The Puzzle of Bond Length Variation in Substituted Cyclobutenes. A

Apr 5, 2010 - 1,2,3,3,4,4-hexafluorocyclobut-1-ene (hexafluorocyclobutene) was measured in gas-phase electron-diffraction (GED) experi- ments by Chang...
0 downloads 0 Views 476KB Size
Download PDF
5358

J. Phys. Chem. A 2010, 114, 5358–5364

The Puzzle of Bond Length Variation in Substituted Cyclobutenes. A New Example: Molecular Structure and Conformations of 1,2-Dimethoxy-3,3,4,4-tetrafluorocyclobut-1-ene Alan D. Richardson,† Kenneth Hedberg,*,† and Bruno Lunelli‡ Department of Chemistry, Oregon State UniVersity, CorVallis, Oregon 97331-4003, and Istituto per lo Studio dei Materiali Nanostrutturati (ISMN) CNR, Bologna Section, 101 Via P. Gobetti, I-40129 Bologna, Italy ReceiVed: April 6, 2009; ReVised Manuscript ReceiVed: February 25, 2010

The structure and composition of 1,2-dimethoxy-3,3,4,4-tetrafluorocyclobut-1-ene (DMCB) have been measured by electron diffraction from the gas at a temperature of 370 K with the help of auxiliary data from molecular orbital and normal coordinate calculations, the former at several levels of theory and basis-set size, most importantly B3LYP/cc-pVTZ. The compound was found to exist primarily as a rotamer of Cs symmetry (ca. 98%; 2σ ) 11%) with the remainder one of C2V symmetry; theory predicts about 88% Cs. Values for some of the more important parameters (rg/Å; ∠R/deg) of the Cs form are r(CdC) ) 1.337(21), r(C1sC4) ) 1.496(8), r(C2sC3) ) 1.501(8), r(C3sC4) ) 1.567(12), r(C1sO) ) 1.318(12), r(C2sO) ) 1.340(12), r(C3sF) ) 1.375(4), r(C4sF) ) 1.368(4), ∠ave(CdCsC) ) 94.4(4), ∠ave(CdCsO) ) 133.5(12), ∠ave(CsOsC) ) 119.6(13), and ∠ave(FsCsF) ) 104.4(7). Surprisingly, although electron-diffraction values for the fluorinated C3sC4 bond in other cyclobutenes are greater than that for cyclobutene itself, that is not the case for DMCB where it is found to be about the same. Details of the DMCB structure, together with possible reasons for the observed variations in the length of the C3sC4 bond in fluorinated cyclobutene-like molecules, are discussed. Introduction A normal sp3-sp3 carbon-carbon bond has a thermal-average length (rg) of about 1.54 Å. Many years ago the structure of 1,2,3,3,4,4-hexafluorocyclobut-1-ene (hexafluorocyclobutene) was measured in gas-phase electron-diffraction (GED) experiments by Chang et al.,1 who found that the sp3-sp3 carbon-carbon bond opposite the double bond was surprisingly long (ra ) 1.595(16) Å) compared to the norm just mentioned and, as well, to the length of this bond in cyclobutene itself (rs ) 1.566(3) Å2). Although rs values are usually slightly smaller than ra ones, the cyclobutene/hexafluorocyclobutene differences seemed too large to be accounted for in this way. We felt this unexpected result deserved a check, and although the ra value of 1.581(11) Å found in our subsequent investigation3 is somewhat less than the one found by Chang et al., the bond still qualifies as unusually long. We pursued the bond-length issue with similar investigations of two related molecules, 1,2-dichloro-3,3,4,4tetrafluorocyclobut-1-ene (dichlorotetrafluorocyclobutene)4 and 2,2,3,3,5,5,6,6-octafluorobicyclo[2.2.0]hex-1(4)-ene (octafluorobicyclohexene),5 the latter comprising two substituted cyclobutene rings fused at the double bond. Not only were the single bonds opposite the double bonds again found to be exceptionally long (ra ) 1.598(10) Å in dichlorotetrafluorocyclobutene and 1.625(5) Å in octafluorobicyclohexene), but their excess lengths were even greater than in the perfluoro compound. The elongation of these fluorinated bonds was attributed in major part to repulsion between C3 and C4, which should accrue substantial positive charges due to the carbonsfluorine electronegativity difference. The sum of the evidence cited above would appear to establish that elongation of the bonds opposite the double bonds in these * To whom correspondence should be addressed. E-mail: kenneth. [email protected]. † Oregon State University. ‡ Istituto per lo Studio dei Materiali Nanostrutturati.

Figure 1. Diagram of the 1,2-dimethoxy-3,3,4,4-tetrafluorocyclobut1-ene (DMCB) molecule with atom numbering. The major component of Cs symmetry is indicated by solid lines; the minor component of C2V symmetry, by dotted lines.

and, presumably, in other cyclobutene molecules of similar type occurs upon fluorine substitution. However, this neat picture is marred by results from studies of hexafluorocyclobutene6 and dichlorotetrafluorocyclobutene7 by MW spectroscopy in which the C3sC4 distance in these molecules was found to be shorter than we reported by 0.01-0.02 Å.8 Moreover, our attribution of C3sC4 bond elongation to charge repulsion, and indeed the accuracy of our measurements, has been called into question by the results of theoretical work at a high level with large basis sets9 on several fluorinated cyclobutenes, which showed no increase in the length (re) of this bond when fluorinated over that in cyclobutene itself. The inconsistencies cited above clearly present a problem that invites further attention. Reported here are our structural results for 1,2-dimethoxy-3,3,4,4-tetrafluorocyclobutene (henceforth DMCB), diagrammed in Figure 1. We had expected that substitution of halogen atoms at the end of the double bond with methoxy groups would have little effect on the length of the opposing single bondsthat is, at least as seen by our experiment, it would be longer than in cyclobutene just as it is in the other halogenated cyclobutenes. Remarkably, this turns

10.1021/jp911185z  2010 American Chemical Society Published on Web 04/05/2010

Bond Length Variation in Substituted Cyclobutenes

J. Phys. Chem. A, Vol. 114, No. 16, 2010 5359

out not to be the case: the bond is roughly the same as in cyclobutene itself and thus considerably shorter than in the other fluorinated molecules we have investigated.. This result for DMCB adds an especially intriguing feature to the structures of compounds of this class and complicates the problem of rationalizing them. We present our structural analysis of DMCB below. It is followed by an examination of the overall picture in terms of the intuitive bonding concepts that have played such an important role in understanding structure in the past. Experimental Section Sample. DMCB, density at 25 °C 1.3409 g mL-1, mp -5.3 °C,10 bp 55 °C at 20 Torr,11 was prepared according to a reported method.12 The IR spectrum of the product showed that it contained about 2% of the monomethoxy compound, 1,3,3,4,4pentafluoro-2-methoxycyclobutene, bp 88 °C, which could not be eliminated by reaction with a solution of sodium methylate in methanol. In spite of the remarkable difference of vapor pressure of the two compounds, separation by fractional distillation was not efficient. However, determination of the previously unknown freezing point of monomethoxypentafluorocyclobutene revealed that it was -43 to -44 °C, i.e., much lower13 than that of DMCB, so that the crystals that separated it from a mixture of DMCB and about 2% of the monomethoxy compound contained less than 0.5% of the latter. By repeated cooling of the mixture and partial distillation under vacuum of the supernatant liquid,14 we obtained as a product DMCB (mp -4 °C) with a purity of about 99.9%. This product was used for all the measurements. Preserving the attained purity was a problem: New bands appeared in the control spectra of the substance kept at -15 °C under argon in containers of borosilicate glass where the vapors could contact the fluorinated polymer-lined commercial gasket of the cap. The substances formed are less volatile than DMCB and by comparison with the spectrum of 1-fluoro-2-chloro-3,3difluorocyclobutenone (C4F2OFCl)15 may include 1,2-dimethoxy3,3-difluorocyclobutenone (C4F2O(OCH3)2). The only satisfactory container was a glass ampule sealed by torch in nitrogen or vacuum. Diffraction Experiments. GED data for DMCB were obtained from the Oregon State apparatus at nozzle-tip temperatures of 98 °C under the following conditions: shape of the rotating sector, r3. nominal long (LC) and middle (MC) camera distances, 30 and 19 cm; nominal accelerating potential, 60 kV with calibration against CO2 with ra(CO) and ra(OO), respectively, equal to 1.1646 and 2.3244 Å; recording medium, Kodak electron-image film developed in D19 developer diluted 1:1; electron-beam current, 0.6-0.7 µA; and exposure times, 45-75 s (LC) and 120-180 s (MC). Four films from each of the two camera distances were selected for analysis. Digitized microdensitometer traces of each were reduced16 to provide the totaland molecular scattered-intensity curves, respectively in the forms s4It(s) and sIm(s). Averages of the latter are shown in Figure 2. Data for the molecular intensity curves are available in the Supporting Information. Structure Analysis Theoretical Calculations. Ab initio and density functional molecular orbital calculations were carried out at several levels of theory and basis sets with use of the Gaussian program set.17 Because DMCB, apart from methyl group rotation, is a double internal rotor, these calculations were extensive. They began with structure optimizations, at the B3LYP/6-31G(d) level, over all unique combinations of the two torsion angles C1dC2sOsC

Figure 2. Scattered molecular intensity curves. The experimental average curves consist of data from four films made at the LC distance and four from the MC distance. The difference curves are experimental minus the theoretical for the final model.

and C2dC1sOsC at intervals of 30°. These calculations led to the predicted existence of two conformers of minimum energy with these angles either equal to, or very close to 0° and 180° (equal to or close to Cs symmetry), and 180° and 180° (symmetry C2V), as illustrated in Figure 1. Full optimizations of the two conformers starting from torsion-angle combinations at or near those cited, followed by frequency calculations, were also carried out at the HF/6-31G(d), B3LYP/6-311++g(d,p), B3LYP/cc-pVTZ, B3PW91/6-311++G(dp), and B3PW91/ccpVTZ levels of theory. Stable minima were found in all calculations except B3PW91/6-311++G(dp) where convergence was elusive, and where one imaginary frequency occurred for both conformations. From these calculations the two overall optimized structures were variously predicted to have either exactly Cs and C2V symmetry (B3PW91/6-311++G(d,p), B3LYP/ 6-31G(d), and B3LYP/cc-pVTZ); a C2V form and a C1 form that deviated from Cs by several degrees (B3LYP/6311++G(d,p)); or two C1 forms that deviated from Cs and C2V symmetry each by a small amount, chiefly in the orientation of the methyl groups (RHF/6-31G(d)). The theoretical conformational internal and Gibbs free energies are given in Table 1, the latter for the system at the temperature of the experiments. The theoretical compositions of the system are shown in the last column where account has been taken of the multiplicity of the Cs form relative to the C2V (2:1). Some of the more important parameter values from the calculations are available in the Supporting Information. Since the bulk of these results suggest that, with the possible exception of the methyl group orientations, the molecular symmetries of the two stable conformers of DMCB are Cs and C2V, we felt we could adopt this assumption in our analyses of the diffraction data without fear of it significantly compromising the results. The reasons are two: any deviations of the heavy atoms from these symmetries would most likely be very small and thus undetectable from our experiments, and orientations of the methyl groups would in any case be undetectable because of the negligible weight of the longer H · · · M scattering terms that define the orientation. With use of the program ASYM40,18 we also carried out normal coordinate analyses based on the theoretical force fields obtained from the B3LYP/cc-pVTZ molecular orbital calculations for molecules of C2V and Cs symmetries. The results included estimates of amplitudes of vibration and values for

5360

J. Phys. Chem. A, Vol. 114, No. 16, 2010

Richardson et al.

TABLE 1: Theoretical Energies of the Cs and C2W Forms of 1,2-Dimethoxy-3,3,4,4-tetrafluorocyclobut-1-ene and Estimated Gas-Phase System Composition Eha

mol fraction of Cs from

Ghb

theory/basis

Cs

C2V

Cs

C2V

∆Ec

∆G° c

∆E

∆G°

HF/6-31G(d) B3LYP/6-31G(d) B3LYP/6-311++G(d,p) B3LYP/cc-pVTZ B3PW91/cc-pVTZ

-777.965272d -781.850198 -782.106060 -782.284862 -781.865497

-777.963216 -781.849899 -782.105527 -782.284106 -781.864766

-778.021833 -781.908345 -782.165874 -782.222680 -781.923491

-778.095028 -781.907213 -782.164180 -782.221227 -781.922019

1.501 0.248 0.456 0.474 0.550

0.796 0.744 1.113 0.955 0.968

0.925 0.723 0.762 0.782 0.792

0.854 0.845 0.900 0.879 0.881

a Zero-point energy corrections included. b At 100 °C. c C2V-Cs in kilocalories/mole. d This form has C1 symmetry because the heavy atoms deviate slightly from coplanarity.

terms used to interconvert different types of interatomic distance. Experience has shown that amplitudes and conversion terms are not very sensitive to small differences in force fields; the choice of the B3LYP/cc-pVTZ molecular orbital results for this purpose was appropriate. System Model. Preliminary refinements presuming the existence of only one of the conformers gave reasonable fits in each case. This result is due to the similarity between the numbers and magnitudes of the interatomic distances for the two forms: as Figure 1 reveals, except for the distances between methyl groups that contribute a relatively small amount of the total molecular scattering, these are essentially the same. However, the theoretical results uniformly suggest a mixture of the two conformers containing about 10% of the C2V form (Table 1), and we decided to include this form in our models. Because of the structural similarities of the components and the likely small amount of the C2V form, there was no hope of obtaining reliable independent parameter values for each component from unrestricted refinements. However, these observations suggested that a suitable model of the system could be defined by parameters specific to the Cs form only and carrying along the structure of the C2V form in accordance with some assumptions about how it differs from the Cs. For the structure-defining parameters of the system (in rR space) we chose the following set of averages and differences for the Cs form, the differences always taken to be the value of the parameter on the left side of the molecule minus that on the right side (Figure 1). 〈r(CsH)〉; r(CdC); 〈r(CsC)〉 ) [r(C1sC4) + r(C2sC3)]/2; ∆r(CsC) ) r(C1sC4) - r(C2sC3); 〈r(CsF)〉 ) [2r(C4sF9) + 2r(C3sF7)]/4; ∆r(CsF) ) r(C4sF9) - r(C3sF7); 〈r(CsO)〉 ) [r(C1sO5) + r(C2sO6)]/2; ∆r(CsO) ) r(C1sO5) - r(C2sO6); 〈r(OsCH3)〉 ) [r(O5sC11) + r(O6sC12)]/2; ∆r(OsCH3) ) r(O5sC11) - r(O6sC12); 〈∠(CsCsC)〉)[∠(C2sC1sC4)+∠(C1sC2sC3)]/2;∆∠(CsCsC)〉 )∠(C2sC1sC4)-∠(C1sC2sC3);〈∠(CsCsO)〉)[∠(C2sC1sO5) + ∠(C1sC2sO6)]/2; ∆∠(CsCsO)〉 ) ∠(C2sC1sO5) ∠(C1sC2sO6);〈∠(FsCsF)〉)[∠(F9sC4sF10)+∠(F7sC3-F8)]/ 2;∆∠(FsCsF)〉)∠(F9sC4sF10)-∠(F7sC3sF8);〈∠(CsOsC)〉 ) [∠(C1sO5sC11) + ∠(C2sO6sC12)]/2; ∆∠(CsOsC)〉 ) ∠(C1sO5sC11) - ∠(C2sO6sC12); 〈∠(flap)〉 ) [∠(C3sC4sX) +∠(C4sC3sX′)]/2;∆∠(flap)〉)∠(C3sC4sX)-∠(C4sC3sX′), (C4sX and C3sX′ are the bisectors of the FsCsF angles); and tilt; the angle between the (assumed) 3-fold axis of the methyl groups and the CsO bonds where a negative sign moves H13 and H16 nearer the oxygen atoms. Finally, to complete the description of the system a composition parameter is required; this was taken to be the mole fraction of the Cs conformer, χ. Although these parameter designations refer formally only to the Cs conformer, we have noted above that the presence of the presumably small amount of the C2V conformer is included through certain assumptions about how its structure differs from

the Cs. These assumptions are that the experimental differences between the two forms were equal to the theoretical ones obtained by comparisons of the results from the B3LYP/ccpVTZ calculations. For example, the calculations yielded theoretical values for r(C1sC4) and r(C2sC3) for each conformer (symmetrically equivalent in the C2V form) and accordingly for the quantities 〈r(CsC)Cs〉 and r(CsC)C2V. The “experimental” value of r(CsC)C2V was then defined in terms of the Cs form through the formula r(CsC)exp,C2V ) 〈r(CsC)〉exp,Cs + r(C1sC4)theor,C2V - 〈r(CsC)〉theor,Cs. With this device, the structure of the C2V form follows along and varies with the Cs in the course of the refinements. Treatment of the vibrational-amplitude parameters was done in the usual way: those corresponding to similar distances were formed into groups with differences between group members fixed at values taken from the normal coordinate calculations; each group was then refined by a single group parameter. Refinements and Final Model. All terms were included except H · · · H, and M · · · H separated by more than one bond angle; the longer M · · · H distances are of no importance because of their relatively low scattering power and the washing-out effect of methyl group rotation. A large number of refinements characterized by different combinations of parameters were carried out. One of these was a refinement in which all parameters were successfully converged. However, this refinement did not seem an appropriate choice for the final model because the uncertainties associated with some of the values were so large as to make the measurements meaningless. These unreasonable values were all for difference parameters, and were not unexpected: it is well-known that the scattered intensity distribution from interatomic distances of similar value is virtually independent of their splits. Accordingly, some of these difference parameters were held at values taken from the B3LYP/cc-pVTZ calculation and the refinement repeated to give results for our final model of the DMCB system. Table 2 lists experimental and theoretical parameter values of this model. Table 3 is the correlation matrix for the refined parameters, and Table 4 contains the bond distances, bond amplitudes, and the more important angle values for the model; an expanded list is found in the Supporting Information. The intensity- and radial distribution curves are seen in Figures 2 and 3. Results and Discussion Vapor Composition. As Table 2 shows, the experimental and the theoretical results for the isomeric composition of DMCB are in good agreement taking into account the associated uncertainties. Which type is the more accurate is hard to judge because the criteria defining the uncertainties (estimated 2σ) are quite dissimilar; nevertheless, there is no doubt that the Cs form is present in major proportion. The predominance of the

Bond Length Variation in Substituted Cyclobutenes

J. Phys. Chem. A, Vol. 114, No. 16, 2010 5361

TABLE 2: Structure-Defining Parameter Values for the Cs and C2W Forms of 1,2-Dimethoxy-3,3,4,4-tetrafluorocyclobut-1-enea theoretical,b,c re

experimental, rRb

experimental,∠R

theoretical,c ∠e

distances

Cs

C2V

Cs

C2V

angles

Cs

C2V

Cs

C2V

〈r(CsH)〉 r(CdC) 〈r(CsC)〉 ∆r(CsC) 〈r(CsF)〉 ∆r(CsF) 〈r(CsO)〉 ∆r(CsO) 〈r(OsCH3)〉 ∆r(OsCH3) composition χe

1.077 (8) 1.330 (21) 1.493 (8) [-0.003] 1.360 (4) [-0.006] 1.315 (12) [-0.011] 1.350 (13) [-0.002]

1.077 1.329 1.494 0.0 1.363 0.0 1.313 0.0 1.351 0.0

1.088 1.355 1.495 -0.003 1.361 -0.006 1.332 -0.011 1.437 -0.001

1.088 1.353 1.496 0.0 1.364 0.0 1.331 0.0 1.438 0.0

0.98(11)

0.02

0.88(4)f

0.12

〈∠CdCsC〉 ∆∠(CdCsC) 〈∠(CdCsO)〉 ∆∠(CdCsO) 〈∠(FsCsF)〉 ∆∠(FsCsF) 〈∠flapd〉 ∆∠flapd 〈∠(sOsC)〉 ∆∠(CsOsC) HsCsH tilt

94.4 (4) [0.4] 133.5 (12) 4.2 (29) 104.4 (7) [0.6] 133.4 (9) [-0.2] 119.6 (13) [0.3] 105.1 (49) [-3.6]

94.4 0.0 130.0 0.0 104.0 0.0 133.2 0.0 119.7 0.0 106.8 [-3.6]

93.9 0.4 136.2 5.6 106.1 0.6 135.1 -0.2 116.3 0.3 109.9 -3.6

93.8 0.0 132.8 0.0 105.6 0.0 135.7 0.0 116.3 0.0 109.9 -3.7

a

Distances (r) in angstro¨ms, angles (∠) in degrees. Quantities in parentheses are estimated 2σ and include estimates of correlation among data points and possible systematic error. Quantities in square brackets were assumed. b Occasional small disagreements between the experimental and theoretical differences for the Cs and C2V forms are due to rounding errors. c B3LYP/cc-pVTZ. d See text. e Mole fraction. f From ∆G°.

TABLE 3: Correlation Matrix (×100) for Structure-Defining Parameters of 1-Methoxy-2,3,3,4,4-pentafluorocyclobut-1-ene 1 2 3 4 5 6 7 8 9 10 11 12 13

〈r(CsH)〉 r(C1dC2) 〈r(CsC)〉 〈r(CsF)〉 〈r(CsO)〉 r(OsCH3) 〈∠CdCsC〉 〈∠(CdCsO)〉 ∆(CdCsO) 〈∠(FsCsF)〉 〈∠flapb〉 ∠(CsOsC) 〈∠(HsCsH)〉

100 σLSa

1

2

3

4

5

6

7

8

9

10

11

12

13

0.29 0.74 0.28 0.15 0.42 0.45 15.6 43.5 102.4 26.0 30.5 44.8 171.8

100 -26 -23 27 4 15 30 3 10 -21 -27 -9 -5

100 -8 -32 -60 22 -84 -52 18 22 17 13 51

100 -24 43 -78 -40 39 17 42 72 8 14

100 -51 -11 37 -38 26 -48 -52 44 -18

100 -27 33 75 -33 24 37 -44 -25

100 25 -23 4 -9 -45 -32 8

100 24 -11 -26 -48 -16 -43

100 -36 34 31 -35 -11

100 -27 -36 -4 11

100 33 -27 71

100 16 8

100 -20

100

a

Standard deviations from the least-squares fitting. Distances (r) in angstro¨ms; angles (∠) in degrees. FsCsF angle and the C3sC4 bond.

Cs form is mostly due to the higher rotational entropy, but there is also a small contribution from the vibrational entropy difference. Fortunately, due to the structural similarity of the isomers the parameter values for the structure are very little affected by this proportion. Ring Bond Lengths. Our main interest in the DMCB structure concerns the length of the C3sC4 bond compared to those in similar molecules. As is seen in Table 4, the thermal average length (rg) of this bond from our GED experiments is in every case greater than “normal”, i.e., greater than about 1.54 Å, by a few hundredths of an angstro¨m. Unfortunately, a suitable GED measurement of the structure of cyclobutene, which would provide a very useful comparison, is not available: one was done a half century ago19 by a now obsolete method that yielded results of considerably lower precision than is customary today. However, an estimate of rg for C3sC4 can be had from the rs structure for cyclobutene by adding the correction term used to convert rR to rg. The value is listed in Table 4 and fits our characterization of a long bond.20 Table 5 summarizes ring-bond lengths in our comparison molecules obtained by two levels of theory and two types of experiment. Taking into account the estimated uncertainties, it is seen that the GED rg lengths of the C3sC4 bonds in all the substituted cyclobutenes except DMCB are greater than the estimated experimental rg value in cyclobutene itself. As we have cited, this suggests an accounting of the excess length in terms of charge repulsion and orbital rehybridization. However, as Table 5 also shows, the high-level theoretical work mentioned

b

Angle between the bisector of an

earlier, i.e., CCSD(T)/cc-pVTZ, predicts rg values21 smaller than ours for this bond in three of our compounds (hexafluorocyclobutene, dichlorotetrafluorocyclobutene, and octafluorobicyclohexene) and, except for the last, also smaller than in cyclobutene itself (1.5744 Å). Thus, in contrast to our experimental results, theory at this high level suggests that fluorine substitution does not lengthen the C3sC4 bond and hence that charge repulsion has little or no effect on its value. Our B3LYP/cc-pVTZ calculations present a different picture. From this theory the length of the C3sC4 bond is predicted to be greater than that from CC-SD(T)/cc-pVTZ in every case and is in good agreement with experiment except for dichlorotetrafluorocyclobutene where the GED value is still greater by more than the associated uncertainty. After taking into account the uncertainties, the experimental values of the other ring bonds are in satisfactory agreement with both levels of theory for all molecules except for C1dC2 in both dichlorotetrafluorocyclobutene (the MW value is too small) and octafluorobicyclobutene (the GED value is too large). The following observations derive from these comparisons of rg distances. First, according to our criterion, the sp3-sp3 carbon-carbon single bond in all molecules in which a cyclobutene skeleton is present is longer than normal, so that whatever causes the lengthening in cyclobutene (e.g., crossring repulsion) probably plays a similar role in the other molecules. However, the matter of change in the length of this bond upon fluorine substitution is a more difficult question. On the purely theoretical side, data from the coupled-cluster level

5362

J. Phys. Chem. A, Vol. 114, No. 16, 2010

Richardson et al.

TABLE 4: Values of Bond Distances (r/Å), Amplitudes of Vibration (l/Å), and Some Angles (∠/deg) for the Cs and C2W Forms of 1,2-Dimethoxy-3,3,4,4-tetrafluorocyclobut-1-ene

a Quantities in parentheses are 2σ uncertainties and include estimates of the effects of correlation among the data. Amplitudes enclosed by square brackets were refined as a group. Curly brackets indicate symmetry-related distances. b B3LYP/cc-pVTZ. c Uncertainties for rg and ra assumed the same as for rR. d Uncertainties assumed equal to those of the Cs form. e Assumed. See text. f Quality-of-agreement factor. R ) [Σiwi∆i2/Σiwi(siIm,i(obsd))2]1/2 where ∆i ) siIm,i(obsd) - siIm,i(calc).

with the largest bases do not predict a lengthening of the C3sC4 bond upon fluorine substitution, whereas our DFT calculations with similar bases do. The comparisons of MW results from cyclobutene, hexafluorocyclobutene, and dichlorotetrafluorocyclobutene (Table 5) also support the conclusion from coupledcluster theory, but there may be some question about the accuracy of the MW value for the last case. Unfortunately, similar comparisons for GED experiments are complicated by the aforementioned absence of modern GED data for cy-

clobutene. However, although the GED values for the C3sC4 bond length in hexafluorocyclobutene, dichlorotetrafluorocyclobutene, and octafluorobicyclohexene are all greater than both the MW and coupled-cluster values for cyclobutene, they are in good agreement with our results from density-functional theory. Adding to the uncertainty about causes for the lengthening of the C3sC4 bond in the three molecules mentioned above are the results for the molecule of the current study, DMCB.

Bond Length Variation in Substituted Cyclobutenes

J. Phys. Chem. A, Vol. 114, No. 16, 2010 5363

TABLE 5: Comparison of Ring Bond-Distance Values (rg/Å) from Gas-Phase Electron Diffraction (GED), Microwave Spectroscopy (MW), and Theory for Some Molecules with Cyclobutene Skeletons C1dC2 re B3LYP/cc-pVTZa CCSD(T)/cc-pVTZb MWc

1.3342 1.3422

B3LYP/cc-pVTZa CCSD(T)/cc-pVTZb MWd GEDf GEDg

1.3352 1.3374

B3LYP/cc-pVTZa CCSD(T)/cc-pVTZb MWh GEDi

1.3411 1.3430

B3LYP/cc-pVTZa CCSD(T)/cc-pVTZb GEDj B3LYP/cc-pVTZa GEDa

C2sC3 rg

1.3400 1.3480 1.348(4)c

re

r(C3sC4) rg

Cyclobutene 1.5154 1.5154

re

rg

1.5226 1.5226 1.524(3)c

1.5697 1.5654

1.5787 1.5744 1.575(3)c

Hexafluorocyclobutene 1.3406 1.5012 1.3428 1.4962 1.338(6)e 1.324(23) 1.328(24)

1.5077 1.5027 1.484(6)e 1.502(5) 1.503(5)

1.5760 1.5618

1.5836 1.5694 1.560(6)e 1.584(11) 1.585(8)

Dichlorotetrafluorocyclobutene 1.3463 1.5082 1.3482 1.5029 1.316(15) 1.359(9)

1.5148 1.5095 1.494(15) 1.500(6)

1.5722 1.5602

1.5796 1.5676 1.558(15) 1.599(10)

1.3266 1.3186

Octafluorobicyclohexene 1.3323 1.5254 1.3243 1.5286 1.376(14)

1.5318 1.5350 1.530(3)

1.6128 1.5985

1.6229 1.6086 1.627(5)

1.3548

Dimethoxytetrafluorocyclobutene 1.3609 1.4938 1.5022 1.338(21) 1.499(8)

1.5585

1.5679 1.570(11)

This work. rg values made use of rg - re differences given in ref 9. b Reference 9. c Reference 2. d Reference 6. e rg values estimated as indicated in footnote a. f Reference 5. g Reference 8. h Reference 6. i Reference 4. j Reference 5. a

are respectively about 0.06 and 0.11 Å shorter than the sum of the single-bond radii corrected for electronegativity difference.22 These shortenings imply considerably more double-bond character in the C1sO bond than in the C1sF and accordingly higher electron density on C1 (and C2) in DMCB. To the extent that this density tends to “drift” to atoms C3 and C4, any positive charge there (from the polarity of the C3sF and C4sF bonds) would be weaker in DMCB than in hexafluorocyclobutene with consequent reduced repulsion between C3 and C4 and a shorter C3sC4 bond. Acknowledgment. This work was supported by the National Science Foundation under grant CHE-0613298. We are grateful to Lise Hedberg for advice concerning the structure refinements. B.L. gratefully acknowledges the support of the ISMN and ISAC Institutes of CNR at Bologna for allowing access to facilities. Figure 3. Radial distribution curves. The interatomic distances from the final model are indicated by the labeled vertical bars under the experimental curve. This curve was calculated from the experimental intensity curve with unobserved data for s < 2.00 Å-1 from the theoretical curve for the final model. The damping factor was exp(-as2) with a ) 0.0025 Å2. The difference curve is experimental minus theoretical.

Our GED result for the bond in question is insignificantly different from that in cyclobutene itself. Notably, it is in excellent agreement with that from B3LYP/cc-pVTZ theory (a CCSD(T)/cc-pVTZ result is not available), and both values are smaller than the corresponding ones in each of the other molecules. This poses the question of why the replacement of fluorine atoms on bond C1dC2 in hexafluorocyclobutene by methoxy groups shortens the C3sC4 bond, while leaving bonds C1sC4 and C2sC3 essentially unchanged. It seems likely that the cause is connected to the values of the C1sF distance in hexafluorocyclobutene and the C1sO distance in DMCB, which

Supporting Information Available: Table of molecular scattered intensities from each film. Table of selected theoretical parameter values for both forms of DMCB calculated from several levels of theory with different bases. Table of an expanded list of experimental distances and amplitudes, and theoretical distances. This information is available free of charge via the Internet at http://pubs.acs.org. References and Notes (1) Chang, C. H.; Porter, R. F.; Bauer, S. H. J. Mol. Struct. 1971, 7, 89. (2) Bak, B.; Nygaard, L.; Rastrup-Andersen, J.; Sørensen, G. O. J. Mol. Struct. 1969, 3, 369. (3) Csa´sza´r, A. G.; Hedberg, K. J. Phys. Chem. 1990, 94, 3525. (4) Thomassen, H.; Hedberg, K. J. Phys. Chem. 1990, 94, 4847. (5) Richardson, A. D.; Hedberg, K.; Junk, C. P.; Lemal, D. M. J. Phys. Chem. A 2003, 107, 3064. (6) Li-Wei, X.; Klausner, M. E.; Andrews, A. M.; Kuczkowski, R. L. J. Phys. Chem. 1993, 97, 10346.

5364

J. Phys. Chem. A, Vol. 114, No. 16, 2010

(7) Van Wynsberghe, A. W.; Peebles, S. A.; Peebles, R. A.; Kuczkowski, R. L. J. Phys. Chem. A 2000, 104, 8702. (8) We later explored the discrepancy between the MW and GED results for hexafluorocyclobutene by rerefining the GED structure with inclusion of the MW rotational constants as constraints and have been able to fit both the MW B’s and the GED data with a structure essentially the same as our original one that still contains the long C1sC4 bond (ra up slightly to1.583(8) Å from the originally reported 1.581(11) Å from GED alone). See: Hedberg, L.; Hedberg, K. J. Phys. Chem. 1993, 97, 10349. (9) Csa´sza´r, A. G. J. Phys. Chem. A 2004, 108, 2002. (10) Park, J. D.; Sharrah, M. L.; Lacher, J. R. J. Am. Chem. Soc. 1949, 71, 2337. (11) Barr, J. T.; Rapp, K. E.; Pruett, R. L.; Bahner, C. T.; Gibson, J. D.; Lafferty, R. H., Jr. J. Am. Chem. Soc. 1949, 71, 2337. (12) Smart, F. E.; Reddy, G. S. J. Am. Chem. Soc. 1976, 98, 5593. (13) Brown, R. J. C.; Brown, R. F. C. J. Chem. Educ. 2000, 77, 724. (14) Lunelli, B. Inorg. Chim. Acta 2007, 360, 1217. (15) Scherer, O.; Ho¨rlein, G.; Millauer, H. Chem. Ber. 1966, 99, 1966. (16) Data reduction Gundersen, G.; Hedberg, K. J. Chem. Phys. 1969, 51, 2500. Background removel: Hedberg, L. Abstracts of the 5th Austin Symposium on Gas-Phase Molecular Structure; Austin, TX, 1974; p 37. (17) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb, M. A.; Cheeseman, J. R.; Montgomery, J. A., Jr.; Vreven, T.; Kudin, K. N.; Burant, J. C.; Millam, J. M.; Iyengar, S. S.; Tomasi, J.; Barone, V.; Mennucci, B.; Cossi, M.; Scalmani, G.; Rega, N.; Petersson, G. A.; Nakatsuji, H.; Hada, M.; Ehara, M.; Toyota, K.; Fukuda, R.; Hasegawa, J.; Ishida, M.; Nakajima, T.; Honda, Y.; Kitao, O.; Nakai, H.; Klene, M.; Li, X.; Knox, J. E.; Hratchian, H. P.; Cross, J. B. Adamo, C.; Jaramillo, J.; Gomperts, R.; Stratmann, R. E.; Yazyev, O.; Austin, A. J.; Cammi, R.;

Richardson et al. Pomelli, C.; Ochterski, J. W.; Ayala, P. Y.; Morokuma, K.; Voth, G. A.; Salvador, P.; Dannenberg, J. J.; Zakrzewski, V. G.; Dapprich, S.; Daniels, A. D.; Strain, M. C.; Farkas, O.; Malick, D. K.; Rabuck, A. D.; Raghavachari, K.; Foresman, J. B.; Ortiz, Cui, Q.; Baboul, A. G.; Clifford, S.; Cioslowski, J.; Stefanov, B. B.; Liu, G.; Liashenko, A.; Piskorz, P.; Komaromi, I.; Martin, R. L.; Fox, D. J.; Keith, T.; Al-Laham, M. A.; Peng, C. Y.; Nanayakkara, A.; Challacombe, M.; Gill, P. M. W. Johnson, B.; Chen, W.; Wong, M. W.; Gonzalez, C.; Pople, J. A. Gaussian 03, Revision B.05; Gaussian, Inc.: Pittsburgh, PA, 2003. (18) (a) Hedberg, L.; Mills, I. M. J. Mol. Spectrosc. 2000, 203, 82. (b) Hedberg, L.; Mills, I. M. Ibid. 1993, 160, 117. (19) Goldish, E.; Hedberg, K.; Schomaker, V. J. Am. Chem. Soc. 1956, 78, 2714. The ra values for C1dC2 and the single bond average (CsC)ave from this study of cyclobutene were estimated at 1.325(46) Å and 1.571(10) Å, but values for the individual CsC bonds could not be obtained. (20) An rs value for a bond is usually much closer to the re one than any of the electron-diffraction derived types. (See: Robiette, A. G. In Molecular Structure by Diffraction Methods, Vol 1; Specialist Periodical Reports; The Chemical Society: London, 1973; Chapter 4, p 174. ) Our estimate of rg was obtained by adding the rg- re correction to rs. (21) The re to rg conversion factors from CC-SD(T)/cc-pVTZ theory are those given in ref 8. Since such factors are not very sensitive to changes in higher level theory, we judged them to be satisfactory for our B3LYP/ ccc-pVTZ work as well. (22) Pauling, L. The Nature of the Chemical Bond, 3rd ed.; Cornell University Press: Ithaca, New York, 1960; Chapter 7.

JP911185Z