Article pubs.acs.org/JPCA
Microwave Spectrum of Hexafluoroisopropanol and Torsional Behavior of Molecules with a CF3−C−CF3 Group Abhishek Shahi† and Elangannan Arunan*,† †
Department of Inorganic and Physical Chemistry, Indian Institute of Science, Bangalore 560012, India S Supporting Information *
ABSTRACT: This paper presents the first microwave spectroscopic investigation on hexafluoroisopropanol (HFIP). A pulsed nozzle Fourier transform microwave spectrometer has been used to determine the rotational constants for HFIP as A = 2105.12166(18) MHz, B = 1053.99503(12) MHz, and C = 932.33959(13) MHz. In addition, five isotopologues of HFIP have been observed experimentally to determine the accurate structure of HFIP. The observed spectrum could be assigned to the most stable conformer of HFIP, called antiperiplanar. Available spectroscopic information and ab initio calculations on five prototype molecules helped in exploring the torsional behavior of molecules having a CF3−C−CF3 group. Twodimensional potential energy surfaces have been analyzed for all molecules, which explained the presence/absence of doubling in the rotational transitions. With the help of natural bond orbital (NBO) analysis, reasons for the conformational preference of HFIP have been explained.
■
INTRODUCTION Fluorinated alcohols are important solvents and are extensively used by biologists, polymer chemists, and organic chemists. Hexafluoroisopropanol (HFIP; CF3−C(H)(OH)−CF3) is a fluorinated alcohol which possesses many interesting properties. It is a protic solvent which can work as a hydrogen bond donor as well as a hydrogen bond acceptor. Its aqueous binary solution helps in stabilizing the α-helical structure of proteins.1 It has a propensity to dissolve even polymers because of its ability to form strong hydrogen bonds. Therefore, the polymer should possess hydrogen bond acceptor groups. For example, polymers such as polythene terephthalate, polyamides, etc.2 could be dissolved in HFIP. These polymers are hard to dissolve in other solvents. HFIP is a suitable solvent for rearrangements via zwitterionic intermediates, whereas CH3OH is not suitable for such a reaction.3 The HFIP molecule is interesting for spectroscopists as well. The conformational preference of HFIP has been a subject of several studies. IR, Raman, and matrix isolation studies show that the molecule exists in two conformations: antiperiplanar (AP) and synclinal (SC). The AP conformer has been found to be more stable than the SC conformer.4−6 Both conformers exist in CCl4 solution as well as in an N2 matrix and a CO matrix. However, in polar media the IR transitions corresponding to the SC conformer were found to be more intense.5 In an Ar matrix, only the AP conformer could be observed.5 The relative intensity of the conformers also depends on the temperature. These conformers exist because of the internal torsion of the −OH group. The prototype molecule isopropanol also exists in two conformers due to the same © 2015 American Chemical Society
internal torsion. However, for this molecule the SC conformer is more stable than the AP conformer.6−8 Suhm’s group has extensively 6 studied the effect of fluorination on the isopropanol molecule using IR spectroscopy. Our interest in HFIP was triggered by the fact that it is an exceptional solvent with the potential to form different types of hydrogen bonds with other hydrogen bond acceptor/donor molecules. Therefore, we wanted to study hydrogen-bonded complexes of HFIP with molecules such as H2O. However, we found that the rotational spectrum of the HFIP monomer itself was unknown. Somewhat coincidently, rotational spectra of the other prototype molecules hexafluoroisobutene (HFIBE; CF3− C(CH2)−CF3)9 and hexafluoroacetone imine10 (HFAIM; CF3−C(NH)−CF3) were investigated recently. For both of these molecules, the rotational transitions appear as doublets. The reason for this splitting was found to be the counter rotation of the two CF3 groups. Would such a splitting be observed in HFIP? In this paper, we report the first rotational spectroscopic study of the HFIP molecule. The results are compared with those of prototype molecules having a CF3−C−CF3 group. Unlike the case for HFIBE and HFAIM, the rotational transition for HFIP did not show any evidence of splitting. Other related molecules such as hexafluoroacetone (HFA), hexafluoroisobutane (HFIBA; CF3−C(H)(CH3)−CF3), and hexafluoroisopropylamine (HFIPA; CF3−C(H)(NH2)−CF3) Received: April 3, 2015 Revised: May 18, 2015 Published: May 19, 2015 5650
DOI: 10.1021/acs.jpca.5b03240 J. Phys. Chem. A 2015, 119, 5650−5657
Article
The Journal of Physical Chemistry A
O1−H2 dihedral angle at the MP2/6-311+G(d) level. The AP conformer is more stable than SC by 5.0 kJ/mol (Figure 2).
also do not show any splitting. A detailed two-dimensional potential energy surface (2D-PES) analysis on all of these molecules revealed the reason for the presence/absence of splitting. Natural bond orbital (NBO) analysis has been carried out to understand the relative stability of the conformers.
■
EXPERIMENTAL METHODS HFIP (≥99%) was purchased from Aldrich. Its monodeuterated isotope (OD) was prepared by mixing HFIP with D2O (99.9% by Cambridge Isotope Laboratory) in a 1/1 molar ratio. Its doubly deuterated (CD-OD) isotopologue was purchased from Aldrich (99%). None of the commercial samples were further purified. All C-13 isotopologues were observed in natural abundance. The rotational spectrum was recorded using a pulsed nozzle Fourier transform microwave (PNFTMW) spectrometer, the details of which have already been published.11 Initial experiments were carried out using Ar as the carrier gas. However, we observed that the HFIP signals were more intense while using helium as the carrier gas. Therefore, for further experiments, helium was used as the carrier gas. The flow rate of He was kept at 200 SCCM, and 1% of it was flown through a bubbler containing an HFIP sample. The carrier gas seeded with the HFIP was expanded from a backing pressure of 1.5 bar through a 0.8 mm diameter pulse valve into the Fabry−Perot cavity. Multiple free induction decays (FIDs) were recorded per gas pulse. At a sampling rate of 5 MHz, 256 points were collected for each FID during the search for rotational transitions. Once a signal was observed, it was further averaged with 512 or 1024 points to improve the resolution. A microwave pulse of 1.3 μs duration was found to be the optimum for both the b and c type transitions.
Figure 2. Relaxed potential energy surface scan of the OH torsion (dihedral angle H1−C1−O1−H2) for the HFIP molecule at the MP2/6-311+G(d)level. AP, SC, and SP denote antiperiplanar, synclinal, and synperiplanar conformers of HFIP, respectively.
■
Two minima corresponding to the SC conformer are present in the PES. These minima are separated by a small barrier of 1.3 kJ/mol (∼108.4 cm−1), which is slightly less than the ZPE (109.5 cm−1) of the corresponding vibrational motion. The saddle point for this barrier corresponds to the synperiplanar structure (the H1−C1−O1−H2 dihedral angle is 0°). In addition to the torsion about the C−O bond, HFIP has a torsional vibration involving the two CF3 groups. This motion is also present in HFIBE, HFAIM, HFA, HFIPA, and HFIBA. As mentioned earlier, HFIBE and HFAIM show splitting of rotational transitions due to this motion. All of these molecules were optimized at the MP2/6-311++G(d,p) level of theory for comparison. The 2D-PES was constructed for all the molecules, allowing all the other coordinates to optimize. For the HFIP molecule, the dihedral angles selected for the 2D scan are H1−C1−C2− F1 and H1−C1−C3−F6 (see Figure 1). Each angle was scanned for 120° with a step size of 3°, and a total of 1681 (=412) grid points was calculated (Figure 4). Similarly, two dihedral angles for the rest of the molecules were selected. Where a H atom on the center C atom was absent, O, N, and C atoms were selected in place of the H atom for HFA, HFAIM, and HFIBE molecules, respectively. The advantages of performing these expensive and laborious scans were to compare the barrier heights, both small and large, and also to determine the most stable conformer (vide infra). With computation cost in mind, a reasonably good level of theory, B3LYP/6-311G*, was selected to scan the coordinates for 1681 grid points. Motion along the diagonal of the 2D-scan graph corresponds to counter rotation of the two CF3 groups. Orbital properties and their overlap were calculated using the NBO 6.013 program for all molecules at the MP2 level of theory.
THEORETICAL CALCULATIONS Ab initio calculations were performed using the G09 suite of programs.12 Different possible conformers of HFIP were optimized at the MP2/6-311++G(d,p) level. Frequency calculations were performed for the optimized geometries at the same level. These calculations and earlier IR studies6 show that the HFIP molecule has two minima, AP and SC. These two conformers originate due to the internal rotation of −OH group, and the H1−C1−O1−H2 dihedral angles are 180 and 52° for AP and SC, respectively (Figure 1). Vibrational
Figure 1. Antiperiplanar (AP) and synclinal (SC) conformers of HFIP. Geometries were optimized at the MP2/6-311++G(d,p) level of theory.
frequency calculations confirm that both conformers are true minima. Both of the terminal CF3 groups were eclipsed with respect to each other for the AP conformer and slightly staggered for the SC conformer. The dihedral angle F1−C2− C3−F6 for the SC conformer is 18.7°. A relaxed potential energy surface (PES) scan was performed for the H1−C1− 5651
DOI: 10.1021/acs.jpca.5b03240 J. Phys. Chem. A 2015, 119, 5650−5657
Article
The Journal of Physical Chemistry A
■
of the fit was 4 kHz. At this stage, it was not clear whether the observed progression is for the AP or the SC conformer. We were expecting the presence of other progressions corresponding to another conformer for two reasons. First, IR, Raman, and matrix isolation studies showed the presence of two conformers for HFIP, and second, a microwave study on the prototype molecule isopropanol8,18 showed the presence of two isomers because of the torsion of the −OH group. The optimized geometry for the AP conformer has no a dipole component. However, the SC conformer has finite dipole moment components along all three (a, b, c) principal axes. Therefore, for the SC conformer all the three types of transitions were expected. However, we could not observe any a type transition even after performing long searches. These searches were performed on the basis of ab initio rotational constants for both the AP and SC conformers. Absence of the a type transitions suggested that the experimentally observed spectra correspond to the AP conformer. If SC conformers exist in a supersonic molecular beam, a type transitions might not be seen because of the very small barrier (108.4 cm−1) for interconversion. This could lead to a very large tunneling splitting, making the observation and assignment difficult. Frequency calculations do predict one such motion whose zero point energy is 109.5 cm−1, comparable to the barrier height. The saddle point for this motion was the synperiplanar structure (vide supra). This could lead to hindered rotation and levels with internal rotor angular momentum. No transitions could yet be assigned for such levels. However, more evidence favoring the presence of the AP conformer are given in the next paragraphs. Moreover, unlike the rotational spectra of the prototype HFIBE and HFAIM molecules, no doubling in rotational transitions of the HFIP molecule was observed. The reason behind this doubling is different from that of isopropanol and is discussed later in detail. The signal intensity also helped us in identifying the conformer. For the observed progression, b type signals were always stronger than c type. This experimental observation is consistent with the ab initio calculated dipole moment of the AP conformer geometry, for which the b dipole component (0.6 D) is 3 times larger than the c dipole component (0.2 D). In contrast, for the SC conformer, the c dipole component (2.1 D) is larger than both the b dipole (1.8 D) and a dipole (1.0 D) components (Table 1). In addition for the synperiplanar conformer, which is a saddle point, the c dipole component (2.9 D) is larger than the b dipole (1.9 D). Hence, both
RESULTS AND DISCUSSION Calculated rotational constants for the AP and the SC conformers are very close to each other at the MP2/6-311+ +G(d,p) level (Table 1). On the basis of predictions, using Table 1. Experimental and Theoretical Rotational and Distortion Constants for Hexafluoroisopropanol exptl A/MHz B/MHz C/MHz DJ/kHz DJK/kHz DK/kHz d1/kHz d2/kHz RMS/MHz no. of transitions dipole moment (a, b, c)/D
2105.12166(18) 1053.99503(12) 932.33959(13) 0.05713(79) 0.50829(76) −0.4592(41) −0.00731(13) 0.002357(59) 0.0048 111 a=0 b>c≠0
a
calcd (AP)
calcd (SC)
2098.76488 1053.29499 931.98682 0.0522 0.2635 −0.2125 −0.0068 0.0015
2102.51954 1054.01680 932.48540 0.0560 0.2317 −0.1833 −0.0075 0.0017
0.0, 0.6, 0.2
1.2, 1.5, 2.2
a
Numbers in parentheses are signal standard errors in units of least significant figures.
these rotational constants, a search was started for a strong b type 414 ← 303 transition for the AP conformer, which was predicted at ∼8465 MHz. The search was started at 8465 MHz with Ar as the carrier gas. Within a 10 MHz search, a transition was observed at 8473.8905 MHz. Five more signals were searched similarly on the basis of the predictions and were observed readily. These rotational transitions were found at 9125.4651, 7059.1440, 10213.3467, 9526.6928, and 11161.5053 MHz and could be assigned to 505 ← 414, 404 ← 313, 515 ← 404, 321 ← 212, and 606 ← 515, respectively. All of these signals were significantly stronger with He carrier gas and were fitted to a semirigid rotor Hamiltonian. The rotational constants obtained from this tentative fit were used for further predictions. Eventually, a total of 111 transitions could be observed. These transitions included both b and c type pure rotational transitions of R and Q branches (see Table S1 of the Supporting Information). The observed transitions were fitted to Watson’s S reduction Hamiltonian.14 Programs used for the fitting were mainly SPFIT15,16 and also ASFIT.17 As expected, they produced the same results. The experimental rotational constants and distortion constants are given in Table 1. The RMS deviation
Table 2. Fitted Rotational and Distortion Constants, RMS Values, and Numbers of Transitions for Four Different Isotopologues of HFIPa A/MHz B/MHz C/MHz DJ/kHz DJK/kHz DK/kHz d1/kHz d2/kHz RMS #N a
CF3−C(H)(OH)−CF3
CF3−C(H)(OD)−CF3
CF3−C(D)(OD)−CF3
CF3−13C(H)(OH)−CF3
2105.12166(18) 1053.99503(12) 932.33959(13) 0.05713(79) 0.50829(76) −0.4592(41) −0.00731(13) 0.002357(59) 0.0048 111
2066.6444(10) 1052.74254(91) 925.60404(54) 0.0552(55) 0.442(55) −0.483(71) −0.0089(37) 0.0059(35) 0.0056 26
2042.0355(23) 1046.7717(10) 925.2142(10) 0.0486(97) 0.508(74) −0.54(11) −0.00731 0.002357 0.013 28
2099.9217(16) 1053.43442(63) 931.88122(63) 0.0492(57) 0.481(53) −0.460(60) −0.00731 0.002357 0.0057 17
13
CF3−C(H)(OH)−CF3 2105.0638(18) 1050.43361(72) 929.53229(60) 0.0528(59) 0.525(62) −0.411(74) −0.00731 0.002357 0.0076 22
Numbers without parentheses are not included in the fit. Values of the parent isotopologue are used for that particular fitting. 5652
DOI: 10.1021/acs.jpca.5b03240 J. Phys. Chem. A 2015, 119, 5650−5657
Article
The Journal of Physical Chemistry A synperiplanar and SC conformers are ruled out and the observed signals can be unambiguously attributed to AP. Since experimentally observed rotational constants were close to the calculated values, predictions of spectra for the other isotopologues were easy. We took the difference between the experimental and calculated rotational constants of the parent molecule and assumed similar differences for the rotational constants of the other isotopologues. Rotational spectra for four different isotopologues, CF3−C(H)(OD)−CF3 (HFIP(OD)), CF 3 −C(D)(OD)−CF 3 (HFIP(CDOD)), CF3−13C(H)(OH)−CF3 (HFIP-13(center)), and 13CF3−C(H)(OH)−CF3 (HFIP-13(side)), could be observed experimentally. We observed 26 transitions for HFIP(OD), 28 transitions for HFIP(CDOD), 17 transitions for HFIP13(center), and 22 transitions for HFIP-13(side) isotopologues. These are given in Tables S2−S5 of the Supporting Information, respectively. These transitions could be fitted within the experimental uncertainty, and the RMS values are 5.6, 13.0, 5.7, and 7.6 kHz for HFIP(OD), HFIP(CDOD), HFIP-13(center), and HFIP-13(side), respectively. The rotational constants obtained from these fits are given in Table 2. We could observe only one set of transitions for the side C-13 carbon. The signals for the HFIP-13(side) isotopologue were more intense (almost twice) than the signals for the HFIP13(center) isotopologue. This indicated that the side C atoms are equivalent, which is in accord with the ab initio results on the AP conformer. In the calculated structure, the molecule exhibited a plane of symmetry containing H2−O1−C1−H1 atoms. However, the side C atoms are different for the SC conformer in its equilibrium structure and there is no plane of symmetry. This is consistent with the observed spectrum corresponding to the AP conformer. The HFIP(OD) and HFIP(CDOD) isotopologues were comparatively easier to observe experimentally than the C-13 isotopologues. The hyperfine splitting from the quadrupolar D atom could be observed but was not well resolved. The line centers were used in the fit. With the help of a Kraitchman analysis,14 the position of the isotopic substitution can be determined. This analysis assumes that the bond length of the molecule remains unchanged on isotopic substitution. It is important to mention that this assumption is quite reasonable for heavy-atom substitution but not for the H/D substitution in stable molecules. Kisiel’s programs17 were used for this analysis. Using rotational constants of the five isotopologues, Cartesian coordinates of the substituted atoms have been determined. Along with directly bonded parameters, some nonbonded parameters are also given in Table 3. These parameters again support the AP conformer: e.g., the distance between the center of mass and H2 is equal to the calculated distance for the AP conformer and significantly different from that of the SC conformer (Table 3). Moreover, the experimental H1−C1−H2 angle and the C3− C1−H2−C2 dihedral angle are also close to the calculated values of the AP conformer. To examine the behavior of the CF3−C−CF3 group, comparisons of HFIP with other prototype molecules, HFIBE, HFAIM, HFA, HFIBA, and HFIPA, have been carried out. Equilibrium geometries of these molecules are shown in Figure 3. Rotational constants were taken from earlier works,9,10,19 and we fitted them to get the internal coordinates (Table 4). An isotopic study for the HFA monomer20 is not known; therefore, we used the calculated parameters for this molecule for further analysis. The 2D-PES scans showed that
Table 3. Results from the Kraitchman Analysis for HFIP Molecules, using the Rotational Constants of the Different Isotopologues param
exptl
calcd (AP)
calcd (SC)
C1−C2/Å C1−C3/Å C1−H1/Å C1−H2/Å CM−H2/Å C2−C1−C3/deg H1−C1−C2/deg H1−C1−C3/deg H1−C1−H2/deg C2−C1−H1−C3/deg C3−C1−H2−C2/deg
1.530(7) 1.530(7) 1.121(3) 1.935(3) 2.119(1) 113.2(6) 103.9(7) 113.2(7) 137.4(3) 123.2(8) 113.5(5)
1.531 1.531 1.092 1.930 2.109 114.1 107.2 107.2 135.9 122.8 115.1
1.531 1.527 1.097 1.930 2.511 113.9 106.5 106.5 94.5 121.9 134.6
two types of barriers exist for HFIBE, HFA, and HFAIM (denoted large and small). On the other hand, for the HFIP, HFIBA, and HFIPA molecules only one (large) barrier was present. The large barriers, in all of these molecules, are due to the CF3 counter rotation, which interconverts the staggered and the eclipsed forms of these molecules (Figure 4). The smaller barrier is due to the restricted counter rotation of the two CF3 groups just around its equilibrium geometry. Saddle points for the small barrier are marked in Figure 4 by small green dots and also pointed by an arrow. Barrier heights are extracted by following the diagonal path of the 2D scan plot, and a representative one-dimensional PES for all molecules is shown in Figure 5. Compiled results of barrier heights are presented in Table 5 for all of the prototype molecules. The barrier heights corresponding to the staggered−eclipsed interchange are very large, and they are unlikely to be responsible for the doubling observed in the rotational spectra of HFIBE and HFAIM molecules. As discussed earlier, for some molecules the smaller barrier is not present. Position 1 in the PES represents saddle points for the large barrier (Figure 5). At position 1, opposite CF3 groups are eclipsed for all molecules and are shown in Figure S1(a1−f1) in the Supporting Information. In the most stable equilibrium geometries for CF3−C(sp3)−CF3 types of molecules, the terminal CF3 groups are eclipsed with respect to each other and are staggered with respect to the center carbon substituents (Figure 3d−f or Figure S1(d3−f3)). The HFIPA molecule appeared as an exception, and the most stable structure for this molecule has slightly staggered CF3 groups. On the other hand, the most stable equilibrium structures for the CF3−C(sp2)−CF3 group of molecules are slightly staggered (Figure 3a−c or Figure S1(a2/a4,b2/b4,c2/c4)) and are denoted by position 2 or 4 in Figure 5. Position 3 in Figure 5 corresponds to the saddle points for the small barriers. The geometries corresponding to this point are given for all of the molecules in Figure S1(a3−c3). In these geometries both CF3 groups are eclipsed with each other as well as with the double bond. We have classified these molecules into two types, “type I” with small barriers (e.g., HFIBE, HFA, and HFAIM exhibited CF3−C(sp2)−CF3 group) and “type II” without small barriers (e.g., HFIP, HFIBA, and HFIPA exhibited CF3−C(sp3)−CF3 group). At this point, we observed two important structural differences between type I and type II molecules. The first difference is that the two CF3 groups are slightly staggered in type I molecules, while these are eclipsed in type II molecules in their equilibrium geometries. The second difference, as noted 5653
DOI: 10.1021/acs.jpca.5b03240 J. Phys. Chem. A 2015, 119, 5650−5657
Article
The Journal of Physical Chemistry A
Figure 3. Equilibrium geometries of the prototype molecules (a) hexafluoroisobutene (HFIBE), (b) hexafluoroacetone imine (HFAIM), (c) hexafluoroacetone (HFA), (d) hexafluoroisobutane (HFIBA), (e) hexafluoroisopropylamine (HFIPA), and (f) hexafluoroisopropanol (HFIP).
(Figure 3 and Table 4). Interestingly, HFA shows a longer than expected C−C bond (1.55 Å), even though the center C atom is sp2 hybridized and bond angles are close to 120°. Furthermore, HFA did not show splitting in the rotational spectrum, even with a small barrier being predicted, resulting in an equilibrium geometry that is slightly staggered. Overall, HFA turned out to be different from HFIBE and HFAIM. Inspection of the small barrier revealed the reason behind the presence/ absence of doubling in the rotational transitions. For HFA, the height for the “small barrier” was larger (1.6 kJ/mol) in comparison to that for HFIBE (0.2 kJ/mol) and HFAIM (0.8 kJ/mol). In addition, the order of the experimental splitting for HFIBE is 10 MHz9 and that for HFAIM is 10 kHz.10 The higher barrier for the HFA molecule qualitatively suggests that the splitting should be on the order of 100 Hz or less. Resolving splittings of less than 5 kHz is not possible with the resolution of typical microwave spectrometers. In other words, because of the higher barrier, the splitting is too small to be observed for the HFA molecule. Moreover, there was a vibrational motion corresponding to counter rotation of the CF3 groups with frequencies of 31, 37, and 24 cm−1 for HFIBE, HFA, and HFAIM, respectively, at the B3LYP/6-311G(d) level. The ZPE corresponding to these motions are 0.2, 0.2, and 0.1 kJ/mol, respectively. Thus, the small barriers (V2) are always larger than the ZPE corresponding to this vibrational mode for all of the molecules (Table 5). To the best of our knowledge, microwave spectra of the HFIPA and HFIBA have not been reported to date. Our prediction says that both molecules should not show doubling, due to the free counter rotation of the two CF3 groups. NBO Analysis of Conformers. Both CF3 groups are eclipsed for HFIP and slightly staggered for HFIBE in their equilibrium structure. In general, CF3 groups are eclipsed or staggered, depending on the hybridization of the center C atom (vide supra). HFIP and HFIBE center C atoms are sp3 and sp2 hybridized, respectively. As discussed earlier, corresponding to an eclipsed structure in HFIBE, there is a small barrier and one imaginary frequency was found. We thought of identifying some particular orbital overlaps which change on this
Table 4. Geometrical Parameters Obtained from the Kraitchman Analysis for HFIP, HFIBE, HFAIM, and Hexafluoropropanea param
value (V = 0)
value (V = 1)
Hexafluoroisobutene (HFIBE) C1−C2/Å 1.465(2) 1.462(3) C1−C3/Å 1.465(2) 1.462(3) C1−C4/Å 1.339(4) 1.340(4) C2−C1−C3/deg 121.0(2) 121.0(3) C2−C1−C4/deg 116.1(12) 119.5(2) C3−C1−C4/deg 122.9(12) 119.5(2) C2−C1−C4−C3/deg 180(0) 177.3(79) Hexafluoroacetone Imine (HFAIM) C1−C3/Å 1.486(02) 1.487(2) C1−C2/Å 1.490(29) 1.478(2) C1−N1/Å 1.268(03) 1.270(3) C2−C1−C3/deg 120.5(19) 121.4(2) N1−C1−C2/deg 117.5(22) 115.9(9) N1−C1−C3/deg 121.9(11) 122.3(9) C2−C1−N1−C3/deg 177.5(116) 172.8(40) Hexafluoroacetone (HFA) C1−C2/Å C1−C3/Å C1−O1/Å O1−C1−C2/deg O1−C1−C3/deg C2−C1−C3/deg C1−C2−O1−C3/deg a
calcd 1.507 1.507 1.337 116.7 121.6 121.6 180.0 1.528 1.525 1.271 116.4 118.6 125 −179.0 1.547 1.547 1.203 121.8 121.8 116.4 180.0
For HFA ab initio calculated parameters are reported.
above, is the hybridization of the middle carbon atom. Type I has the center carbon in sp2 hybridization, and type II has the center carbon in sp3 hybridization. For the HFIP molecule, C− C bond lengths were 1.53 Å and angles at the center carbon were close to tetrahedral angles, which are the typical features of sp3 C atoms. However, in HFIBE and HFAIM, C−C bond lengths were close to 1.47 Å and angles at the center carbon were close to 120°, which are typical of sp2 carbon atoms 5654
DOI: 10.1021/acs.jpca.5b03240 J. Phys. Chem. A 2015, 119, 5650−5657
Article
The Journal of Physical Chemistry A
Figure 4. 2D-PES scan for the CF3 counter rotation in (a) hexafluoroisopropanol, (b) hexafluoroisopropylamine, (c) hexafluoroisobutane, (d) hexafluoroisobutene, (e) hexafluoroacetone imine, and (f) hexafluoroacetone molecules. In parts d−f, green dots (denoted by arrows) represent the small barrier and the corresponding structures are saddle points. Detailed information about the axis values is given in the text.
conformational exchange in HFIBE (slightly staggered ↔ eclipsed). For this purpose, NBO calculations were performed for slightly staggered (equilibrium geometry) and eclipsed conformers (saddle point geometry). The overlapping between C1−C10 π-bonding orbitals and C2−F4 and C3−F9 antibonding orbitals was larger in the slightly staggered conformer than in the eclipsed conformer, determined by the second-order perturbation energy (E2). However, there are many orbital overlaps which led to stabilization/destabilization on eclipsed−slightly staggered exchange and it was difficult to identify any specific orbital overlaps. Finally, we decided to sum up all E2 values arising from all such overlaps. The total sums for the slightly staggered and the eclipsed conformers were 2713.9 and 2713.2 kJ/mol, respectively. Qualitatively, these values supported that the slightly staggered conformer is more stable than the eclipsed conformer. To the best of our knowledge, this method has been used for the first time to
Figure 5. Barrier for counter rotation of CF3 groups. The energy values for different positions are given in Table 5. Their structures are presented in Figure S1 in the Supporting Information.
Table 5. Barrier Height, Angle, and ZPE for All Prototype Moleculesa property
position
HFIBE
HFAIM
HFA
HFIP
HFIBA
HFIPA
energy for large barrier (V1) (kJ/mol) energy for small barrier(V2) (kJ/mol)
1 2 3 4 1 2 3
25 0 0.2 0 −30 0 30 31 0.2 10 MHz
21 0 0.8 0 −42 0 42 24 0.1 10 kHz
15 0 1.6 0 −52 0 52 37 0.2 not obsd
43
45
47
0
0
0
0b
0b
25b
22 0.1 n/a
26 0.2 n/a
40 0.2 n/a
dihedral angle F−C−C−X/deg
vibrational freq (cm−1) ZPE (kJ/mol) order of exptl splitting a
The numbers 1−4 are the positions in the PES scan of Figure 5. All calculations were performed at the B3LYP/6-311G* level of theory. bF−C−C− F dihedral angle. 5655
DOI: 10.1021/acs.jpca.5b03240 J. Phys. Chem. A 2015, 119, 5650−5657
Article
The Journal of Physical Chemistry A
Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.jpca.5b03240.
confirm the conformational preference. The electronic energy difference between the two conformers was 1.3 kJ/mol, and the total E2 difference was 0.7 kJ/mol. Clearly, half of the net stabilization arises due to favorable orbital overlap. To verify the above approach, calculations were carried out for AP and SC conformers. The sums of E2 for AP and SC conformers are 2557.4 and 2539.0 kJ/mol, respectively, which predicts that AP is more stable than SC. The electronic energy difference between the two conformers is only 5.0 kJ/mol, in comparison to the difference of 18.4 kJ mol−1 in E2. Clearly, in this case, the net stabilization energy is less than that predicted by comparing the E2 values. In the AP conformer, we could identify three overlaps which are not found in SC: first, the lone pair on the O atom overlapped with the C1−C2 and C1−C3 antibonding orbitals equally, second, another lone pair of the O atom overlapped well with the C−H antibonding orbital, and third, the O−H bonding MO overlapped with the C−H antibonding MO. Overall, orientations of oxygen lone pairs are the key points for conformational preferences. In these two examples, we were not expecting the exact match between the net stabilization energy from electronic structure calculations and the difference between the sum of E2 values, as dispersion and exchange repulsion also contribute. It was interesting to note that the relative stability of the various conformers could be predicted by looking at the sum of all E2 values.
■
Corresponding Author
*E-mail for E.A.: [email protected]. Notes
The authors declare no competing financial interest.
■
ACKNOWLEDGMENTS The authors thank the Indo-French Centre for Promotion of Advanced Scientific Research for partial support. The authors also thank the Super Computer Research Center and Inorganic and Physical Chemistry department at the Indian Institute of Science for providing good computational facilities. A.S. thanks the Council for Research and Industrial Research (CSIR) of India for a fellowship and Dr. Devendra Mani for his help during experiments.
■
ABBREVIATIONS pulsed nozzle Fourier transform microwave spectrometer,PNFTMW; hexafluoroisopropanol,HFIP; hexafluoroisobutane,HFIBA; hexafluoroisopropylamine,HFIPA; hexafluoroisobutene,HFIBE; hexafluoroacetoneimine,HFAIM; hexafluoracetone,HFA
■
■
CONCLUSION HFIP and four of its isotopologues have been observed experimentally using a PNFTMW spectrometer. Observed transitions could be fitted well to Watson’s S reduction Hamiltonian within the experimental uncertainties. Experimental rotational constants for the parent themselves could not distinguish between AP and SC conformers. The absence of a type transitions and the fact that b type transitions are stronger than c type transitions provide evidence in support of the AP conformer as the experimentally observed structure. Moreover, isotopic substitution results also suggest the AP structure. We performed 2D-PES scans for several prototype molecules. Our results can be summarized as follows. (a) If the center C atom is sp2, a small barrier will be present, leading to doubling of energy levels: e.g. HFIBE, HFAIM, and HFA. Depending on the barrier height and actual splitting in energy levels, doubling of rotational transitions may be obtained experimentally. (b) If the center C atom is sp3, there would be no small barrier and hence no splitting of rotational transition due to a torsional motion: e.g. HFIP, HFIPA, and HFIBA. With the help of NBO calculations, reasons for a particular conformational preference can be rationalized. Total secondorder perturbation energy for a molecule can give an idea of the conformational preference, a new approach proposed here. We hope that this approach can be generalized and used extensively.
■
AUTHOR INFORMATION
REFERENCES
(1) Hong, D. P.; Hoshino, M.; Kuboi, R.; Goto, Y. Clustering of Fluorine-Substituted Alcohols as a Factor Responsible for Their Marked Effects on Proteins and Peptides. J. Am. Chem. Soc. 1999, 121, 8427−8433. (2) Cazes, J. Encyclopedia of Chromatography; Taylor & Francis: Abigdon, U.K., 2005; Vol. 2. (3) Burdisso, M.; Gandolfi, R.; Toma, L.; Obertib, R.; Oberti, R. Hexafluoroisopropanol as a Suitable Solvent for Rearrangements via Zwitterionic Intermediates. Tetrahedron 1991, 47, 6725−6736. (4) Cyvin, S. J.; Brunvoll, J.; Perttilä, M. Substituted Profanes IX. Vibrational Frequencies and Calculated Mean Amplitude for Hexalfuoroisopropanol-2. J. Mol. Struct. 1973, 17, 17−21. (5) Barnes, A. J.; Murto, J. Infra-red Cryogenic Studies. Part 10. Conformational Isomerism of 1,1,1,3,3,3-Hexafluoropropan-2-ol. J. Chem. Soc., Faraday Trans. 2 1972, 68, 1642−1651. (6) Schaal, H.; Häber, T.; Suhm, M. A. Hydrogen Bonding in 2Propanol. The Effect of Fluorination. J. Phys. Chem. A 2000, 104, 265− 274. (7) Hirota, E.; Kawashima, Y. Internal Rotation of the Hydroxyl Group in Isopropanol and the Chirality of the Gauche Form: Fourier Transform Microwave Spectroscopy of (CH(3))(2)CHOD. J. Mol. Spectrosc. 2001, 207, 243−253. (8) Kondo, S.; Hirota, E. Microwave Spectrum and Internal Rotation of Isopropyl Alcohol. J. Mol. Spectrosc. 1970, 34, 97−107. (9) Grubbs, G. S.; Novick, S. E.; Pringle, W. C.; Laane, J.; Ocola, E. J.; Cooke, S. A.; Ii, G. S. G.; Esther, J.; Road, A. H. Bis-trifluoromethyl Effect: Doubled Transitions in the Rotational Spectra of Hexafluoroisobutene, (CF 3) 2 CCH 2. J. Phys. Chem. A 2012, 116, 8169− 8175. (10) Obenchain, D. A.; Frohman, D. J.; Grubbs II, G. S.; Long, B. E.; Pringle, W. C.; Novick, S. E. A Pure Rotational Study of Two NearlyEquivalent Structures of Hexafluoroacetone Imine. 67th International Symposium on Molecular Spectroscopy; Ohio State University, Columbus, OH, USA, 2012; Talk TC04. (11) Arunan, E.; Tiwari, A. P.; Mandal, P. K.; Mathias, P. C. Pulsed Nozzle Fourier Transform Microwave Spectrometer: Ideal to Define Hydrogen Bond Radius. Curr. Sci. 2002, 82, 533−540. (12) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb, M. A.; Cheeseman, J. R.; Scalmani, G.; Barone, V.; Mennucci,
ASSOCIATED CONTENT
* Supporting Information S
Text giving the complete author list of ref 12, Figure S1 showing structures at different maxima and minima of the 2DPES given in Figure 4, and Tables S1−S5 presenting all measured transition frequencies and quantum number assignments for HFIP and all isotopologues. The Supporting 5656
DOI: 10.1021/acs.jpca.5b03240 J. Phys. Chem. A 2015, 119, 5650−5657
Article
The Journal of Physical Chemistry A B.; Petersson, G. A.; et al. Gaussian 09, revision D.01; Gaussian, Inc., Wallingford, CT, 2009. (13) Glendening, E. D.; Badenhoop, J. K.; Reed, A. E.; Carpenter, J. E.; Bohmann, J. A.; Morales, C. M.; Landis, C. R.; Weinhold, F. NBO6.0; Theoretical Chemistry Institute, University of Wisconsin, Madison, WI, USA, 2013. (14) Gordy, W.; Cook, R. L. Microwave Molecular Spectra, 2nd ed.; Wiley-Interscience: New York, 1984. (15) Pickett, H. M. The Fitting and Prediction of Vibration-Rotation Spectra with Spin Interactions. J. Mol. Spectrosc. 1991, 148, 371−377. (16) Pickett, H. M. SPFIT/SPCAT Program; http://spec.jpl.nasa. gov/ftp/pub/calpgm. (17) Kisiel, Z. PROSPE-Programs for Rotational Spectroscopy; http:// info.ifpan.edu.pl/~kisiel/prospe.htm. (18) Hirota, E. Internal Rotation in Isopropyl Alcohol Studied by Microwave Spectroscopy. J. Phys. Chem. 1979, 83, 1457−1465. (19) Onda, M.; Tsuda, K.; Sakamoto, E. The Microwave Spectrum and Structure of 1,1,1,3,3,3-Hexafluoropropane, Freon R236fa. J. Mol. Struct. 2006, 780−781, 222−224. (20) Grabow, J.-U.; Heineking, N.; Stahl, W. The Microwave Spectrum and Dipole Moment of Hexafluoropropanone. Z. Naturforsch., A: Phys. Sci. 1991, 46a, 229−232.
5657
DOI: 10.1021/acs.jpca.5b03240 J. Phys. Chem. A 2015, 119, 5650−5657
Microwave Spectrum of Hexafluoroisopropanol and Torsional Behavior of Molecules with a CF3−C−CF3 Group Abhishek Shahi† and Elangannan Arunan*,† †
Department of Inorganic and Physical Chemistry, Indian Institute of Science, Bangalore 560012, India S Supporting Information *
ABSTRACT: This paper presents the first microwave spectroscopic investigation on hexafluoroisopropanol (HFIP). A pulsed nozzle Fourier transform microwave spectrometer has been used to determine the rotational constants for HFIP as A = 2105.12166(18) MHz, B = 1053.99503(12) MHz, and C = 932.33959(13) MHz. In addition, five isotopologues of HFIP have been observed experimentally to determine the accurate structure of HFIP. The observed spectrum could be assigned to the most stable conformer of HFIP, called antiperiplanar. Available spectroscopic information and ab initio calculations on five prototype molecules helped in exploring the torsional behavior of molecules having a CF3−C−CF3 group. Twodimensional potential energy surfaces have been analyzed for all molecules, which explained the presence/absence of doubling in the rotational transitions. With the help of natural bond orbital (NBO) analysis, reasons for the conformational preference of HFIP have been explained.
■
INTRODUCTION Fluorinated alcohols are important solvents and are extensively used by biologists, polymer chemists, and organic chemists. Hexafluoroisopropanol (HFIP; CF3−C(H)(OH)−CF3) is a fluorinated alcohol which possesses many interesting properties. It is a protic solvent which can work as a hydrogen bond donor as well as a hydrogen bond acceptor. Its aqueous binary solution helps in stabilizing the α-helical structure of proteins.1 It has a propensity to dissolve even polymers because of its ability to form strong hydrogen bonds. Therefore, the polymer should possess hydrogen bond acceptor groups. For example, polymers such as polythene terephthalate, polyamides, etc.2 could be dissolved in HFIP. These polymers are hard to dissolve in other solvents. HFIP is a suitable solvent for rearrangements via zwitterionic intermediates, whereas CH3OH is not suitable for such a reaction.3 The HFIP molecule is interesting for spectroscopists as well. The conformational preference of HFIP has been a subject of several studies. IR, Raman, and matrix isolation studies show that the molecule exists in two conformations: antiperiplanar (AP) and synclinal (SC). The AP conformer has been found to be more stable than the SC conformer.4−6 Both conformers exist in CCl4 solution as well as in an N2 matrix and a CO matrix. However, in polar media the IR transitions corresponding to the SC conformer were found to be more intense.5 In an Ar matrix, only the AP conformer could be observed.5 The relative intensity of the conformers also depends on the temperature. These conformers exist because of the internal torsion of the −OH group. The prototype molecule isopropanol also exists in two conformers due to the same © 2015 American Chemical Society
internal torsion. However, for this molecule the SC conformer is more stable than the AP conformer.6−8 Suhm’s group has extensively 6 studied the effect of fluorination on the isopropanol molecule using IR spectroscopy. Our interest in HFIP was triggered by the fact that it is an exceptional solvent with the potential to form different types of hydrogen bonds with other hydrogen bond acceptor/donor molecules. Therefore, we wanted to study hydrogen-bonded complexes of HFIP with molecules such as H2O. However, we found that the rotational spectrum of the HFIP monomer itself was unknown. Somewhat coincidently, rotational spectra of the other prototype molecules hexafluoroisobutene (HFIBE; CF3− C(CH2)−CF3)9 and hexafluoroacetone imine10 (HFAIM; CF3−C(NH)−CF3) were investigated recently. For both of these molecules, the rotational transitions appear as doublets. The reason for this splitting was found to be the counter rotation of the two CF3 groups. Would such a splitting be observed in HFIP? In this paper, we report the first rotational spectroscopic study of the HFIP molecule. The results are compared with those of prototype molecules having a CF3−C−CF3 group. Unlike the case for HFIBE and HFAIM, the rotational transition for HFIP did not show any evidence of splitting. Other related molecules such as hexafluoroacetone (HFA), hexafluoroisobutane (HFIBA; CF3−C(H)(CH3)−CF3), and hexafluoroisopropylamine (HFIPA; CF3−C(H)(NH2)−CF3) Received: April 3, 2015 Revised: May 18, 2015 Published: May 19, 2015 5650
DOI: 10.1021/acs.jpca.5b03240 J. Phys. Chem. A 2015, 119, 5650−5657
Article
The Journal of Physical Chemistry A
O1−H2 dihedral angle at the MP2/6-311+G(d) level. The AP conformer is more stable than SC by 5.0 kJ/mol (Figure 2).
also do not show any splitting. A detailed two-dimensional potential energy surface (2D-PES) analysis on all of these molecules revealed the reason for the presence/absence of splitting. Natural bond orbital (NBO) analysis has been carried out to understand the relative stability of the conformers.
■
EXPERIMENTAL METHODS HFIP (≥99%) was purchased from Aldrich. Its monodeuterated isotope (OD) was prepared by mixing HFIP with D2O (99.9% by Cambridge Isotope Laboratory) in a 1/1 molar ratio. Its doubly deuterated (CD-OD) isotopologue was purchased from Aldrich (99%). None of the commercial samples were further purified. All C-13 isotopologues were observed in natural abundance. The rotational spectrum was recorded using a pulsed nozzle Fourier transform microwave (PNFTMW) spectrometer, the details of which have already been published.11 Initial experiments were carried out using Ar as the carrier gas. However, we observed that the HFIP signals were more intense while using helium as the carrier gas. Therefore, for further experiments, helium was used as the carrier gas. The flow rate of He was kept at 200 SCCM, and 1% of it was flown through a bubbler containing an HFIP sample. The carrier gas seeded with the HFIP was expanded from a backing pressure of 1.5 bar through a 0.8 mm diameter pulse valve into the Fabry−Perot cavity. Multiple free induction decays (FIDs) were recorded per gas pulse. At a sampling rate of 5 MHz, 256 points were collected for each FID during the search for rotational transitions. Once a signal was observed, it was further averaged with 512 or 1024 points to improve the resolution. A microwave pulse of 1.3 μs duration was found to be the optimum for both the b and c type transitions.
Figure 2. Relaxed potential energy surface scan of the OH torsion (dihedral angle H1−C1−O1−H2) for the HFIP molecule at the MP2/6-311+G(d)level. AP, SC, and SP denote antiperiplanar, synclinal, and synperiplanar conformers of HFIP, respectively.
■
Two minima corresponding to the SC conformer are present in the PES. These minima are separated by a small barrier of 1.3 kJ/mol (∼108.4 cm−1), which is slightly less than the ZPE (109.5 cm−1) of the corresponding vibrational motion. The saddle point for this barrier corresponds to the synperiplanar structure (the H1−C1−O1−H2 dihedral angle is 0°). In addition to the torsion about the C−O bond, HFIP has a torsional vibration involving the two CF3 groups. This motion is also present in HFIBE, HFAIM, HFA, HFIPA, and HFIBA. As mentioned earlier, HFIBE and HFAIM show splitting of rotational transitions due to this motion. All of these molecules were optimized at the MP2/6-311++G(d,p) level of theory for comparison. The 2D-PES was constructed for all the molecules, allowing all the other coordinates to optimize. For the HFIP molecule, the dihedral angles selected for the 2D scan are H1−C1−C2− F1 and H1−C1−C3−F6 (see Figure 1). Each angle was scanned for 120° with a step size of 3°, and a total of 1681 (=412) grid points was calculated (Figure 4). Similarly, two dihedral angles for the rest of the molecules were selected. Where a H atom on the center C atom was absent, O, N, and C atoms were selected in place of the H atom for HFA, HFAIM, and HFIBE molecules, respectively. The advantages of performing these expensive and laborious scans were to compare the barrier heights, both small and large, and also to determine the most stable conformer (vide infra). With computation cost in mind, a reasonably good level of theory, B3LYP/6-311G*, was selected to scan the coordinates for 1681 grid points. Motion along the diagonal of the 2D-scan graph corresponds to counter rotation of the two CF3 groups. Orbital properties and their overlap were calculated using the NBO 6.013 program for all molecules at the MP2 level of theory.
THEORETICAL CALCULATIONS Ab initio calculations were performed using the G09 suite of programs.12 Different possible conformers of HFIP were optimized at the MP2/6-311++G(d,p) level. Frequency calculations were performed for the optimized geometries at the same level. These calculations and earlier IR studies6 show that the HFIP molecule has two minima, AP and SC. These two conformers originate due to the internal rotation of −OH group, and the H1−C1−O1−H2 dihedral angles are 180 and 52° for AP and SC, respectively (Figure 1). Vibrational
Figure 1. Antiperiplanar (AP) and synclinal (SC) conformers of HFIP. Geometries were optimized at the MP2/6-311++G(d,p) level of theory.
frequency calculations confirm that both conformers are true minima. Both of the terminal CF3 groups were eclipsed with respect to each other for the AP conformer and slightly staggered for the SC conformer. The dihedral angle F1−C2− C3−F6 for the SC conformer is 18.7°. A relaxed potential energy surface (PES) scan was performed for the H1−C1− 5651
DOI: 10.1021/acs.jpca.5b03240 J. Phys. Chem. A 2015, 119, 5650−5657
Article
The Journal of Physical Chemistry A
■
of the fit was 4 kHz. At this stage, it was not clear whether the observed progression is for the AP or the SC conformer. We were expecting the presence of other progressions corresponding to another conformer for two reasons. First, IR, Raman, and matrix isolation studies showed the presence of two conformers for HFIP, and second, a microwave study on the prototype molecule isopropanol8,18 showed the presence of two isomers because of the torsion of the −OH group. The optimized geometry for the AP conformer has no a dipole component. However, the SC conformer has finite dipole moment components along all three (a, b, c) principal axes. Therefore, for the SC conformer all the three types of transitions were expected. However, we could not observe any a type transition even after performing long searches. These searches were performed on the basis of ab initio rotational constants for both the AP and SC conformers. Absence of the a type transitions suggested that the experimentally observed spectra correspond to the AP conformer. If SC conformers exist in a supersonic molecular beam, a type transitions might not be seen because of the very small barrier (108.4 cm−1) for interconversion. This could lead to a very large tunneling splitting, making the observation and assignment difficult. Frequency calculations do predict one such motion whose zero point energy is 109.5 cm−1, comparable to the barrier height. The saddle point for this motion was the synperiplanar structure (vide supra). This could lead to hindered rotation and levels with internal rotor angular momentum. No transitions could yet be assigned for such levels. However, more evidence favoring the presence of the AP conformer are given in the next paragraphs. Moreover, unlike the rotational spectra of the prototype HFIBE and HFAIM molecules, no doubling in rotational transitions of the HFIP molecule was observed. The reason behind this doubling is different from that of isopropanol and is discussed later in detail. The signal intensity also helped us in identifying the conformer. For the observed progression, b type signals were always stronger than c type. This experimental observation is consistent with the ab initio calculated dipole moment of the AP conformer geometry, for which the b dipole component (0.6 D) is 3 times larger than the c dipole component (0.2 D). In contrast, for the SC conformer, the c dipole component (2.1 D) is larger than both the b dipole (1.8 D) and a dipole (1.0 D) components (Table 1). In addition for the synperiplanar conformer, which is a saddle point, the c dipole component (2.9 D) is larger than the b dipole (1.9 D). Hence, both
RESULTS AND DISCUSSION Calculated rotational constants for the AP and the SC conformers are very close to each other at the MP2/6-311+ +G(d,p) level (Table 1). On the basis of predictions, using Table 1. Experimental and Theoretical Rotational and Distortion Constants for Hexafluoroisopropanol exptl A/MHz B/MHz C/MHz DJ/kHz DJK/kHz DK/kHz d1/kHz d2/kHz RMS/MHz no. of transitions dipole moment (a, b, c)/D
2105.12166(18) 1053.99503(12) 932.33959(13) 0.05713(79) 0.50829(76) −0.4592(41) −0.00731(13) 0.002357(59) 0.0048 111 a=0 b>c≠0
a
calcd (AP)
calcd (SC)
2098.76488 1053.29499 931.98682 0.0522 0.2635 −0.2125 −0.0068 0.0015
2102.51954 1054.01680 932.48540 0.0560 0.2317 −0.1833 −0.0075 0.0017
0.0, 0.6, 0.2
1.2, 1.5, 2.2
a
Numbers in parentheses are signal standard errors in units of least significant figures.
these rotational constants, a search was started for a strong b type 414 ← 303 transition for the AP conformer, which was predicted at ∼8465 MHz. The search was started at 8465 MHz with Ar as the carrier gas. Within a 10 MHz search, a transition was observed at 8473.8905 MHz. Five more signals were searched similarly on the basis of the predictions and were observed readily. These rotational transitions were found at 9125.4651, 7059.1440, 10213.3467, 9526.6928, and 11161.5053 MHz and could be assigned to 505 ← 414, 404 ← 313, 515 ← 404, 321 ← 212, and 606 ← 515, respectively. All of these signals were significantly stronger with He carrier gas and were fitted to a semirigid rotor Hamiltonian. The rotational constants obtained from this tentative fit were used for further predictions. Eventually, a total of 111 transitions could be observed. These transitions included both b and c type pure rotational transitions of R and Q branches (see Table S1 of the Supporting Information). The observed transitions were fitted to Watson’s S reduction Hamiltonian.14 Programs used for the fitting were mainly SPFIT15,16 and also ASFIT.17 As expected, they produced the same results. The experimental rotational constants and distortion constants are given in Table 1. The RMS deviation
Table 2. Fitted Rotational and Distortion Constants, RMS Values, and Numbers of Transitions for Four Different Isotopologues of HFIPa A/MHz B/MHz C/MHz DJ/kHz DJK/kHz DK/kHz d1/kHz d2/kHz RMS #N a
CF3−C(H)(OH)−CF3
CF3−C(H)(OD)−CF3
CF3−C(D)(OD)−CF3
CF3−13C(H)(OH)−CF3
2105.12166(18) 1053.99503(12) 932.33959(13) 0.05713(79) 0.50829(76) −0.4592(41) −0.00731(13) 0.002357(59) 0.0048 111
2066.6444(10) 1052.74254(91) 925.60404(54) 0.0552(55) 0.442(55) −0.483(71) −0.0089(37) 0.0059(35) 0.0056 26
2042.0355(23) 1046.7717(10) 925.2142(10) 0.0486(97) 0.508(74) −0.54(11) −0.00731 0.002357 0.013 28
2099.9217(16) 1053.43442(63) 931.88122(63) 0.0492(57) 0.481(53) −0.460(60) −0.00731 0.002357 0.0057 17
13
CF3−C(H)(OH)−CF3 2105.0638(18) 1050.43361(72) 929.53229(60) 0.0528(59) 0.525(62) −0.411(74) −0.00731 0.002357 0.0076 22
Numbers without parentheses are not included in the fit. Values of the parent isotopologue are used for that particular fitting. 5652
DOI: 10.1021/acs.jpca.5b03240 J. Phys. Chem. A 2015, 119, 5650−5657
Article
The Journal of Physical Chemistry A synperiplanar and SC conformers are ruled out and the observed signals can be unambiguously attributed to AP. Since experimentally observed rotational constants were close to the calculated values, predictions of spectra for the other isotopologues were easy. We took the difference between the experimental and calculated rotational constants of the parent molecule and assumed similar differences for the rotational constants of the other isotopologues. Rotational spectra for four different isotopologues, CF3−C(H)(OD)−CF3 (HFIP(OD)), CF 3 −C(D)(OD)−CF 3 (HFIP(CDOD)), CF3−13C(H)(OH)−CF3 (HFIP-13(center)), and 13CF3−C(H)(OH)−CF3 (HFIP-13(side)), could be observed experimentally. We observed 26 transitions for HFIP(OD), 28 transitions for HFIP(CDOD), 17 transitions for HFIP13(center), and 22 transitions for HFIP-13(side) isotopologues. These are given in Tables S2−S5 of the Supporting Information, respectively. These transitions could be fitted within the experimental uncertainty, and the RMS values are 5.6, 13.0, 5.7, and 7.6 kHz for HFIP(OD), HFIP(CDOD), HFIP-13(center), and HFIP-13(side), respectively. The rotational constants obtained from these fits are given in Table 2. We could observe only one set of transitions for the side C-13 carbon. The signals for the HFIP-13(side) isotopologue were more intense (almost twice) than the signals for the HFIP13(center) isotopologue. This indicated that the side C atoms are equivalent, which is in accord with the ab initio results on the AP conformer. In the calculated structure, the molecule exhibited a plane of symmetry containing H2−O1−C1−H1 atoms. However, the side C atoms are different for the SC conformer in its equilibrium structure and there is no plane of symmetry. This is consistent with the observed spectrum corresponding to the AP conformer. The HFIP(OD) and HFIP(CDOD) isotopologues were comparatively easier to observe experimentally than the C-13 isotopologues. The hyperfine splitting from the quadrupolar D atom could be observed but was not well resolved. The line centers were used in the fit. With the help of a Kraitchman analysis,14 the position of the isotopic substitution can be determined. This analysis assumes that the bond length of the molecule remains unchanged on isotopic substitution. It is important to mention that this assumption is quite reasonable for heavy-atom substitution but not for the H/D substitution in stable molecules. Kisiel’s programs17 were used for this analysis. Using rotational constants of the five isotopologues, Cartesian coordinates of the substituted atoms have been determined. Along with directly bonded parameters, some nonbonded parameters are also given in Table 3. These parameters again support the AP conformer: e.g., the distance between the center of mass and H2 is equal to the calculated distance for the AP conformer and significantly different from that of the SC conformer (Table 3). Moreover, the experimental H1−C1−H2 angle and the C3− C1−H2−C2 dihedral angle are also close to the calculated values of the AP conformer. To examine the behavior of the CF3−C−CF3 group, comparisons of HFIP with other prototype molecules, HFIBE, HFAIM, HFA, HFIBA, and HFIPA, have been carried out. Equilibrium geometries of these molecules are shown in Figure 3. Rotational constants were taken from earlier works,9,10,19 and we fitted them to get the internal coordinates (Table 4). An isotopic study for the HFA monomer20 is not known; therefore, we used the calculated parameters for this molecule for further analysis. The 2D-PES scans showed that
Table 3. Results from the Kraitchman Analysis for HFIP Molecules, using the Rotational Constants of the Different Isotopologues param
exptl
calcd (AP)
calcd (SC)
C1−C2/Å C1−C3/Å C1−H1/Å C1−H2/Å CM−H2/Å C2−C1−C3/deg H1−C1−C2/deg H1−C1−C3/deg H1−C1−H2/deg C2−C1−H1−C3/deg C3−C1−H2−C2/deg
1.530(7) 1.530(7) 1.121(3) 1.935(3) 2.119(1) 113.2(6) 103.9(7) 113.2(7) 137.4(3) 123.2(8) 113.5(5)
1.531 1.531 1.092 1.930 2.109 114.1 107.2 107.2 135.9 122.8 115.1
1.531 1.527 1.097 1.930 2.511 113.9 106.5 106.5 94.5 121.9 134.6
two types of barriers exist for HFIBE, HFA, and HFAIM (denoted large and small). On the other hand, for the HFIP, HFIBA, and HFIPA molecules only one (large) barrier was present. The large barriers, in all of these molecules, are due to the CF3 counter rotation, which interconverts the staggered and the eclipsed forms of these molecules (Figure 4). The smaller barrier is due to the restricted counter rotation of the two CF3 groups just around its equilibrium geometry. Saddle points for the small barrier are marked in Figure 4 by small green dots and also pointed by an arrow. Barrier heights are extracted by following the diagonal path of the 2D scan plot, and a representative one-dimensional PES for all molecules is shown in Figure 5. Compiled results of barrier heights are presented in Table 5 for all of the prototype molecules. The barrier heights corresponding to the staggered−eclipsed interchange are very large, and they are unlikely to be responsible for the doubling observed in the rotational spectra of HFIBE and HFAIM molecules. As discussed earlier, for some molecules the smaller barrier is not present. Position 1 in the PES represents saddle points for the large barrier (Figure 5). At position 1, opposite CF3 groups are eclipsed for all molecules and are shown in Figure S1(a1−f1) in the Supporting Information. In the most stable equilibrium geometries for CF3−C(sp3)−CF3 types of molecules, the terminal CF3 groups are eclipsed with respect to each other and are staggered with respect to the center carbon substituents (Figure 3d−f or Figure S1(d3−f3)). The HFIPA molecule appeared as an exception, and the most stable structure for this molecule has slightly staggered CF3 groups. On the other hand, the most stable equilibrium structures for the CF3−C(sp2)−CF3 group of molecules are slightly staggered (Figure 3a−c or Figure S1(a2/a4,b2/b4,c2/c4)) and are denoted by position 2 or 4 in Figure 5. Position 3 in Figure 5 corresponds to the saddle points for the small barriers. The geometries corresponding to this point are given for all of the molecules in Figure S1(a3−c3). In these geometries both CF3 groups are eclipsed with each other as well as with the double bond. We have classified these molecules into two types, “type I” with small barriers (e.g., HFIBE, HFA, and HFAIM exhibited CF3−C(sp2)−CF3 group) and “type II” without small barriers (e.g., HFIP, HFIBA, and HFIPA exhibited CF3−C(sp3)−CF3 group). At this point, we observed two important structural differences between type I and type II molecules. The first difference is that the two CF3 groups are slightly staggered in type I molecules, while these are eclipsed in type II molecules in their equilibrium geometries. The second difference, as noted 5653
DOI: 10.1021/acs.jpca.5b03240 J. Phys. Chem. A 2015, 119, 5650−5657
Article
The Journal of Physical Chemistry A
Figure 3. Equilibrium geometries of the prototype molecules (a) hexafluoroisobutene (HFIBE), (b) hexafluoroacetone imine (HFAIM), (c) hexafluoroacetone (HFA), (d) hexafluoroisobutane (HFIBA), (e) hexafluoroisopropylamine (HFIPA), and (f) hexafluoroisopropanol (HFIP).
(Figure 3 and Table 4). Interestingly, HFA shows a longer than expected C−C bond (1.55 Å), even though the center C atom is sp2 hybridized and bond angles are close to 120°. Furthermore, HFA did not show splitting in the rotational spectrum, even with a small barrier being predicted, resulting in an equilibrium geometry that is slightly staggered. Overall, HFA turned out to be different from HFIBE and HFAIM. Inspection of the small barrier revealed the reason behind the presence/ absence of doubling in the rotational transitions. For HFA, the height for the “small barrier” was larger (1.6 kJ/mol) in comparison to that for HFIBE (0.2 kJ/mol) and HFAIM (0.8 kJ/mol). In addition, the order of the experimental splitting for HFIBE is 10 MHz9 and that for HFAIM is 10 kHz.10 The higher barrier for the HFA molecule qualitatively suggests that the splitting should be on the order of 100 Hz or less. Resolving splittings of less than 5 kHz is not possible with the resolution of typical microwave spectrometers. In other words, because of the higher barrier, the splitting is too small to be observed for the HFA molecule. Moreover, there was a vibrational motion corresponding to counter rotation of the CF3 groups with frequencies of 31, 37, and 24 cm−1 for HFIBE, HFA, and HFAIM, respectively, at the B3LYP/6-311G(d) level. The ZPE corresponding to these motions are 0.2, 0.2, and 0.1 kJ/mol, respectively. Thus, the small barriers (V2) are always larger than the ZPE corresponding to this vibrational mode for all of the molecules (Table 5). To the best of our knowledge, microwave spectra of the HFIPA and HFIBA have not been reported to date. Our prediction says that both molecules should not show doubling, due to the free counter rotation of the two CF3 groups. NBO Analysis of Conformers. Both CF3 groups are eclipsed for HFIP and slightly staggered for HFIBE in their equilibrium structure. In general, CF3 groups are eclipsed or staggered, depending on the hybridization of the center C atom (vide supra). HFIP and HFIBE center C atoms are sp3 and sp2 hybridized, respectively. As discussed earlier, corresponding to an eclipsed structure in HFIBE, there is a small barrier and one imaginary frequency was found. We thought of identifying some particular orbital overlaps which change on this
Table 4. Geometrical Parameters Obtained from the Kraitchman Analysis for HFIP, HFIBE, HFAIM, and Hexafluoropropanea param
value (V = 0)
value (V = 1)
Hexafluoroisobutene (HFIBE) C1−C2/Å 1.465(2) 1.462(3) C1−C3/Å 1.465(2) 1.462(3) C1−C4/Å 1.339(4) 1.340(4) C2−C1−C3/deg 121.0(2) 121.0(3) C2−C1−C4/deg 116.1(12) 119.5(2) C3−C1−C4/deg 122.9(12) 119.5(2) C2−C1−C4−C3/deg 180(0) 177.3(79) Hexafluoroacetone Imine (HFAIM) C1−C3/Å 1.486(02) 1.487(2) C1−C2/Å 1.490(29) 1.478(2) C1−N1/Å 1.268(03) 1.270(3) C2−C1−C3/deg 120.5(19) 121.4(2) N1−C1−C2/deg 117.5(22) 115.9(9) N1−C1−C3/deg 121.9(11) 122.3(9) C2−C1−N1−C3/deg 177.5(116) 172.8(40) Hexafluoroacetone (HFA) C1−C2/Å C1−C3/Å C1−O1/Å O1−C1−C2/deg O1−C1−C3/deg C2−C1−C3/deg C1−C2−O1−C3/deg a
calcd 1.507 1.507 1.337 116.7 121.6 121.6 180.0 1.528 1.525 1.271 116.4 118.6 125 −179.0 1.547 1.547 1.203 121.8 121.8 116.4 180.0
For HFA ab initio calculated parameters are reported.
above, is the hybridization of the middle carbon atom. Type I has the center carbon in sp2 hybridization, and type II has the center carbon in sp3 hybridization. For the HFIP molecule, C− C bond lengths were 1.53 Å and angles at the center carbon were close to tetrahedral angles, which are the typical features of sp3 C atoms. However, in HFIBE and HFAIM, C−C bond lengths were close to 1.47 Å and angles at the center carbon were close to 120°, which are typical of sp2 carbon atoms 5654
DOI: 10.1021/acs.jpca.5b03240 J. Phys. Chem. A 2015, 119, 5650−5657
Article
The Journal of Physical Chemistry A
Figure 4. 2D-PES scan for the CF3 counter rotation in (a) hexafluoroisopropanol, (b) hexafluoroisopropylamine, (c) hexafluoroisobutane, (d) hexafluoroisobutene, (e) hexafluoroacetone imine, and (f) hexafluoroacetone molecules. In parts d−f, green dots (denoted by arrows) represent the small barrier and the corresponding structures are saddle points. Detailed information about the axis values is given in the text.
conformational exchange in HFIBE (slightly staggered ↔ eclipsed). For this purpose, NBO calculations were performed for slightly staggered (equilibrium geometry) and eclipsed conformers (saddle point geometry). The overlapping between C1−C10 π-bonding orbitals and C2−F4 and C3−F9 antibonding orbitals was larger in the slightly staggered conformer than in the eclipsed conformer, determined by the second-order perturbation energy (E2). However, there are many orbital overlaps which led to stabilization/destabilization on eclipsed−slightly staggered exchange and it was difficult to identify any specific orbital overlaps. Finally, we decided to sum up all E2 values arising from all such overlaps. The total sums for the slightly staggered and the eclipsed conformers were 2713.9 and 2713.2 kJ/mol, respectively. Qualitatively, these values supported that the slightly staggered conformer is more stable than the eclipsed conformer. To the best of our knowledge, this method has been used for the first time to
Figure 5. Barrier for counter rotation of CF3 groups. The energy values for different positions are given in Table 5. Their structures are presented in Figure S1 in the Supporting Information.
Table 5. Barrier Height, Angle, and ZPE for All Prototype Moleculesa property
position
HFIBE
HFAIM
HFA
HFIP
HFIBA
HFIPA
energy for large barrier (V1) (kJ/mol) energy for small barrier(V2) (kJ/mol)
1 2 3 4 1 2 3
25 0 0.2 0 −30 0 30 31 0.2 10 MHz
21 0 0.8 0 −42 0 42 24 0.1 10 kHz
15 0 1.6 0 −52 0 52 37 0.2 not obsd
43
45
47
0
0
0
0b
0b
25b
22 0.1 n/a
26 0.2 n/a
40 0.2 n/a
dihedral angle F−C−C−X/deg
vibrational freq (cm−1) ZPE (kJ/mol) order of exptl splitting a
The numbers 1−4 are the positions in the PES scan of Figure 5. All calculations were performed at the B3LYP/6-311G* level of theory. bF−C−C− F dihedral angle. 5655
DOI: 10.1021/acs.jpca.5b03240 J. Phys. Chem. A 2015, 119, 5650−5657
Article
The Journal of Physical Chemistry A
Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.jpca.5b03240.
confirm the conformational preference. The electronic energy difference between the two conformers was 1.3 kJ/mol, and the total E2 difference was 0.7 kJ/mol. Clearly, half of the net stabilization arises due to favorable orbital overlap. To verify the above approach, calculations were carried out for AP and SC conformers. The sums of E2 for AP and SC conformers are 2557.4 and 2539.0 kJ/mol, respectively, which predicts that AP is more stable than SC. The electronic energy difference between the two conformers is only 5.0 kJ/mol, in comparison to the difference of 18.4 kJ mol−1 in E2. Clearly, in this case, the net stabilization energy is less than that predicted by comparing the E2 values. In the AP conformer, we could identify three overlaps which are not found in SC: first, the lone pair on the O atom overlapped with the C1−C2 and C1−C3 antibonding orbitals equally, second, another lone pair of the O atom overlapped well with the C−H antibonding orbital, and third, the O−H bonding MO overlapped with the C−H antibonding MO. Overall, orientations of oxygen lone pairs are the key points for conformational preferences. In these two examples, we were not expecting the exact match between the net stabilization energy from electronic structure calculations and the difference between the sum of E2 values, as dispersion and exchange repulsion also contribute. It was interesting to note that the relative stability of the various conformers could be predicted by looking at the sum of all E2 values.
■
Corresponding Author
*E-mail for E.A.: [email protected]. Notes
The authors declare no competing financial interest.
■
ACKNOWLEDGMENTS The authors thank the Indo-French Centre for Promotion of Advanced Scientific Research for partial support. The authors also thank the Super Computer Research Center and Inorganic and Physical Chemistry department at the Indian Institute of Science for providing good computational facilities. A.S. thanks the Council for Research and Industrial Research (CSIR) of India for a fellowship and Dr. Devendra Mani for his help during experiments.
■
ABBREVIATIONS pulsed nozzle Fourier transform microwave spectrometer,PNFTMW; hexafluoroisopropanol,HFIP; hexafluoroisobutane,HFIBA; hexafluoroisopropylamine,HFIPA; hexafluoroisobutene,HFIBE; hexafluoroacetoneimine,HFAIM; hexafluoracetone,HFA
■
■
CONCLUSION HFIP and four of its isotopologues have been observed experimentally using a PNFTMW spectrometer. Observed transitions could be fitted well to Watson’s S reduction Hamiltonian within the experimental uncertainties. Experimental rotational constants for the parent themselves could not distinguish between AP and SC conformers. The absence of a type transitions and the fact that b type transitions are stronger than c type transitions provide evidence in support of the AP conformer as the experimentally observed structure. Moreover, isotopic substitution results also suggest the AP structure. We performed 2D-PES scans for several prototype molecules. Our results can be summarized as follows. (a) If the center C atom is sp2, a small barrier will be present, leading to doubling of energy levels: e.g. HFIBE, HFAIM, and HFA. Depending on the barrier height and actual splitting in energy levels, doubling of rotational transitions may be obtained experimentally. (b) If the center C atom is sp3, there would be no small barrier and hence no splitting of rotational transition due to a torsional motion: e.g. HFIP, HFIPA, and HFIBA. With the help of NBO calculations, reasons for a particular conformational preference can be rationalized. Total secondorder perturbation energy for a molecule can give an idea of the conformational preference, a new approach proposed here. We hope that this approach can be generalized and used extensively.
■
AUTHOR INFORMATION
REFERENCES
(1) Hong, D. P.; Hoshino, M.; Kuboi, R.; Goto, Y. Clustering of Fluorine-Substituted Alcohols as a Factor Responsible for Their Marked Effects on Proteins and Peptides. J. Am. Chem. Soc. 1999, 121, 8427−8433. (2) Cazes, J. Encyclopedia of Chromatography; Taylor & Francis: Abigdon, U.K., 2005; Vol. 2. (3) Burdisso, M.; Gandolfi, R.; Toma, L.; Obertib, R.; Oberti, R. Hexafluoroisopropanol as a Suitable Solvent for Rearrangements via Zwitterionic Intermediates. Tetrahedron 1991, 47, 6725−6736. (4) Cyvin, S. J.; Brunvoll, J.; Perttilä, M. Substituted Profanes IX. Vibrational Frequencies and Calculated Mean Amplitude for Hexalfuoroisopropanol-2. J. Mol. Struct. 1973, 17, 17−21. (5) Barnes, A. J.; Murto, J. Infra-red Cryogenic Studies. Part 10. Conformational Isomerism of 1,1,1,3,3,3-Hexafluoropropan-2-ol. J. Chem. Soc., Faraday Trans. 2 1972, 68, 1642−1651. (6) Schaal, H.; Häber, T.; Suhm, M. A. Hydrogen Bonding in 2Propanol. The Effect of Fluorination. J. Phys. Chem. A 2000, 104, 265− 274. (7) Hirota, E.; Kawashima, Y. Internal Rotation of the Hydroxyl Group in Isopropanol and the Chirality of the Gauche Form: Fourier Transform Microwave Spectroscopy of (CH(3))(2)CHOD. J. Mol. Spectrosc. 2001, 207, 243−253. (8) Kondo, S.; Hirota, E. Microwave Spectrum and Internal Rotation of Isopropyl Alcohol. J. Mol. Spectrosc. 1970, 34, 97−107. (9) Grubbs, G. S.; Novick, S. E.; Pringle, W. C.; Laane, J.; Ocola, E. J.; Cooke, S. A.; Ii, G. S. G.; Esther, J.; Road, A. H. Bis-trifluoromethyl Effect: Doubled Transitions in the Rotational Spectra of Hexafluoroisobutene, (CF 3) 2 CCH 2. J. Phys. Chem. A 2012, 116, 8169− 8175. (10) Obenchain, D. A.; Frohman, D. J.; Grubbs II, G. S.; Long, B. E.; Pringle, W. C.; Novick, S. E. A Pure Rotational Study of Two NearlyEquivalent Structures of Hexafluoroacetone Imine. 67th International Symposium on Molecular Spectroscopy; Ohio State University, Columbus, OH, USA, 2012; Talk TC04. (11) Arunan, E.; Tiwari, A. P.; Mandal, P. K.; Mathias, P. C. Pulsed Nozzle Fourier Transform Microwave Spectrometer: Ideal to Define Hydrogen Bond Radius. Curr. Sci. 2002, 82, 533−540. (12) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb, M. A.; Cheeseman, J. R.; Scalmani, G.; Barone, V.; Mennucci,
ASSOCIATED CONTENT
* Supporting Information S
Text giving the complete author list of ref 12, Figure S1 showing structures at different maxima and minima of the 2DPES given in Figure 4, and Tables S1−S5 presenting all measured transition frequencies and quantum number assignments for HFIP and all isotopologues. The Supporting 5656
DOI: 10.1021/acs.jpca.5b03240 J. Phys. Chem. A 2015, 119, 5650−5657
Article
The Journal of Physical Chemistry A B.; Petersson, G. A.; et al. Gaussian 09, revision D.01; Gaussian, Inc., Wallingford, CT, 2009. (13) Glendening, E. D.; Badenhoop, J. K.; Reed, A. E.; Carpenter, J. E.; Bohmann, J. A.; Morales, C. M.; Landis, C. R.; Weinhold, F. NBO6.0; Theoretical Chemistry Institute, University of Wisconsin, Madison, WI, USA, 2013. (14) Gordy, W.; Cook, R. L. Microwave Molecular Spectra, 2nd ed.; Wiley-Interscience: New York, 1984. (15) Pickett, H. M. The Fitting and Prediction of Vibration-Rotation Spectra with Spin Interactions. J. Mol. Spectrosc. 1991, 148, 371−377. (16) Pickett, H. M. SPFIT/SPCAT Program; http://spec.jpl.nasa. gov/ftp/pub/calpgm. (17) Kisiel, Z. PROSPE-Programs for Rotational Spectroscopy; http:// info.ifpan.edu.pl/~kisiel/prospe.htm. (18) Hirota, E. Internal Rotation in Isopropyl Alcohol Studied by Microwave Spectroscopy. J. Phys. Chem. 1979, 83, 1457−1465. (19) Onda, M.; Tsuda, K.; Sakamoto, E. The Microwave Spectrum and Structure of 1,1,1,3,3,3-Hexafluoropropane, Freon R236fa. J. Mol. Struct. 2006, 780−781, 222−224. (20) Grabow, J.-U.; Heineking, N.; Stahl, W. The Microwave Spectrum and Dipole Moment of Hexafluoropropanone. Z. Naturforsch., A: Phys. Sci. 1991, 46a, 229−232.
5657
DOI: 10.1021/acs.jpca.5b03240 J. Phys. Chem. A 2015, 119, 5650−5657
