters Derived from the Analyses of the 19F Shielding of SF6 in Various Thermotropic Liquid Crystals at Variable Temperatures parameter/LC
Phase 4
HAB
ZLI 2806
ZLI 3125
σo (ppm) ∆ σo (ppm) R (10-4 K-1)a R (10-4 K-1)b Fo (g cm-3) χd (10-9 m3 mol-1)c ∆χd (10-9 m3 mol-1)c yd z
-8.33 -1.42 17.8 14.9 0.94 -2.40 0.43 0.997 0.183
-9.44 -1.09 18.6 11.4 0.9 -2.85 0.50 0.997 0.177
-7.28 -0.48 15.6 16.0 0.98 -3.06 -0.13 0.997 0.180
-7.36 -0.85 18.4 21.7 0.94 -2.36 -0.10 0.997 0.209
a In the isotropic phase. b In the nematic phase. c Kept fixed. These values were derived from those given in ref 26 for benzene and cyclohexane based liquid crystals. d Kept fixed.
Figure 3. 19F shielding of SF6 as a function of temperature in the Phase 4 liquid crystal. The diamonds are experimental points whereas the solid lines are the results of least-squares fits to eqs 3 and 4 as described in the text: I ) isotropic phase, N ) nematic phase.
Figure 5. (a) 19F shielding of SF6 as a function of temperature in the HAB liquid crystal. The diamonds are experimental points whereas the solid lines are the results of least-squares fits to eqs 3 and 4 as described in the text: I ) isotropic phase, N ) nematic phase, and SmA ) smectic A phase. No attempt was made to model the data in the smectic A phase. (b) Stack plot of the 19F NMR spectra of SF6 in HAB at variable temperatures. The small peak on the right-hand side of the main peak arises from the 34SF6 isotopomers. On the horizontal axis, 19F chemical shift values are shown whereas in (a) the vertical axis shows the 19F shielding values.
Figure 4. 19F shielding of SF6 as a function of temperature in the ZLI 3125 liquid crystal. The diamonds are experimental points whereas the solid lines are the results of least-squares fits to eqs 3 and 4 as described in the text: I ) isotropic phase, N ) nematic phase.
There is one fundamental difference between the 19F and 129Xe shielding of SF and xenon, respectively, in liquid crystal 6 solutions. Namely, the bulk susceptibility contribution to the 19F shielding is as high as over 65% in Phase 4, ZLI 2806, and ZLI 3125 LC and ca. 30% in HAB, while for 129Xe this contribution is on the order of 1% in various LCs.7,17 Figures 6 and 7 display the effect of the bulk susceptibility on the 19F shielding in Phase 4, which orients along the external magnetic field, and in ZLI 2806, which orients perpendicular to the external magnetic field. The sudden change at the isotropicnematic phase transition arises from the anisotropy of the diamagnetic susceptibility tensor. The analyses of the experimental shielding data reveal that the 19F shielding tensor becomes slightly anisotropic in a liquidcrystalline environment; the anisotropy is from -0.5 to -1.4 ppm at the isotropic-nematic phase transition temperature as given in Table 1. The molecular structure of SF6 is not deformed by the anisotropic forces acting in a liquid-crystalline environment, as will be discussed below, but these forces are strong
Figure 6. Bulk susceptibility effect, σb, on the 19F shielding of SF6 in Phase 4 liquid crystal as a function of temperature.
Figure 7. Bulk susceptibility effect, σb, on the 19F shielding of SF6 in the ZLI 2806 liquid crystal as a function of temperature.
enough to affect the electron distribution leading to an anisotropic shielding tensor. B. SF6 Self-Diffusion. In the present self-diffusion experiments of SF6 with gradient pulses,9,10 the diffusion coefficient
16236 J. Phys. Chem. B, Vol. 110, No. 33, 2006
Tervonen et al. TABLE 3: Activation Energy, Ea, Derived from the Fit of the Arrhenius Equation to the Experimental SF6 Self-diffusion Coefficients Determined at Variable Temperatures LC HAB Phase 4 ZLI3125
Figure 8. A/A0 as a function of the square of the field gradient strength (Gz) in the PGSE experiment at 341 K (SF6 in the Phase 4). Circles are experimental points and the solid line fit to eq 5.
ZLI2806
Figure 10. and 3 atm.
Figure 9. Self-diffusion coefficient of SF6 in HAB as a function of reciprocal temperature. The circles are experimental points whereas the solid lines are fits to the Arrhenius equation.
in the direction of the external magnetic field was determined from eq 5 when using the PGSE pulse sequence
A(4τ,Gz) A0(4τ,0)
{
[( )
) exp -γ2Gz2D δ2 ∆ -
]}
δ d3 δd2 + 3 30 6
(5)
and from eq 6 when using the DSE pulse sequence
A(4τ,Gz) A0(4τ,0)
{
[( )
) exp -2γ2Gz2D δ2 ∆ -
]}
δ d3 δd2 + 3 30 6
(6)
where the notations of Figure 2 have been used. A(4τ,Gz) and A0(4τ,0) are the signal amplitudes with and without field gradient pulses, respectively, and γ is the gyromagnetic ratio of the 19F isotope. The diffusion was encoded in each temperature by making the strength of the gradient pulses variable. Figure 8 illustrates the behavior of the normalized echo amplitude as a function of the square of the amplitude of the field gradient pulse and the result of the fit to eq 5. The resulting values of the SF6 self-diffusion coefficient D in the parallel direction with the external magnetic field B0 are plotted on a logarithmic scale as a function of reciprocal temperature for SF6 in HAB in Figure 9. The diffusion activation energies, Ea, were obtained by fits to the Arrhenius equation, D ) D0 exp(Ea/RT), where R is the universal gas constant (8.3145 J mol-1 K-1) and D0 is the preexponential factor. The phase transition temperatures were derived from the 19F shielding data. The activation energies for each sample in the isotropic and in the liquid-crystalline phases are given in Table 3. The results reveal a clear difference in the activation energies of the isotropic and oriented phases; energy increases in the order isotropic < nematic < smectic A. The activation energies are very similar to those derived from the 129Xe self-diffusion experiments for xenon in a mixture of
19F-33S
phase
Ea/kJ mol-1
isotropic nematic smectic A isotropic nematic isotropic nematic isotropic nematic
21.8 37.3 102.9 27.8 29.9 23.8 34.4 18.7 29.2
1D HMQC spectrum of gaseous SF6 at 223 K
two thermotropic liquid crystals.16 This study also revealed a small anisotropy (defined as D⊥/D|) in the 129Xe diffusion tensor. The almost nonexistent change of the diffusion coefficient at the isotropic-nematic phase transition (see, Figure 9) is an indication of a negligible anisotropy in the SF6 diffusion tensor in the nematic phase. On the other hand, when approaching the smectic A phase, the anisotropy seems to increase similarly as earlier observed for 129Xe.2 C. 19F-33S Spin-Spin Coupling. The indirect spin-spin coupling between the 19F and 33S (spin 3/2) isotopes leads to the quartet structure in the 19F NMR spectrum of the 33SF6 isotopomer. The value of the coupling constant in liquid SF6 has been determined by Wasylishen et al., J(19F,33S) ) 251.8 Hz.17 In the present case, we measured the coupling constant for gaseous SF6 using a one-dimensional HMQC sequence at two temperatures. This technique had to be applied because lines in a conventional 19F NMR spectrum are so wide that the dominating peak from the most abundant 32SF6 isotopomer completely masks the 33S satellites. The spectrum shown in Figure 10 also displays fairly broad lines, but the splitting due to the J(19F,33S) coupling is clearly visible. The coupling constant value was derived by line shape fitting: at 297 K, J(19F,33S) ) 242 Hz, while at 223 K, it is 245 Hz. These values, measured for a sample in which the pressure is 3 atm, are practically equal because their error limits are fairly large due to broad lines in the spectra. Jackowski et al. report a value of 250.1 ( 0.4 Hz for a high pressure (ca. 20 atm) gas sample.8 The J(19F,33S) values in liquid crystal solutions at variable temperatures were derived from the 33S satellites of 1D 19F spectra (see, Figure 11). In one case, however, a 2D 19F-33S HMQC spectrum was recorded. It is displayed in Figure 12. No systematic temperature dependence of J(19F,33S) could be detected in any liquid crystal; the possible small change is masked by the relatively large uncertainty in the values which arises from the low signal-to-noise ratio of the 33S satellites in the 19F spectra. The average values of the coupling constants are the following: 253.3 Hz in ZLI 2806 and HAB, 253.9 Hz in ZLI 3125, and 255.5 Hz in Phase 4. These are slightly bigger than the corresponding value in liquid SF6 but very similar to those measured for SF6 in various isotropic liquids.8 In all, the coupling seems to be fairly insensitive to solvent properties.
19F
SQ and
19F-33S
HMQC NMR of SF6
J. Phys. Chem. B, Vol. 110, No. 33, 2006 16237 the splitting in the 33S satellites in the 19F spectra of SF6 is practically equal in the isotropic and liquid-crystalline phases, meaning that the RDC is diminishingly small. D. Secondary Isotope Shift of the 19F Shielding. The replacement of an atom by its heavier or lighter isotope in a molecule causes changes in the nuclear shielding of the neighboring atoms. This phenomenon is called a secondary isotope effect. In the present case, the effect of sulfur isotopes on the 19F shielding was measured in the liquid crystal solutions mentioned above as a function of temperature. The change of 19F shielding is defined as 1
Figure 11. 19F NMR spectrum of SF6 dissolved in HAB at 315 K. The most intensive signal stems from the 32SF6 isotopomers whereas the signal at 7.53 ppm arises from the 34SF6 isotopomers. The inset shows the quartet of the 33SF6 isotopomers.
∆19F(m′/mS) ) δ(19F,m′S) - δ(19F,mS) ) σ(19F,mS) - σ(19F,m′S) (7)
Sulfur has four stable isotopes: 32S (mass, 31.97207; natural abundance, 95.0%), 33S (32.97146; 0.76%), 34S (33.96782; 4.22%), and 36S (35.96709; 0.014%). The abundance of the last one is so low that the 19F spectrum of the 36SF6 isotopomer is not detectable within a reasonable amount of time. It has been shown that, to a good approximation, the isotope shift linearly depends on the relative mass factor24
∆19F(m′/mS) ) C
1
Figure 12. 19F-33S 2D HMQC spectrum of SF6 dissolved in Phase 4 liquid crystal.
TABLE 4: Values of the Coefficient C in eq 8 as Derived from the Mean Values of the Experimental 19F Isotope Shifts in Various Liquid Crystals
a
liquid crystal
C (ppm)a
ZLI 2806 ZLI 3125 HAB Phase 4
-0.88 -0.86 -0.86 -0.85
Estimated error (0.01 ppm.
The 1H and 13C NMR spectra of methane in liquid crystal solutions display residual dipolar couplings (RDCs) of a few hertz.18-20 Due to its high symmetry, the average orientation of methane in liquid crystals is zero, and consequently, the dipolar couplings should average to zero as well. The observed RDC is due to the fact that the various vibrational modes do exhibit orientation resulting in a correlation between vibrational and reorientational motions.21-23 Generally, the RDC between nuclei K and L is proportional to 〈sKL/R3KL〉, where sKL is a momentary orientational order parameter of the RKL vector and RKL is the distance between the nuclei. If no correlation is present, 〈sKL/R3KL〉 ) 〈sKL〉〈1/R3KL〉 ) SKL〈1/R3KL〉, where SKL ) 1/2〈3 cos2 θB,KL - 1〉 and θB,KL is the angle between the external magnetic field and RKL. For example, in the 13C NMR spectrum of the methane molecule, the quintet splitting (distance between two successive resonance peaks) |J(13C,1H) + 2D(13C,1H)| detectable in liquid crystal phases is clearly different from the J value measured in isotropic phases. On the contrary,
m′ - m m′
(8)
where C is a constant and is derived from a least-squares fit to experimental results, m ) 31.97207 is the mass of the lightest isotope, 32S, and m′ ) 31.97207, 32.97146, and 33.96782. The experimental results yield a slight temperature dependence in the isotope shift; the shift increases with decreasing temperature by ca. (2-6) × 10-5 ppm/K. The shifts are (2-5) × 10-2 ppm, and therefore, the temperature dependence was neglected and the coefficient C was determined using the mean values of the isotope shifts. The C values are shown in Table 4. The C values are practically independent of the liquidcrystalline environment, indicating that the anisotropic forces do not significantly affect the vibrations of the SF6 molecule. The factor C was also determined for liquid and gaseous SF6 using the data reported in refs 25 and 26. For the liquid, C ) -0.90 ppm whereas for gas C ) -0.87 ppm. These values are very similar to those derived in this work for SF6 in liquidcrystalline solutions. Consequently, one may conclude that the structure of the SF6 molecule, in particular the S-F bond length, is independent of whether SF6 appear in a gas, liquid, or solution state. Conclusions We have used 19F NMR of SF6 gas to study properties of four thermotropic liquid crystals, HAB, Phase 4, ZLI3125, and ZLI2806. The 19F shielding data were analyzed by using an earlier developed theoretical model. Although some assumptions in applying the model had to be made, it is obvious that the liquid-crystalline environment deforms the electron system of the molecule, leading to an anisotropic shielding tensor. On the contrary, no sign of residual dipolar coupling could be detected, indicating that the deformation effects, that is, the interactions between intramolecular vibrations and reorientational motion, are negligible. The self-diffusion experiments reveal the discontinuous behavior of SF6 at the isotropic-nematic phase transition in all of the four studied LCs (also at the nematicsmectic A phase transition in the HAB LC). However, it can be concluded that the anisotropy of the self-diffusion tensor is small or even negligible in nematic phase but becomes significant in the smectic A phase. Great advantages in using
16238 J. Phys. Chem. B, Vol. 110, No. 33, 2006 SF6 as a probe compared to xenon are the facts that the diffusion measurements can be performed much faster and within a larger dynamic range. On the contrary, xenon is superior in its probing properties, such as orientational order parameters in various mesophases and thermal expansion coefficients. Consequently, in most cases, it would be beneficial to use both xenon and SF6 as probes. Acknowledgment. The authors are grateful to the Academy of Finland (Grants 43979 and 203278) and to the Finnish Cultural Foundation (J.S.) for financial support. References and Notes (1) Jokisaari, J. Progr. NMR Spectrosc. 1994, 26, 1. (2) Cifelli, M.; Saunavaara, J.; Jokisaari, J.; Veracini, C. A. J. Phys. Chem. A 2004, 108, 3973. (3) Happer, W.; Miron, E.; Schaefer, D.; Wijngaarden, W. A. v.; Zeng, X. Phys. ReV. A: At., Mol., Opt. Phys. 1984, 29, 3092. (4) Kuethe, D. O.; Caprihan, A.; Gach, H. M.; Lowe, I. J.; Fukushima, E. J. Appl. Physiol. 2000, 88, 2279. (5) Beyea, S. D.; Codd, S. L.; Kuethe, D. O.; Fukushima, E. Magn. Reson. Imaging 2003, 21, 201. (6) Terekhov, M.; Neutzler, S.; Aluas, M.; Hoepfel, D.; Oellrich, L. R. Magn. Reson. Chem. 2005, 43, 926. (7) Ylihautala, M.; Lounila, J.; Jokisaari, J. J. Chem. Phys. 1999, 110, 6381.
Tervonen et al. (8) Jackowski, K.; Wilczek, M.; Makulski, W.; Kozminski, W. J. Phys. Chem. A 2002, 106, 2829. (9) Stejskal, E. O.; Tanner, J. E. J. Chem. Phys. 1965, 42, 288. (10) Ruohonen, J.; Ylihautala, M.; Jokisaari, J. Mol. Phys. 2001, 99, 711. (11) Lounila, J.; Oikarinen, K.; Ingman, P.; Jokisaari, J. J. Magn. Reson. A 1996, 118, 50. (12) Wedler, W. In Handbook of Liquid Crystals; Demus, D., Goodby, J., Gray, G. W., Spiess, H.-W., Vill, V., Eds.; Wiley-VCH: Weinheim, Germany, 1998; Vol. 1, p 334. (13) Haller, I. Prog. Solid State Chem. 1975, 10, 103. (14) Buckingham, A. D.; Burnell, E. D. J. Am. Chem. Soc. 1967, 89, 3341. (15) Lounila, J.; Muenster, O.; Jokisaari, J. J. Chem. Phys. 1992, 97, 8977. (16) Ruohonen, J.; Jokisaari, J. Phys. Chem. Chem. Phys. 2001, 3, 3208. (17) Wasylishen, R. E.; Connor, C.; Fiedrich, J. O. Can. J. Chem. 1984, 62, 981. (18) Jokisaari, J.; Hiltunen, Y. Mol. Phys. 1983, 50, 1013. (19) Jokisaari, J.; Hiltunen, Y.; Va¨a¨na¨nen, T. Mol. Phys. 1984, 51, 779. (20) Hiltunen, Y.; Jokisaari, J.; Pulkkinen, A.; Va¨a¨na¨nen, T. Chem. Phys. Lett. 1984, 109, 509. (21) Lounila, J.; Diehl, P. J. Magn. Res. 1984, 56, 254. (22) Lounila, J.; Diehl, P. Mol. Phys. 1984, 52, 827. (23) Lounila, J. Mol. Phys. 1986, 58, 897. (24) Jameson, C.; Osten, H.-J. J. Chem. Phys. 1984, 81, 4293. (25) Gillespie, R. J.; Quail, J. W. J. Chem. Phys. 1963, 39, 2555. (26) Makulski, W. Mol. Phys. Rep. 2001, 33, 82.
