Gas-Phase Structures of Ketene and Acetic Acid from Acetic

Gas-Phase Structures of Ketene and Acetic Acid from Acetic...

0 downloads 98 Views 956KB Size


Gas-Phase Structures of Ketene and Acetic Acid from Acetic Anhydride Using Very-High-Temperature Gas Electron Diffraction Sandra J. Atkinson,† Robert Noble-Eddy,‡ and Sarah L. Masters*,† †

Department of Chemistry, University of Canterbury, Private Bag 4100, Christchurch 8140, New Zealand School of Chemistry, University of Edinburgh, David Brewster Road, Edinburgh, United Kingdom, EH9 3FJ

S Supporting Information *

ABSTRACT: The gas-phase molecular structure of ketene has been determined using samples generated by the pyrolysis of acetic anhydride (giving acetic acid and ketene), using one permutation of the very-high-temperature (VHT) inlet nozzle system designed and constructed for the gas electron diffraction (GED) apparatus based at the University of Canterbury. The gas-phase structures of acetic anhydride, acetic acid, and ketene are presented and compared to previous electron diffraction and microwave spectroscopy data to show improvements in data extraction and manipulation with current methods. Acetic anhydride was modeled with two conformers, rather than a complex dynamic model as in the previous study, to allow for inclusion of multiple pyrolysis products. The redetermined gas-phase structure of acetic anhydride (obtained using the structure analysis restrained by ab initio calculations for electron diffraction method) was compared to that from the original study, providing an improvement on the description of the low vibrational torsions compared to the dynamic model. Parameters for ketene and acetic acid (both generated by the pyrolysis of acetic anhydride) were also refined with higher accuracy than previously reported in GED studies, with structural parameter comparisons being made to prior experimental and theoretical studies.

INTRODUCTION Gas electron diffraction (GED) is a technique that can be used to study the structure of molecules in the gas phase. Work in Canterbury has focused on the millisecond-to-second time scale with molecules generated using pyrolysis. In these experiments, molecules are generated using in situ pyrolysis techniques utilizing a very hot inlet nozzle. Previous work with this very hot inlet nozzle includes the study of a low-volatility species at high temperatures.1 To enable the generation of short-lived species and their study by GED, it was necessary to construct a new very-hightemperature (VHT) nozzle. Testing of the nozzle with antimony(III) oxide showed the capability of the VHT nozzle for studying stable low-volatility species,1 but it did not test the capability and setup of the combined VHT−GED system to generate and study short-lived species. Ketene (CH2CO) was chosen as the test case because it is an unstable molecule, dimerizing rapidly to form diketene (OCCH2OCCH2) at room temperature.2 Ketene itself has been studied before by GED by carefully storing a cold sample,3,4 and it can also be generated from a variety of precursors. The successful generation and study of ketene by the described VHT−GED method will allow the study of other substituted ketenes, including asymmetrically substituted ketenes. Several ketene derivatives have already been studied by GED, including diketene and methyldiketene;5 bis(trimethylgermyl)ketene;6 bis(trifluoromethylthio)ketene;7 and perhaps most interestingly, dichloroketene,8 which was © 2016 American Chemical Society

generated from the precursor trichloroacetyl chloride using a gas−solid reaction. Acetic anhydride, [(CH3CO)2O], was chosen as a precursor for the test system because pyrolysis generates acetic acid and ethenone (CH2CO), the simplest example of a ketene, in good yield by flash vacuum pyrolysis (FVP)9 and the pyrolysis of acetic anhydride has not been studied by GED. Previous theoretical and kinetic studies suggest the decomposition mechanism to be an intramolecular rearrangement,10−12 with the most obvious route involving one methyl proton moving to the central O and subsequent breaking of the C−O bond (Scheme 1).10 Acetic anhydride has been previously studied by GED and IR spectroscopy.13 The authors employed a dynamic model for the GED refinement because of the relatively low barrier of rotation around the central C−O−CO dihedral angles. In this work, a dynamic model is too complex to incorporate into Scheme 1. Mechanism for the Pyrolysis of Acetic Anhydride to Give Acetic Acid and Ketene

Received: January 21, 2016 Revised: February 26, 2016 Published: February 26, 2016 2041

DOI: 10.1021/acs.jpca.6b00704 J. Phys. Chem. A 2016, 120, 2041−2048


The Journal of Physical Chemistry A

Figure 1. VHT nozzle inlet system used in this work showing (a) the general setup and (b) the cross section of the path of the electron beam through the end of the nozzle. The outer inlet system cooling is not shown in panel a.

Previous work with the VHT nozzle has involved redetermination of the gas-phase structure of antimony(III) oxide.1 The configuration of the VHT nozzle allowed for the study of relatively involatile samples over a wider range of temperatures than are normally accessible using the Canterbury GED apparatus. The setup used in the study of antimony(III) oxide is one of the four potential configurations for the VHT nozzle system. Because this study is looking at the generation of acetic acid and ketene by pyrolysis of acetic anhydride, a different setup of the nozzle is required and is presented in detail here. Specific details regarding the working of the VHT nozzle, such as the types of materials used, are covered in more detail elsewhere1 and are only briefly discussed here. As shown in Figure 1a, the VHT nozzle setup consists of an external sample inlet which is used to allow the sample to pass into the VHT oven in close proximity to the diffraction zone. With this setup design, changing the sample at the external gas inlet is simple, so benzene calibration can be obtained without removing the VHT nozzle from under vacuum. Using an external gas inlet system is applicable only if the sample of interest has sufficient vapor pressure. Acetic anhydride has a vapor pressure of 4 Torr at 293 K,17 which is sufficient for this study by GED with this nozzle setup. For the pyrolysis experiments in this paper, the sample of acetic anhydride was passed into the VHT oven and heated, using a tungsten wire element inlaid into the molybdenum oven-surround, by thermocouples connected to an external controller.1 The steel case surrounding the nozzle system was cooled with liquid nitrogen.1 Electron diffraction occurred perpendicular to the gaseous sample beam through an orifice at the end of the enclosed high-temperature cell (Figure 1).1 The asymmetric orifice cut-out allows for detection of diffracted electrons at wide angles because the cut-out is larger on the detection side (Figure 1b).1 No fouling of the photographic film from the glow of the high-temperature oven was observed. Computational Methods. Geometry and frequency calculations were performed using the resources of the EPSRC National Service for Computational Chemistry Software18 and the Gaussian09 program.19 All MP220 methods were frozen core [MP2(fc)]. Geometry optimizations were performed using C2v symmetry for ketene and Cs symmetry for acetic acid using the MP2 method, whereas the C1 and C2 conformers for acetic anhydride were optimized using the M062X21 method. Analytic second derivatives of the energy with respect to nuclear coordinates were calculated at the M06-2X/6-311+ +G** level for acetic anhydride and at MP2/6-311++G** for both ketene and acetic acid. These determine the nature of the potential energy surface and were used in the program

a refinement that also includes decomposition products. A simpler model was generated for acetic anhydride, and to test its validity, the data obtained in the previous study were also reanalyzed using this simpler model. This paper presents the redetermination of the gas-phase structure of acetic anhydride and the pyrolysis of acetic anhydride (to give acetic acid and ketene) using one permutation of the new Canterbury VHT nozzle system. The original GED data of acetic anhydride has been re-evaluated in this work using a two-conformer model by implementing the structure analysis restrained by ab initio calculations for electron diffraction (SARACEN) method14−16 and is compared to the data refined using a dynamic model in the previous work. The SARACEN method is shown to give an improvement on the description of the low vibrational torsions compared to the dynamic model.13 The GED structure of ketene obtained from the pyrolysis of acetic anhydride is presented for the first time as, even though the structure of ketene is well-known, the gasphase structure of ketene generated from acetic anhydride has not been reported. The parameters of ketene are compared to those given by previous electron diffraction studies.4

EXPERIMENTAL SECTION Synthesis. Acetic anhydride (>99%) was obtained from Sigma-Aldrich and used without further purification in the GED experiments. Acetic acid and ketene were generated by pyrolysis of acetic anhydride using the very-high-temperature nozzle system as described below. The Very-High-Temperature Nozzle System. To enable the generation of short-lived species for gas-phase electron diffraction experiments, a new very-high-temperature nozzle system has been constructed to work with the Canterbury GED apparatus.1 The nozzle system combines the techniques of flash vacuum pyrolysis with GED. The construction was undertaken by the group of Prof. Georgiy Girichev, who have substantial experience with VHT−GED experiments. A FVP system usually consists of an oven to vaporize the precursor, a second hotter pyrolysis zone, and a trap to collect the product.1 To incorporate this type of nozzle system with the Canterbury GED apparatus, the VHT nozzle system allows the generated product to travel into the GED apparatus, meaning the main chamber of the GED apparatus replaces the traditional cold trap for the FVP-generated product. This also means that the VHT nozzle system is fully self-contained within the GED apparatus under standard experimental vacuum conditions.1 Advantages of a self-contained system include allowing for the use of a much smaller pyrolysis oven within the nozzle than would be used with external FVP equipment, thus allowing for better control in maintaining a high temperature. 2042

DOI: 10.1021/acs.jpca.6b00704 J. Phys. Chem. A 2016, 120, 2041−2048


The Journal of Physical Chemistry A

Figure 2. Lowest-energy structures of the (sp, sp) and (sp, ac) conformers of acetic anhydride with atom numbering.

SHRINK22,23 to provide estimates of the amplitudes of vibration (u) and perpendicular distance corrections (k) for use in the GED refinement. Because harmonic force constants were calculated, the refinement type is denoted rh1.24 Gas Electron Diffraction. The GED structure of acetic anhydride was redetermined using the electron diffraction data as provided by Wu et al.13 in their Supporting Information, which used an average experimental temperature of 338 K. Data were collected for the pyrolysis of acetic anhydride to acetic acid and ketene using the GED apparatus now based at the University of Canterbury1 and the VHT nozzle setup as described above. An accelerating voltage of around 40 keV was used (electron wavelength ca. 6.0 pm) with a maintained pyrolysis temperature of 823 K. Scattering intensities were recorded at a nozzle-to-plate distance of 212.2 mm on Kodak Electron Image films. An Epson Expression 1680 Pro flatbed scanner was used to convert the electron-scattering intensities to mean optical densities as a function of the scattering variable, s, using an established program.25 The data-reduction and leastsquares-refinement processes were carried out using the [email protected] program26 (version 2.3) employing the scattering factors of Ross et al.27 The scattering intensities of the GED data of acetic anhydride from Wu et al.13 were scaled up by 50 for use with the in-house refinement program.26 This scaling is arbitrary and does not affect the results. The weighting points for the offdiagonal weight matrices, correlation parameters, and scale factors for the nozzle-to-camera distance are given in Tables S1 and S2, together with the electron wavelengths, which were determined from the scattering patterns of benzene vapor.

Figure 3. Lowest-energy structures of ketene (left) and acetic acid (right) with atom numbering.

Analytic second derivatives of the energy with respect to the nuclear coordinates were calculated at the M06-2X level with the 6-311++G** basis set for acetic anhydride and at the MP2/ 6-311++G** level for the pyrolysis products (ketene and acetic acid), giving force fields which were then used to provide estimates of the amplitudes of vibration (u) for use in the GED refinements. These force fields were also used to calculate the vibrational frequencies, which provided information about the nature of the stationary points. Gas Electron Diffraction. GED refinements were carried out for acetic anhydride, based on the ab initio calculations described above (optimized at the M06-2X/6-311++G** level). The gas-phase structure for acetic anhydride was redetermined by reprocessing the raw GED data from a previous experiment,13 using the SARACEN method.14−16 The original investigation used a dynamic model to describe the two nonplanar minima of acetic anhydride (sp, ac) and (sp, sp) implementing eight pseudoconformers; five were clustered around the (sp, sp) conformer, and three were around the (sp, ac) conformer. In the present study, a two-conformer model was employed rather than a dynamic model because a model with multiple species was to be employed for the GED study of the acetic anhydride pyrolysis products. The SARACEN refinement provides an improvement on the refined values of dihedral angles and provides a suitable proof-of-concept for applying the method to the pyrolysis experiment. If the SARACEN method was not used it would be too complicated to include acetic anhydride (described by a dynamic model) in a model which also includes multiple pyrolysis products. The structure was defined in terms of 32 independent geometric parameters and 20 dependent parameters, as described in the Supporting Information. Theoretical Cartesian force fields were generated at the M062X/6-311++G** level and converted to a force field described by a set of symmetry coordinates. Root-mean-square (RMS) amplitudes of vibration were obtained from the SHRINK program.22,23 All independent geometric parameters were refined using a least-squares method, and restraints were applied using the SARACEN method14−16 (Table 1). The

RESULTS AND DISCUSSION Computational Studies. A previous ab initio study of acetic anhydride13 identified two nonplanar minima, labeled (sp, sp) and (sp, ac), related to each other by large amplitude motions. Conformer names are based on the OC−O−C dihedral angles. The abbreviation sp refers to synperiplanar (a dihedral angle of 0 ± 30°); ac refers to anticlinal (a dihedral angle of −90 to −150°). The lowest-energy structures of the acetic anhydride conformers [(sp, sp) and (sp, ac)] and atom numbering are given in Figure 2. The acetic anhydride conformer geometries were optimized at the HF level with the 6-31G* basis set,28−30 and at the M062X level with 6-31G*, 6-311G*,31,32 6-311+G*, and 6-311+ +G** basis sets. The results of calculations performed at the M06-2X level are presented in the Supporting Information (Table S3). Ab initio calculations were also undertaken in a similar manner for ketene and acetic acid from HF/3-21G* to MP2/6311++G** with the MP2 results presented in Table S4. The atom numbering and structures of acetic acid and ketene are given in Figure 3. 2043

DOI: 10.1021/acs.jpca.6b00704 J. Phys. Chem. A 2016, 120, 2041−2048


The Journal of Physical Chemistry A

Table 1. Refined (rh1) and Calculated (re) Geometric Parameters for Acetic Anhydride, [CH3C(O)O]2, from the SARACEN GED Refinementa parameters


p1 p2 p3 p4 p5 p6 p7 p8 p9 p10 p11 p12 p13 p14 p15 p16 p17 p18 p19 p20 p21 p22 p23 p24 p25 p26 p27 p28 p29 p30 p31 p32

rC−C (sp, sp) rC−C diff 1 rC−C diff 2 rCO (sp, sp) rCO diff 1 rCO diff 2 rC−O (sp, sp) rC−O diff 1 rC−O diff 2 rC−H av ∠H(11)−C−C (sp, sp) ∠H−C−C diff 1 ∠H−C−C diff 2 ∠H(25)−C−H(24) (sp, ac) ∠H(20)−C−H(19) (sp, ac) ∠C−CO (sp, sp) ∠C−CO diff 1 ∠C−CO diff 2 ∠C−C−O (sp, sp) ∠C−C−O diff 1 ∠C−C−O diff 2 ∠C−O−C (sp, sp) ∠C−O−C (sp, ac) ϕOC−C−H(11) (sp, sp) ϕOC−C−H(25) (sp, ac) ϕOC−C−H(20) (sp, ac) ϕO(9)C(1)−C(10)−O(2) (sp, sp) ϕOC−C−O diff 1 ϕOC−C−O diff 2 ϕC(10)−C−O−C (sp, sp) ϕC(23)−C−O−C (sp, ac) ϕC(17)−C−O−C (sp, ac)

dp1 dp2 dp3 dp4 dp5 dp6 dp7 dp8 dp9 dp10 dp11 dp12 dp13 dp14 dp15 dp16 dp17 dp18 dp19 dp20

rC(14)−C(23) (sp, ac) rC(16)−C(17) (sp, ac) rC(14)O(22) (sp, ac) rC(16)O(21) (sp, ac) rC(14)−O(15) (sp, ac) rC(16)−O(15) (sp, ac) ∠H(25)−C−C (sp, ac) ∠H(20)−C−C (sp, ac) ∠C−CO(22) (sp, ac) ∠C−CO(21) (sp, ac) ∠O−CO (sp, sp) ∠O−CO(21) (sp, ac) ∠O−CO(22) (sp, ac) ∠C(17)−C−O (sp, ac) ∠C(23)−C−O (sp, ac) ϕO(21)C(16)−C(17)−O(15) (sp, ac) ϕO(22)C(14)−C(23)−O(15) (sp, ac) ϕO(8/9)C−O−C (sp, sp) ϕO(21)C−O−C (sp, ac) ϕO(22)C−O−C (sp, ac)

GED (rh1)

M06-2X/6-311++G** (re)


149.8 0.1 0.3 118.7 0.8 0.1 138.7 2.2 1.3 108.9 109.5 1.8 0.1 110.5 109.9 127.2 1.6 0.9 110.0 7.2 0.1 120.3 122.5 5.7 −1.0 21.3 177.6 1.9 1.0 −153.2 172.7 −38.6

− 0.1(1) 0.3(1) − 0.8(1) 0.1(1) − 2.2(1) 1.3(1) 108.9(6) 109.5(1) 1.8(1) 0.1(1) 110.5(1) 109.9(1) − 1.6(1) 0.9(1) − 7.2(2) 0.1(1) 120.3(1) 122.5(2) 5.7(20) −1.0(10) 21.3(20) 177.6(1) 1.9(11) 1.0(1) −153.2(20) 172.7(22) −38.6(20)

149.9 150.1 119.5 118.6 136.5 140.0 109.4 107.7 126.3 125.6 122.7 117.2 123.5 117.2 110.1 175.7 −178.6 29.1 145.3 −8.7

− − − − − − − − − − − −

Independent 148.5(3) 0.1(1) 0.3(1) 118.3(1) 0.8(1) 0.1(1) 138.7(2) 2.2(1) 1.3(1) 108.4(4) 109.5(1) 1.8(1) 0.1(1) 110.5(1) 109.9(1) 126.3(5) 1.6(1) 0.9(1) 111.2(5) 7.2(2) 0.1(1) 120.3(1) 122.4(2) 5.7(20) −1.1(10) 21.7(19) 177.6(1) 1.4(11) 1.0(1) −155.7(15) 173.5(20) −41.5(15) Dependent 148.6(3) 148.8(3) 119.0(1) 118.2(1) 136.5(2) 140.0(2) 109.4(1) 107.7(1) 125.4(5) 124.7(5) 122.5(2) 116.8(2) 123.3(2) 118.4(5) 111.3(5) 176.2(11) −178.6(1) 26.6(16) 142.0(18) −7.8(20)

− − − − − − −

Distances (rh1) are in picometers, and bond angles (∠) and dihedral angles (ϕ) are in degrees.

refinement can be assessed numerically using the final R factor,33 which was RG = 0.073 (RD = 0.067), and visually using the radial-distribution and difference curves (Figure 4) and the

restraints were based on values calculated at the M06-2X/6311++G** level. In addition, the corresponding amplitudes of vibration were refined (see Table S5). The success of the 2044

DOI: 10.1021/acs.jpca.6b00704 J. Phys. Chem. A 2016, 120, 2041−2048


The Journal of Physical Chemistry A

Table 2. Refined and Calculated Parameters for GED Refinement of Pyrolysis of Acetic Anhydride, [CH3C( O)O]2, giving Ketene and Acetic Acida parametersb p1 p2 p3 p4 p5 p6 p7 p8 p9 p10 p11 p12 p13 p14 p15

Figure 4. Experimental and difference (experimental − theoretical) radial distribution curves for acetic anhydride.

molecular-scattering intensity curves (Figure S1). The refined conformer weighting for (sp, sp) of 40.7(±9.4)% was comparable to the refined weighting in the original study (37 ± 15%).13 Calculations at the M06-2X level21 showed that MP2 incorrectly predicted the (sp, sp) conformer as being lower in energy. In the previous GED study13 56.8% was obtained as the predicted conformer weighting from the sum of the (sp, sp) pseudoconformers, compared to a weighting of 28.8% as predicted by M06-2X/6-311++G**. A trial refinement was conducted at the MP2/6-311++G** level and was used as a starting point for the M06-2X refinement. The energies and calculated conformer weightings at the MP2 and M06-2X level for a range of basis sets are presented in Table S6. The leastsquares correlation matrix is given in Table S7, and coordinates for the final GED structure and for the calculated structure (M06-2X/6-311++G**) are in Tables S8 and S9, respectively. GED refinement of the pyrolysis products of acetic anhydride was carried out based on ab initio calculations at the MP2/6-311++G** level with C2v and Cs symmetry for ketene and acetic acid, respectively. From the mechanism for the pyrolysis of acetic anhydride,9 it is expected that ketene and acetic acid (Figure 3) are generated with 1:1 stoichiometry, so this was fixed in the refinement. Rather than setting the amount of ketene present to be 50%, the ratio of ketene to acetic acid was fixed at 67:33% to reflect the ratio of enantiomers present in the gas-phase vapor. A trial refinement of the data showed that the contribution of acetic anhydride was c ∠O(2)−C(1)O(9) ∠O(2)−C(3)O(8) ∠O(2)−C(1)−C(10) ∠O(2)−C(3)−C(4) ∠C(1)−O(2)−C(3) ϕO(9)C(1)−O(2)−C(3) ϕO(8)C(3)−O(2)−C(1)

148.5(3) 148.5(3) 118.3(1) 118.3(1) 138.7(2) 138.7(2) 108.4(4) 122.5(2) 122.5(2) 111.2(5) 111.2(5) 120.3(1) 26.6(16) 26.6(16)

C(14)−C(23) C(16)−C(17) C(14)O(22) C(16)O(21) C(14)−O(15) C(16)−O(15) ⟨C−H⟩c ∠O(15)−C(14)O(22) ∠O(15)−C(16)O(21) ∠O(15)−C(14)−C(23) ∠O(15)−C(16)−C(17) ∠C(14)−O(15)−C(16) ϕO(21)C(16)−O(15)−C(14) ϕO(22)C(14)−O(15)−C(16)

148.6(3) 148.8(3) 119.0(1) 118.2(1) 136.5(2) 140.0(2) 108.4(4) 123.3(2) 116.8(2) 111.3(5) 118.4(5) 122.4(2) 142.0(18) −7.8(20)

148.9(2)b − 118.2(3) 118.2(3) 137.0(15) 137.0(15) 109.9(4) 124.8(20) 124.8(20) 114.6(23) 114.6(23) 116.5(20) 30.9(67) 30.9(67)

149.8 149.8 118.7 118.7 138.7 138.7 108.9 122.7 122.7 110.0 110.0 120.3 29.1 29.1

148.9(2)b − 119.4(3) 118.2(3) 137.0(13) 140.6(6) 109.9(4) 117.1(10) 124.2(18) 110.9(17) 111.1(22) 121.0(15) 122.0(39) −27.4(53)

149.9 150.1 119.5 118.6 136.5 140.0 108.9 117.2 123.5 110.1 117.2 122.5 145.3 −8.7

(sp, ac)


The distances are given in picometers; the bond angles (∠) and dihedral angles (ϕ) are in degrees. bThe C−C distances were averaged over both conformers. cAn average C−H distance from both conformers was used.

level for the C−C parameter, with the value in the original study, 148.9(2) pm, being comparable to the C−C distances that were individually described for the (sp, sp) and (sp, ac) conformers in this study [148.5(3) pm for (sp, sp), and 148.6(3) and 148.8(3) pm for (sp, ac)]. The C−O bond lengths for the (sp, sp) conformer were underestimated in the original study compared to the ab initio results [137.0(15) pm compared to 138.7 pm], with the C−O distance from this study [138.7(2) pm] in close agreement with the M06-2X/6-311+ +G** value. A difference was also observed between the ab initio and GED values for the (sp, ac) conformer, with values for C(14)−O(15) and C(16)−O(15) from the original study being higher and lower than the corresponding ab initio values [137.0(13) pm compared to 136.5 pm, and 140.6(6)pm compared to 140.0 pm]. As before, the C−O values obtained for (sp, ac) in this study [136.5(2) pm and 140.0(2) pm] were in close agreement with the ab initio values. The ∠O−C−C parameters from both studies agreed reasonably well with the ab initio values, but this was not the case for ∠OC−O in the original study, which were wider for the (sp, sp) conformer [124.8(10)°] and overestimated for the (sp, ac) conformer [117.1(10)° and 124.2(18)°] compared to the results of this study [122.5(2)° for (sp, sp) and 116.8(2)° and 123.3(2)° for (sp, ac), respectively]. Similarly ∠C−O−C also differed between the two studies and was underestimated for (sp, sp) and (sp, ac) [116.5(20)° and 121.0(15)° compared to 120.3(1)° and 122.4.(2)° in this study]. The differences

between these angles and the ab initio results can be attributed to the conformational flexibility of acetic anhydride, specifically due to ϕOC−O−C, which significantly differs between the conformers. Because the SARACEN method14−16 was not used in the original study, these parameters were not previously refined with high accuracy or precision. For the (sp, sp) conformer, ϕOC−O−C agreed reasonably well with the ab initio calculations, but for ϕO(22)C(14)−O(15)−C(16) and ϕO(21)C(16)−O(15)−C(14) in (sp, ac), the previous GED study overestimated and underestimated the magnitude of the dihedral angles [−27.4(3)° and 122.0(39)°, compared to −7.8(20)° and 142.0(18)° in this study]. ϕOC−O−C was shown to vary depending on the ab initio calculation used, as indicated in Table S3. This suggested that these parameters would be poorly refined by GED data alone; therefore, in this study a restraint was applied that encompassed this range of values. The refined ϕOC−O−C from the previous GED study13 falls outside this range. The rest of the parameter comparisons are given in Table 3, and the numbering scheme is given in Figure 3. The fact that the parameters in the original study did not describe the flexible dihedral angles overly well was reflected in the refinement of the corresponding large amplitudes of vibration. This resulted in the refined amplitudes for the O··· O, O···C, and C···C interatomic distances differing greatly from the calculated values. This is more apparent for the (sp, ac) conformer, such as for O(21)···O(22), where the calculated 2046

DOI: 10.1021/acs.jpca.6b00704 J. Phys. Chem. A 2016, 120, 2041−2048


The Journal of Physical Chemistry A

parameters which normally would not be well refined from GED data alone. In this case, using the SARACEN method was particularly important for refining parameters defining the flexible dihedral angles of acetic anhydride and the parameters involving hydrogen for ketene and acetic acid. Overall, the refined parameter values from this study are in close agreement with the predicted ab initio values, and this work successfully demonstrates the advantages of using the SARACEN method in such refinements. Overall, this work highlights a permutation of the University of Canterbury VHT nozzle and demonstrates its capability to study pyrolysis products in situ by GED. The pyrolysis of acetic anhydride (giving acetic acid and ketene) was chosen as a test system and has been successfully studied by the Canterbury VHT−GED setup. The original GED data of acetic anhydride has also been reanalyzed using the SARACEN method, providing improvements on the original refinement.

amplitudes of vibration were 7.8 pm (MP2/6-31G*) compared to the refined value of 14.2(24) pm.13 In this current study, the same amplitude of vibration was calculated to be 26.3 pm (M06-2X/6-311++G**) and refined to 30.0(18) pm (GED). For these flexible dihedral angles, and their corresponding amplitudes of vibration, applying restraints with the SARACEN method14−16 was of great benefit to the success of the refinement. A full comparison of these amplitudes against the theoretically calculated values is given in Table S13. Ketene and Acetic Acid. The structures of ketene and acetic acid were compared to previous GED4,36 and microwave spectroscopy (MW)37,38 studies (Table 4). The theoretical ab initio values from the MP2/6-311++G** calculations are included alongside for comparison. Table 4. Selected Structural Parameters for the LowestEnergy Structures of Ketene and Acetic Acida parameters

GED (rh1)

rCO rCC rC−H ∠H−C−H

115.4(3) 132.3(2) 108.5(1) 121.3(10)

rCO rC−O rC−C rO−H rC−H av ∠C−CO ∠C−C−O ∠C−O−H

119.6(4) 136.5(6) 150.0(3) 96.9(3) 108.6(8) 126.4(6) 110.4(7) 105.7(27)

previous GED studies4,34 Ketene 116.0(20) 130.0(20) 107.0(20)c 117.5(125)c Acetic Acid 121.4(3) 136.4(3) 152.0(5) 97.0c 110.2(10) 126.6(6) 110.6(6) 107.0c

previous MW studies35,36

MP2/6311++G** (re)

116.1b 131.4b 108.3b 122.4b

116.8 132.2 108.5 121.8

120.9(6) 135.7(5) 149.4(10) 97.0(3) 110.2(12) 126.2(7) 112.0(6) 105.9(5)

121.0 135.9 150.4 96.8 109.1 126.4 111.0 105.8


S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.jpca.6b00704. Additional tables and figures as described in the text (PDF)


Corresponding Author

*E-mail: [email protected] Notes

The authors declare no competing financial interest.

ACKNOWLEDGMENTS S.J.A. thanks the Department of Chemistry, University of Canterbury for the award of a doctoral scholarship and the Betty Wignall scholarship. The authors acknowledge the use of the EPSRC UK National Service for Computational Chemistry Software (NSCCS) at Imperial College London in carrying out this work.


Bond lengths are given in picometers, and bond angles (∠) in degrees. bNo estimated standard deviations were quoted. cThis parameter was assumed.

The parameters from the MW study closely agree with the MP2/6-311++G** ab initio results. The GED results from this study and the previous work mostly agree with the MW and ab initio results as well, although some parameters differ as described below. For ketene and acetic acid, the CO bond length was shorter in this study [115.4(3) and 119.6(4) pm] when compared to the previous GED studies [116.0(20) and 121.4(3) pm respectively]. However, these distances occur in a region of considerable peak overlap in the radial distribution curve (Figure 5), when considering that the parameters of ketene and acetic acid being compared were obtained from independent GED studies. For ketene, ∠H−C−H from this study [121.3(10)°] was closer to the MW and ab initio values [122.4° and 121.8°] than the assumed parameter in the original GED study [117.5(125)°]. For acetic acid, ∠C−O−H was assumed in the previous GED study with a value of 107.0° and hence differed from the refined value of 105.7(27)° in this study. It is important to note that the previous studies did not apply restraints to the parameters involving hydrogen, choosing instead to assume the parameter values in the refinement. This GED study, for both the acetic anhydride redetermination and pyrolysis study to form ketene and acetic acid, uses the SARACEN method14−16 to apply suitable restraints to these


(1) Masters, S. L.; Girichev, G. V.; Shylkov, S. A. The ReDetermination of the Molecular Structure of Antimony(III) Oxide Using Very-High-Temperature Gas Electron Diffraction (VHT-GED). Dalton Trans. 2013, 42, 3581−3586. (2) Chick, F.; Wilsmore, N. T. M. LXXXIX. - Acetylketen: A Polymeride of Keten. J. Chem. Soc., Trans. 1908, 93, 946−950. (3) Beach, J. Y.; Stevenson, D. P. The Electron Diffraction Investigation of the Molecular Structures of Ketene and Thiophosphoryl Chloride. J. Chem. Phys. 1938, 6, 75−80. (4) Broun, T. T.; Livingston, R. L. An Electron Diffraction Investigation of the Molecular Structures of Ketene, Carbonyl Fluoride and Tetrafluoroethylene. J. Am. Chem. Soc. 1952, 74, 6084−6091. (5) Bregman, J.; Bauer, S. H. An Electron Diffraction Study of Ketene Dimer, Methylketene Dimer and β-propiolactone. J. Am. Chem. Soc. 1955, 77, 1955−1965. (6) Rozsondai, B.; Hargittai, I. Electron Diffraction Study of the Molecular Structure of Bis(trimethylgermyl)ketene. J. Mol. Struct. 1973, 17, 53−64. (7) Korn, M.; Mack, H.-G.; Praas, H.-W.; Della Védova, C. O.; Oberhammer, H. Structure and Conformation of Bis(trifluoromethylthio)ketene, (CF3S)2CCO. J. Mol. Struct. 1995, 352−353, 145−151. (8) Rozsondai, B.; Tremmel, J.; Hargittai, I.; Khabashesku, V. N.; Kagramanov, N. D.; Nefedov, O. M. Molecular Structures of Unstable


DOI: 10.1021/acs.jpca.6b00704 J. Phys. Chem. A 2016, 120, 2041−2048


The Journal of Physical Chemistry A Dichloroketene and its Precursor, Trichloroacetyl Chloride, From Electron Diffraction. J. Am. Chem. Soc. 1989, 111, 2845−2849. (9) Fisher, G. J.; Maclean, A. F.; Schnizer, A. W. Apparatus for the Preparation of Ketene by the Pyrolysis of Acetic Anhydride. J. Org. Chem. 1953, 18, 1055−1057. (10) Blake, P. G.; Davies, H. H. Reactions of Keten. Part I. Kinetics of the Gas-phase Reaction With Acetic Acid. J. Chem. Soc. B 1971, 1727−1728. (11) Lee, I.; Cha, O. J.; Lee, B. S. Theoretical Studies on the Gasphase Pyrolysis of Esters. J. Phys. Chem. 1990, 94, 3926−3930. (12) Blake, P. G.; Speis, A. Kinetics of Thermal Decomposition of Acetic Anhydride in Flow and Static Systems. J. Chem. Soc. B 1971, 1877−1878. (13) Wu, G.; Van Alsenoy, C.; Geise, H. J.; Sluyts, E.; Van Der Veken, B. J.; Shishkov, I. F.; Khristenko, L. V. Acetic Anhydride in the Gas Phase, Studied by Electron Diffraction and Infrared Spectroscopy, Supplemented with ab Initio Calculations of Geometries and Force Fields. J. Phys. Chem. A 2000, 104, 1576−1587. (14) Blake, A. J.; Brain, P. T.; McNab, H.; Miller, J.; Morrison, C. A.; Parsons, S.; Rankin, D. W. H.; Robertson, H. E.; Smart, B. A. Structure Analysis Restrained by ab Initio Calculations: The Molecular Structure of 2,5-Dichloropyrimidine in Gaseous and Crystalline Phases. J. Phys. Chem. 1996, 100, 12280−12287. (15) Brain, P. T.; Morrison, C. A.; Parsons, S.; Rankin, D. W. H. Tetraborane(10), B4H10: Structures in Gaseous and Crystalline Phases. J. Chem. Soc., Dalton Trans. 1996, 4589−4596. (16) Mitzel, N. W.; Rankin, D. W. H. SARACEN - Molecular Structures from Theory and Experiment: The Best of Both Worlds. Dalton Trans. 2003, 3650−3662. (17) Young, J. A. Acetic Anhydride. J. Chem. Educ. 2001, 78, 1176. (18) EPSRC UK National Service for Computational Chemistry Software. (19) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb, M. A.; Cheeseman, J. R.; Scalmani, G.; Barone, V.; Mennucci, B.; Petersson, G. A. et al. Gaussian 09, revision D.01; Gaussian, Inc.: Wallingford, CT, 2013. (20) Møller, C.; Plesset, M. S. Note on an Approximation Treatment for Many-Electron Systems. Phys. Rev. 1934, 46, 618−622. (21) Zhao, Y.; Truhlar, D. G. Comparative DFT Study of van der Waals Complexes: Rare-gas Dimers, Alkaline-earth Dimers, Zinc Dimer and Zinc-rare-gas Dimers. J. Phys. Chem. A 2006, 110, 5121− 5129. (22) Sipachev, V. A. Calculation of Shrinkage Corrections in Harmonic Approximation. J. Mol. Struct.: THEOCHEM 1985, 22, 143−151. (23) Sipachev, V. A. Local Centrifugal Distortions Caused by Internal Motions of Molecules. J. Mol. Struct. 2001, 567−568, 67−72. (24) Wann, D. A.; Less, R. J.; Rataboul, F.; McCaffrey, P. D.; Reilly, A. M.; Robertson, H. E.; Lickiss, P. D.; Rankin, D. W. H. Accurate Gas-phase Experimental Structures of Octasilsesquioxanes (Si8O12X8; X = H, Me). Organometallics 2008, 27, 4183−4187. (25) Fleischer, H.; Wann, D. A.; Hinchley, S. L.; Borisenko, K. B.; Lewis, J. R.; Mawhorter, R. J.; Robertson, H. E.; Rankin, D. W. H. Molecular Structures of Se(SCH3)2 and Te(SCH3)2 Using Gas-phase Electron Diffraction and ab Initio and DFT Geometry Optimisations. Dalton Trans. 2005, 3221−3228. (26) Hinchley, S. L.; Robertson, H. E.; Borisenko, K. B.; Turner, A. R.; Johnston, B. F.; Rankin, D. W. H.; Ahmadian, M.; Jones, J. N.; Cowley, A. H. The Molecular Structure of Tetra-tert-butyldiphosphine: An Extremely Distorted, Sterically Crowded Molecule. Dalton Trans. 2004, 2469−2476. (27) Ross, A. W.; Fink, M.; Hilderbrandt, R. In International Tables for Crystallography, Wilson, A. J. C., Ed.; Kluwer Academic Publishers: Dordrecht, Netherlands, 1992; Vol. C, p 245. (28) Gordon, M. S. The Isomers of Silacyclopropane. Chem. Phys. Lett. 1980, 76, 163−168. (29) Hariharan, P. C.; Pople, J. A. The Influence of Polarization Functions on Molecular Orbital Hydrogenation Energies. Theor. Chem. Acc. 1973, 28, 213−222.

(30) Hehre, W. J.; Ditchfield, R.; Pople, J. A. Self-Consistent Molecular Orbital Methods. XII. Further Extensions of Gaussian-Type Basis Sets for Use in Molecular Orbital Studies of Organic Molecules. J. Chem. Phys. 1972, 56, 2257−2261. (31) McLean, A. D.; Chandler, G. S. Contracted Gaussian-basis Sets for Molecular Calculations. 1. 2nd Row Atoms, Z = 11−18. J. Chem. Phys. 1980, 72, 5639−5648. (32) Krishnan, R.; Binkley, J. S.; Seeger, R.; Pople, J. A. Selfconsistent Molecular Orbital Methods. XX. A Basis Set for Correlated Wave Functions. J. Chem. Phys. 1980, 72, 650−654. (33) Masters, S. L.; Atkinson, S. J.; Hölbling, M.; Hassler, K. Gasphase Molecular Structure of 1,1,1,2-Tetrabromo-2,2-dimethyldisilane: theoretical and Experimental Investigation of a Super-halogenated Disilane and Computational Investigation of the F, Cl and I Analogues. Struct. Chem. 2013, 24, 1201−1206. (34) Derissen, J. L. A reinvestigation of the molecular structure of acetic acid monomer and dimer by gas electron diffraction. J. Mol. Struct. 1971, 7, 67−80. (35) Cox, A. P.; Thomas, L. F.; Sheridan, J. Internuclear Distances in Keten from Spectroscopic Measurements. Spectrochim. Acta 1959, 15, 542−543. (36) van Eijck, B. P.; van Opheusden, J.; van Schaik, M. M. M.; van Zoeren, E. Acetic acid: Microwave Spectra, Internal Rotation and Substitution Structure. J. Mol. Spectrosc. 1981, 86, 465−479.


DOI: 10.1021/acs.jpca.6b00704 J. Phys. Chem. A 2016, 120, 2041−2048