Electron attachment to indole and related molecules Alberto Modelli, Derek Jones, and Stanislav A. Pshenichnyuk Citation: The Journal of Chemical Physics 139, 184305 (2013); doi: 10.1063/1.4829057 View online: http://dx.doi.org/10.1063/1.4829057 View Table of Contents: http://scitation.aip.org/content/aip/journal/jcp/139/18?ver=pdfcov Published by the AIP Publishing Articles you may be interested in Dissociative electron attachment to pentaerythritol tetranitrate: Significant fragmentation near 0 eV J. Chem. Phys. 132, 134305 (2010); 10.1063/1.3386386 High resolution dissociative electron attachment to gas phase adenine J. Chem. Phys. 125, 084304 (2006); 10.1063/1.2336775 Electron attachment and detachment and the electron affinity of cyclo- C 4 F 8 J. Chem. Phys. 120, 7024 (2004); 10.1063/1.1683082 Electron attachment to 5-chloro uracil J. Chem. Phys. 118, 4107 (2003); 10.1063/1.1540108 Dissociative electron attachment in cyclopentanone, -butyrolactone, ethylene carbonate, and ethylene carbonate- d 4 : Role of dipole-bound resonances J. Chem. Phys. 110, 11376 (1999); 10.1063/1.479078
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP: 195.99.248.194 On: Tue, 08 Apr 2014 09:51:22
THE JOURNAL OF CHEMICAL PHYSICS 139, 184305 (2013)
Electron attachment to indole and related molecules Alberto Modelli,1,2,a) Derek Jones,3,b) and Stanislav A. Pshenichnyuk4,c) 1
Dipartimento di Chimica “G. Ciamician”, Universitá di Bologna, via Selmi 2, 40126 Bologna, Italy Centro Interdipartimentale di Ricerca in Scienze Ambientali (CIRSA), Universitá di Bologna, via S. Alberto 163, 48123 Ravenna, Italy 3 ISOF, Istituto per la Sintesi Organica e la Fotoreattività, C.N.R., via Gobetti 101, 40129 Bologna, Italy 4 Institute of Molecule and Crystal Physics, Ufa Research Centre, Russian Academy of Sciences, Prospeκt Oktyabrya 151, 450075 Ufa, Russia 2
(Received 24 September 2013; accepted 23 October 2013; published online 11 November 2013) Gas-phase formation of temporary negative ion states via resonance attachment of low-energy (0–6 eV) electrons into vacant molecular orbitals of indoline (I), indene (II), indole (III), 2-methylen1,3,3-trimethylindoline (IV), and 2,3,3-trimethyl-indolenine (V) was investigated for the first time by electron transmission spectroscopy (ETS). The description of their empty-level structures was supported by density functional theory and Hartree-Fock calculations, using empirically calibrated linear equations to scale the calculated virtual orbital energies. Dissociative electron attachment spectroscopy (DEAS) was used to measure the fragment anion yields generated through dissociative decay channels of the parent molecular anions of compounds I-V, detected with a mass filter as a function of the incident electron energy in the 0–14 eV energy range. The vertical and adiabatic electron affinities were evaluated at the B3LYP/6-31+G(d) level as the anion/neutral total energy difference. The same theoretical method is also used for evaluation of the thermodynamic energy thresholds for production of the negative fragments observed in the DEA spectra. The loss of a hydrogen atom from the parent molecular anion ([M-H]− ) provides the most intense signal in compounds I-IV. The gasphase DEAS data can provide support for biochemical reaction mechanisms in vivo involving initial hydrogen abstraction from the nitrogen atom of the indole moiety, present in a variety of biologically important molecules. © 2013 AIP Publishing LLC. [http://dx.doi.org/10.1063/1.4829057] I. INTRODUCTION 1
Since the original paper by Gregory in 1966, the two “dissimilar fields of research” (activity in vivo of molecular species and their gas-phase interactions with electrons) have offered a number of important studies aimed at a more complete understanding of molecular activities in living organisms.2–10 Many molecular species or their analogues can, in particular, be studied using gas phase electron transmission and dissociative attachment spectroscopies (ETS, DEAS) which allow experimental determination of empty level structures of these molecular species through electron attachment and the subsequent destiny of their temporary anion states.11–13 Recent studies by this group using these techniques have also shed light on the mechanisms involved in the specific activities of commonly used drugs including antipyretics14 (aspirin, paracetamol, phenacetin, and ibuprofen), the anti-malarial drug artemisinin and its derivatives15, 16 and flavonoids.17 Here we turn our attention to one of the most ubiquitous chemical moieties in living organisms. The indole moiety is not only present in almost all living organisms but also plays
a) Author to whom correspondence should be addressed. Electronic mail:
[email protected]
b) E-mail: [email protected] c) E-mail: [email protected]
0021-9606/2013/139(18)/184305/9/$30.00
a decisive role in a fascinatingly wide variety of important biochemical reactions.18 Its interaction, for example, with iron in heme-containing oxygenase enzymes19 is an important step in the insertion of oxygen into organic substrates, crucial for aerobic organisms, as in the case of human indoleamine 2,3-dioxygenase (IDO) which cleaves the indole pyrrole ring of L-tryptophan to incorporate both atoms of oxygen following the first step of this process which involves the interaction of the 2,3 bond of the indole ring with the iron-bound dioxygen.20 Both of the proposed mechanisms (ionic or radical) depend upon the free form of the indole NH and its electronic properties.19 The indole moiety also plays a vital role in radical scavenging and antioxidant activities of melatonin21 and its precursors serotonin and tryptophan where once again the 2,3 bond of the indole pyrrole ring is protagonist.22 The importance of the indole moiety has been confirmed by calculations showing that the 5-methoxy and N-acetyl groups of melatonin do not significantly affect its thermodynamic capacity for free-radical trapping.23 It is, in fact, well-known that the 3-position on the pyrrole ring of indole is the preferred site for electrophilic attack,24 even in the presence of 3-substituents, the 2-position being energetically unfavourable as it would involve an intermediate with perturbation of the π -electron system of the benzene ring. This is an important aspect for understanding some of the mechanisms involved in indole alkaloid biosynthesis.
139, 184305-1
© 2013 AIP Publishing LLC
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP: 195.99.248.194 On: Tue, 08 Apr 2014 09:51:22
184305-2
Modelli, Jones, and Pshenichnyuk
J. Chem. Phys. 139, 184305 (2013)
CHART 1.
Although the filled-level electronic structure of indole,25, 26 indoline,27 indene,28, 29 and 2-methylen-1,3,3trimethylindoline30 have been published, experimental determination of their energies of electron attachment and the various pathways of the resulting anion decay, with possible molecular dissociation of such states, has not been previously investigated, except for the temporary anion states of indole formed in the 0–3.5 eV energy range.10 The electron transmission spectroscopy (ETS) technique devised by Sanche and Schulz11 is still one of the most suitable means for detecting gas-phase temporary anion states, which appear as sharp structures, “resonances,” in their electron-molecule scattering cross sections. Here this technique is applied for the first time to detect the formation of temporary anion states and describe the empty-level structure of indoline (I), indene (II), indole (III), 2-methylen-1,3,3trimethylindoline (IV), and 2,3,3-trimethyl-indolenine (V). Chart 1 reports their structural formulae. Additional information on the fate of temporary molecular anions can be supplied by dissociative electron attachment (DEA) spectroscopy12, 13 which measures the yield of massselected negative ions as a function of the impact electron energy. When suitable energetic conditions apply, the decay of unstable molecular anions formed by resonance can follow a dissociative channel which generates a negative fragment and a neutral radical, in kinetic competition with simple reemission of the extra electron. The negative ion currents measured in the DEA spectra of compounds I-V, in the 0–14 eV energy range, are presented and the structures of the corresponding fragments are proposed with the support of density functional theory (DFT) calculations of the thermodynamic energy thresholds required for their production. The same theoretical method is employed to evaluate the vertical (EAv ) and adiabatic (EAa ) electron affinities of compounds I-V.
II. EXPERIMENTAL METHODS
Our electron transmission apparatus (Bologna Laboratory) is in the format devised by Sanche and Schulz11 and has been previously described.31, 32 To enhance the visibility of the sharp resonance structures, the impact energy of the electron beam is modulated with a small ac voltage, and the derivative of the electron current transmitted through the gas sample is measured directly by a synchronous lock-in amplifier. Each resonance is characterized by a minimum and a maximum in the derivative signal. The energy of the midpoint between these features is assigned as the vertical attachment energy (VAE). The ET spectra were obtained using the apparatus in the “high-rejection” mode33 and the signal is, there-
fore, related to the nearly total scattering cross section. The electron beam resolution was about 50 meV full width at half maximum (FWHM). The energy scale was calibrated with reference to the (1s1 2s2 ) 2 S anion state of He. The estimated accuracy is ±0.05 or ±0.1 eV, depending on the number of decimal digits reported. In order to detect the weakest signals from the negative fragments produced by DEA decay channels, although with a lower energy resolution of the incident electron beam, a negative ion magnetic mass spectrometer (Ufa Laboratory), described in detail previously34, 35 was used. Briefly, an electron beam of defined energy is passed through a collision cell containing a vapor of the substance under investigation, under single-collision conditions. A current of magnetically massselected negative ions is recorded as a function of the incident electron energy in the 0–14 eV energy range. The electron energy scale is calibrated with the zero energy SF6 − signal. The FWHM of the electron energy distribution is 0.4 eV, and the accuracy of the measured peak positions is estimated to be ±0.1 eV. The analyzed samples were commercially available from Aldrich and used without further purification. The collision cells of both instruments were held at 60◦ C–80◦ C. Calculations were carried out with the GAUSSIAN 09 set of programs.36 Evaluation of the virtual orbital energies (VOEs) of the neutral molecules was performed with the 6-31G(d) basis set at the Hartree-Fock (HF), with the MP2 optimized geometries, and B3LYP levels of theory.37 The vertical electron affinity (EAv ) was calculated as the difference between the total energy of the neutral and the lowest anion state, both in the optimized geometry of the neutral state, using the B3LYP hybrid functional with the standard 6-31+G(d) basis set, which includes the minimum addition of diffuse functions (s and p type diffuse functions to the nonhydrogen atoms). The adiabatic electron affinity (EAa ) was obtained as the energy difference between the neutral and the lowest anion state, each in its optimized geometry. The same theoretical method was also used to evaluate the thermodynamic energy thresholds for production of fragment anions, obtained as the total energy of the (geometrically relaxed) ground-state fragments relative to the ground neutral state molecule. Zero-point vibrational energy corrections were also calculated. III. RESULTS AND DISCUSSION A. Electron transmission spectrum and calculated VAEs
The present family of bicyclic compounds II-V possesses four empty π * MOs, deriving from the three benzene π * MOs
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP: 195.99.248.194 On: Tue, 08 Apr 2014 09:51:22
184305-3
Modelli, Jones, and Pshenichnyuk
J. Chem. Phys. 139, 184305 (2013)
and an ethene π * MO of the pentacyclic ring, while indoline (I), where the latter MO is not present, possesses only the three π * MOs with mainly benzene character. The π * VAEs measured in the ET spectra of the reference molecules benzene31 (1.12 eV, e2u ; 4.82 eV, b2g ) and ethene38 (1.73 eV) indicate that electron capture into the empty π * MOs of compounds I-V is expected to occur at energies lower than 6 eV. Systematic ETS studies39 have demonstrated that hydrocarbons not substituted by third-row or heavier heteroatoms generally do not give rise to distinct low-energy resonances associated with empty MOs of (local) σ symmetry, so that detection of signals associated with electron capture into empty σ * MOs are not expected in this series, although a broad and weak feature associated with a σ * resonance was observed in the ET spectrum of the hexacycle piperidine.40 The ET spectra of compounds I-V in the 0–6 eV energy range are reported in Figure 1. The measured VAEs, together with calculated virtual orbital energies (VOEs), are given in Table I. In indoline, the degeneracy of the e2u lowest unoccupied molecular orbital (LUMO) is removed by mixing with the adjacent nitrogen lone pair. The component (here denoted as π *A ) with a node at the substituted carbon atom is expected to be only slightly perturbed, whereas the component (π *S ) with maximum atomic coefficient on that carbon atom should
FIG. 1. Derivative of transmitted current, as a function of electron energy, in gas-phase indoline (I), indene (II), indole (III), 2-methylen-1,3,3trimethylindoline (IV), and 2,3,3-trimethyl-indolenine (V). Vertical lines locate the VAEs.
TABLE I. VOEs calculated at the HF/6-31G(d)//MP2/6-31G(d) and B3LYP/6-31G(d) levels, scaleda VOEs and measured VAEs. All values in eV. HF//MP2
Indoline (I)
Indene (II)
Indole (III)
2-methylen-1,3,3trimethyl-indoline (IV)
2,3,3-trimethylindolenine (V)
a b
B3LYP
Expt.
Orbital
VOE
Sc. VOE
VOE
Sc. VOE
VAE
π *O σ *NH π *S - n N π *A π *O σ *CH π *CC - π *S π *A π *S + π *CC π *O π *CC - nN σ *NH π *- nN π *S + π *CC π *O σ *CH π *CC - nN π *S - π *CC π *A π *O σ *CH π *CN - π *S π *A π *S + π *CN
9.662 6.393 4.526 3.944 9.887b 6.764 6.027 4.013 3.286 10.663 6.526 6.129 4.882 3.610 9.397 6.051 5.789 4.296 3.856 9.219 5.715 5.624 3.935 3.216
4.83
5.083 2.434 0.707 0.228 4.993b 2.850 1.614 0.159 − 0.483 5.160 2.092 2.183 0.948 − 0.066 4.263 2.304 1.515 0.415 0.097 4.424 2.061 1.263 0.084 − 0.704
5.02
4.93
1.49 1.10 5.23
1.76 1.10 4.7
2.22 1.05 0.53 5.08 2.61
2.47 1.20 0.78 4.7 2.78
1.68 0.87 4.36
1.88 0.99 4.44
2.14 1.25 1.00 4.49
2.28 1.36 1.04 4.55
1.94 0.99 0.35
2.26 1.05 0.46
1.50 1.13 4.98 2.48 1.17 0.70 5.48 2.80 1.73 0.91 4.66 2.32 1.35 1.07 4.54 2.21 1.12 0.65
See text for the linear equation used to scale the VOEs. Average between the VOEs of two MOs with the same symmetry.
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP: 195.99.248.194 On: Tue, 08 Apr 2014 09:51:22
184305-4
Modelli, Jones, and Pshenichnyuk
be sizeably destabilized, its VAE being 1.70 eV in aniline.41 In agreement, the first two measured VAEs (1.10 and 1.76 eV) are ascribed to electron capture into these two MOs, respectively, while the third resonance (4.93 eV) is associated with the mainly benzene b2g MO (totally antibonding), denoted π *O in Table I. To give further support to this qualitative interpretation of the spectral features of indoline and that of the ET spectra of derivatives II-V, we carried out theoretical calculations. An approach adequate for describing unstable anion states involves difficulties not encountered for neutral or cation states.42–45 The most correct method is, in principle, the calculation of the total scattering cross section with the use of continuum functions, but complications arise from the lack of an accurate description of the electron-molecule interaction.46 The VAE corresponding to the ground state anion can be obtained as the difference between the energy of the anion and that of the neutral state, both at the optimized geometry of the neutral species. A proper description of the spatially diffuse electron distributions of anions normally requires a basis set with diffuse functions.47, 48 However, calculated anion state energies decrease as the basis set is expanded, so the choice of an appropriate basis set is a delicate task.42, 43, 49, 50 The Koopmans’ theorem (KT) approximation51 neglects correlation and relaxation effects. However, it has been demonstrated42, 44 that good linear correlations exist between the π *C=C VAEs measured in a large number of alkenes and benzenoid hydrocarbons and the corresponding VOEs of the neutral molecules obtained with simple HF calculations, using basis sets which do not include diffuse functions. More recently it has been shown49 that the neutral state π * VOEs obtained with B3LYP/6-31G(d) calculations also supply a good linear correlation with the corresponding VAEs measured over a variety of different families of unsaturated compounds, including hetero-substituted compounds. A more accurate correlation is expected using “training” compounds structurally closer to the subject molecule. Here we use two linear correlations to scale the π * VOEs of compounds I-V. One (VAE = 0.64795 VOE − 1.429842 ) was derived from HF/6-31G(d) VOEs obtained for the geometries optimized at the MP2/6-31G(d) level. The second correlation (VAE = 0.8065 VOE + 0.919452 ) was derived from the B3LYP/631G(d) VOEs of π * MOs in a series of alternating phenyl and ethynyl groups. The VAEs of indoline (I) are closely reproduced by both sets of scaled VOEs (see Table I), although the destabilizing effect of the nitrogen lone pair on π *S is somewhat underestimated. The first feature displayed by the ET spectrum of indene (II) is due to the (not completely resolved) contributions of the mainly π *S and π *A benzene MOs, located at 0.78 and 1.20 eV, respectively. The first resonance, as observed in styrene,53 displays evidence of a vibrational structure, with a spacing of about 150 meV. This resonance results from electron capture into the LUMO, composed of the bonding combination of the π *S component of the benzene e2u MO with the ethene π * MO. Its antibonding counterpart gives rise to the third resonance (2.47 eV), while the second resonance (1.20 eV) is ascribed to the (only slightly perturbed) benzene π *A component. A representation of the localiza-
J. Chem. Phys. 139, 184305 (2013)
FIG. 2. Representation of the first three empty π * MOs of indene, as supplied by the HF/6-31G(d)//MP2/6-31G(d) calculations.
tion properties of the first three empty π * MOs of indene, as supplied by the HF/6-31G(d)//MP2/6-31G(d) calculations, is given in Figure 2 (essentially equal representations are obtained with the B3LYP/6-31G(d) method). The mainly benzene π *O MO gives rise to the fourth resonance, located at 4.7 eV. The scaled VOEs nicely reproduce the measured VAEs. The main discrepancy is concerned with an overevaluation of the π *O VAE. This finding is expected, due to possible significant mixing of the latter shape resonance with higher-lying core-excited resonances (electron capture accompanied by simultaneous excitation), obviously not accounted for by KT calculations. Comparison of the ET spectra of indene and its “opencycle” analogue styrene leads to an interesting observation: while the second and third π * anion states lie at the same energies, within experimental limits, the energy of the first anion state of indene (0.63 eV quoted at the first vibrational structure) is significantly larger than that (0.25 eV53 ) of its styrene counterpart. The smaller EAv of indene cannot be easily traced back to geometrical factors. According to the calculations, the distance between the substituted benzene carbon atom and the adjacent ethene carbon atom (where the LUMO is bonding) is somewhat shorter (0.005 Å) in indene than in styrene, while the ethene double bond (where the LUMO is antibonding) is shorter (0.008 Å) in styrene. These geometrical variations should result in a stabilization of the first anion state of indene relative to styrene, in contrast with observation. A possible explanation to the observed destabilization can be supplied by the calculated MO localization properties. As shown in Figure 2, the LUMO of indene receives a significant antibonding contribution from σ orbitals (of π symmetry) localized on the CH2 group, obviously not present in styrene. Replacement of the CH2 group of indene with an NH group leads to indole (III). Again, the scaled VOEs closely reproduce the VAEs. The first three values are equal within experimental limits to those previously reported by Aflatooni et al.10 Mixing with the nitrogen lone pair destabilizes the indene π * anion states. In particular, the mainly π *A anion which possesses the highest wave function coefficient on the adjacent carbon atom is destabilized by nearly 0.7 eV. In 2-methylen-1,3,3-trimethylindoline (IV), the benzene and ethene groups are not adjacent, and conjugation between their empty π * systems occurs through mixing with the intermediate nitrogen lone pair. The first two anion states possess mainly π *A and π *S character, respectively, while the third anion state is essentially localized only on the ethene group and the nitrogen atom, in an antibonding manner. Finally,
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP: 195.99.248.194 On: Tue, 08 Apr 2014 09:51:22
184305-5
Modelli, Jones, and Pshenichnyuk
2,3,3-trimethyl-indolenine (V), except for the methyl groups which only slightly perturb the π * anion energies, can be thought of as an indene molecule where an ethene CH group has been replaced by a more electronegative nitrogen atom. Such replacement is known to stabilize anion states localized at the site of substitution. For instance, the π *S component of pyridine54 is 0.4 eV more stable relative to benzene and the π *C=N anion of ditertbutylimine is stabilized by the same amount relative to the π *C=C anion of ethene.55 In agreement, the first VAE of V is 0.32 eV lower than that of indene, so that this compound is the most electron acceptor of the series. B. DEA spectra
The mass-selected negative ion yields measured in the DEA spectra of compounds I-III and IV-V in the 0–14 eV energy range are displayed in Figures 3 and 4, respectively. Their m/e ratios, peak energies, and relative intensities, as evaluated from the peak heights, are given in Table II. As a general comment, except for V, the number of negative fragments observed is small, and the intensities of these signals are relatively low. The thermodynamic energy thresholds obtained with B3LYP/6-31+G(d) calculations for the produc-
FIG. 3. Mass-selected signals of negative ions formed by DEA to gas-phase indoline (I), indene (II), and indole (III) as a function of the incident electron energy.
J. Chem. Phys. 139, 184305 (2013)
tion of the fragments are given in Table III, where the numerical labels refer to the fragment structures displayed in Chart 2. Table III also reports the energies of the vertical and adiabatic anion states relative to the corresponding neutral molecules. It can be noted that the first VAEs obtained with this method for I-V are in good agreement with the corresponding scaled VOEs (see Table I). In compounds I-IV, the most intense negative fragment (labeled [M-H]− ) corresponds to the loss of a hydrogen atom from the parent molecular anion. In compounds I-III, this signal peaks in the 1.5–2 eV energy range, close to the energy of the second resonance observed in their ET spectra and to the corresponding B3LYP/6-31+G(d) thermodynamic energy thresholds (see Table III) for the loss of a H atom from the NH (I and III) or CH2 (II) groups. In III, the [M-H]− species, resulting from the loss of hydrogen from the pyrrolic nitrogen, is virtually the only anion fragment detected below 10 eV electron impact energy. Breaking of a σ * N-H or C-H bond through electron addition to a π * MO can be explained in terms of σ */π * mixing caused by out-of-plane vibrations of the hydrogen atom. In fact, according to the calculated localization properties, the second empty π * MO of compounds I-III possesses a considerable wave-function coefficient on the benzene carbon atom adjacent to the N-H or C-H bond. However, the possibility that the [M-H]− signals arise directly from occupation of short-lived N-H or C-H σ * resonances (although too broad
FIG. 4. Mass-selected signals of negative ions formed by DEA to gas-phase 2-methylen-1,3,3-trimethylindoline (IV) and 2,3,3-trimethyl-indolenine (V). Isotopic contribution from the m/e = 143 negative fragment to the m/e = 144 signal in V is shown by a dotted line.
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP: 195.99.248.194 On: Tue, 08 Apr 2014 09:51:22
184305-6
Modelli, Jones, and Pshenichnyuk
J. Chem. Phys. 139, 184305 (2013)
TABLE II. Anion currents observed in the DEA spectra of compounds I-V, probable structures of the corresponding fragment negative ions, peak energies (eV), and relative intensities evaluated from the peak heights. m/e
Anion structure
Peak energy
TABLE III. B3LYP/6-31+G(d) energies (eV) relative to the neutral ground state of compounds I-V. The values in parentheses include zero-point vibrational energy corrections. Numbering refers to the fragment structures displayed in Chart 2.
Relative intensity m/e
Indoline (I)
119 119 118 93 93 93 92 91 91
Vertical anion Adiabatic anion [M–H]− + H• (from NH) [aniline]− + HC≡CH [M–CN]− + • C≡N (1) [M–CN]− + • C≡N (2) [toluene]− + HC≡N [C6 H4 NH]− + H2 C = CH2 [C6 H5 CH2 ]− + HCNH•
1.072 (0.959) 1.033 (0.930) 2.544 (2.106) 3.366 (3.013) 4.895 (4.487) 4.727 (4.266) 2.012 (1.505) 2.757 (2.397) 3.119 (2.669)
116 116 115 114 114 25
Indene (II) Vertical anion Adiabatic anion [M –H]− + H• (from CH2 ) [M–2H]− + H2 (3) [M–2H]− + H2 (4) HCC− + C6 H5 CH2 •
0.874 (0.679) 0.687 (0.492) 1.972 (1.575) 3.558 (3.086) 3.633 (3.162) 3.365 (3.023)
117 117 116 116 115 115
Indole (III) Vertical anion Adiabatic anion [M-H]− + H• (from NH) [M-H]− + H• (from C adj. to N) [M–2H]− + H2 (5) [M–2H]− + H2 (6)
0.888 (0.831) 0.867 (0.814) 1.800 (1.408) 3.375 (2.981) 3.642 (3.173) 3.680 (3.194)
Indoline (I) 118
[M–H]−
93
[M–HCCH]−
92
[M–CNH]−
91
[M–C2 H4 ]−
115
[M–H]−
2.0 4.8 7.8 3.8 5.9 4.0 8.0 3.3
100 36 29 27 36 39 15 17
1.5 5.0 7.8 9.0 8.5
100 5.7 90 13 3.1
1.6 10.0
100 0.3
Indene (II)
114 25
[M–2H]− C2 H−
116 115
[M–H]− [M–2H]−
Indole (III)
172 116 158 156 144 143 142 131
1,3,3-trimethyl-2-methyleneindoline (IV) 4.0 [M–H]− 8.4 9.0 [M–HN(CH3 )C(CH2 )H]− 2,3,3-trimethylindolenine (V) 8.0 [M–H]− 8.4 [M–3H]− [M–CH3 ]− 4.8 5.6 0.0 [M–CH4 ]− 8.5 8.2 [M–NH3 ]− 1.0 [M–C2 H4 ]−
19 100 24 19 100 7.7 7.7 0.8 11 8.2 9.5
to be detected in the ET spectra) cannot be excluded, as discussed for the dissociation of O-H56 and N-H57 bonds. The DEA spectrum of IV, where the lowest energy threshold for the loss of a H atom (from the N-CH3 group) is calculated to be >3 eV, displays [M-H]− signals at 4.0 and 8.0 eV, the latter being thus likely associated with coreexcited resonances. The [M-H]− negative ion is observed also in compound V, but only at high energy (8.0 eV) although the thermodynamic threshold for its formation is calculated to be much smaller (around 2 eV, see Table III). In addition, in contrast with compounds I-IV, in compound V the [M-H]− signal is not the most intense. The [M-3H]− fragment, formed by loss of a H atom and a H2 molecule, is by far the most abundant. The thermodynamic energy threshold for formation of this fragment anion is calculated to be large (>5.3 eV, see Table III), and in fact the corresponding signal peaks at 8.4 eV. The DEA spectrum of V displays two signals at low energy. Although very weak, a negative current with m/e = 143 is detected at zero energy, a more intense signal peaking at 8.5 eV. We assign this fragment negative ion to the loss of a methane neutral molecule from the parent molecular anion. In
173 173 172 172 143 116 159 159 158 156 156 144 144 143 142 131 131 26
2-methylen-1,3,3-trimethyl-indoline (IV) Vertical anion Adiabatic anion [M-H]− + H• (from NCH3 (7) [M-H]− + H• (from C = CH2 ) [M-C2 H6 ]− + C2 H6 (8) [M-57]− + HN(CH3 )C(CH2 )H (9) 2,3,3-trimethyl-indolenine (V) Vertical anion Adiabatic anion [M-H]− + H• (10) [M-3H]− + H2 + H• (11) [M-3H]− + H2 + H• (12) [M-CH3 ]− + CH3 • (13) [M-CH3 ]− + CH3 • (14) [M-CH4 ]− + CH4 (15) [M-NH3 ]− + NH3 (16) [M-C2 H4 ]− + H2 C = CH2 (17) [M-NCH2 ]− + NCH2 • (18) CN− + [M-CN]• (19)
0.901 (0.834) 0.894 (0.823) 3.684 (3.253) 3.819 (3.395) 1.421 (1.176) 2.509 (2.244)
0.560 (0.366) 0.330 (0.153) 2.300 (1.891) 6.113 (5.341) 6.569 (5.739) 0.782 (0.423) 2.987 (2.621) − 0.214 (−0.511) 3.197 (2.949) 1.348 (1.216) 3.786 (3.401) 1.019 (0.774)
agreement, the calculations predict the products of this decay channel to be more stable than the neutral molecule V (see Table III). The zero energy signal could be associated with dissociation of the low-energy portion of the first resonance observed in the ET spectrum (VAE = 0.46 eV), or alternatively with formation and dissociation of a dipole bound state, in line with a large dipole moment (2.557 D) calculated for this molecule. The other low-energy signal (m/e = 131) peaks at 1.0 eV, the same energy of the second resonance displayed by the ET spectrum, and is likely due to the loss of an ethene
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP: 195.99.248.194 On: Tue, 08 Apr 2014 09:51:22
184305-7
Modelli, Jones, and Pshenichnyuk
J. Chem. Phys. 139, 184305 (2013)
CH3
CH2
H
NH
H
(1)
(3)
(2)
(4)
(5)
CH2
CH2
N
N
CH2
NH
(6)
CH2
(8)
(7)
N
(10)
(9)
CH2
(11)
CH2
N
N
(12)
(13)
(14)
(17)
N
(15)
CH
N
(16)
N
N
(18)
(19) CHART 2.
neutral molecule from the parent anion. In this regard, the calculated localization properties predict a significant delocalization of the LUMO (and to a lesser extent of the second LUMO) on the carbon atoms of the two geminal methyl groups of V, but a node on the ring carbon atom to which they are attached. It can be noted that the above mentioned m/e = 143 and 131 signals in V and the [M-H]− negative ions observed in I-III are the only fragment anions produced at low energy (
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP: 195.99.248.194 On: Tue, 08 Apr 2014 09:51:22
THE JOURNAL OF CHEMICAL PHYSICS 139, 184305 (2013)
Electron attachment to indole and related molecules Alberto Modelli,1,2,a) Derek Jones,3,b) and Stanislav A. Pshenichnyuk4,c) 1
Dipartimento di Chimica “G. Ciamician”, Universitá di Bologna, via Selmi 2, 40126 Bologna, Italy Centro Interdipartimentale di Ricerca in Scienze Ambientali (CIRSA), Universitá di Bologna, via S. Alberto 163, 48123 Ravenna, Italy 3 ISOF, Istituto per la Sintesi Organica e la Fotoreattività, C.N.R., via Gobetti 101, 40129 Bologna, Italy 4 Institute of Molecule and Crystal Physics, Ufa Research Centre, Russian Academy of Sciences, Prospeκt Oktyabrya 151, 450075 Ufa, Russia 2
(Received 24 September 2013; accepted 23 October 2013; published online 11 November 2013) Gas-phase formation of temporary negative ion states via resonance attachment of low-energy (0–6 eV) electrons into vacant molecular orbitals of indoline (I), indene (II), indole (III), 2-methylen1,3,3-trimethylindoline (IV), and 2,3,3-trimethyl-indolenine (V) was investigated for the first time by electron transmission spectroscopy (ETS). The description of their empty-level structures was supported by density functional theory and Hartree-Fock calculations, using empirically calibrated linear equations to scale the calculated virtual orbital energies. Dissociative electron attachment spectroscopy (DEAS) was used to measure the fragment anion yields generated through dissociative decay channels of the parent molecular anions of compounds I-V, detected with a mass filter as a function of the incident electron energy in the 0–14 eV energy range. The vertical and adiabatic electron affinities were evaluated at the B3LYP/6-31+G(d) level as the anion/neutral total energy difference. The same theoretical method is also used for evaluation of the thermodynamic energy thresholds for production of the negative fragments observed in the DEA spectra. The loss of a hydrogen atom from the parent molecular anion ([M-H]− ) provides the most intense signal in compounds I-IV. The gasphase DEAS data can provide support for biochemical reaction mechanisms in vivo involving initial hydrogen abstraction from the nitrogen atom of the indole moiety, present in a variety of biologically important molecules. © 2013 AIP Publishing LLC. [http://dx.doi.org/10.1063/1.4829057] I. INTRODUCTION 1
Since the original paper by Gregory in 1966, the two “dissimilar fields of research” (activity in vivo of molecular species and their gas-phase interactions with electrons) have offered a number of important studies aimed at a more complete understanding of molecular activities in living organisms.2–10 Many molecular species or their analogues can, in particular, be studied using gas phase electron transmission and dissociative attachment spectroscopies (ETS, DEAS) which allow experimental determination of empty level structures of these molecular species through electron attachment and the subsequent destiny of their temporary anion states.11–13 Recent studies by this group using these techniques have also shed light on the mechanisms involved in the specific activities of commonly used drugs including antipyretics14 (aspirin, paracetamol, phenacetin, and ibuprofen), the anti-malarial drug artemisinin and its derivatives15, 16 and flavonoids.17 Here we turn our attention to one of the most ubiquitous chemical moieties in living organisms. The indole moiety is not only present in almost all living organisms but also plays
a) Author to whom correspondence should be addressed. Electronic mail:
[email protected]
b) E-mail: [email protected] c) E-mail: [email protected]
0021-9606/2013/139(18)/184305/9/$30.00
a decisive role in a fascinatingly wide variety of important biochemical reactions.18 Its interaction, for example, with iron in heme-containing oxygenase enzymes19 is an important step in the insertion of oxygen into organic substrates, crucial for aerobic organisms, as in the case of human indoleamine 2,3-dioxygenase (IDO) which cleaves the indole pyrrole ring of L-tryptophan to incorporate both atoms of oxygen following the first step of this process which involves the interaction of the 2,3 bond of the indole ring with the iron-bound dioxygen.20 Both of the proposed mechanisms (ionic or radical) depend upon the free form of the indole NH and its electronic properties.19 The indole moiety also plays a vital role in radical scavenging and antioxidant activities of melatonin21 and its precursors serotonin and tryptophan where once again the 2,3 bond of the indole pyrrole ring is protagonist.22 The importance of the indole moiety has been confirmed by calculations showing that the 5-methoxy and N-acetyl groups of melatonin do not significantly affect its thermodynamic capacity for free-radical trapping.23 It is, in fact, well-known that the 3-position on the pyrrole ring of indole is the preferred site for electrophilic attack,24 even in the presence of 3-substituents, the 2-position being energetically unfavourable as it would involve an intermediate with perturbation of the π -electron system of the benzene ring. This is an important aspect for understanding some of the mechanisms involved in indole alkaloid biosynthesis.
139, 184305-1
© 2013 AIP Publishing LLC
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP: 195.99.248.194 On: Tue, 08 Apr 2014 09:51:22
184305-2
Modelli, Jones, and Pshenichnyuk
J. Chem. Phys. 139, 184305 (2013)
CHART 1.
Although the filled-level electronic structure of indole,25, 26 indoline,27 indene,28, 29 and 2-methylen-1,3,3trimethylindoline30 have been published, experimental determination of their energies of electron attachment and the various pathways of the resulting anion decay, with possible molecular dissociation of such states, has not been previously investigated, except for the temporary anion states of indole formed in the 0–3.5 eV energy range.10 The electron transmission spectroscopy (ETS) technique devised by Sanche and Schulz11 is still one of the most suitable means for detecting gas-phase temporary anion states, which appear as sharp structures, “resonances,” in their electron-molecule scattering cross sections. Here this technique is applied for the first time to detect the formation of temporary anion states and describe the empty-level structure of indoline (I), indene (II), indole (III), 2-methylen-1,3,3trimethylindoline (IV), and 2,3,3-trimethyl-indolenine (V). Chart 1 reports their structural formulae. Additional information on the fate of temporary molecular anions can be supplied by dissociative electron attachment (DEA) spectroscopy12, 13 which measures the yield of massselected negative ions as a function of the impact electron energy. When suitable energetic conditions apply, the decay of unstable molecular anions formed by resonance can follow a dissociative channel which generates a negative fragment and a neutral radical, in kinetic competition with simple reemission of the extra electron. The negative ion currents measured in the DEA spectra of compounds I-V, in the 0–14 eV energy range, are presented and the structures of the corresponding fragments are proposed with the support of density functional theory (DFT) calculations of the thermodynamic energy thresholds required for their production. The same theoretical method is employed to evaluate the vertical (EAv ) and adiabatic (EAa ) electron affinities of compounds I-V.
II. EXPERIMENTAL METHODS
Our electron transmission apparatus (Bologna Laboratory) is in the format devised by Sanche and Schulz11 and has been previously described.31, 32 To enhance the visibility of the sharp resonance structures, the impact energy of the electron beam is modulated with a small ac voltage, and the derivative of the electron current transmitted through the gas sample is measured directly by a synchronous lock-in amplifier. Each resonance is characterized by a minimum and a maximum in the derivative signal. The energy of the midpoint between these features is assigned as the vertical attachment energy (VAE). The ET spectra were obtained using the apparatus in the “high-rejection” mode33 and the signal is, there-
fore, related to the nearly total scattering cross section. The electron beam resolution was about 50 meV full width at half maximum (FWHM). The energy scale was calibrated with reference to the (1s1 2s2 ) 2 S anion state of He. The estimated accuracy is ±0.05 or ±0.1 eV, depending on the number of decimal digits reported. In order to detect the weakest signals from the negative fragments produced by DEA decay channels, although with a lower energy resolution of the incident electron beam, a negative ion magnetic mass spectrometer (Ufa Laboratory), described in detail previously34, 35 was used. Briefly, an electron beam of defined energy is passed through a collision cell containing a vapor of the substance under investigation, under single-collision conditions. A current of magnetically massselected negative ions is recorded as a function of the incident electron energy in the 0–14 eV energy range. The electron energy scale is calibrated with the zero energy SF6 − signal. The FWHM of the electron energy distribution is 0.4 eV, and the accuracy of the measured peak positions is estimated to be ±0.1 eV. The analyzed samples were commercially available from Aldrich and used without further purification. The collision cells of both instruments were held at 60◦ C–80◦ C. Calculations were carried out with the GAUSSIAN 09 set of programs.36 Evaluation of the virtual orbital energies (VOEs) of the neutral molecules was performed with the 6-31G(d) basis set at the Hartree-Fock (HF), with the MP2 optimized geometries, and B3LYP levels of theory.37 The vertical electron affinity (EAv ) was calculated as the difference between the total energy of the neutral and the lowest anion state, both in the optimized geometry of the neutral state, using the B3LYP hybrid functional with the standard 6-31+G(d) basis set, which includes the minimum addition of diffuse functions (s and p type diffuse functions to the nonhydrogen atoms). The adiabatic electron affinity (EAa ) was obtained as the energy difference between the neutral and the lowest anion state, each in its optimized geometry. The same theoretical method was also used to evaluate the thermodynamic energy thresholds for production of fragment anions, obtained as the total energy of the (geometrically relaxed) ground-state fragments relative to the ground neutral state molecule. Zero-point vibrational energy corrections were also calculated. III. RESULTS AND DISCUSSION A. Electron transmission spectrum and calculated VAEs
The present family of bicyclic compounds II-V possesses four empty π * MOs, deriving from the three benzene π * MOs
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP: 195.99.248.194 On: Tue, 08 Apr 2014 09:51:22
184305-3
Modelli, Jones, and Pshenichnyuk
J. Chem. Phys. 139, 184305 (2013)
and an ethene π * MO of the pentacyclic ring, while indoline (I), where the latter MO is not present, possesses only the three π * MOs with mainly benzene character. The π * VAEs measured in the ET spectra of the reference molecules benzene31 (1.12 eV, e2u ; 4.82 eV, b2g ) and ethene38 (1.73 eV) indicate that electron capture into the empty π * MOs of compounds I-V is expected to occur at energies lower than 6 eV. Systematic ETS studies39 have demonstrated that hydrocarbons not substituted by third-row or heavier heteroatoms generally do not give rise to distinct low-energy resonances associated with empty MOs of (local) σ symmetry, so that detection of signals associated with electron capture into empty σ * MOs are not expected in this series, although a broad and weak feature associated with a σ * resonance was observed in the ET spectrum of the hexacycle piperidine.40 The ET spectra of compounds I-V in the 0–6 eV energy range are reported in Figure 1. The measured VAEs, together with calculated virtual orbital energies (VOEs), are given in Table I. In indoline, the degeneracy of the e2u lowest unoccupied molecular orbital (LUMO) is removed by mixing with the adjacent nitrogen lone pair. The component (here denoted as π *A ) with a node at the substituted carbon atom is expected to be only slightly perturbed, whereas the component (π *S ) with maximum atomic coefficient on that carbon atom should
FIG. 1. Derivative of transmitted current, as a function of electron energy, in gas-phase indoline (I), indene (II), indole (III), 2-methylen-1,3,3trimethylindoline (IV), and 2,3,3-trimethyl-indolenine (V). Vertical lines locate the VAEs.
TABLE I. VOEs calculated at the HF/6-31G(d)//MP2/6-31G(d) and B3LYP/6-31G(d) levels, scaleda VOEs and measured VAEs. All values in eV. HF//MP2
Indoline (I)
Indene (II)
Indole (III)
2-methylen-1,3,3trimethyl-indoline (IV)
2,3,3-trimethylindolenine (V)
a b
B3LYP
Expt.
Orbital
VOE
Sc. VOE
VOE
Sc. VOE
VAE
π *O σ *NH π *S - n N π *A π *O σ *CH π *CC - π *S π *A π *S + π *CC π *O π *CC - nN σ *NH π *- nN π *S + π *CC π *O σ *CH π *CC - nN π *S - π *CC π *A π *O σ *CH π *CN - π *S π *A π *S + π *CN
9.662 6.393 4.526 3.944 9.887b 6.764 6.027 4.013 3.286 10.663 6.526 6.129 4.882 3.610 9.397 6.051 5.789 4.296 3.856 9.219 5.715 5.624 3.935 3.216
4.83
5.083 2.434 0.707 0.228 4.993b 2.850 1.614 0.159 − 0.483 5.160 2.092 2.183 0.948 − 0.066 4.263 2.304 1.515 0.415 0.097 4.424 2.061 1.263 0.084 − 0.704
5.02
4.93
1.49 1.10 5.23
1.76 1.10 4.7
2.22 1.05 0.53 5.08 2.61
2.47 1.20 0.78 4.7 2.78
1.68 0.87 4.36
1.88 0.99 4.44
2.14 1.25 1.00 4.49
2.28 1.36 1.04 4.55
1.94 0.99 0.35
2.26 1.05 0.46
1.50 1.13 4.98 2.48 1.17 0.70 5.48 2.80 1.73 0.91 4.66 2.32 1.35 1.07 4.54 2.21 1.12 0.65
See text for the linear equation used to scale the VOEs. Average between the VOEs of two MOs with the same symmetry.
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP: 195.99.248.194 On: Tue, 08 Apr 2014 09:51:22
184305-4
Modelli, Jones, and Pshenichnyuk
be sizeably destabilized, its VAE being 1.70 eV in aniline.41 In agreement, the first two measured VAEs (1.10 and 1.76 eV) are ascribed to electron capture into these two MOs, respectively, while the third resonance (4.93 eV) is associated with the mainly benzene b2g MO (totally antibonding), denoted π *O in Table I. To give further support to this qualitative interpretation of the spectral features of indoline and that of the ET spectra of derivatives II-V, we carried out theoretical calculations. An approach adequate for describing unstable anion states involves difficulties not encountered for neutral or cation states.42–45 The most correct method is, in principle, the calculation of the total scattering cross section with the use of continuum functions, but complications arise from the lack of an accurate description of the electron-molecule interaction.46 The VAE corresponding to the ground state anion can be obtained as the difference between the energy of the anion and that of the neutral state, both at the optimized geometry of the neutral species. A proper description of the spatially diffuse electron distributions of anions normally requires a basis set with diffuse functions.47, 48 However, calculated anion state energies decrease as the basis set is expanded, so the choice of an appropriate basis set is a delicate task.42, 43, 49, 50 The Koopmans’ theorem (KT) approximation51 neglects correlation and relaxation effects. However, it has been demonstrated42, 44 that good linear correlations exist between the π *C=C VAEs measured in a large number of alkenes and benzenoid hydrocarbons and the corresponding VOEs of the neutral molecules obtained with simple HF calculations, using basis sets which do not include diffuse functions. More recently it has been shown49 that the neutral state π * VOEs obtained with B3LYP/6-31G(d) calculations also supply a good linear correlation with the corresponding VAEs measured over a variety of different families of unsaturated compounds, including hetero-substituted compounds. A more accurate correlation is expected using “training” compounds structurally closer to the subject molecule. Here we use two linear correlations to scale the π * VOEs of compounds I-V. One (VAE = 0.64795 VOE − 1.429842 ) was derived from HF/6-31G(d) VOEs obtained for the geometries optimized at the MP2/6-31G(d) level. The second correlation (VAE = 0.8065 VOE + 0.919452 ) was derived from the B3LYP/631G(d) VOEs of π * MOs in a series of alternating phenyl and ethynyl groups. The VAEs of indoline (I) are closely reproduced by both sets of scaled VOEs (see Table I), although the destabilizing effect of the nitrogen lone pair on π *S is somewhat underestimated. The first feature displayed by the ET spectrum of indene (II) is due to the (not completely resolved) contributions of the mainly π *S and π *A benzene MOs, located at 0.78 and 1.20 eV, respectively. The first resonance, as observed in styrene,53 displays evidence of a vibrational structure, with a spacing of about 150 meV. This resonance results from electron capture into the LUMO, composed of the bonding combination of the π *S component of the benzene e2u MO with the ethene π * MO. Its antibonding counterpart gives rise to the third resonance (2.47 eV), while the second resonance (1.20 eV) is ascribed to the (only slightly perturbed) benzene π *A component. A representation of the localiza-
J. Chem. Phys. 139, 184305 (2013)
FIG. 2. Representation of the first three empty π * MOs of indene, as supplied by the HF/6-31G(d)//MP2/6-31G(d) calculations.
tion properties of the first three empty π * MOs of indene, as supplied by the HF/6-31G(d)//MP2/6-31G(d) calculations, is given in Figure 2 (essentially equal representations are obtained with the B3LYP/6-31G(d) method). The mainly benzene π *O MO gives rise to the fourth resonance, located at 4.7 eV. The scaled VOEs nicely reproduce the measured VAEs. The main discrepancy is concerned with an overevaluation of the π *O VAE. This finding is expected, due to possible significant mixing of the latter shape resonance with higher-lying core-excited resonances (electron capture accompanied by simultaneous excitation), obviously not accounted for by KT calculations. Comparison of the ET spectra of indene and its “opencycle” analogue styrene leads to an interesting observation: while the second and third π * anion states lie at the same energies, within experimental limits, the energy of the first anion state of indene (0.63 eV quoted at the first vibrational structure) is significantly larger than that (0.25 eV53 ) of its styrene counterpart. The smaller EAv of indene cannot be easily traced back to geometrical factors. According to the calculations, the distance between the substituted benzene carbon atom and the adjacent ethene carbon atom (where the LUMO is bonding) is somewhat shorter (0.005 Å) in indene than in styrene, while the ethene double bond (where the LUMO is antibonding) is shorter (0.008 Å) in styrene. These geometrical variations should result in a stabilization of the first anion state of indene relative to styrene, in contrast with observation. A possible explanation to the observed destabilization can be supplied by the calculated MO localization properties. As shown in Figure 2, the LUMO of indene receives a significant antibonding contribution from σ orbitals (of π symmetry) localized on the CH2 group, obviously not present in styrene. Replacement of the CH2 group of indene with an NH group leads to indole (III). Again, the scaled VOEs closely reproduce the VAEs. The first three values are equal within experimental limits to those previously reported by Aflatooni et al.10 Mixing with the nitrogen lone pair destabilizes the indene π * anion states. In particular, the mainly π *A anion which possesses the highest wave function coefficient on the adjacent carbon atom is destabilized by nearly 0.7 eV. In 2-methylen-1,3,3-trimethylindoline (IV), the benzene and ethene groups are not adjacent, and conjugation between their empty π * systems occurs through mixing with the intermediate nitrogen lone pair. The first two anion states possess mainly π *A and π *S character, respectively, while the third anion state is essentially localized only on the ethene group and the nitrogen atom, in an antibonding manner. Finally,
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP: 195.99.248.194 On: Tue, 08 Apr 2014 09:51:22
184305-5
Modelli, Jones, and Pshenichnyuk
2,3,3-trimethyl-indolenine (V), except for the methyl groups which only slightly perturb the π * anion energies, can be thought of as an indene molecule where an ethene CH group has been replaced by a more electronegative nitrogen atom. Such replacement is known to stabilize anion states localized at the site of substitution. For instance, the π *S component of pyridine54 is 0.4 eV more stable relative to benzene and the π *C=N anion of ditertbutylimine is stabilized by the same amount relative to the π *C=C anion of ethene.55 In agreement, the first VAE of V is 0.32 eV lower than that of indene, so that this compound is the most electron acceptor of the series. B. DEA spectra
The mass-selected negative ion yields measured in the DEA spectra of compounds I-III and IV-V in the 0–14 eV energy range are displayed in Figures 3 and 4, respectively. Their m/e ratios, peak energies, and relative intensities, as evaluated from the peak heights, are given in Table II. As a general comment, except for V, the number of negative fragments observed is small, and the intensities of these signals are relatively low. The thermodynamic energy thresholds obtained with B3LYP/6-31+G(d) calculations for the produc-
FIG. 3. Mass-selected signals of negative ions formed by DEA to gas-phase indoline (I), indene (II), and indole (III) as a function of the incident electron energy.
J. Chem. Phys. 139, 184305 (2013)
tion of the fragments are given in Table III, where the numerical labels refer to the fragment structures displayed in Chart 2. Table III also reports the energies of the vertical and adiabatic anion states relative to the corresponding neutral molecules. It can be noted that the first VAEs obtained with this method for I-V are in good agreement with the corresponding scaled VOEs (see Table I). In compounds I-IV, the most intense negative fragment (labeled [M-H]− ) corresponds to the loss of a hydrogen atom from the parent molecular anion. In compounds I-III, this signal peaks in the 1.5–2 eV energy range, close to the energy of the second resonance observed in their ET spectra and to the corresponding B3LYP/6-31+G(d) thermodynamic energy thresholds (see Table III) for the loss of a H atom from the NH (I and III) or CH2 (II) groups. In III, the [M-H]− species, resulting from the loss of hydrogen from the pyrrolic nitrogen, is virtually the only anion fragment detected below 10 eV electron impact energy. Breaking of a σ * N-H or C-H bond through electron addition to a π * MO can be explained in terms of σ */π * mixing caused by out-of-plane vibrations of the hydrogen atom. In fact, according to the calculated localization properties, the second empty π * MO of compounds I-III possesses a considerable wave-function coefficient on the benzene carbon atom adjacent to the N-H or C-H bond. However, the possibility that the [M-H]− signals arise directly from occupation of short-lived N-H or C-H σ * resonances (although too broad
FIG. 4. Mass-selected signals of negative ions formed by DEA to gas-phase 2-methylen-1,3,3-trimethylindoline (IV) and 2,3,3-trimethyl-indolenine (V). Isotopic contribution from the m/e = 143 negative fragment to the m/e = 144 signal in V is shown by a dotted line.
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP: 195.99.248.194 On: Tue, 08 Apr 2014 09:51:22
184305-6
Modelli, Jones, and Pshenichnyuk
J. Chem. Phys. 139, 184305 (2013)
TABLE II. Anion currents observed in the DEA spectra of compounds I-V, probable structures of the corresponding fragment negative ions, peak energies (eV), and relative intensities evaluated from the peak heights. m/e
Anion structure
Peak energy
TABLE III. B3LYP/6-31+G(d) energies (eV) relative to the neutral ground state of compounds I-V. The values in parentheses include zero-point vibrational energy corrections. Numbering refers to the fragment structures displayed in Chart 2.
Relative intensity m/e
Indoline (I)
119 119 118 93 93 93 92 91 91
Vertical anion Adiabatic anion [M–H]− + H• (from NH) [aniline]− + HC≡CH [M–CN]− + • C≡N (1) [M–CN]− + • C≡N (2) [toluene]− + HC≡N [C6 H4 NH]− + H2 C = CH2 [C6 H5 CH2 ]− + HCNH•
1.072 (0.959) 1.033 (0.930) 2.544 (2.106) 3.366 (3.013) 4.895 (4.487) 4.727 (4.266) 2.012 (1.505) 2.757 (2.397) 3.119 (2.669)
116 116 115 114 114 25
Indene (II) Vertical anion Adiabatic anion [M –H]− + H• (from CH2 ) [M–2H]− + H2 (3) [M–2H]− + H2 (4) HCC− + C6 H5 CH2 •
0.874 (0.679) 0.687 (0.492) 1.972 (1.575) 3.558 (3.086) 3.633 (3.162) 3.365 (3.023)
117 117 116 116 115 115
Indole (III) Vertical anion Adiabatic anion [M-H]− + H• (from NH) [M-H]− + H• (from C adj. to N) [M–2H]− + H2 (5) [M–2H]− + H2 (6)
0.888 (0.831) 0.867 (0.814) 1.800 (1.408) 3.375 (2.981) 3.642 (3.173) 3.680 (3.194)
Indoline (I) 118
[M–H]−
93
[M–HCCH]−
92
[M–CNH]−
91
[M–C2 H4 ]−
115
[M–H]−
2.0 4.8 7.8 3.8 5.9 4.0 8.0 3.3
100 36 29 27 36 39 15 17
1.5 5.0 7.8 9.0 8.5
100 5.7 90 13 3.1
1.6 10.0
100 0.3
Indene (II)
114 25
[M–2H]− C2 H−
116 115
[M–H]− [M–2H]−
Indole (III)
172 116 158 156 144 143 142 131
1,3,3-trimethyl-2-methyleneindoline (IV) 4.0 [M–H]− 8.4 9.0 [M–HN(CH3 )C(CH2 )H]− 2,3,3-trimethylindolenine (V) 8.0 [M–H]− 8.4 [M–3H]− [M–CH3 ]− 4.8 5.6 0.0 [M–CH4 ]− 8.5 8.2 [M–NH3 ]− 1.0 [M–C2 H4 ]−
19 100 24 19 100 7.7 7.7 0.8 11 8.2 9.5
to be detected in the ET spectra) cannot be excluded, as discussed for the dissociation of O-H56 and N-H57 bonds. The DEA spectrum of IV, where the lowest energy threshold for the loss of a H atom (from the N-CH3 group) is calculated to be >3 eV, displays [M-H]− signals at 4.0 and 8.0 eV, the latter being thus likely associated with coreexcited resonances. The [M-H]− negative ion is observed also in compound V, but only at high energy (8.0 eV) although the thermodynamic threshold for its formation is calculated to be much smaller (around 2 eV, see Table III). In addition, in contrast with compounds I-IV, in compound V the [M-H]− signal is not the most intense. The [M-3H]− fragment, formed by loss of a H atom and a H2 molecule, is by far the most abundant. The thermodynamic energy threshold for formation of this fragment anion is calculated to be large (>5.3 eV, see Table III), and in fact the corresponding signal peaks at 8.4 eV. The DEA spectrum of V displays two signals at low energy. Although very weak, a negative current with m/e = 143 is detected at zero energy, a more intense signal peaking at 8.5 eV. We assign this fragment negative ion to the loss of a methane neutral molecule from the parent molecular anion. In
173 173 172 172 143 116 159 159 158 156 156 144 144 143 142 131 131 26
2-methylen-1,3,3-trimethyl-indoline (IV) Vertical anion Adiabatic anion [M-H]− + H• (from NCH3 (7) [M-H]− + H• (from C = CH2 ) [M-C2 H6 ]− + C2 H6 (8) [M-57]− + HN(CH3 )C(CH2 )H (9) 2,3,3-trimethyl-indolenine (V) Vertical anion Adiabatic anion [M-H]− + H• (10) [M-3H]− + H2 + H• (11) [M-3H]− + H2 + H• (12) [M-CH3 ]− + CH3 • (13) [M-CH3 ]− + CH3 • (14) [M-CH4 ]− + CH4 (15) [M-NH3 ]− + NH3 (16) [M-C2 H4 ]− + H2 C = CH2 (17) [M-NCH2 ]− + NCH2 • (18) CN− + [M-CN]• (19)
0.901 (0.834) 0.894 (0.823) 3.684 (3.253) 3.819 (3.395) 1.421 (1.176) 2.509 (2.244)
0.560 (0.366) 0.330 (0.153) 2.300 (1.891) 6.113 (5.341) 6.569 (5.739) 0.782 (0.423) 2.987 (2.621) − 0.214 (−0.511) 3.197 (2.949) 1.348 (1.216) 3.786 (3.401) 1.019 (0.774)
agreement, the calculations predict the products of this decay channel to be more stable than the neutral molecule V (see Table III). The zero energy signal could be associated with dissociation of the low-energy portion of the first resonance observed in the ET spectrum (VAE = 0.46 eV), or alternatively with formation and dissociation of a dipole bound state, in line with a large dipole moment (2.557 D) calculated for this molecule. The other low-energy signal (m/e = 131) peaks at 1.0 eV, the same energy of the second resonance displayed by the ET spectrum, and is likely due to the loss of an ethene
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP: 195.99.248.194 On: Tue, 08 Apr 2014 09:51:22
184305-7
Modelli, Jones, and Pshenichnyuk
J. Chem. Phys. 139, 184305 (2013)
CH3
CH2
H
NH
H
(1)
(3)
(2)
(4)
(5)
CH2
CH2
N
N
CH2
NH
(6)
CH2
(8)
(7)
N
(10)
(9)
CH2
(11)
CH2
N
N
(12)
(13)
(14)
(17)
N
(15)
CH
N
(16)
N
N
(18)
(19) CHART 2.
neutral molecule from the parent anion. In this regard, the calculated localization properties predict a significant delocalization of the LUMO (and to a lesser extent of the second LUMO) on the carbon atoms of the two geminal methyl groups of V, but a node on the ring carbon atom to which they are attached. It can be noted that the above mentioned m/e = 143 and 131 signals in V and the [M-H]− negative ions observed in I-III are the only fragment anions produced at low energy (
