THE JOURNAL OF CHEMICAL PHYSICS 127, 144312 共2007兲
2-amino-1-propanol versus 1-amino-2-propanol: Valence band and C 1s core-level photoelectron spectra D. Catone,a兲 S. Turchini, G. Contini, N. Zema, S. Irrera, and T. Prosperi Istituto di Struttura della Materia (ISM), Consiglio Nazionale delle Ricerche (CNR), Via del Fosso del Cavaliere 100, 00133 Rome, Italy
M. Stener, D. Di Tommaso, and P. Decleva Dipartimento di Scienze Chimiche, Università di Trieste, Via L. Giorgieri 1, Trieste, Italy, Consorzio Interuniversitario Nazionale per la Scienza e Tecnologia dei Materiali (INSTM), Unità di Trieste, Trieste, Italy and INFM DEMOCRITOS National Simulation Center, Trieste, Italy
共Received 19 July 2007; accepted 19 September 2007; published online 12 October 2007兲 Valence band and C 1s core-level photoelectron spectra of S-共+兲-2-amino-1-propanol 共alaninol兲 and S-共+兲-1-amino-2-propanol 共isopropanolamine兲 have been studied by means of synchrotron radiation photoelectron spectroscopy in gas phase. The alaninol, the reduced derivative of the alanine, is a good test system of amino acid-like structures. The isopropanolamine, presenting the inversion of the two functional groups of the alaninol at the chiral carbon, offers the opportunity to study the effect of –OH and –NH2 structural position on the photoelectron spectra. The influence of the conformational contribution on the electronic structure and the photoelectron spectra has been interpreted using density functional and ab initio theoretical calculations. Agreement has been achieved by taking into account the presence, in gas phase, of two conformers with different population ratios in both chiral systems. The C 1s core-level spectra of alaninol and isopropanolamine are reported and the peak positions of the three carbon atoms of the molecules are assigned. © 2007 American Institute of Physics. 关DOI: 10.1063/1.2798113兴 I. INTRODUCTION
Chiral amino alcohol derivatives are important constituents of molecules which play a role in life-related processes such as transmission of nerve impulse or building of cell membranes. These particular systems possess a flexible backbone whose conformation strongly depends on its stereochemistry and on the environment. A special role is played by an intramolecular hydrogen bond between the –OH and –NH2 groups, which characterize several conformers, whose relative stability is dictated by intra- as well as intermolecular forces, and which in turn may determine the interaction mode of the molecule with its biological surroundings. In fact, the simultaneous presence of the amino and hydroxyl groups in the same molecule, both able to act as a H-bond donor and acceptor, opens a wide range of possible hydrogen bond interactions, in particular, when the groups are close to each other. To have a better understanding on this class of molecules, 2-amino-1-propanol 共alaninol兲 and 1-amino-2propanol 共isopropanolamine兲 were studied in gas phase by means of photoelectron spectroscopy 共PES兲 technique. Alaninol can be used as a good test system of amino acidlike structures and, with respect to alanine, has the advantage of being liquid at room temperature, allowing an easier evaporation. Isopropanolamine, presenting the inversion of the two functional groups of alaninol at the chiral carbon, offers the opportunity to study the influence of –OH and a兲
Author to whom correspondence should be addressed. Electronic mail:
[email protected]
0021-9606/2007/127共14兲/144312/10/$23.00
–NH2 structural position on the photoelectron spectra. Figure 1 reports the structures of the first two most stable conformers of both chiral systems. The PES technique provides important spectroscopic information, which should be helpful for other kinds of promising investigation methodologies and this paper intends to be the first step toward a deeper understanding on these particular chiral systems. One of the most innovative spectroscopy is definitely the circular dichroism in angular distribu-
FIG. 1. The calculated structures of the two lowest energy conformers for alaninol 共A1 and A2兲 and isopropanolamine 共I1 and I2兲.
127, 144312-1
© 2007 American Institute of Physics
Downloaded 15 Oct 2007 to 140.105.55.84. Redistribution subject to AIP license or copyright, see http://jcp.aip.org/jcp/copyright.jsp
144312-2
Catone et al.
tion 共CDAD兲,1–6 which requires a detailed understanding of the distribution of valence and core states of chiral molecules in order to explain how the circular polarized light in the VUV and soft x-ray ranges can interact with such asymmetric systems, shedding light on the origin of homochirality in the living matter. There is in fact the real opportunity to guide the discovery of bio-organic species in the interstellar medium by interpretation of the molecular structures of these systems in gas phase conditions. The knowledge of the electronic properties in solvent-free conditions is also an important prelude to interpret structural and electronic modifications due to the liquid environment or to follow the adsorption of chiral molecules on a substrate. The nanoscopic understanding, afforded by advanced surface science experiments, opens a new way to the heterogeneous enantioselective catalysis, nonlinear optical devices, biosensors, or smart coatings.7,8 Moreover the weak interactions in the van der Waals diastereomeric clusters,9,10 which simulate well the molecular and chiral recognition processes, so important in regulating several biological activities, can be related to the electronic properties of the same systems in the isolated conditions and can be explained on the basis of the information obtained from PES studies. In this paper the synchrotron radiation photoelectron spectra of alaninol and isopropanolamine, acquired at different photon energies, are presented and discussed with the help of theoretical calculations, and used to evaluate the most stable conformers in the experimental conditions, to estimate the vertical ionization energies for both chiral systems, and to simulate the valence band photoelectron spectra. II. EXPERIMENTAL DETAILS AND DATA ANALYSIS
The experimental data have been recorded on the POLAR beamline at Elettra synchrotron radiation facility with the angular resolved photoelectron spectroscopy end station.6 A variable polarization wiggler/undulator11 provides photons in the range between 5 and 900 eV. The beamline is equipped with a grazing incident spherical grating monochromator 共SGM兲 and a normal incident monochromator 共NIM兲 sharing the same entrance and exit slits.12 This setup provides both linearly and circularly polarized light over the entire spectral range. All the data in this work have been collected with circularly polarized light and with two hemispherical electron analyzers mounted at the magic angle of 54.7° in the forward/backward geometry with respect to the synchrotron radiation propagation direction. These particular conditions assure that the measurements should be insensitive to the  asymmetry parameter. Commercial samples of S-共+兲-2-amino-1-propanol and S-共+兲-1-amino-2-propanol 共Aldrich兲 have been purified by repeated sublimation cycles before introducing them in the vacuum experimental chamber. The sample pressure has been kept constant at 2 ⫻ 10−4 Pa, thermalizing the effusive jet source at 330 K. To ascertain that at the used temperature no molecule decomposition occurs, the experiment was monitored by mass spectrometry showing no modification upon evaporation. Valence photoelectron spectra have been recorded at two photon energies: h = 22.5 eV with NIM and h = 50.0 eV
J. Chem. Phys. 127, 144312 共2007兲
with SGM. The valence band spectra have been obtained with the beamline slits set at 200 m and the electron analyzer pass energy at 25 eV, corresponding to an overall resolution of 120 meV. C 1s core-level spectra have been recorded at h = 310.0 eV photon energy in SGM mode, with slit aperture at 100 m and pass energy at 25 eV; these conditions yield a resolution of about 250 meV. The kinetic energy scale has been calibrated against the ionization line of Ar 3p atomic state and of the CO C 1s for the valence and core-level acquisitions, respectively. All the spectra shown in this paper are the result of the sum of several scans sequentially recorded with right and left circularly polarized light, each one normalized to the incident photon intensity in order to cancel any dichroic intensity contributions which might arise with an enantiomerically pure sample. III. COMPUTATIONAL DETAILS
For calculation, the molecular structures of the two lowest energy conformers both for alaninol 关g⬘Gg⬘共A1兲, the most stable, and gG⬘g共A2兲 from Fausto et al.13兴 and isopropanolamine 关the most stable gG⬘g共I1兲 and g⬘Gg⬘共I2兲 from Cacela et al.14兴 have been considered 共see Fig. 1兲. For both molecules and conformers these geometries have been optimized with the ADF program,15 employing a density functional method with the Vosko-Wilk-Nusair 共VWN兲 local density approximation16 exchange correlation energy functional and a basis set of Slater-type orbitals of double zeta potential 共DZP兲 type. The resulting internal coordinates have been found to be in excellent agreement with the MP2 results.13,14 In order to have a better estimation of the energy differences between the conformers, an additional single point calculation at the optimized geometry has been performed, but employing the Becke-Perdew generalized gradient approximation17,18 exchange and correlation energy functionals. Therefore energy gaps between the two lowest energy conformers of 2.05 and 4.80 kJ/ mol, respectively, for alaninol and isopropanolamine have been obtained. These energy gaps have been employed to calculate the relative population according to Boltzmann distribution; at 330 K the ratio 1:0.48 for A1 : A2 共68%:32%兲 concentration of alaninol and the ratio 1:0.18 for the I1 : I2 共85%:15%兲 concentration of isopropanolamine were found. The ionization energies 共IEs兲 have been calculated with a DZP basis set as the Kohn-Sham 共KS兲 eigenvalues at the optimized geometries, employing the LB94 exchange correlation potential,19 which has the correct asymptotic Coulomb behavior. The same electron density obtained with the LB94 potential is also employed to build the KS Hamiltonian represented with multicentric 关linear combination of atomic orbitals 共LCAO兲兴 B-spline basis functions, employed to calculate continuum unbound orbitals from which the cross section profiles are calculated via conventional dipole transition moments. Such cross section profiles have been successively employed to simulate the experimental valence band photoelectron spectra. The computational scheme of the continuum B-spline method has been already described elsewhere in detail.20,21 The crucial choice for the LCAO con-
Downloaded 15 Oct 2007 to 140.105.55.84. Redistribution subject to AIP license or copyright, see http://jcp.aip.org/jcp/copyright.jsp
144312-3
PES of amino-propanols
FIG. 2. Experimental valence photoelectron spectrum of alaninol, recorded at h = 50 eV. The calculated IEs for the two more stable conformers 共A1 and A2兲 are reported with two series of bars. The 11 theoretical IEs at lower energies have been calculated with the OVGF method, while the other 5 have been calculated with the KS-LB94 method. The calculated data are summarized in Table I.
tinuum B-spline calculations is the maximum angular momentum 共Lmaxo兲 employed in the expansion on the origin, which has been set on the nucleus of the asymmetric carbon atom. Lmaxo = 16 has proven an optimum choice for converging results with respect to the basis set size for both molecules considered in this work. The maximum angular momentum 共Lmaxi兲 employed in the off-center expansions has been fixed to Lmaxi = 2 for C and O and Lmaxi = 1 for N and H. The expansion on the origin is divided with a radial grid with step size of 0.2 a.u. up to 20 a.u.; the intervals so obtained are supplemented with additional knots at the nucleus positions in order to make the basis more flexible in the region of core orbitals. The expansion on the off-center nuclei is divided with a five-step radial linear grid up to 1.2 a.u. for C, N, and O, while for H the expansion ranges from 0.678 to 0.88 a.u. depending on the distance of the atom from the one center expansion and from adjacent atoms. Additional Hartree-Fock 共HF兲 and outer valence Green function 共OVGF兲 calculations, to discuss the nature of the outer valence region, have been performed with the GAUSSIAN program,22 employing a Dunning cc-pVDZ basis set. In the case of core IE calculation the KS-LB94 results have been supplemented with the more appropriate ⌬SCF results employing the VWN exchange correlation potential with the same DZP basis set, obtained as energy differences between the neutral molecule and the system with the core hole at the given atomic site. This procedure has the advantage of taking into account the large relaxation energy involved during the core hole formation. IV. RESULTS AND DISCUSSION A. Alaninol: Valence band and C 1s core-level spectra
The experimental valence band photoelectron spectrum of alaninol, recorded at photon energy h = 50 eV and displayed in Fig. 2, shows a detailed overview of the outer valence region, together with the calculated IEs for the two
J. Chem. Phys. 127, 144312 共2007兲
more stable conformers 共A1 and A2兲, marked with two series of bars. The 11 theoretical IEs at lower energies have been calculated with the OVGF method, while the other 5 have been calculated with the KS-LB94 method. All the calculated and experimental energy positions are summarized in Table I. The experimental spectrum shows two sharp and well resolved peaks at 9.45± 0.05 and 10.30± 0.05 eV, highest occupied molecular orbital 共HOMO兲 and HOMO-1, respectively, followed by an extended region of overlapped photoelectron bands between 11.00 and 18.50 eV and five resolved peaks at IE higher than 18.50 eV 共19.30± 0.05, 21.70± 0.05, 23.60± 0.05, 27.60± 0.05, and 31.40± 0.05 eV兲, completing all the 16 valence states of alaninol. The experimental valence photoelectron spectrum of alaninol recorded at 22.5 eV photon energy is reported in Fig. 3共a兲 together with the fit results using 11 Voigt functions. The ionization energy values of each Voigt peak were obtained by fitting at the same time several photoelectron spectra acquired at different photon energies 共not shown here兲 and using the indications of the theoretical results. In Fig. 3, the theoretical simulations of the valence photoelectron spectrum obtained in the cases of bare A1 关Fig. 3共c兲兴, bare A2 关Fig. 3共d兲兴, and mixing of the two conformers A1 + A2 with abundance of 68% and 32%, respectively 关Fig. 3共b兲兴, are reported. The theoretical simulated spectra have been generated using the IEs calculated by the OVGF method and the corresponding cross sections calculated by the KS-LB94 method 共reported in Fig. 3 as vertical bars兲, convoluted with Voigt functions to achieve a realistic simulation of the experimental data. The comparison between theory and experiment shows that the simulated spectrum obtained for A1 + A2 reproduces the experimental data better than those obtained for A1 and A2, although some similarities can be found in the A1 spectrum. Important differences between the simulated spectra of A1 and A1 + A2 can be pointed out in the region between the 12.5 and 13.5 eV, where A1 does not present states while A1 + A2 shows an envelope of three states assigned to A2, and in the region around 17.0 eV where A1 presents no states while the A1 + A2 shows two separated peaks due to A2. Afterward all the simulations confirm that the experimental spectrum is strongly characterized by the more abundant conformer A1; for this reason the fit procedure has been performed with one series of bands that mainly reflects the energetics of A1. The state overlapping in the region between 11.0 and 18.5 eV of ionization energy prevents from unambiguously identifying every single state of the two conformers, separating their different contributions. In particular, the fitted bands from the 5th to the 11th do not represent a single electronic state of the molecule, but they were used only to point out the main features of the photoelectron spectrum and are characterized by an indiscernible superposition of electronic states due to A1 and A2, as theoretical calculations predict. Table I reports the calculated ionization energies for A1 and A2, together with the experimental IEs obtained from the fitted spectrum reported in Fig. 3; the energies relative to the lower IE peak is reported in parentheses. Three different levels of theory have been considered for calculating IEs: The KS-LB94 column reports the negative eigenvalues calculated at the DFT
Downloaded 15 Oct 2007 to 140.105.55.84. Redistribution subject to AIP license or copyright, see http://jcp.aip.org/jcp/copyright.jsp
144312-4
J. Chem. Phys. 127, 144312 共2007兲
Catone et al.
TABLE I. Theoretical KS-LB94, KT, and OVGF for A1 and A2 conformers and experimental valence state IEs 共eV兲 of alaninol. Theory KS-LB94a MO HOMO HOMO-1 HOMO-2 HOMO-3 HOMO-4 HOMO-5 HOMO-6 HOMO-7 HOMO-8 HOMO-9 HOMO-10 HOMO-11 HOMO-12 HOMO-13 HOMO-14 HOMO-15
KTa
OVGF
A1
A2
A1
A2
A1
A2
11.15 共0.00兲c 11.54 共0.39兲 12.69 共1.54兲 13.45 共2.30兲 14.48 共3.33兲 14.77 共3.62兲 15.11 共3.96兲 15.93 共4.78兲 16.32 共5.17兲 16.77 共5.62兲 18.22 共7.07兲 19.62 共8.47兲 21.39 共10.24兲 23.57 共12.42兲 27.36 共16.21兲 30.31 共19.16兲
11.15 共0.00兲c 11.51 共0.36兲 12.70 共1.55兲 13.74 共2.59兲 14.00 共2.85兲 14.16 共3.01兲 14.91 共3.76兲 16.34 共5.19兲 16.52 共5.37兲 17.22 共6.07兲 17.64 共6.49兲 19.30 共8.15兲 21.52 共10.37兲 23.42 共12.77兲 27.32 共16.17兲 30.33 共19.18兲
11.02 共0.00兲c 11.46 共0.44兲 12.61 共1.59兲 13.42 共2.40兲 14.69 共3.67兲 15.00 共3.98兲 15.57 共4.55兲 16.66 共5.64兲 16.99 共5.97兲 17.31 共6.29兲 19.22 共8.20兲 21.79 共10.77兲 24.79 共13.77兲 27.74 共16.72兲 32.59 共21.57兲 36.53 共25.51兲
10.98 共0.00兲c 11.45 共0.47兲 12.64 共1.66兲 13.79 共2.81兲 14.11 共3.13兲 14.47 共3.49兲 15.22 共4.24兲 16.92 共5.94兲 17.16 共6.18兲 18.26 共7.28兲 18.59 共7.61兲 21.38 共10.40兲 27.93 共13.95兲 27.56 共16.58兲 32.54 共21.56兲 36.56 共25.58兲
9.67 共0.00兲c 10.12 共0.45兲 11.45 共1.78兲 12.43 共2.76兲 13.48 共3.81兲 14.02 共4.35兲 14.48 共4.81兲 15.30 共5.63兲 15.75 共6.08兲 16.02 共6.35兲 17.89 共8.22兲 ¯
9.70 共0.00兲c 10.13 共0.43兲 11.48 共1.78兲 12.54 共2.84兲 13.18 共3.48兲 13.44 共3.74兲 14.04 共4.34兲 15.82 共6.12兲 15.85 共6.15兲 16.92 共7.22兲 17.20 共7.50兲 ¯
¯
¯
¯
¯
¯
¯
¯
¯
Expt. 共±0.05 eV兲 9.45 共0.00兲c 10.33b 共0.88兲 11.61b 共2.16兲 12.42b 共2.97兲 13.30b 共3.85兲 13.87b 共4.42兲 14.47b 共5.02兲 15.17b 共5.72兲 15.90b 共6.45兲 16.71b 共7.26兲 17.74b 共8.29兲 19.30d 共9.85兲 21.70d 共12.25兲 23.60d 共14.15兲 27.60d 共18.15兲 31.41d 共21.95兲
a
The opposite orbital eigenvalues. Experimental IE resulting from the fit of the photoelectron spectrum acquired at 22.5 eV photon energy 关see Fig. 3共a兲兴. c The relative energy with respect to HOMO is reported in parentheses 共eV兲. d Experimental IE obtained from the photoelectron spectrum acquired at 50.0 eV photon energy 共see Fig. 2兲. b
KS level with LB94 exchange correlation potential, the KT column reports Koopmans’ theorem results, and finally OVGF results are considered for orbitals from HOMO down to HOMO-10. Considering the energy differences for inner orbital IEs with respect to the HOMO, the overall quality between the different levels of theory is similar; in fact for HOMO-1 orbital, all methods give an error of about 0.4 eV, while for HOMO-2, the error is around 0.4 eV for OVGF and 0.6 eV for KS-LB94 and KT. However, the comparison between the HOMO absolute IE and the experimental data indicates that, as expected, the OVGF method gives the best results, with a deviation of less than 0.3 eV. KS and KT results, on the other hand, deviate by about 1.6 eV. In Table II the KS-LB94 and HF molecular orbital Mulliken populations of the valence states of alaninol for A1 and A2 conformers are reported only for atomic orbitals with a population higher than 10%. In the case of the KS-LB94 level of theory, the HOMO and HOMO-1 states in A1 and A2 conformers are strongly characterized by the 2p orbitals
of hydroxyl oxygen and amino nitrogen, respectively. The states from HOMO-2 to HOMO-10 are largely marked out by the chain carbon atoms orbitals. The 2p oxygen orbitals present a stronger delocalization on a large number of states with respect to 2p nitrogen orbitals that participate only at HOMO-1, HOMO-8, and HOMO-10 for both conformers. Comparing the Mulliken population found for the two conformers A1 and A2, only very small differences for the three outer orbitals 共from HOMO to HOMO-2兲 and for the five innermost valence orbitals with main 2s contribution 共from HOMO-11 to HOMO-15兲 can be observed; on the other hand the nature of the molecular orbitals between HOMO-3 and HOMO-9 is strongly affected by conformational changes. This can be rationalized by considering that the molecular orbitals, which are delocalized on the whole structure, are more sensitive to conformational changes than molecular orbitals which lie on moieties and are not affected by structural changes. It is quite interesting to compare the KS-LB94 results with the Hartree-Fock population reported in Table II.
Downloaded 15 Oct 2007 to 140.105.55.84. Redistribution subject to AIP license or copyright, see http://jcp.aip.org/jcp/copyright.jsp
144312-5
J. Chem. Phys. 127, 144312 共2007兲
PES of amino-propanols
FIG. 3. Experimental valence photoelectron spectrum of alaninol, recorded at h = 22.5 eV, together with the results of the fit 共a兲. The theoretical simulations of the photoelectron spectrum of alaninol obtained in the cases of the bare A1 共c兲, the bare A2 共d兲, and mixing of the two conformers 共A1 + A2兲 with the weight of 68% for A1 and 32% for A2, respectively 共b兲. The calculated and experimental data are summarized in Table I.
The most striking finding is the inversion of the nature of the HOMO and HOMO-1 orbitals, which at the Hartree-Fock level have predominant N 2p and O 2p contributions, respectively. This observation indicates that the assignment of
HOMO and HOMO-1 remains uncertain at present. Direct comparison of the CDAD experimental behavior of these two states with the computational results, obtained for KSLB94 and Hartree-Fock wave functions, may result in a promising way to solve this assignment uncertainty. In fact the dichroism is a dynamical property that is very sensitive to the orbital character, and therefore a comparison between experimental and calculated dichroisms may give some helpful information on band assignment. The comparison between the KS-LB94 and Hartree-Fock molecular orbital populations for the other valence orbitals does not show relevant differences between the two methods for both A1 and A2 conformations, concluding that the most pronounced and important difference is limited to the two outermost orbitals. The experimental C 1s photoelectron spectrum of alaninol acquired at 310.0 eV of incident photon energy and the fit performed using three Voigt functions with equal widths 共Gaussian and Lorentzian兲 are reported in Fig. 4, together with the C 1s state theoretical IE values calculated with the KS-LB94 method both for A1 and A2 conformers 共marked with two series of bars兲. The experimental spectrum appears as a broad band centered at 291.5 eV and was fitted using three peaks, ascribed to the more stable conformer A1, principally because the contribution of the less abundant conformer A2 can be neglected for the same reasons already discussed in the valence case. The C 1s IEs predicted by KS-LB94 calculations for A1 and A2, compared with the experimental IEs, are summarized in Table III.
TABLE II. Calculated KS-LB94 and Hartree-Fock molecular orbital Mulliken populations of alaninol for A1 and A2 conformers. KS-LB94a MO HOMO HOMO-1 HOMO-2 HOMO-3 HOMO-4 HOMO-5 HOMO-6 HOMO-7 HOMO-8
A1
Hartree-Focka A2
A1
A2
22% O 2p, 43% N 2p 24% N 2p, 36% O 2p 29% O 2p, 24% H 1s, 14% C2-2p, 19% C1-2p 44% H 1s, 10% O 共2p兲 +20% C22p, 16% C3-2p 16% C3-2p, 23% H 1s, 22% O 共2p兲 28% C3-2p, 14% O 2p, 40% H 1s
16% O 2p, 49% N 2p 15% N 2p, 42% O 2p 28% O 2p, 15% C2-2p + 17% C1-2p 28% C3-2p, 26% C2-2p, 13% O 2p 29% C3-2p, 13% N2p, 44% H 1s 32% C3-2p, 16% O 2p, 41% H 1s 33% O 2p, 23% H 1s, 19% C1-2p 12% C1-2p, 27% H 1s, 22% N 2p 16% O 2p, 11% C2-2p, 27% C1-2p, 30% H 24% C2-2p, 24% N 2p, 31% H 1s 12% C2-2p, 15% O 2p, 22% N 2p, 30% H 1s 14% C2-2s, 11% N 2p 24% C1-2s, 19% C3-2s 23% C3-2s, 18% C2-2s, 12% N 2s 49% N 2s, 18% C2-2s 69% O 2s
57% O 2p, 11% N 2p 52% N 2p, 14% O 2p 35% O 2p, 17% H 1s, 13% C3-2p, 12% C2-2p 38% H 1s, 18% C2-2p, 12% C32p 26% C3-2p, 23% H 1s, 15% O 共2p兲 17% C3-2p, 15% O 2p, 28% H 1s
52% O 2p, 16% N 2p 48% N 2p, 17% O 2p 35% O 2p, 13% C2-2p + 12% C3-2p 26% C3-2p, 22% C2-2p, 11% O 2p 30% C3-2p, 43% H 1s
30% C3-2p, 11% O 2p, 22% H 共Me兲 1s 19% C3-2p, 19% C2-2p, 15% O 2p 26% N 2p, 11% C1-2p
29% O 2p, 21% H 1s, 20% C1-2p 30% C3-2p, 12% O 2p, 37% H
33% C3-2p, 36% H 1s
22% C1-2p, 31% H 1s, 12% C32p 26% O 2p, 20% N 2p
HOMO-9
25% O 2p, 16% N 2p, 15% C22p + 11% C1-2p HOMO-10 29% N 2p, 19% C2-2p
35% C2-2p, 10% C3-2p, 20% H 1s 37% N 2p, 14% H 1s
HOMO-11 17% C2-2s, 15% N 2p HOMO-12 23% C1-2s, 23% C3-2s HOMO-13 24% C3-2s, 19% C2-2s, 11% N 2s HOMO-14 51% N 2s, 17% C2-2s HOMO-15 68% O 2s
15% C2-2s, 14% N 2p 24% C1-2s, 21% C3-2s 22% C3-2s, 21% C2-2s, 10% N 2s 51% N 2s, 17% C2-2s 68% O 2s
12% C3-2p, 11% C2-2p, 19% C1-2p, 23% O 2p 24% N 2p, 12% C2-2p, 12% C12p, 12% O 2p 23% O 2p, 20% N 2p, 12% C22p 22% N 2p, 15% C2-2p, 12% Cl2p, 13% O 2p 14% C2-2s, 12% N 2p 23% C1-2s, 21% C3-2s 23% C3-2s, 18% C2-2s, 13% N 2s 49% N 2s, 18% C2-2s 69% O 2s
a
The atomic orbitals with a population higher than 10% are reported.
Downloaded 15 Oct 2007 to 140.105.55.84. Redistribution subject to AIP license or copyright, see http://jcp.aip.org/jcp/copyright.jsp
144312-6
J. Chem. Phys. 127, 144312 共2007兲
Catone et al.
FIG. 4. Experimental C 1s core-level photoelectron spectrum of alaninol, recorded at h = 310.0 eV, together with the results of the fit 共G = 0.41; ⌫L = 0.20兲. The C 1s theoretical IE values 共KS-LB94 method兲 for both A1 and A2 are marked with two series of bars. The calculated and experimental data are summarized in Table III.
The assignment of the three signals to each carbon state can be obtained by comparison with previous PES studies on alanine,23 2-butanol,24 ethanol,25 glycine,26,27 and other amino acids28,29 in gas phase. This assignment has been also corroborated by the theoretical calculations which reproduce well the experimental absolute and relative energies 共see Table III兲. The three different peaks at 290.7± 0.1, 291.4± 0.1, and 292.0± 0.1 eV were assigned to the carbon atom of the methyl group 共C3兲, the chiral carbon bound to the amino group 共C2兲, and the carbon bound to the hydroxyl group 共Cl兲, respectively 共see Fig. 1 for carbon atom numbering兲. Comparing the ⌬SCF results for A1 with the experimental data, a good agreement for the energy shift between C2 and C3 共calculated: 0.84 eV; measured: 0.7 eV兲 has been found. However, the calculated energy shift between C1 and C2 is very small 共less than 0.1 eV兲 while the experimental value is 0.6 eV. Because for A2 similar calculated IE values are obtained, it is unlikely that an energy shift of 0.6 eV between C1 and C2 is due to the presence of two conform-
FIG. 5. Experimental valence photoelectron spectrum of isopropanolamine, recorded at h = 50 eV. The calculated IEs for the two more stable conformers 共I1 and I2兲 are reported with two series of bars. The 11 theoretical IEs at lower energies have been calculated with the OVGF method, while the other 5 have been calculated with the KS-LB94 method. The calculated data are summarized in Table IV.
ers. The KS-LB94 results give a slightly larger energy shift between C2 and C3, but in this case the IE of C2 is higher than the IE of C1; anyhow the differences are so small that it is possible to assume that both theoretical methods 共⌬SCF and KS-LB94兲 give in practice the same IE for both C1 and C2. B. Isopropanolamine: Valence band and C 1s core-level spectra
The experimental valence band photoelectron spectrum, recorded at h = 50 eV photon energy, and the calculated IEs for the two more stable conformers of isopropanolamine 共I1 and I2兲, marked with two series of bars, are reported in Fig. 5. The 11 theoretical IEs at lower energies have been calculated with the OVGF method, while the other 5 have been calculated with the KS-LB94 method, as in the alaninol case. The experimental photoelectron spectrum, including all the 16 valence states of isopropanolamine, shows at low IE two well resolved peaks at 9.65± 0.05 eV 共HOMO兲 and
TABLE III. Theoretical KS-LB94 and ⌬SCF 共VWN兲 for A1 and A2 conformers and experimental C 1s, IEs 共eV兲 of alaninol. ⌬SCF
KS-LB94a A1
A2
A1
A2
Expt.b 共±0.1 eV兲
c
290.52 共0.00兲d
290.29 共0.00兲d
294.48 共0.00兲d
294.29 共0.00兲d
290.7 共0.0兲d
C2c
291.54 共1.02兲
291.53 共1.24兲
295.32 共0.84兲
295.32 共1.03兲
291.4 共0.7兲
C1c
291.42 共0.90兲
291.44 共1.15兲
295.37 共0.89兲
295.37 共1.08兲
292.0 共1.3兲
C 1s C3
a
The opposite orbital eigenvalues. Experimental IE resulting from the fit of the photoelectron spectrum acquired at 310.0 eV photon energy 共see Fig. 4兲. c For numbering of the carbon atoms, see Fig. 1. d The relative energy with respect to the lowest value is reported in parentheses 共eV兲. b
Downloaded 15 Oct 2007 to 140.105.55.84. Redistribution subject to AIP license or copyright, see http://jcp.aip.org/jcp/copyright.jsp
144312-7
PES of amino-propanols
FIG. 6. Experimental valence photoelectron spectrum of isopropanolamine, recorded at h = 22.5 eV, together with the results of the fit 共a兲. The theoretical simulations of the photoelectron spectrum of isopropanolamine obtained in the cases of the bare I1 共c兲, the bare I2 共d兲, and mixing of the two conformers 共I1 + I2兲 with the weight of 85% for I1 and 15% for A2 共b兲. The calculated and experimental data are summarized in Table IV.
10.40± 0.05 eV 共HOMO-1兲, then a first group of overlapped bands between 11.00 and 14.50 eV, separated by a gap at about 15.00 eV from a second group that extends to 18.50 eV, followed by five resolved peaks at 19.35± 0.05, 21.65± 0.05, 24.05± 0.05, 27.65± 0.05, and 31.70± 0.05 eV. The experimental valence photoelectron spectrum of isopropanolamine taken at 22.5 eV photon energy is reported in Fig. 6共a兲, together with the fit performed with Voigt functions. In the same figure the theoretical simulations of the photoelectron spectrum obtained in the cases of bare I1 关Fig. 6共c兲兴, bare I2 关Fig. 6共d兲兴, and thermal mixing of the two conformers I1 + I2 with the abundance of 85% and 15%, respectively 关Fig. 6共b兲兴, are reported. The method used to obtain the fit and simulated spectra has already been discussed for alaninol 共Sec. IV A兲. The theoretical simulated valence spectrum of I1 conformer looks particularly close to the experiment with respect to that obtained for I2. In fact, the calculated profile for I1 reproduces well the gap at about 15.0 eV that is not present in the simulation of I2. The simulation of the mixed conformers I1 + I2, which reproduces better the spectral envelope of the experimental data, is very similar to the bare I1 one because the low abundance of I2 cannot heavily affect the I1 + I2 profile. An important difference can be pointed out only in the gap region between 14.0 and 16.0 eV, where I1 does not present any state while I1 + I2 shows a little contribution of two states assigned to I2. The shape of the experimental spectrum is less congested with respect to the alaninol one, probably because isopropanolamine has a different conformational population that affects in an unequal way the overlapping of the states in the two spectra 关Fig. 3共a兲 and
J. Chem. Phys. 127, 144312 共2007兲
Fig. 6共a兲兴. As a matter of fact, the I1 energetics strongly characterizes the main features of the isopropanolamine photoelectron valence spectrum, as predicted by the theoretical simulations. For this reason the fit procedure has been performed with one series of peaks that preferentially reflects the energetics of conformer I1. The general agreement between the experimental spectrum and the theoretical calculation suggests a clearer correspondence between the deconvoluted peaks and the spectroscopic features of the more abundant conformer I1, with respect to the already discussed alaninol case. Table IV reports the calculated ionization energies for I1 and I2, together with the experimental IEs obtained from the fitted peaks in the photoelectron spectrum of isopropanolamine; the energies relative to the lower IE peak are reported in parentheses. In Table V the KS-LB94 and HF molecular orbital Mulliken populations of the valence states of isopropanolamine for I1 and I2 conformers are reported only for atomic orbitals with a population higher than 10%. At the KS-LB94 level of theory the HOMO and HOMO-1 states in I1 and I2 are strongly characterized by the 2p orbitals of hydroxyl oxygen and by the 2p orbitals of amino nitrogen, respectively, as already found in the case of alaninol, whereas the states from HOMO-2 to HOMO-10 are largely marked out by the chain carbon atom 2p orbitals. The oxygen presents a stronger orbital delocalization on a large number of electronic states of both conformers, with respect to nitrogen that is delocalized only in the inner valence states, i.e., from HOMO-7 to HOMO-10 and from HOMO-6 to HOMO-10 in the cases of I1 and I2, respectively. In analogy with the alaninol case, very small differences are observed for the three outermost HOMO, HOMO-1, and HOMO-2 orbitals in I1 and I2, as well as for the inner valence orbitals with main 2s character. As in the case of alaninol, the comparison between the KSLB94 and Hartree-Fock molecular orbital populations for the HOMO-2 to HOMO-15 states does not show relevant differences between the two methods for both I1 and I2 conformations, whereas the HOMO and HOMO-1 states still present inverted localization for the 2p oxygen and nitrogen orbitals, preventing a clear assignment for these two states. The experimental C 1s photoelectron spectrum of isopropanolamine acquired at 310.0 eV of incident photon energy and the fit performed using three Voigt functions with equal widths 共Gaussian and Lorentzian兲 are reported in Fig. 7, together with the C 1s state theoretical IE values calculated with the KS-LB94 method both for I1 and I2 conformers and marked with two series of bars. The experimental spectrum presents two separated photoelectron bands centered at 290.3 and 291.5 eV. The theory at both KS-LB94 and ⌬SCF levels predicts that the C 1s states 共Table VI兲 for I1 and I2 have very close IEs, such as to be not clearly distinguishable. Moreover, the low abundance of the less stable conformer I2 permits to fit the photoelectron profile with a single series of three Voigts, principally ascribed to I1. The three peaks at 290.3± 0.1, 291.2± 0.1, and 291.7± 0.1 eV were assigned to the carbon of the methyl group 共C3兲, the carbon bound to the amino group 共C1兲, and the chiral carbon bound to the hydroxyl group 共C2兲, respec-
Downloaded 15 Oct 2007 to 140.105.55.84. Redistribution subject to AIP license or copyright, see http://jcp.aip.org/jcp/copyright.jsp
144312-8
J. Chem. Phys. 127, 144312 共2007兲
Catone et al.
TABLE IV. Theoretical KS-LB94, KT, and OVGF for I1 and I2 conformers and experimental valence state IEs 共eV兲 of isopropanolamine. Theory KS-LB94a
KTa
OVGF
I1
I2
I1
I2
I1
I2
Expt. 共±0.05 eV兲
HOMO
11.10 共0.00兲c
11.11 共0.00兲c
10.92 共0.00兲c
10.94 共0.00兲c
9.64 共0.00兲c
9.62 共0.00兲c
9.56b 共0.00兲c
HOMO-1
11.59 共0.49兲
11.60 共0.49兲
11.48 共0.56兲
11.56 共0.62兲
10.18 共0.54兲
10.15 共0.53兲
10.40 共0.75兲
HOMO-2
12.78 共1.68兲
12.61 共1.50兲
12.68 共1.76兲
12.47 共1.53兲
11.42 共1.79兲
11.37 共1.75兲
11.62b 共1.97兲
HOMO-3
13.51 共2.41兲
13.41 共2.30兲
13.73 共2.81兲
13.49 共2.55兲
12.57 共2.93兲
12.54 共2.83兲
12.60b 共2.95兲
HOMO-4
14.11 共3.01兲
13.77 共2.66兲
14.48 共3.56兲
15.01 共4.07兲
13.34 共3.70兲
14.03 共4.41兲
13.87b 共4.22兲
HOMO-6
14.44 共3.34兲
15.36 共4.25兲
14.65 共3.37兲
15.64 共4.70兲
13.57 共3.94兲
14.53 共4.91兲
14.52b 共4.87兲
HOMO-7
16.41 共5.31兲
15.56 共4.45兲
16.98 共6.06兲
16.19 共5.25兲
15.71 共6.07兲
14.82 共5.21兲
15.47b 共5.82兲
HOMO-8
16.73 共5.63兲
16.58 共5.47兲
17.37 共6.45兲
17.14 共6.20兲
16.23 共6.59兲
15.93 共6.23兲
16.03b 共6.38兲
HOMO-9
16.80 共5.70兲
16.96 共5.85兲
17.81 共6.89兲
17.91 共6.97兲
16.33 共6.69兲
16.52 共6.91兲
16.57b 共6.92兲
HOMO-10
17.99 共6.89兲
17.97 共6.86兲
18.88 共7.96兲
18.96 共8.02兲
17.62 共7.98兲
17.63 共8.01兲
17.48b 共7.83兲
HOMO-11
19.17 共8.07兲
19.25 共8.14兲
21.33 共10.41兲
21.31 共10.37兲
¯
¯
19.35d 共9.70兲
HOMO-12
21.35 共10.25兲
21.49 共10.38兲
24.66 共13.74兲
24.87 共13.93兲
¯
¯
21.65d 共12.00兲
HOMO-13
23.62 共12.52兲
23.45 共12.34兲
27.88 共16.96兲
27.68 共16.74兲
¯
¯
24.05d 共14.40兲
HOMO-14
27.22 共16.12兲
27.22 共16.11兲
32.37 共21.45兲
32.37 共21.43兲
¯
¯
27.65d 共18.00兲
HOMO-15
30.32 共19.22兲
30.28 共19.17兲
36.54 共25.62兲
36.51 共25.57兲
¯
¯
31.07d 共22.05兲
MO
a
The opposite orbital eigenvalues. Experimental IE resulting from the fit of the photoelectron spectrum acquired at 22.5 eV photon energy 关see Fig. 6共a兲兴. c The relative energy with respect to HOMO is reported in parentheses 共eV兲. d Experimental IE obtained from the photoelectron spectrum acquired at 50.0 eV photon energy 共see Fig. 5兲. b
tively 共see Fig. 1 for carbon atom numbering兲, following the same arguments discussed in Sec. IV A. Comparing the ⌬SCF results for I1 with the experimental data, a good agreement for the energy shift between C1 and C3 共calculated: 1.08 eV; measured: 0.9 eV兲 and for C2 and C3 共calculated: 1.37 eV; measured: 1.4 eV兲 has been found. The ⌬SCF results for I2 are very close to those found in the I1 case: 1.10 eV for the energy shift between C1 and C3, and 1.48 eV for the energy shift between C2 and C3. The energy separation among the three states, the fine overlap of the bands of I1 and I2, and the high abundance of the more stable conformer lead to a better resolved photoelectron spectrum with respect to the alaninol one. V. CONCLUSIONS
The synchrotron radiation gas phase photoelectron spectra of alaninol and isopropanolamine were measured for the
first time and discussed on the basis of DFT and ab initio calculations. The satisfactory agreement between the experimental valence spectra and the theoretical calculations for both molecular systems offers the possibility to discuss the PES profiles but not to give a definitive identification of the complete electronic structure. The presence of the valence state ionization energy overlapping due to the conformational contribution prevents a clear assignment. The different shapes of the valence band photoelectron spectra found for alaninol and isopropanolamine can be mostly ascribed to the different conformational populations of the two systems. In fact, alaninol has a different weight for the first two more stable conformers 共68% for A1 versus 32% for A2兲 with respect to isopropanolamine 共85% for I1 versus 15% for I2兲, which presents a quasimonoconformational population and characterizes its spectra with a less pronounced overlapped profile with respect to alaninol. The theoretical calculations have
Downloaded 15 Oct 2007 to 140.105.55.84. Redistribution subject to AIP license or copyright, see http://jcp.aip.org/jcp/copyright.jsp
144312-9
J. Chem. Phys. 127, 144312 共2007兲
PES of amino-propanols
TABLE V. Calculated KS-LB94 and Hartree-Fock molecular orbital Mulliken populations of isopropanolamine for I1 and I2 conformers. KS-LB94a I1
MO
Hartree-Focka I2
I1
I2
HOMO
50% O 2p, 15% N 2p
52% O 2p, 16% N 2p
37% N 2p, 24% O 2p
43% N 2p, 22% O 2p
HOMO-1
48% N 2p, 18% O 2p
48% N 2p, 20% O 2p
29% N 2p, 31% O 2p
23% N 2p, 38% O 2p
HOMO-2
35% O 2p, 13% C1-2p, 13% C22p, 11% C3-2p
30% O 2p, 23% H 1s, 14% C12p, 13% C2-2p
26% O 2p, 18% C2-2p, 16% C32p, 16% C1-2p
16% C1-2p, 17% C2-2p, 21% O 2p, 27% H 1s
HOMO-3
38% H 1s, 16% C3-2p, 19% O 2p
39% H 1s, 16% C1-2p, 18% C32p
17% C3-2p, 30% O 2p, 35% H 1s
17% C1-2p, 18% C3-2p, 15% O 2p, 40% H 1s
HOMO-4
35% C3-2p, 33% H 1s of Me
26% O 2p, 26% H 1s of Me
30% C3-2p, 43% H 1s
20% C3-2p, 32% O 2p, 31% H 1s
HOMO-5
28% H 1s, 14% C1-2p, 21% C32p
34% C3-2p, 25% H 1s of Me
13% C1-2p, 28% C3-2p, 43% H 1s
32% C3-2p, 11% O 2p, 40% H1s
HOMO-6
27% O 2p, 19% C2-2p, 22% C32p
17% N 2p, 18% C3-2p, 13% C22p
16% C-2p, 18% C3-2p, 33% O 2p
20% N 2p, 14% C1-2p, 18% C3-2p, 29% H 1s
HOMO-7
21% C2-2p, 19% C1-2p, 12% N 2p
18% C2-2p, 16% C3-2p, 10% O 2p
26% N 2p, 19% C1-2p, 13% C22p, 12% C3-2p
18% C2-2p, 18% C3-2p, 19% O 2p, 25% H 1s
HOMO-8
22% N 2p, 16% O 2p
26% O 2p, 14% N 2p
23% N 2p, 15% C2-2p, 15% O 2p
14% N 2p, 17% C2-2p, 27% O 2p, 24% H 1s
HOMO-9
26% C1-2p, 21% N 2p, 9% O 2p
26% C2-2p, 20% N 2p, 13% O 2p
16% N 2p, 28% C1-2p, 16% O 2p
24% N 2p, 24% C1-2p, 12% O 2p, 30% H 1s
HOMO-10
26% N 2p, 14% O 2p, 10% C12p
37% N 2p, 17% C1-2p
21% N 2p, 21% O 2p
27% N 2p, 16% C1-2p, 32% H 1s
HOMO-11
17% C2-2s, 14% O 2p
15% C2-2s, 23% O 2p
15% C2-2s, 12% O 2p
15% C2-2s, 15% O 2p
HOMO-12
25% C1-2s, 19% C3-2s
25% C1-2s, 21% C3-2s
23% C1-2s, 19% C3-2s
23% C1-2s, 19% C3-2s
HOMO-13
25% C2-2s, 24% C3-2s
25% C2-2s, 23% C3-2s
25% C3-2s, 25% C2-2s
24% C3-2s, 25% C2-2s
HOMO-14
55% N 2s, 13% C1-2s
55% N 2s, 13% C1-2s
55% N 2s, 14% C1-2s
54% N 2s, 13% C-2s
HOMO-15
67% O 2s
67% O 2s
68% O 2s
68% O 2s
a
The atomic orbitals with a population higher than 10% are reported.
allowed us to produce simulations of the experimental photoelectron valence band spectra, corroborating this interpretation. The C 1s core-level photoelectron spectra of alaninol and isopropanolamine point out similar experimental energy shift. The energy shift between the carbon of the methyl
group 共C3 for alaninol and isopropanolamine兲 and the carbon bound to the amino group is 0.7 eV 共C2 for alaninol兲 and 0.9 eV 共C1 for isopropanolamine兲, and that between the carbon of the methyl group and the carbon bound to the hydroxyl group is 1.3 eV 共C1 for alaninol兲 and 1.4 eV 共C2 for isopropanolamine兲. In conclusion, the photoelectron spectroscopy measurements presented in this paper does not show marked differences in the energetics and in the electronic structure, inTABLE VI. Theoretical KS-LB94 and ⌬SCF 共VWN兲 for I1 and I2 conformers and experimental C 1s IEs 共eV兲 of isopropanolamine. ⌬SCF
KS-LB94a C 1s
I1
I2
I1
I2
Expt.b 共±0.1 eV兲
C3c
290.09 共0.00兲d 291.29 共1.20兲 291.62 共1.53兲
290.11 共0.00兲d 291.30 共1.19兲 291.62 共1.51兲
294.11 共0.00兲d 295.19 共1.08兲 295.48 共1.37兲
294.10 共0.00兲d 295.20 共1.10兲 295.48 共1.48兲
290.3 共0.0兲d 291.2 共0.9兲 291.7 共1.4兲
C1c C2c a
FIG. 7. Experimental C 1s core-level photoelectron spectrum of isopropanolamine, recorded at energy h = 310.0 eV, together with the results of the fit 共G = 0.27; ⌫L = 0.25兲. The C 1s theoretical IE values 共KS-LB94 method兲 for both I1 and I2 are marked with two series of bars. The calculated and experimental data are summarized in Table VI.
The opposite orbital eigenvalues. Experimental IE resulting from the fit of the photoelectron spectrum acquired at 310.0 eV photon energy 共see Fig. 7兲. c For numbering of the carbon atoms, see Fig. 1. d The relative energy with respect to the lowest value is reported between in parentheses 共eV兲. b
Downloaded 15 Oct 2007 to 140.105.55.84. Redistribution subject to AIP license or copyright, see http://jcp.aip.org/jcp/copyright.jsp
144312-10
Catone et al.
duced by the swap of the two functional groups –OH and –NH2. All the differences observed in the valence band and C 1s core-level photoelectron spectra of the two chiral molecules should principally be ascribed to the unequal conformational content. Despite the capability of both DFT and ab initio theoretical schemes to reproduce the experimental IE sequence, it must be pointed out that the nature of HOMO and HOMO-1 bands of both molecules is still uncertain. Future studies based on the circular dichroism in the photoelectron angular distribution, a spectroscopy sensitive to the orbital character, may help in clarifying this point. In spite of the difficulties in the assignment of the electronic states for both systems, the present work succeeds in providing some important elements to understand the energetics of conformers and their influence on the photoelectron profile. Besides, the theoretical predictions offer an important tool in the interpretation of all the experimental results, pointing out the differences between the two studied structural isomers. 1
A. G. Garcia, N. Laurent, L. Mogens, H. Jean-Christophe, D. Danielle, and I. Powis, J. Chem. Phys. 119, 8781 共2003兲. 2 N. Böwering, T. Lischke, B. Schmidtke, N. Müller, T. Khalil, and U. Heinzmann, Phys. Rev. Lett. 86, 1187 共2001兲. 3 A. Giardini, D. Catone, S. Stranges, M. Satta, M. Tacconi, S. Piccirillo, S. Turchini, N. Zema, G. Contini, and T. Prosperi, ChemPhysChem 6, 1164 共2005兲. 4 T. Lischke, N. Böwering, B. Schmidtke, N. Müller, T. Khalil, and U. Heinzmann, Phys. Rev. A 70, 022507 共2004兲. 5 S. Stranges, S. Turchini, M. Alagia, G. Alberti, G. Contini, P. Decleva, G. Fronzoni, M. Stener, N. Zema, and T. Prosperi, J. Chem. Phys. 122, 244303 共2005兲. 6 S. Turchini, N. Zema, G. Contini, G. Alberti, M. Alagia, S. Stranges, G.
J. Chem. Phys. 127, 144312 共2007兲 Fronzoni, M. Stener, P. Decleva, and T. Prosperi, Phys. Rev. A 70, 014502 共2004兲. 7 R. Raval, Nature 共London兲 425, 463 共2003兲. 8 R. Raval and S. M. Barlow, Surf. Sci. Rep. 50, 201 共2003兲. 9 D. Catone, A. Giardini, A. Paladini, S. Piccirillo, F. Rondino, M. Satta, D. Scuderi, and M. Speranza, Angew. Chem., Int. Ed. 43, 1868 共2004兲. 10 A. Giardini, A. Paladini, D. Catone, S. Piccirillo, F. Rondino, M. Satta, A. Filippi, M. Speranza, S. Turchini, and N. Zema, Chirality 18, 562 共2006兲. 11 R. P. Walker and B. Diviacco, Rev. Sci. Instrum. 63, 332 共1992兲. 12 A. Derossi, F. Lama, M. Piacentini, T. Prosperi, and N. Zema, Rev. Sci. Instrum. 66, 1718 共1995兲. 13 R. Fausto, C. Cacela, and M. L. Duarte, J. Mol. Struct. 550–551, 365 共2000兲. 14 C. Cacela, R. Fausto, and M. L. Duarte, Vib. Spectrosc. 26, 113 共2001兲. 15 E. Baerends, D. E. Ellis, and P. Ros, Chem. Phys. 2, 41 共1973兲. 16 S. H. Vosko, L. Wilk, and M. Nusair, Can. J. Phys. 58, 1200 共1980兲. 17 A. D. Becke, Phys. Rev. A 38, 3098 共1988兲. 18 J. P. Perdew, Phys. Rev. B 33, 8822 共1986兲. 19 R. van Leeuwen and E. J. Baerends, Phys. Rev. A 49, 2421 共1994兲. 20 D. Toffoli, M. Stener, G. Fronzoni, and P. Decleva, Chem. Phys. 276, 25 共2002兲. 21 M. Stener, G. Fronzoni, D. D. Tommaso, and P. Decleva, J. Chem. Phys. 120, 3284 共2004兲. 22 M. J. Frisch, G. W. Trucks, H. B. Schlegel, G. E. Scuseria, M. A. Robb, J. R. Cheeseman, J. A. Montgomery, Jr., T. Vreven, K. N. Kudin, and J. C. Burant, Gaussian Inc., Wallingford, CT, 2004. 23 I. Powis, E. E. Rennie, U. Hergenhahn, O. Kugeler, and R. Bussy-Socrate, J. Phys. Chem. A 107, 25 共2003兲. 24 E. E. Rennie, I. Powis, U. Hergenhahn, O. Kugeler, S. Marburger, and T. M. Watson, J. Phys. Chem. A 106, 12221 共2002兲. 25 A. Schweig and W. Thiel, Mol. Phys. 27, 265 共1974兲. 26 U. Gelius, P. F. Hedén, J. Hedman, B. J. Lindberg, R. Manne, R. Nordberg, C. Nordling, and K. Siegbahn, Phys. Scr. 2, 70 共1970兲. 27 A. R. Slaughter and M. S. Banna, J. Phys. Chem. 92, 2165 共1988兲. 28 D. P. Chong, Can. J. Chem. 74, 1005 共1996兲. 29 D. T. Clark, J. Peeling, and L. Colling, Biochim. Biophys. Acta 453, 533 共1976兲.
Downloaded 15 Oct 2007 to 140.105.55.84. Redistribution subject to AIP license or copyright, see http://jcp.aip.org/jcp/copyright.jsp
