Structure and Stability of Phenoxide and Fluorophenoxide Anions

Article pubs.acs.org/JPCA

Structure and Stability of Phenoxide and Fluorophenoxide Anions Investigated with Infrared Multiple-Photon Dissociation and Detachment Spectroscopy and Tandem Mass Spectrometry Jeffrey D. Steill,†,# Amanda L. May,† Shawn R. Campagna,† Jos Oomens,‡,§ and Robert N. Compton*,†,∥ †

Department of Chemistry, University of Tennessee, Knoxville, Tennessee 37996, United States Institute for Molecules and Materials, FELIX Facility, Radboud University Nijmegen, Toernooiveld 7, 6525 ED Nijmegen, The Netherlands § University of Amsterdam, Science Park 904, 1098XH Amsterdam, The Netherlands ∥ Department of Physics, University of Tennessee, Knoxville, Tennessee 37996, United States ‡

ABSTRACT: The gas-phase infrared multiple-photon dissociation and detachment (IRMPD) vibrational action spectra of the unsubstituted phenoxide anion and a series of fluorine- and trifluoromethyl-substituted phenoxide anions in the spectral region between 600 and 1800 cm−1 are presented along with density functional theory (DFT) harmonic vibrational frequency calculations to establish the characteristic vibrations of the phenoxide functionality. The fluorophenoxide anions studied include the conjugate bases of o-, m-, and p-fluorophenol (C6H4FO−) as well as o-, m-, and p-α,α,α-trifluorocresol (CF3C6H4O−). The influence of the substituent on the characteristic vibrational frequencies is interpreted in terms of inductive and resonance shifts. In addition to the dissociation induced by infrared multiple-photon excitation, the electron detachment is also shown to play an important role in the decomposition of the unsubstituted phenoxide. It is demonstrated that the amount of electron detachment relative to dissociation is strongly mitigated by fluorination, and interpretations aided by DFT energy calculations suggest this is primarily due to the increased availability of low-energy dissociation pathways in the substituted phenoxides. Collision-induced dissociation (CID) mass spectrometry of the parent ions is used to estimate relative energies of the dissociation processes, and particular fragmentation motifs are elucidated. In particular, overall HF and CO losses provide facile decomposition pathways, yielding interesting fragment ions such as C6H− or C3H2FO− from the CF3C6H4O− parent anions. stretching frequency of the π-conjugated carboxylic acid carbonyl group. In that study,4 the phenoxide Ph−O− stretch frequency was determined to be approximately 1350 cm−1, which is intuitively sensible as bracketed by typical C−O stretch frequencies of 1000−1100 cm−1 and CO stretch frequencies of 1700−1800 cm−1. These initial investigations of the gas-phase bonding characteristics of this fundamental molecular anion deserve further development as the phenoxide moiety serves as an important negative charge stabilization site and will be subjected to continued investigation of physical and biochemical relevance. The vibrational spectroscopy of this ion serves to elucidate the underlying bonding characteristics and allow for a more thorough interpretation of calculated and observed properties. It is important to spectroscopically characterize this functionality in an isolated manner to avoid the strong perturbations present in condensed phase and complexes, especially for negative ions. Gas-phase action spectroscopy methods are ideal for this purpose, and in particular infrared multiple-photon dissociation and electron

I. INTRODUCTION Phenols are strong gas-phase acids due to the stabilization of the excess charge in the aromatic system. The importance of the aromatic resonance in stabilizing the negative charge of the conjugate base is apparent by comparison to benzyl alcohol, which exhibits a gas-phase acidity similar to that of ethanol; deprotonation from these alcohols requires approximately 100 kJ/mol more energy than from phenol.1 This stability is apparent in the overall gas-phase chemistry of the phenoxides, as they show limited reactivity and a large stability toward electron detachment greater than 2 eV.2,3 The conjugation between the aromatic ring and the oxygen can significantly perturb (and be perturbed by) the bonding of other substituents on the aromatic system, resulting in changes in stability and spectroscopy. An illustrative example of this effect was demonstrated in recent studies of the deprotonation isomers present from electrospray ionization (ESI) of a phydroxybenzoic acid (p-HBA) solution4−7 with similar effects observed in related molecules.8−10 The vibrational action spectrum of the gas-phase phenoxide isomer of the deprotonated p-HBA anion shows an IR-active band that is due to the phenoxide substituent stretch, as well as two other bands strongly coupled to ring-stretching vibrations. The significant conjugation effect in the phenoxide deprotonation isomer was apparent in the strongly red-shifted CO © XXXX American Chemical Society

Special Issue: A. W. Castleman, Jr. Festschrift Received: March 28, 2014 Revised: May 5, 2014

A

dx.doi.org/10.1021/jp503103a | J. Phys. Chem. A XXXX, XXX, XXX−XXX

The Journal of Physical Chemistry A

Article

phenoxide anions were an ESI needle voltage of −3 kV and temperatures of 90 and 50 °C on the needle and cone assemblies, respectively. Ions were accumulated for approximately 5 s in a hexapole trap with a trapping well potential of 5 V before subsequent storage and isolation in the Penning trap of a home-built Fourier-transform ion cyclotron resonance mass spectrometer (FTICR-MS).12 Infrared radiation in the frequency range 600−1800 cm−1 was generated by the freeelectron laser (FEL) located at the FOM Institute for Plasma Physics, Rijnhuizen.13 The bandwidth of the FEL was less than 1% of the central wavelength over the stated frequency range. The ions were irradiated for 3−5 s at a macropulse repetition rate of 5 Hz, with 3−5 mass spectra averaged per spectral data point. The decomposition yields are not corrected for laserpower variation across the investigated spectral range. Electron detachment from the irradiated anions was detected via electron capture by an inert ‘electron scavenger” molecule (SF6), as described in previous studies.14,15 The background pressure in the cell without an electron capture gas was 5 × 10−8 Torr. With addition of the background electron scavenger gas SF6, the ICR cell pressure increased to approximately 1 × 10−7 Torr. The IRMPD yield spectrum was produced by dividing the sum of the fragment ion(s) by the sum of parent and fragment ion(s) as a function of FEL wavelength. Multiple IRMPD yield spectra were averaged to generate a final action spectrum for each isomer of each anion. B. Collision-Induced Dissociation Mass Spectrometry. Energy-resolved collision-induced dissociation mass spectrometry experiments were performed at The University of Tennessee Center for Mass Spectrometry using an ABI Sciex Qstar Elite quadrupole−time-of-flight mass spectrometer. The phenoxide ions were generated by ESI of a few micromolar solution of the conjugate acid phenol, and N2 was used as the collision gas for fragmentation in the collision cell. The applied voltage generated an ion with a lab energy in the range of 0−7 eV; this lab energy was adjusted to the center of mass collision energy by a factor equal to MWN2/(MWion + MWN2). Thirty mass-spectra were recorded per individual collision energy and fragment intensities were determined by averaged mass peak heights. Two individual experiments were performed and averaged for each molecular anion studied. The specifications of the Qstar Elite instrument indicate that a maximal mass accuracy of 5 ppm is achievable. For the mass ranges observed in these experiments, this level of mass accuracy would lead to precision of four decimal places for full scans. Therefore, the masses of the spectral features observed during tandem massspectrometry experiments were corrected by adjusting the fragment ion masses by the difference between the calculated and observed parent ion mass and are reported to four decimal places. This led to a −0.0090 amu mass correction for the fluoro-substituted parent and fragment ions. C. Computational Methods. Minimum energy geometries and harmonic vibrational frequencies for neutral phenols, deprotonated phenoxide anions, their dissociative neutral and ionic fragments and radical neutral phenoxyl species were computed by the B3LYP hybrid DFT method implemented within Gaussian03,16 using the 6-311++G(2d,2p) basis set, with additional calculations specifically for the phenoxide anion using a range of basis sets from 6-311++G** to aug-cc-pVTZ. Harmonic frequencies were scaled by a factor of 0.975 and convoluted with a Gaussian line shape function of 20 cm−1 fwhm for comparison to experimental IR spectra.

detachment (IRMPD) spectroscopies provide a means to probe the vibrations of the isolated anion. Previous efforts with gas-phase benzoate have demonstrated that comparison of vibrational spectra within a series of fluorinated analogues of an aromatic molecule can provide insight into the contribution of electron withdrawing effects through inductive and resonance influences on the vibrational characteristics.11 By employing aromatic substitution, it is possible to utilize the similarities and shifts in vibrational band frequencies and intensities to elucidate the peculiarities of the phenoxide system. Systematic spectral shifts within a series of substituted aromatic anions provide guidelines for chemical interpretation, and comparison to density-functional theory (DFT) molecular orbital computations allows for a detailed interpretation in terms of the distribution of electron density. The fluorine substituent serves as a classic case of the balance between inductive and resonant aromatic effects, and vibrational spectra can be expected to manifest this interaction with the phenoxide. The frequency shift of the phenoxide stretching mode is related to the withdrawal of the electron density away from the C−O− moiety to achieve greater CO character and, as such, is a sensitive probe of inductive and resonance effects. The CO containing resonance structure that can be drawn by placing the excess charge on the ortho- and/or para-site carbon is stabilized by the addition of an electronegative substituent such as −F or −CF3. The addition of the fluorine atoms can be expected to stabilize the negative charge on the molecular system as a whole, and the meta-position of the substituent gives a different possible influence on the phenoxide bond than the ortho- or para-position substituent. An advantage of the current methodology results from the possibility to simultaneously investigate the effects of substitution upon the unimolecular dissociation dynamics. The addition of the fluorine substituent can be expected to increase the electron detachment energy and thus lower the detachment rate upon intense resonant IR irradiation. However, this is expected to be minimal relative to the larger change in dissociation rate because fluorination results in entropically and energetically more accessible dissociation channels relative to the unsubstituted phenoxide anion. The calculation of detachment and dissociation energies using DFT methods can provide a means for interpretation of this unimolecular rate competition. These calculations can be validated by experimental investigations of bond dissociation energies through collision-induced dissociation (CID) massspectrometry methods. In addition to providing an experimental basis to examine the overall decomposition dynamics of the phenoxide anions, the fragmentation patterns from CID provide valuable insight into the relative stabilities of both parent and fragment species. Understanding the contribution of the aromatic substitution to the spectroscopic shifts and decomposition dynamics can provide an improved understanding of the role of resonance in phenoxides that is relevant to aromatic anions in general and useful for detailed application to more complex systems.

II. EXPERIMENTAL METHODS A. IRMP Dissociation and Detachment Spectroscopy. Phenoxide anions were generated by deprotonation of the corresponding phenol. A 1 mM solution of the phenol in 50:50 MeOH:H2O solvent was introduced into a Micromass “Zspray” electrospray ionization (ESI) source at a flow rate of 10 μL/min. Typical conditions optimized to produce the B



Article

III. RESULTS AND DISCUSSION A. IRMPD Spectroscopy. All of the phenoxide anions show electron detachment to some degree upon on-resonance IR irradiation as well as dissociation. The unsubstituted phenoxide anion shows a competition between decomposition pathways upon multiple-photon absorption, but as demonstrated in Figure 1, the electron detachment channel dominates.

reached in the irradiated anion. Once the threshold for the higher energy channel is reached, an increase of laser power or absorption cross section results in a disproportionate increase of the higher energy channel relative to the lower.17,18 As shown in Figure 1, the increase of the dissociation signal as a function of increased laser power and/or irradiation time is in greater proportion than the concomitant increase in electron detachment signal, in agreement with the expectations from this simple unimolecular rate picture of the competition between dissociation and detachment processes. This competition is similar to that seen from IRMP excitation of nitrobenzene anions19 and SF5− anions.14 The results of DFT electronic structure calculations used to aid in the interpretation of the vibrational spectrum are shown in the lower panel of Figure 1. Comparison of the calculated to observed infrared activity results in a reasonable match, but with some significant discrepancies. Certain key features of the computed spectrum are observed, including the benzene mode 19a near 1500 cm−1 and the C−O− stretch mode (which is strongly coupled to the benzene modes 19a and 7a) that appears between 1300 and 1400 cm−1 with sufficiently high FEL fluence. However, there is clearly a lack of detail in the spectrum, as many predicted low-frequency modes are missing and the highest frequency mode shown is apparently calculated to be significantly too high and/or too intense. This is likely due to the comparatively low vibrational density of states and large thermodynamic stability of the anion, both factors that do not promote the IRMPD process. However, it is clear by examination of the attenuated and extended IRMPD scans in Figure 1 that the large feature centered at 1500 cm−1 is not symmetrically broadened, and the larger width of the peak on the blue side of the spectrum is a result of the contribution from an unresolved benzene 8a mode. The scan with increased irradiation time demonstrates the asymmetry of the large band and more clearly reveals the band predicted between 1300 and 1400 cm−1. Thus, the lack of agreement is not unreasonable, and comparison to other phenoxide systems can be expected to provide further insight. The calculated harmonic vibrational frequencies for the phenoxide anions are presented in Table 1, along with the concomitant description in terms of benzene modes. The determination of the relationship of the phenoxide anion modes to the component benzene modes is determined by a mode projection analysis implemented with the ViPA

Figure 1. Infrared multiple-photon dissociation and detachment spectra of phenoxide anions (C6H5O−) with additional spectra acquired using different FEL powers and irradiation durations. One spectrum is at the same power with increased irradiation time (red traces) and the other is at the same irradiation time with FEL power decreased by ∼50% (blue traces). The upper panel shows the electron detachment yield as determined by SF6 electron scavenger yield, and the middle panel shows dissociation to m/z 65 anionic fragment. The lower panel shows the results of DFT scaled harmonic frequency calculations convolved with a 20 cm−1 Gaussian line shape function.

The observed dissociation to a fragment of m/z 65 is presumably due to CO loss from the phenoxide anion to produce the C5H5− cyclopentadiene anion. Dissociation to this fragment from the phenoxide anion is expected to be energetically and entropically costly, as it involves breaking the aromatic ring to remove the CO molecule. Thus, it is not surprising that the detachment process dominates in the unsubstituted phenoxide relative to this dissociation process. Typically, the higher energy dissociation channels are significantly accessed only as very high internal energies are

Table 1. Comparison of Calculated and Observed Vibrational Frequencies (cm−1) for Selected IR-Active Modes Related to the Phenoxide Moietya phenoxide C6H5O− fluorophenoxides p-FC6H4O− o-FC6H4O− m-FC6H4O− trifluorocresoxides p-CF3C6H4O− o-CF3C6H4O− m-CF3C6H4O−

8a

19a

C−O− str

1570 (1550)

1494 (1490)

1364 (1370)

0 (0)

1574 (1550) 1573 (1560) 1585 (1570)

1494 (1480) 1511 (1510) 1498 (1510)

1397 (1370) 1358 (1340) 1360 (1340)

33 (0) −6 (−30) −4 (−30)

1587 (1570) 1592 (1570) 1573 (1555)

1530 (1530) 1514 (1510) 1497 (1490)

1391 (1380) 1366 (N/A) 1365 (1360)

27 (10) 2 (N/A) 1 (−10)

ΔE

a

Calculated frequency results are from calculations using Hybrid-DFT B3LYP with the 6-311++G(2d,2p) basis set, and the experimental results are given in parentheses. Although there is significant coupling between the ring and CO modes, the header denotes the dominant mode. The ΔE column shows the band shifts of the CO stretch relative to the unsubstituted phenoxide. C



Article

program.20 The assignment of the vibrational modes for the fluorophenoxides is performed by visual inspection and comparison to the phenoxide anion vibrations. The three modes shown in Table 1 are ring-stretching and C−O− stretching modes that are consistent for all of the anions studied. Other features in the IRMPD spectra are not quantitatively analyzed. Addition of a fluorine substituent to the molecular anion provides an increased complexity in decomposition behavior and spectral response. The IRMPD spectra for o-, m-, and pfluorophenoxide acquired from both the dissociation and electron detachment channels are shown in Figure 2 along

Figure 3. Infrared multiple-photon dissociation and detachment spectra of p-, o-, and m-perfluoromethylphenoxide anions (CF3C6H4O−). The solid line shows the dissociation yield, and the dotted line shows the detachment yield. Overlaid with the experimental results are the results of DFT scaled harmonic frequency calculations convolved with a 20 cm−1 Gaussian line shape function.

observed for the fluorophenoxides, the dissociation process dominates over the electron detachment yield due to the lower energy dissociation channels available. The calculations account for the three bands related to the ring stretching and phenoxide stretching reasonably well. However, the poor agreement between the simulated and observed spectra below 1200 cm−1 demonstrates that the DFT harmonic frequency calculations are unsuccessful in modeling the coupling of the −CF3 vibrations with the phenyl ring. Nonetheless, the key spectroscopic features of the phenoxide moiety are present and allow for positive spectroscopic identification. The results of the spectroscopic assignments are summarized in Table 1. Within this series of substituted phenoxides, a comparison of the vibrational frequencies (as shown in Table 1) reveals qualitative trends in the band shifts for the phenoxide stretching mode that are related to aromatic resonance effects. Using the C−O− stretching frequency of the unsubstituted phenoxide anion as the basis for comparison reveals no clear overall red or blue shift of this vibrational mode in the substituted phenoxides. This may be due to significant coupling of this mode to the substituent and ring modes. However, the calculations and data show the same directional trend within the series for the effect of the −F substituent and the −CF3 substituent; in general, there is a red shift of the meta-isomer relative to the ortho- and/or para-isomers. This is because the resonance structure that results in partial CO character and partial negative charge on the ortho- and para-substituted carbons is stabilized by the electronegative substituent. The meta-isomer lacks the conjugation between the substituent site and the C−O bond. This suggests that the primary influence of the electron withdrawing substituents is through resonance effects. Interestingly, there is an overall blue shift of the phenoxide stretch for the −CF3-substituted phenoxides relative to the fluorophenoxides. This can be understood in terms of both resonance and direct inductive effects; the more strongly electron withdrawing −CF3 group allows for excess electron density at the ring carbons, better stabilizing the negative charge within the ring and increasing the CO bond character.

Figure 2. Infrared multiple-photon dissociation and detachment spectra of p-, o-, and m-fluorophenoxide anions (FC6H4O−). The solid line shows the dissociation yield, and the dotted line shows the detachment yield. Overlaid with the experimental results are the results of DFT scaled harmonic frequency calculations convolved with a 20 cm−1 Gaussian line shape function.

with the results of DFT calculated frequencies. The IRMPD spectra of the substituted phenoxides also shows a competition between the possible decomposition processes, including multiple fragmentation channels, but the primary dissociation channel is loss of 20 Da, clearly HF loss, which dominates over the electron detachment process. Comparison of the calculated and observed IR spectra shown in Figure 2 provides a means for a more detailed comparison of the characteristic vibrations of the phenoxide system. The three dominant vibrational modes in the wavelength region between 1300 and 1700 cm−1 are consistent with the phenoxide anion. In addition, the C−F stretching mode at ∼1100 cm−1 is clear. There is an overall agreement with the data and the calculations, and an analysis of the band shifts of the C−O− stretching frequency between the different isomers can be used to analyze trends in the bonding interactions. The vibrational spectroscopy analysis for the fluorophenoxides is summarized in Table 1. IRMPD experiments and computations regarding the series of o, m-, and p-trifluorocresoxide anions (CF3C6H4O−) provide a means for further investigation of vibrational frequency shifts involved in substituted phenoxides. Extension of the analysis from a monofluoro substituent to a −CF3 moiety allows for verification of the assigned bands and improved accounting of the relevance of the site of substitution on the aromatic moiety. The IRMPD experimental and calculated vibrational spectra for the trifluorocresol conjugate bases are shown in Figure 3. As D



Article

Table 2. Computed Zero-Kelvin Zero-Point Energy Corrected Dissociation and Detachment Energies of the Phenoxide Systems under Experimental Investigationa relative stability

dissociation channels

phenoxide

phenoxyl

rel E (kJ/mol)

rel E (kJ/mol)

o-FC6H4O− m-FC6H4O− trifluorocresoxides p-CF3C6H4O− o-CF3C6H4O− m-CF3C6H4O−

rel E (kJ/mol)

diss (eV)

CO loss yields C6H5−

phenoxide C6H5O− fluorophenoxides p-FC6H4O−

fragment isomer AEA (eV) 2.22

2.40 2.51

HF loss yields C6H3O− o-C6H3O m-C6H3O o-C6H3O o-C6H3O

0 34 0 0

2.70 3.05 2.73 2.89

2.92 2.84 2.74

HF loss yields C6H3CF2O− (2-dehydro-4-phenoxy) (6-dehydro-2-phenoxy) (6-dehydro-3-phenoxy)

0 21 92

3.24 3.31 4.03

0

0

2.26

−3 −18

10 6

−16 −2 0

1 8 0

1.56

a

Results shown are from calculations using Hybrid-DFT B3LYP with the 6-311++G(2d,2p) basis set. Only the lowest energy fragment isomer found from each of the trifluorocresoxides’ dissociation pathways is shown.

Figure 4. Collision-induced dissociation of the phenoxide, fluorophenoxide, and trifluorocresoxide anions. (a) shows the entire collision energy range investigated, and (b) shows a zoomed-in energy region of the experiment. Note that the seven labeled m/z peak intensities of the middle panels refer to all the FC6H4O− isomers, and the nine labeled m/z peak intensities of the lower panels refer to all the CF3C6H4O− isomers. The suggested chemical identities for the m/z values are given in Table 3. The observed fragments for the aromatic isomers were the same, but observed in different abundances. In all fluorophenoxides, the lowest energy pathway observed was HF loss.

B. Dissociation and Detachment Energetics. In addition to aiding spectroscopic interpretation, the DFT

computations provide a foundation for analysis of the decomposition pathways in terms of calculated adiabatic E



Article

Table 3. Observed Ionic Fragmentation from the Phenoxide and Substituted Phenoxide Anionsa

phenoxide C6H5O−

fluorophenoxides o,m,p-FC6H4O−

trifluorocresoxides o,m,p- CF3C6H4O−

fragment m/z

loss m/z

93 77 65

16 28

111 91 83 65 63 49

20 28 46 48 62

41

70

161 141 121 113 101 93

20 40 48 60 68

73

88

71 43

90 118

proposed leaving group(s)

proposed fragment ion(s)

CO

C5H5−

HF CO HF, C2H2 HF,CO HF, C2H2O C5H2 C4H2, HF

C6H3O− C5H4F− HC4O− C5H3− C4H− CH2FO− HCCO−

HF HF(×2) HF, CO HF(×3) HF(×2), HC4F HF(×3), C4H2F2 HF(×2), HF(×2),

C7H3F2O− C7H2FO− C6H3F2− HC7O− C6H2F− C3H3F2O− C6H− C3H2FO− FC3O− FC2−

CO COb C4H2 CO, C4H2

observed m/z

monoisotopic mass(es) (amu)

93.0276 76.9651 65.0341

93.034 588

6.99

75.1

65.039 673

5.57

85.7

111.0252 91.0191 83.0306 65.0044 63.0253 49.0096

111.025 166 91.018 390 83.030 252 65.003 288 63.024 204 49.008 374 49.009 516 41.003 288

−0.03 −0.71 −0.35 −1.11 −1.10 −1.23 −0.08 −1.31

−0.3 −7.8 −4.2 −17.1 −17.4 −25.0 −1.7 −32.0

161.021 973 141.015 745 121.009 516 113.020 830 101.003 288 93.014 602 93.015 745 73.008 374 73.009 516 70.993 866 42.998 952

0.97 0.15 −0.08 −0.27 −0.81 −0.10 1.04 −0.23 0.92 −1.13 −1.15

6.0 1.0 −0.7 −2.4 −8.0 −1.1 11.2 −3.1 12.5 −16.0 −26.7

41.0046 161.0210 141.0156 121.0096 113.0211 101.0041 93.0147 73.0086 70.9950 43.0001

difference (mDa)

error (ppm)

a

The identities of the fragment ions are proposed on the basis of exact masses that were corrected on the basis of the masses of the known parent ions. Three of the fragment anions (m/z 49, 93,73) are given two possible assignments. Formulas were calculated using the “MS from accurate mass” tool at ChemCalc (http://www.chemcalc.org/mf_finder/mfFinder_em_new). Identifications do not account for the possibility of a radical anion and assume only even electron molecules or complexes. The m/z 77 ion from phenoxide is unidentified. bFor m/z 73: with C6H− the neutral fragment products HCF3 + H2O are also possible.

endothermicity of only 1.6 eV as compared to the greater than 3 eV adiabatic dissociation energy of the fluorophenoxides. Although the qualitative trend of increasing detachment energy with degree of fluorination plays a part in the increasing importance of the dissociation channel over detachment, it is likely less relevant than the increased availability of low-energy dissociation channels. As shown in Table 2, the calculated adiabatic energy for detachment from the phenoxide gives a good agreement with the 2.2 eV experimental determination by photoelectron spectroscopy.2,3 However, the calculated endothermicity for dissociation provides a poor reference to the experiment due to the neglect of the energy barrier. To better understand the trends in decomposition of the anions, the dissociation energies of the phenoxide and fluorophenoxide anions were probed quantitatively by energyresolved collision-induced dissociation (CID) tandem mass spectrometry to determine the degree of difference between the actual dissociation barrier and the calculated adiabatic dissociation energies. The investigation of the phenoxide anion demonstrates the large energy barrier to dissociation over the thermodynamic endothermicity. As shown in the upper panel (a) of Figure 4, the CID of the phenoxide anion undergoes fragmentation to C5H5− at a much higher threshold than 1.6 eV. However, in contrast to the IRMPD experiment where the only dissociation channel observed from IR photoexcitation was CO loss, there was also an additional channel observed at very high energies to produce an unidentified fragment anion of m/z 77. From examination of the same data as plotted over a smaller energy window (shown

electron affinities and dissociation energies. The increasing importance of dissociation over electron detachment upon addition of these strongly electron withdrawing groups is directly related to the energetic barriers to the two processes, but the thermodynamic energies of the reactants and products can give a reasonable insight into the trends for closely related analogues. The energy of electron detachment from the anions is related to the adiabatic electron affinity of the corresponding radical neutrals. For the dissociation energies, the reaction enthalpy to produce the proposed neutral and ionic fragments is calculated. The endothermicities of dissociation for the substituted phenoxide anions are determined by the calculated energy of the lowest-energy isomer produced from HF loss. The endothermicity of dissociation for the unsubstituted phenoxide anion is the energy of CO loss. The results of these calculations are presented in Table 2. There is a clear trend apparent in the increase of the adiabatic electron detachment energy with the degree of electron withdrawal that is experimentally correlated with the decreasing importance of the photodetachment process for the substituted anions. However, the difference in thermodynamic dissociation energies between the unsubstituted and fluorinated phenoxides is misleading, because the dissociation of the unsubstituted phenoxide anion is expected to involve a large barrier. In contrast, for the fluorophenoxides it is expected that the endothermicity for HF loss is not extremely different from the actual barrier to dissociation. As seen in Table 2, the significant thermodynamic stabilities of the phenoxide fragments, namely the C5H5− anion and the CO molecule, result in a relatively low F



Article

of the parameters that influence the dissociation and detachment competition and reveal interesting fragmentation behavior in these ions. As shown in Table 2, the trifluoromethylphenoxides show a trend similar to that for the fluorophenoxides relative to the bare phenoxide anion in that the calculated energy for electron detachment is higher, and the dissociation energy is expected to be significantly lowered. The CID data in the lower panels of Figure 4 show that, indeed, the onset of CF3C6H4O− dissociation observed at energies lower than 3 eV is much lower than the phenoxide dissociation threshold and similar to the onset in the monofluorinated phenoxide anions. Thus, with the available low-energy pathways for dissociation, the reduced role of electron detachment in the IRMPD studies is not surprising. As is the case with the fluorophenoxide anions, the lowest energy pathway for the trifluorocresoxides is HF loss, but the increased degree of fluorination gives a greater variety of dissociation pathways for these molecules. Again, the overall dissociation pathways for ortho-, meta-, and para-isomers are similar, which mirrors previous results for studies involving the set of nonfluorinated cresols.23 As shown in the lower panels of Figure 4 and summarized in Table 3, the HF molecule makes a very good gas-phase leaving group, and loss of one or two HF molecules is a facile process. The loss of 20 or 40 Da was the dominant process in the IRMPD process of these species, as well. Other dissociation pathways become available at energies higher than about 4 eV that involve more substantial dissociation, which could be through expulsion of HF, CO, and 1,3-butadiene. The basic behavior of these fragmentation pathways is proposed to be similar to that observed for the monofluorophenoxides; therefore, the data in Figure 4 are roughly color-coded according to analogous proposed leaving groups. As seen for the monofluorophenoxides, the fragmentation patterns for the trifluorocresoxides indicate that considerable rearrangement of the molecular structure occurs at the highest collision energies studied. Again, either highly unsaturated hydrocarbon anions or oxygen- and fluorine-containing fragments are produced. The dominant high-energy fragment produced from all of the isomers is m/z 73, which is of interest because its production is independent of the position of trifluoromethyl substitution and due to the potential structures of the fragment anion. The formulas C6H− and C3H2FO− are both candidates for this mass, which, incidentally, was also produced in the IRMPD experiments. Again, the production of C6H− would be of particular interest as this was the first molecular anion to be detected in interstellar space.22 Support for the assignment of the formula C6H− comes from the excellent fit of the observed (m/z 73.0086) with the calculated exact mass of the fragment (m/z 73.008374). The formation of an anion (m/z 101.0041) consistent with HC7O− also suggests that the loss of three HF equivalents is possible in this system, and this anion would then only need to lose CO to generate C6H−. However, the agreement of the observed m/z with the calculated exact mass for C3H2FO− (m/z 73.009516) is also within the error of the measurement. The potential presence of FC3O− (m/z 70.9950), which could have arisen from H2 loss from C3H2FO−, also prevents unambiguous assignment for the m/z 73 fragment. The neutral loss from the parent ion (m/z 161.0210) to the m/z 73.0086 fragment is 88.0124 Da, which is consistent with the formula C2H4F2. Due to the interesting possibility that the m/z 49 fragment anion from the fluorophenoxides and the m/z 73 fragment

in upper panel (b) of Figure 4), the observed CID threshold for the C5H5− formation (via CO loss) is approximately 4 eV. Thus, the CID study demonstrates that the dissociation energy for the phenoxide anion is much higher than the electron detachment energy and thus provides a reasonable experimental basis for an interpretation of the trend of decreasing detachment upon fluorination. The dissociation energies for the substituted phenoxides were also studied by tandem mass spectrometry to investigate the agreement between calculated and observed dissociation thresholds and to identify the dominant fragmentation behavior upon collisionally induced dissociation. Shown in the middle panels of Figure 4 are the results for the fluorophenoxide anions. For both the CID and IRMPD experiments, the lowest energy fragmentation pathway observed was HF loss to make the m/z 91 anion (C6H3O−). The CID experiments for all three isomers show an onset of dissociation at energies near 3.0 eV. The observed threshold is in rough agreement with the calculated dissociation energies for HF loss shown in Table 2, except for the meta-isomer. The dissociation energy calculated for this isomer in the table is given assuming that H atom migration on the ring does not occur. As shown in Figure 4, there is no significant difference between the observed dissociation threshold for the meta-isomer as compared to the ortho- and para-isomers. Allowing for the formation of the most stable CF2C6H3O− fragment isomer, the calculated dissociation endothermicity of o-FC6H4O− is lowered to 3.29 eV, similar to that for the other isomers and in better agreement with the experiment. This implies that H atom migration is possible in the dissociation of these species. As seen by comparison to the upper panels of Figure 4, the threshold for dissociation of the fluorophenoxides is considerably lower than the threshold for dissociation of the unsubstituted phenoxide ion. The substituted phenoxides are expected to show a much lowered energy barrier for dissociation relative to the phenoxide anion because HF loss does not need to fragment the aromatic ring. Thus, the decreasing relevance of the IRMPD electron detachment pathway for these anions compared to the bare phenoxide is due primarily to the strongly lowered dissociation energies and not due to the slightly increased electron affinities. At higher collision energies multiple fragmentation pathways were observed, with proposed fragmentation pathways including CO and acetylene loss. Although all of the fragments could not be unambiguously identified, it is interesting to note that the dissociation pathways for the ortho-, meta-, and paraisomers were similar. Furthermore, the fragments at m/z 63 and 49 also indicate that considerable rearrangement of the starting structures occurred to generate either unsaturated hydrocarbon anions (C5H3− and C4H−) or both oxygenated and fluorinated product ions (CH2FO−). The formation of either set of product ions is chemically unprecedented. The possible formation of C4H− is particularly interesting as this was the second anion to be detected in interstellar space,21 after C6H−.22 At the highest collision energies of the study, the aromatic system of the fluorophenoxide anions was completely disrupted, and a small fragment consistent with the ethynol anion was observed at m/z 41. The accurate masses of the collision products were used to aid in the proposed assignments of the fragment ions; these results are summarized in Table 3, as are the CID results for the other anions under investigation. Finally, CID studies of the series of o-, m-, and ptrifluorocresol anions provide a means for further investigation G



Article

barrier in the reaction channel), and thus in contrast to the substituted phenoxides, electron detachment is an especially important channel for the phenoxide ion due to the existence of only high-energy dissociation channels. The addition of the fluorine substituent increases the electron affinity of the radical neutral; however, the stability of the anion is primarily due to aromatic stabilization, and the addition of the electronegative fluorine atom does not have a large effect. The observation of a much-decreased propensity for detachment in the IRMPD studies is therefore a result of the much higher entropy and lower enthalpy of dissociation relative to detachment. The marked difference in this ratio observed from the fluorophenoxides compared to the unsubstituted phenoxide is indicative of a significant change in the dissociation vs detachment rates at the internal energies accessed. The competition between these unimolecular processes is strongly influenced by the availability of dissociation channels, and at high energies the dissociation pathways produce a wide variety of unsaturated anion fragments due in part to the weak gas-phase acidity of the fluoride anion.

anion from the trifluorocresoxides could correspond to the astronomically observed species C4H− and C6H−, a preliminary investigation of the thermodynamics of these dissociation pathways was undertaken. As shown in Table 3, both of these fragment ions have very reasonable alternative assignments. A set of calculations at a consistent level of theory (B3LYP/6311++G(2d,2p)) was used to verify the suitability of the proposed neutral and ionic fragment structures and determine the reaction enthalpies. For the fluorophenoxides, the m/z 49 anion fragment could be either due to loss of H2C2O to make C4H− or due to loss of C5H2 to make CH2OF−. For the first pathway, the ketene isomer of H2C2O is more than 1 eV lower in energy than the alcohol and the C4H− species is calculated as a linear singlet state anion. For the second pathway to form m/z 49, the neutral C5H2 molecule is a linear triplet state and the oxide isomer of CH2FO− is ∼2 eV lower in energy than the enolate isomer. From the o-FC6H4O− anion, the dissociation reaction enthalpy is 4.8 eV to make the C4H− anion and 5.9 eV to make the CH2FO− anion. A similar comparison was performed for the m/z 73 fragment, which could result from loss of 3HF + CO (or HCF3 + H2O) to make C6H−, or alternatively, loss of C4H2F2 to make C3H2OF−. For the first pathway, the C6H− was calculated as a linear singlet state, and the two possible neutral fragment products gave essentially equivalent energies. For the second pathway to make m/z 73, the structure of the neutral fragment was assumed to be HCCC(H)CF2, and the acid halide enolate isomer of C3H2OF− was calculated to be 0.5 eV lower in energy than the fluoroacrolein enolate isomer. With the lowest-energy fragment structures, the dissociation reaction enthalpy from oCF3C6H4O− was 5.6 eV to make C6H−, and 4.5 eV to make C3H2OF−. Although in both cases the lowest-energy pathway that was proposed agrees well with the experimentally observed fragment threshold (i.e., C4H− for the onset of the m/z 49 and C3H2OF− for the m/z 73 ion signals), the energy calculations are not of sufficient accuracy, nor are the barriers accounted for, to conclusively rule out the possibility of the other pathway. Although the m/z 49 fragment is of relatively low yield, the high yield and specificity of the m/z 73 fragmentation pathway is of interest and warrants further investigation, and work is being performed to determine the structure of this anion. If the observed fragment ion at m/z 49 is C4H− and/or the fragment ion at m/z 73 is C6H−, CID and IRMPD could provide a new facile means of producing these interesting anionic species for further study.

■

AUTHOR INFORMATION

Present Address #

Sandia National Laboratories, 7011 East Ave., Livermore, California 94550, United States Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS Parts of this research are based on work supported by the NSF Grant number CHE-0848487. We acknowledge the support of Drs. Britta Redlich and Lex van der Meer as well as others of the FELIX staff. This work is part of the research program of FOM, which is financially supported by the Nederlandse Organisatie voor Wetenschappelijk Onderzoek (NWO).

■

REFERENCES

(1) NIST Chemistry WebBook, http://webbook.nist.gov/chemistry/ (2) Gunion, R. F.; Gilles, M. K.; Polak, M. L.; Lineberger, W. C. Ultraviolet photoelectron spectroscopy of the phenide, benzyl and phenoxide anions, with ab-initio calculations. Int. J. Mass. Spectrom. 1992, 117, 601−620. (3) Kim, J. B.; Yacovitch, T. I.; Hock, C.; Neumark, D. M. Slow photoelectron velocity-map imaging spectroscopy of the phenoxide and thiophenoxide anions. Phys. Chem. Chem. Phys. 2011, 13, 17378− 17383. (4) Steill, J. D.; Oomens, J. Gas-Phase Deprotonation of pHydroxybenzoic Acid Investigated by IR Spectroscopy: SolutionPhase Structure Is Retained upon ESI. J. Am. Chem. Soc. 2009, 131, 13570−13571. (5) McMahon, T. B.; Kebarle, P. Intrinsic Acidities of Substituted Phenols and Benzoic Acids Determined by Gas-Phase Proton-Transfer Equilibria. J. Am. Chem. Soc. 1977, 99, 2222−2230. (6) Tian, Z.; Kass, S. R. Does Electrospray Ionization Produce GasPhase or Liquid-Phase Structures? J. Am. Chem. Soc. 2008, 130, 10842−10843. (7) Schroder, D.; Budesinsky, M.; Roithova, J. Deprotonation of pHydroxybenzoic Acid: Does Electrospray Ionization Sample Solution or Gas-Phase Structures? J. Am. Chem. Soc. 2012, 134, 15897−15905. (8) Almasian, M.; Grzetic, J.; van Maurik, J.; Steill, J. D.; Berden, G.; Ingemann, S.; Buma, W. B.; Oomens, J. Non-Equilibrium Isomer Distribution of the Gas-Phase Photoactive Yellow Protein Chromophore. J. Phys. Chem. Lett. 2012, 3, 2259−2263. (9) Tian, Z. X.; Wang, X. B.; Wang, L. S.; Kass, S. R. Are Carboxyl Groups the Most Acidic Sites in Amino Acids? Gas-Phase Acidities,

IV. CONCLUSIONS The IRMPD results demonstrate that the characteristic phenoxide vibrations are influenced by electron-withdrawing resonance effects, namely through the blue shift of the phenoxide stretch mode for the para- or ortho-substituted species relative to the meta-substituted species. The characteristic modes of these systems are reasonably well-modeled by DFT harmonic frequency calculations for the unsubstituted phenoxide and monofluorinated phenoxides, but somewhat less so for the trifluorocresoxides. In particular, the vibrations of the −CF3 group are poorly modeled by the DFT calculations. The CID results show that the thermodynamic reaction enthalpies provide a poor basis for accounting for the relative yields of the dissociation and electron detachment processes from the IRMPD experiments. Clearly for the bare phenoxide anion, the threshold to dissociation is more than 2 eV higher than the calculated thermodynamic energy (which ignores a substantial H



Article

Photoelectron Spectra, and Computations on Tyrosine, p-Hydroxybenzoic Acid, and Their Conjugate Bases. J. Am. Chem. Soc. 2009, 131, 1174−1181. (10) Rocha-Rinza, T.; Christiansen, O.; Rajput, J.; Gopalan, A.; Rahbek, D. B.; Andersen, L. H.; Bochenkova, A. V.; Granovsky, A. A.; Bravaya, K. B.; Nemukhin, A. V.; Christiansen, K. L.; Nielsen, M. B. Gas Phase Absorption Studies of Photoactive Yellow Protein Chromophore Derivatives. J. Phys. Chem. A 2009, 113, 9442−9449. (11) Oomens, J.; Steill, J. D. Free Carboxylate Stretching Modes. J. Phys. Chem. A 2008, 112, 3281−3283. (12) Valle, J. J.; Eyler, J. R.; Oomens, J.; Moore, D. T.; van der Meer, A. F. G.; von Helden, G.; Meijer, G.; Hendrickson, C. L.; Marshall, A. G.; Blakney, G. T. Free electron laser-Fourier transform ion cyclotron resonance mass spectrometry facility for obtaining infrared multiphoton dissociation spectra of gaseous ions. Rev. Sci. Instrum. 2005, 76, 023103. (13) Oepts, D.; van der Meer, A. F. G.; van Amersfoort, P. W. The Free-Electron-Laser User Facility FELIX. Infrared Phys. Technol. 1995, 36, 297−308. (14) Steill, J. D.; Oomens, J.; Eyler, J. R.; Compton, R. N. Gas-phase infrared multiple photon dissociation spectroscopy of isolated SF6− and SF5− anions. J. Chem. Phys. 2008, 129, 244302. (15) Steill, J. D.; Oomens, J. Action Spectroscopy of Gas-Phase Carboxylate Anions by Multiple Photon IR Electron Detachment/ Attachment. J. Phys. Chem. A 2009, 113, 4941−4946. (16) 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.; et al. Gaussian 03, revision C.02; Gaussian, Inc.: Wallingford, CT, 2004. (17) Oomens, J.; Sartakov, B. G.; Meijer, G.; von Helden, G. Gasphase infrared multiple photon dissociation spectroscopy of massselected molecular ions. Int. J. Mass Spectrom. 2006, 254, 1−19. (18) Parniex, P.; Basire, M.; Calvo, F. Accurate Modeling of Infrared Multiple Photon Dissociation Spectra: The Dynamical Role of Anharmonicities. J. Phys. Chem. A 2013, 117, 3954−3959. (19) Steill, J. D.; Oomens, J. Spectroscopically resolved competition between dissociation and detachment from nitrobenzene radical anion. Int. J. Mass Spectrom. 2011, 308, 239−252. (20) Grafton, A. K.; Wheeler, R. A. Vibrational Projection Analysis: New Tool for Quantitatively Comparing Vibrational Normal Modes of Similar Molecules. J. Comput. Chem. 1998, 19, 1663−1674. (21) Cernicharo, J.; Guélin, M.; Agúndez, M.; Kawaguchi, K.; McCarthy, M.; Thaddeus, P. Astronomical detection of C4H−, the second interstellar anion. Astrono. Astrophys. 2007, 467, L37−L40. (22) McCarthy, M. C.; Gottlieb, C. A.; Gupta, H.; Thaddeus, P. Laboratory and Astronomical Detection of the Negative Ion C6H−. Astrophys. J. 2006, 652, L141−L144. (23) Eichinger, P. C. H.; Bowie, J. H.; Hayes, R. N. Stable NegativeIon Isomers in the Gas Phase. C7H7O- Species. Aust. J. Chem. 1989, 42, 865−874.

I


Structure and Stability of Phenoxide and Fluorophenoxide Anions

Recommend Documents