Low-lying electronic structure of EuH, EuOH, and EuO neutrals and anions determined by anion photoelectron spectroscopy and DFT calculations Jared O. Kafader, Manisha Ray, and Caroline Chick Jarrold

Citation: The Journal of Chemical Physics 143, 034305 (2015); doi: 10.1063/1.4926663 View online: http://dx.doi.org/10.1063/1.4926663 View Table of Contents: http://aip.scitation.org/toc/jcp/143/3 Published by the American Institute of Physics

Articles you may be interested in Photoelectron spectrum of PrO The Journal of Chemical Physics 143, 064305 (2015); 10.1063/1.4928371 Photoelectron spectra of CeO and Ce(OH)2 The Journal of Chemical Physics 142, 064305 (2015); 10.1063/1.4907714 Mixed cerium-platinum oxides: Electronic structure of [CeO]Ptn (n = 1, 2) and [CeO2]Pt complex anions and neutrals The Journal of Chemical Physics 145, 044317 (2016); 10.1063/1.4959279 Molecular and electronic structures of cerium and cerium suboxide clusters The Journal of Chemical Physics 145, 154306 (2016); 10.1063/1.4964817 Explaining the photoelectron spectrum: Rationalization of geometric and electronic structure The Journal of Chemical Physics 146, 104301 (2017); 10.1063/1.4977418 Role of weakly bound complexes in temperature-dependence and relative rates of MxOy + H2O (M = Mo, W) reactions The Journal of Chemical Physics 144, 074307 (2016); 10.1063/1.4941829

THE JOURNAL OF CHEMICAL PHYSICS 143, 034305 (2015)

Low-lying electronic structure of EuH, EuOH, and EuO neutrals and anions determined by anion photoelectron spectroscopy and DFT calculations Jared O. Kafader, Manisha Ray, and Caroline Chick Jarrolda) Department of Chemistry, Indiana University, Bloomington, Indiana 47405, USA

(Received 21 May 2015; accepted 30 June 2015; published online 17 July 2015) The anion photoelectron (PE) spectra of EuH− and the PE spectrum of overlapping EuOH− and EuO− anions are presented and analyzed with supporting results from density functional theory calculations on the various anions and neutrals. Results point to ionically bound, high-spin species. EuH and EuOH anions and neutrals exhibit analogous electronic structures: Transitions from 8Σ− anion ground states arising from the 4f 7σ6s 2 superconfiguration to the close-lying neutral 9Σ− and 7Σ− states arising from the 4f 7σ6s superconfiguration are observed spaced by an energy interval similar to the free Eu+ [4f 76s] 9S - 7S splitting. The electron affinities (EAs) of EuH and EuOH are determined to be 0.771 ± 0.009 eV and 0.700 ± 0.011 eV, respectively. Analysis of spectroscopic features attributed to EuO− photodetachment is complicated by the likely presence of two energetically competitive electronic states of EuO− populating the ion beam. However, based on the calculated relative energies of the close-lying anion states arising from the 4f 7σ6s and 4f 6σ6s 2 configurations and the relative energies of the one-electron accessible 4f 7 and 4f 6σ6s neutral states based on ligand-field theory [M. Dulick, E. Murad, and R. F. Barrow, J. Chem. Phys. 85, 385 (1986)], the remaining features are consistent with the 4f 6σ6s 2 7Σ− and 4f 7σ6s 7Σ− anion states lying very close in energy (the former was calculated to be 0.15 eV lower in energy than the latter), though the true anion ground state and neutral EA could not be established unambiguously. Calculations on the various EuO anion and neutral states suggest 4f -orbital overlap with 2p orbitals in species with 4f 6 occupancy. C 2015 AIP Publishing LLC. [http://dx.doi.org/10.1063/1.4926663]

I. INTRODUCTION

Lanthanide-based materials have unique electronic properties because of the high number of electronic degrees of freedom arising from partial occupation of 4f orbitals. Europium, in particular, is found in an interesting array of applications: Europium monoxide, with divalent Eu2+ ions, is a ferromagnetic semiconductor with a Curie temperature that can be manipulated by both pressure1,2 and doping.2–5 The colossal magnetoresistance arising from the half-filled 4f subshell in divalent EuO makes it an attractive potential material for use in spintronic devices.3,6,7 Europium reacts with hydrogen to form the divalent Eu2+ material EuH2, which also exhibits ferromagnetic ordering,8,9 and with higher H2 pressure, EuH x >2 complexes form.10,11 From a recent infrared spectroscopy-density functional theory (DFT) study on lanthanide hydrides trapped in solid hydrogen, photosensitive D2 stretching bands were observed when Eu atoms were present, raising the possibility that Eu could be exploited in hydrogen storage materials.12 The electronic spectra of small lanthanide atom-containing molecules can lend further insight into the nature of chemical bonding in lanthanide materials. It turns out that the electronic structure of lanthanide diatomics, in general, is very complex, as evinced in the extensive work by Field a)Author to whom correspondence should be addressed. Electronic mail:

[email protected] 0021-9606/2015/143(3)/034305/9/$30.00

and coworkers predicting the numerous close-lying states of LnO13 and LnX 14 (Ln = lanthanide atom, X = halide) based on ligand field theory calculations. On the other hand, our recently reported anion photoelectron (PE) spectrum of CeO− was very simple15 and provided a distilled picture of the molecular orbitals occupied in the ground state of neutral CeO. In this report, we present the anion photoelectron spectrum of EuH− and the combined PE spectra of EuOH− and EuO−, which could not be reliably mass-resolved due to difficulties in generating these species. However, DFT calculations on the anions and neutrals of these species are sufficiently helpful, and we are able to confidently assign spectroscopic features in the combined EuOH−/EuO− spectrum. Our findings indicate that EuH and EuOH anions and neutrals favor maintaining the stable, half-filled 4f 7 orbital occupancy on the Eu center, but for both EuO and EuO−, an alternative occupancy becomes competitive.

II. METHODS A. Experimental methods

A detailed description of the apparatus used in this study has been published previously.16,17 Briefly, Eu-containing anions were generated using a laser ablation/pulsed molecular beam valve cluster source. Approximately, 5 mJ/pulse of the second harmonic (532 nm) output of a Nd:YAG laser, operating

143, 034305-1

© 2015 AIP Publishing LLC

034305-2

Kafader, Ray, and Jarrold

J. Chem. Phys. 143, 034305 (2015)

at a repetition rate of 30 Hz, was focused onto the surface of a target composed of compressed Eu (ESPI Metals) and 98 Mo (trace elements) powders in a mole ratio 0.93:0.07. The 98 Mo powder was added both to help bind the sample during compression and to calibrate the mass spectrum, which was necessary due to the low-abundance and instability Eucontaining anions. The resulting plasma was swept through a 0.5 cm-long, 0.3-cm diameter channel into a vacuum chamber by a pulse of ultra-high purity helium buffer gas (40 psi (gauge) stagnation pressure) introduced from a solenoid-type molecular beam valve. No H2 was added to the gas mixture; outgassing from the numerous stainless steel surfaces in the gas lines, beam valve, and vacuum chamber was sufficient to produce the EuH−. The gas mixture was skimmed, and the anions were accelerated on axis into a beam-modulate timeof-flight mass spectrometer. Prior to colliding with a dual microchannel plate ion detector, the anions were photodetached with the second (532 nm, 2.33 eV) harmonic output of a second Nd:YAG laser at the intersection of the ion drift tube and a 1-m field-free drift tube situated perpendicular to the ion drift path. Drift times of the small fraction (ca. 10−4) of photoelectrons that traveled the length of the 1-m drift tube and collided with a second dual microchannel detector assembly were recorded on a digitizing oscilloscope. The drift times were converted to electron kinetic energy (e−KE) by identifying common transitions observed in spectra of various lanthanide-based ions with similar electron affinities (EAs), collected using both photon energies, and setting the difference in the e−KEs to the fundamental energy [the difference between the energies of the second and third harmonics, 1.1650(1) eV for the laser system used in this study] using the relationship 2  2  ℓ ℓ 2(1.1650 eV) − , (1) = (t 3ν − t o ) (t 2ν − t o ) me where me is the electron mass, t 3ν is the drift time of electrons associated with a selected transition observed in the spectrum obtained using 3.49 eV photon energy that can readily be correlated with a transition in the spectrum obtained with 2.33 eV, appearing at t 2ν . The equation is solved for ℓ and plotted as a function of t o . The intersection between this line and other lines generated from several sets of transitions observed in 3.49 eV and 2.33 eV gives a unique ℓ and to, the calibration parameters necessary to compute the e−KE values from electron drift times. The e−KE values are related to the anion and neutral states via e−KE = hν − EA − Teneutral + Teanion.

(2)

The data presented below show electron counts plotted as a function of e−BE, e−BE = hν − e−KE. −

(3)

The e BE values reflect the energy difference between the final neutral state and the initial anion state and are independent of the photon energy used. Laboratory to center-of-mass frame corrections were made to the e−KE (and e−BE) values. Spectra were accumulated over 320 000 and 2 000 000 laser shots for EuH− and EuOH−/EuO−, respectively, and were measured

with laser polarizations parallel (θ = 0◦) and perpendicular (θ = 90◦) to the electron drift tube in order to determine the asymmetry parameter, β(E), β (E) =

I0 − I90 1 2 I0 + I90

(4)

which can be related to the symmetry of the molecular orbital associated with the detachment transition in atomic spectra and can give insight into the nature of molecular orbitals. B. Computational methods

The molecular and electronic structures of EuH, EuOH, and EuO anions and neutrals were calculated using density functional theory with the primary goal of determining a general map of the energies of different electronic superconfigurations, in particular, for the anions.18 Calculations were performed with the Gaussian 09 package of electronic structure program using the B3LYP hybrid density functional method, which was chosen because of helpful results obtained for the CeO and CeO2H2 anion and neutral species in previous studies.15 To incorporate relativistic effects on Eu metal atom, the Stuttgart RSC ANO/ECP basis set with 28 core electrons and contraction of 14s13p10d8f3g/[6s6p5d4f3g] type, as developed by Cao and Dolg,19 was employed. The Dunningstyle correlation consistent basis set, aug-cc-pVTZ, was used on the H and O atoms. Full geometry optimization and frequency calculations were performed for all the species in multiple spin states. For the sole triatomic molecular species, hydroxide and oxo-hydride structures were considered. The reported relative energies for the species are zero point corrected energies. Adiabatic and vertical detachment energies (ADE and VDE) of 1-e− transitions between anion and neutral states were calculated for both sets of species. The ADE is the energy difference between the zero point corrected energies of the optimized anion and neutral species, and the VDE is the difference between the ground state energy of anion and single point energy of neutral confined to the optimized structure of the anion. Spectral simulations were generated using a home-written LabView code that calculates peak positions, based on adiabatic electron affinity, and anion and neutral frequencies, along with peak intensities from Franck-Condon factors, assuming harmonic oscillator wavefunctions for both anion and neutral. The peak widths are commensurate with the experimental resolution (ca. 12 meV at e−KE = 1 eV); the program varies the peak width with the experimental bandwidth dependence on e−KE.20

III. RESULTS AND ANALYSIS

The mass spectrum of molecular anions generated from the 98Mo/Eu target is shown in Figure 1. The ions are dominated by a peak at 146 amu/e− which is readily identified as 98 MoO3−, in spite of the relatively low abundance of 98Mo in the ablation target. 98MoO2− (120 amu/e−) is also present in significant quantities. Using 98MoO2− and 98MoO3− to

034305-3

Kafader, Ray, and Jarrold

FIG. 1. Mass spectrum of anions generated from laser ablation of the 93 mol. % Eu, 7 mol. % 98Mo target showing the relative intensities of the Eu-containing species.

calibrate the mass spectrum and considering the two nearly equally abundant isotopes of Eu with 151 and 153 amu, the well-resolved doublet at 152 and 154 amu/e− is assigned to EuH−, and the poorly resolved, less intense doublet in the 167–170 amu/e− range is assigned to overlapping EuO− and EuOH− ions. By peak-fitting the ion signal in this range, the EuO−:EuOH− ratio was determined to be 0.6, though the ion signal underwent considerable fluctuation over the course of data acquisition, preventing cleanly mass-resolved PE spectra. A. EuH− PE spectrum and computational results

The PE spectrum of EuH− obtained using 2.33 eV photon energy is shown in Figure 2(a). The spectrum exhibits two intense transitions labeled X0 and A0 at e−BE values of 0.771 ± 0.010 eV and 0.954 ± 0.015 eV, which is an interval of 0.183 ± 0.018 eV (1480 ± 140 cm−1). Additional peak positions and relative energies are summarized in Table I. Peaks labeled X1 and A1 are 1250 cm−1 to higher binding energy of X0 and A0, respectively. The simulation shown in the Fig. 2(b) was generated assuming two electronic transitions with origins coinciding with the positions of X0 and A0, and harmonic oscillator wavefunctions to calculate Franck-Condon factors associated with an Eu—H bondlength change of 0.15 Å for both transitions. Using a harmonic frequency of 1270 cm−1 and an anharmonicity of 10 cm−1, the positions of higher lying low-intensity signal was also reproduced (e.g., X2 and A2). The anion frequency used in the simulation was 1063 cm−1 (T = 1000 K for the spectrum shown) based on computational results (vide infra), though it does not match the position of x′. Figure 3 shows a simple molecular orbital diagram for EuH and EuH− (along with EuOH and EuOH−, vide infra). Neutral Eu has a [Xe] 4f 7 6s2 8 S (JEu = 7/2) ground state.21 The contracted 4f orbitals, while close in energy to the 6s valence orbital, are minimally involved in bonding. In a mixed ionic-covalent bonding scheme, the σ bond formed via overlap between the 1s orbital on H and the 6s, 5d z2, or a 6s-5d z2 hybrid would be more localized on H, and the singly occupied σ orbital would be more localized on the Eu 6s (or 6s-5d z2 hybrid); we consequently label this Eu-local orbital σ6s . In this ionic Eu+H− picture, the neutral ground state would nominally be described as a 9Σ− state. The σ6s orbital is doubly occupied

J. Chem. Phys. 143, 034305 (2015)

in the anion, resulting in a 8Σ− state, and detachment of the σ6s electron with spin antiparallel to the 4f electron spins (red electron in Fig. 3) accesses the 9Σ− state, while detachment of the σ6s electron with the spin aligned with the 4f electrons will access the 7Σ− state. The asymmetry parameter measured for the spectrum is 1.5(0.4) as determined using Eq. (4). This value is comparable to the asymmetry parameter measured for the CeO− PE spectrum, which involved detachment of an electron from a 6s-like orbital localized on the lanthanide atom, resulting in predominantly p-wave photoelectrons. Considering the seminal work of Field and coworkers on lanthanide oxide13 and halide14 diatomics, the angular momentum, JEu+, of Eu+ is largely uncoupled to the molecular axis. The energy between the 9Σ− (Ω = JEu+ = 4) and 7 − Σ (Ω = JEu+ = 3) states should, therefore, be comparable to the [Xe] 4f 7 6s Eu+ 9 S (JEu+ = 4) and 7 S (JEu+ = 3) levels, which is 1669.21 cm−1,20 an interval that is illustrated in Fig. 2(b). Therefore, peaks X0 and A0, which are separated by 0.183 eV (1480 cm−1), are assigned to transitions to these two states, as summarized in Table I. Likewise, 8Σ− (Ω = JEu = 3.5) state would be the lowest energy state arising from 4f 7σ6s 2 occupancy in EuH−. For both the anion and neutral, there are additional close-lying states arising from different projections of JEu+ onto the internuclear axis, Ω = JEu (or JEu+), Ω = JEu (or JEu+) − 1, Ω = JEu (or JEu+) − 2, etc. Calculations on EuF, a molecule that can similarly be described as Eu+ F−, suggest that the different values of Ω associated with each JEu+ are nearly degenerate for both JEu+ = 4 and JEu+ = 3.22

FIG. 2. (a) PES of EuH− obtained using 2.33 eV photon energy; (b) FranckCondon simulations of the 9Σ− − 8Σ− and 7Σ− − 8Σ− transitions. See text for details.

034305-4

Kafader, Ray, and Jarrold

J. Chem. Phys. 143, 034305 (2015)

TABLE I. Peak positions and tentative assignments for the PE spectra of EuH− [Fig. 2(a)] and EuOH−/EuO− [Fig. 4(a)]. For binding energy, values in parentheses represent the uncertainty in the last one or two digits.

Peak

Position e −BE/eV

Relative energy (cm−1)

x′

0.678(7)

−750

X0 X1 A0 X2 A1

0.771(9) 0.926(10) 0.954(7) 1.067(10) 1.106(10)

0 1250 1480 2390 2700

EuO (4f 7σ 6s ) EuO (4f 7σ 6s )

x′ X0

0.502(10) 0.567(10)

−520 0

EuO (4f 7σ 6s )/EuOH

X1

0.650(10)

670

EuOH EuOH EuOH

A0 A1

EuOH EuO (4f 6σ 6s 2) EuO (4f 6σ 6s 2) EuO (4f 6σ 6s 2)

B1 Y0 Y1

0.700(11) 0.761(8) 0.860(8) 0.888(8) 0.933(7) 1.063(7) 1.150(15)

1070 1560 2360 2590 2950 4000 4700

Z0

1.194(12)

5060

Z1

1.300(15)

5910

EuH−

Tentative assignment (v′ = 0) ← (tentative) 9Σ − (v′ = 0) ← 9Σ − (v′ = 1) ← 7Σ − (v′ = 0) ← 9Σ − (v′ = 2) ← 7Σ − (v′ = 1) ←

9Σ −

8Σ −

(v′′ = 1)

(v′′ = 0) (v′′ = 0) 8Σ − (v′′ = 0) 8Σ − (v′′ = 0) 8Σ − (v′′ = 0)

8Σ −

8Σ −

EuOH−/EuO−

EuO (4f 6σ 6s 2)

B0

While our DFT calculations on EuH− and EuH did not include spin-orbit splitting, the results, summarized in Table II, are consistent with the simple molecular orbital diagram, and depictions of several orbitals from the calculations are included in Fig. 3. The relative energies of the 8Σ− anion ground state and the 9Σ− and 7Σ− neutral states (both of which have 4f 7σ6s

FIG. 3. Simplified molecular orbital diagram illustrating orbital occupancy in EuH/EuH− and EuOH/EuOH− ground states. The “red” electron schematically indicates the occupancy change associated with the photodetachment transition to the ground neutral state. Depictions of EuH orbitals are situated to the left of depictions of EuOH orbitals. The Eu atom is pointing left in all depictions.

(v′ = 0) ← 9Σ− (v′′ = 1) (v′ = 0) ← 9Σ− (v′′ = 0) 8Σ − (v′ = 1) ← 9Σ − (v′′ = 1); 9Σ − (v′ = 0) ← 8Σ − (v′′ = 1) 9Σ − (v′ = 0) ← 8Σ − (v′′ = 0) 9Σ − (v′ = 1) ← 8Σ − (v′′ = 0) 9Σ − (v′ = 0) ← 8Σ − (v′′ = 0) Not assigned 9Σ − (v′ = 1) ← 8Σ − (v′′ = 0) 8Σ − (v′ = 0) ← 7Σ − (v′′ = 0) 8Σ − (v′ = 1) ← 7Σ − (v′′ = 0) 6Σ − (v′ = 0) ← 7Σ − (v′′ = 0) (tentative) 6Σ − (v′ = 1) ← 7Σ − (v′′ = 0) (tentative) 8Σ −

8Σ −

orbital occupancy) are consistent with the relative positions of peaks X0 and A0. Given the overall agreement, we assert that peak X0 can be assigned to the transition between the anion and neutral ground electronic states, and the position, 0.771 eV, is the EA of neutral EuH; the calculated value is 0.715 eV, which is in good agreement. Note that the ⟨S 2⟩ = 13 value for the 7Σ− state is larger than S(S + 1) = 12 because all eight electrons are unpaired; the nominally 6s electron is antiferromagnetically coupled to the 4f electrons. Overall, our calculations are in close agreement with higher-level calculations by Infante et al. on the neutral EuH ground electronic state, bondlength, and vibrational frequency,12 along with their reported experimental Eu—H stretch frequency of 1189.4 cm−1 for Eu atoms co-deposited cryogenically with H2. This frequency is also comparable to IR absorption lines measured in solid argon matrices and assigned to EuH2 at 1156 cm−1 (antisymmetric stretch) and 1212 cm−1 (symmetric stretch).23 Our own calculations underestimate the difference in bondlength between the anion and neutral ground states. As pointed out above, the calculated anion vibrational frequency, 1063 cm−1, which was used in the spectral simulation shown in Fig. 2(b), does not reproduce peak x′, which, if a vibrational hot band, gives an anion vibrational frequency of 750 cm−1. We acknowledge shortcomings in the calculations and cede that the anion frequency may be significantly lower than the calculated value. Certainly, the spectrum suggests a larger bondlength change between the anion and neutral states than what was calculated, which is consistent with a lower anion frequency than what was calculated. Based on the MO diagram of EuH− (Fig. 3), it seems unlikely that there is a low-lying

034305-5

Kafader, Ray, and Jarrold

J. Chem. Phys. 143, 034305 (2015)

TABLE II. Summary of calculations on the lowest energy electronic states found for EuH and EuH−, along with experimentally determined values. Uncertainty in the last two digits is given in parentheses. r /Å

ω/cm−1

0.842

2.122

1190

20.007

0.640

2.116

1209

15.764

0

2.218

1063

EuH

⟨S 2⟩

7Σ −

13.005a

9Σ −

Relative ZPE/eV

Experimental values To = 0.183(15) eV ∆G1/2 = 1250(100) cm−1 EAa = 0.771(9) eV ∆G1/2 = 1250(100) cm−1



EuH 8Σ −

a Spin

0 r anion = r neutral + 0.15(2) Å

contaminated; S = 3; S(S + 1) = 12.

excited electronic state of EuH− that may be contributing to the spectrum. B. EuOH−/ EuO− PE spectrum and computational results

The PE spectrum of EuOH−/EuO− obtained using 2.33 eV photon energy shown in Figure 4(a) exhibits numerous irregularly spaced transitions in the 0.5–1.5 eV e−BE range. Peak positions and relative energies are summarized in Table I. The asymmetry parameter is 1.6 ± 0.2 over the range of the spectrum, similar to the asymmetry parameter determined for the EuH− spectrum. Because there are two distinct anionic species contributing to the spectrum, computational results are relied upon to inform attempts to assign the spectrum. We first consider contributions made to the spectrum by EuOH− photodetachment. Assuming a hydroxide molecular structure, the neutral electronic structure would be similar to Eu—H neutral, as well as Eu—X (X = halogen) neutrals, states that have been considered by Kaledin et al.,14 Dolg and Stoll,24 and others at varying levels of theory.22,25 The lowest energy superconfiguration in an ionic Eu+(OH)−, by comparison with EuF and EuH, would be 4f 7 6s, as is illustrated in the molecular orbital diagram shown in Fig. 3. Results from our DFT calculations on anion and neutral EuOH, summarized in Table III, predict linear hydroxide structures to be the most stable. The neutral ground state has a 9Σ− term, with the 7Σ− state 0.18 eV higher in energy, very similar to the 0.183 eV splitting observed in the EuH− spectrum, as well as the 0.176 eV splitting measured for the nonet-septet splitting in EuF by Kaledin and coworkers,14 and the 0.22 eV splitting in EuF calculated by Yamamoto et al., using a fourcomponent relativistic configuration interaction method.22 Additionally, the adiabatic electron affinity is calculated to be 0.65 eV, from the energy difference between the 8Σ− anion ground state and the neutral ground state. Molecular orbitals from the calculations closely resemble the EuH molecular orbitals, with some minor variations due to the availability of orbitals with π symmetry on –OH, and are shown in Fig. 3. Based on the calculated structures of EuOH− and the two neutral states of EuOH (Table III), the only vibrational mode that is expected to be active in the spectrum is the Eu—OH stretch, which is 523 cm−1. Given the similarity between the transition energies and vibrational spacings, we assign bands A and B to transitions from the EuOH− 8Σ− ground state to the 9 − Σ and 7Σ− neutral states, respectively, analogous to the EuH−

photodetachment transitions. This assignment establishes the EA of EuOH as 0.700(11) eV. Fig. 4(b) shows simulations generated using the calculated frequencies and activation of the Eu—OH stretch. The low bend frequency of the anion and the disparity between the anion and neutral bend frequencies result in significant contributions from sequence bands. To better match the FranckCondon profile in the experimental spectrum, the change in Eu—OH bondlength was increased from the calculated value of 0.06 Å to 0.10 Å (0.305 amu1/2 Å). The origins of the simulated transitions to the 9Σ− and 7Σ− transitions were set to 0.695 eV and 0.863 eV (an interval of 0.168 eV, or 1140 cm−1). The vibrational temperature of the anion was set to 500 K. We note the general similarity between the simulated EuOH− and EuH− transitions (in particular, what appears to be a systematic computational underestimation of the bondlength change between anions and neutrals), and the similarity in the nonetseptet splitting, which reflects the JEu+ = 4, 3 splitting, also indicated in Fig. 4(b).21 We now address the remaining features, first assuming they are associated with EuO− photodetachment. With the ionic bonding picture, Eu2+O2−, the particular stability of the 4f 7 orbital occupancy of Eu2+ maintains the 4f 7 configuration in the EuO diatomic, as predicted by Field.13 Figure 5(a) shows a simplified molecular orbital diagram based on this configuration, along with the likely anion EuO− configuration with the additional electron occupying the σ6s orbital. In the case of Eu2+, the ligand field from O2− is not sufficient to render the 4f 6 6s Eu2+ configuration (i.e., 4f 6 σ6s ) more stable than the 4f 7 configuration. However, Dulick and coworkers26 reported ligand field theory calculations predicting the 4f 6 σ6s state lying 0.409 eV above the ground state 4f 7 configuration, also giving an alternative estimate of 0.6 eV based on thermodynamic values. Fig. 5(b) illustrates this ligand-field stabilized occupancy. The relatively low energy of this state is remarkable considering that in the free Eu2+ cation, the 4f 6 6s 8F state is 5.7 eV higher in energy than the 4f 7 9S ground state!21 In agreement with the ligand field theory treatment of EuO, the neutral ground state was calculated to be the 4f 7 8Σ− state. The calculated vibrational frequency of the 4f 7 8Σ− state is 676 cm−1, which is in excellent agreement with the 667.8 cm−1 frequency measured in solid rare gas matrices.27 The relatively low frequency (for comparison, the CeO frequency is 824 cm−1)28 is attributed to enhanced shielding of the nuclear charge by the 4f 7 (4f n+1) occupancy versus the 4f 6 σ6s (4f n σ6s ) ligand-field stabilized occupancy.13

034305-6

Kafader, Ray, and Jarrold

J. Chem. Phys. 143, 034305 (2015)

for sim EuOH 9N band EuOH 7N band

for sim EuO 9A 8N calc 4f6 6s2 tweak X 4f6 6s2 tweak A

for sim sum of sims

FIG. 4. (a) PES collected from unresolved EuOH− and EuO− anions using 2.33 eV photon energy; (b) simulations of the EuOH− 9Σ− − 8Σ− and 7Σ − − 8Σ − transitions, based on calculated vibrational frequencies and sole activation of the Eu—OH stretch; (c) calculation-based simulation of the 8Σ −(4f 7) − 9Σ −(4f 7σ ) transition, along with hypothetical simulations of the 6s 8Σ −(4f 6σ ) − 7Σ −(4f 6σ 2) and 6Σ −(4f 6σ ) − 7Σ −(4f 6σ 2) transitions. See 6s 6s 6s 6s text for details.

Results of DFT calculations on the 4f 7σ6s and 4f 6σ6s 2 anion configurations, summarized in Table III, predict that the 7Σ− anion state arising from the latter is more stable by 0.15 eV than the 9Σ− state arising from 4f 7σ6s occupancy. Our treatment of the 7Σ− state is somewhat oversimplified, since the 4f 6σ6s 2 superconfiguration, depending on which 4f orbital is unoccupied in the molecular environment, will lead to different values of Λ and Ω in the Russell-Saunders coupling scheme, but the states will be heavily mixed and spread over an energy range comparable to the Eu+ 4f 6 6s2 states (not assigned on the NIST database).21 However, since the spectra appear to be dominated by detachment from the σ6s orbital, the ∆Ω = ±1/2 selection rule narrowly limits which states can be observed.

We consider the possibility that transitions from both states of the anion are contributing to the spectrum. Figure 6 shows a schematic of the relative energies of the lowest lying states of EuO and EuO− based on our calculations, as well as relative neutral energies based on ligand field theory calculations by Dulick and coworkers.26 The overlapping EuOH and EuOH− transitions are also included for reference. Transitions between the various anion and neutral states are indicated with arrows labelled and colored to correspond with the peak labels and simulations shown in Fig. 4, as described next. The oneelectron transition indicated with the gray dashed arrow and asterisk involves detachment of a 4f electron; the relative cross section of this transition would be small compared to all other transition in Fig. 4.29 Based on the orbital occupancies of the two close-lying anion states, a one-electron transition to the neutral ground state is only possible from the 4f 7σ6s 9Σ− anion state. The transition energy is calculated to be 0.48 eV, which is the lowest of all the one-electron allowed transitions calculated in this study. We therefore assign X0 to the 4f 7 8Σ− + e− ← 4f 7σ6s 9Σ− transition. Fig. 4(c) shows a simulation of the 4f 7 8Σ− + e − ← 4f 7 σ6s 9Σ− transition (blue trace) based on the calculated anion and neutral vibrational frequencies and difference in EuO bondlength. The origin of the transition was set to 0.567 eV, which is within 0.1 eV of the calculated transition energy. Since the higher-lying vibrational levels overlap with the EuOH− transitions, we did not attempt to adjust any of the spectroscopic parameters from the calculated values. Our attempts to calculate the octet neutral state with 4f 6σ6s occupancy, which, based on the relative energies of the degeneracy-broken 4f orbitals in all other calculations would be an 8Σ− state, were unsuccessful because the 4f 7 8Σ− state is close to, but lower in energy than the excited state with the same spin multiplicity. However, using the energy of the 4f 6σ6s predicted by Dulick and our calculated relative energies of the 4f 6σ6s 2 and 4f 7σ6s anion states (Table III and Fig. 6), the 4f 6σ6s octet state + e− ← 4f 6σ6s 2 7Σ− transition should be in the 1–1.5 eV range. We, therefore, assign the transitions labeled Y0 and Z0 to the 4f 6σ6s 8Σ− + e− ← 4f 6σ6s 2 7Σ− and 4f 6σ6s 6Σ− + e− ← 4f 6σ6s 2 7Σ− transitions, respectively. We consider the terms of the excited neutral states very tentative because we lack computational results to support the assignments. Further, the energy interval between the 8F and 6F states of Eu2+ with the 4f 6 6s occupancy is 0.27 eV, whereas peaks Y0 and Z0 are separated by a 0.13 eV (1050 cm−1). The simulations shown in Fig. 4(c) assume band origins of 1.061 eV and 1.194 eV (Y0 and Z0, respectively), a neutral vibrational frequency of 720 cm−1, an anion vibrational frequency of 624 cm−1 (from calculations, see Table III), and a bondlength change of 0.04 Å (0.16 amu1/2 Å). Figure 4(d) shows the sum of the simulations of the EuOH− transitions and the EuO− transitions. While the sum does not match the spectrum perfectly, the assignments are supported by (1) the combination of DFT results and previous ligand field theory calculations26 and (2) the uniform asymmetry parameter over the spectrum, which is consistent with all transitions involving detachment of electrons from σ6s orbitals.

034305-7

Kafader, Ray, and Jarrold

J. Chem. Phys. 143, 034305 (2015)

TABLE III. Summary of DFT calculations on EuOH and EuO anion and neutral states, with experimentally determined values. Uncertainty in the last two digits is given in parentheses. The calculated Eu—OH stretch frequency was used directly in the simulations shown in Fig. 4(b).

r EuO/Å r OH/Å

Eu—OH stretch; EuO—H stretch Bend/cm−1

2.105 0.957 2.109 0.956

523, 3909 329 (degenerate) 523, 3909 329 (degenerate)

⟨S 2⟩

Relative ZPE/eV

7Σ −

13.002a

0.83

9Σ −

20.007

0.65

8Σ −

15.761

0

2.169 0.956

449, 3897 188 (degenerate)

r EuO,anion = r EuO,neutral + 0.10(2) Å

EuO

⟨S 2⟩

Relative ZPE/eV

r EuO/Å

ω/cm−1

Experimental values

... 15.792

ca. 1.03 (From Dulick et al.)b 0.63 = EA (1 − e− transition energy = 0.48 eV)

... 1.884

... 676

1.063(7) eV 0.57 + ε (∆G1/2 = 668)c

20.000 12.007

0.15 0

1.960 1.906

617 625

ε ...

EuOH

Experimental values Te = 0.160(13) EA = 0.700(10)

EuOH−

8Σ − 4f 6 6s 8Σ − 4f 7

EuO− 9Σ − 4f 7 6s 7Σ − 4f 6 6s 2

contaminated; S = 3; S(S + 1) = 12. 25. c Reference 26. a Spin

b Reference

The EA of EuO cannot be unambiguously established. The position of peak X0, 0.567 eV, is a lower limit and equals the EA only if the 4f 7σ6s 9Σ− anion state is the ground state, contrary to the calculated ground state. However, DFT calculations have proven reliable to within several tenths of an eV, so we conservatively approximate the EA of EuO to be in the 0.57–0.8 eV range. As a last point, we address the possibility that EuOH− may alternatively have an oxo-hydride structure, e.g., H—Eu==O. Considering the electronic structure based on ionic bonding, Eu in HEuO− is in the +2 oxidation states, favoring the 4f 7 orbital occupancy.21 The Eu+3 4f 6 orbital occupancy in neutral HEuO would yield numerous close-lying states associated with the JEu = 0 through 6 Eu(IV) ionic states, which span an energy range of 0.613 eV,21 similar to the span of the observed spectrum. However, the photodetachment cross section for 4f detachment transitions will be very small relative to transitions involving 6s electron detachment.29 In contrast, the integrated electron counts divided by laser shots for the EuOH−/EuO− spectrum are nearly identical to the electron per laser shot ratio for the EuH− spectrum; the photodetachment cross sections are similar. Further, the asymmetry parameter for all transitions is very similar, ca. 1.5. Finally, computational results predict that the hydroxide anion structure is more than 1.67 eV lower in energy than the oxo-hydride structure.

ions, the simple PE spectrum of EuH−, in combination with the good agreement between the calculated anion and neutral state energies and features in the spectrum, provides a strong basis for approaching the overlapping EuOH−/EuO− spectrum. At the same time, we note that difficulties in ion production and ambiguous mass assignments may have plagued previous attempts to measure photodetachment spectra of Eu−. The PE spectrum of Eu− reported by Davis and Thompson in 196130 exhibited two broad peaks separated by 0.189 ± 0.035 eV. Converting their reported e−KE energy scale to e−BE using their photon energy of hv = 1.165 eV (Eq. (3)), the peak positions closely coincide with our spectrum, leading us to suggest that the ion identity in their study was EuH− as well, and the Eu− PE spectrum has not yet been published. Calculations suggest that the EA of Eu is on the order of 0.05 eV31 versus the 1.053 eV value reported by Davis and Thompson.30 The very low EA of Eu is consistent with the absence of Eu− generated in our laser ablation source. The agreement between the PE spectrum of EuH− and the results of DFT calculations points to simple ionic bonding in the neutral diatomic. The Eu+ cation atomic orbital occupancy, 4f 7 6s, is largely preserved in the neutral EuH molecule, as evidenced by the similarity between the 9S − 7S splitting in the Eu+ cation and the 9Σ − 7Σ splitting in EuH neutral. The Eu—H bond, however, is weak compared to numerous transition metal hydride diatomics (e.g., on the high frequency end, Pt—H vibrational frequency is 2295 cm−1).32

IV. DISCUSSION A. EuH electronic structure and reassignment of previously reported PES of Eu−

While there are challenges with assigning the PE spectrum of low-intensity and unstable overlapping EuOH− and EuO−

B. Electron affinity versus oxidation state

The EA of a neutral molecule is a direct measure of the relative stability of the ground state of the neutral and the ground state of the associated anion. Generally, as metal

034305-8

Kafader, Ray, and Jarrold

J. Chem. Phys. 143, 034305 (2015)

C. EuO− electronic structure

FIG. 5. Simplified molecular orbital diagrams illustrating the two close-lying orbital occupancies of EuO and EuO−.

species become sequentially more oxidized and the metal centers become sequentially more positive, the stability of the “extra” electron in the anion occupying metal-local orbitals is enhanced, thereby increasing the electron affinity. This effect has been shown for both transition metal oxides33 and p-block metal oxides.34 Based on DFT calculations, EuO has a lower electron affinity than EuOH, though this cannot be confirmed unambiguously from the PE spectrum (vide supra). Eberhardt and coworkers reported that the EA of LaOn molecules decreased with increasing n,35 and we have measured similar effects from the PES of cerium-based anions.36 The EA of EuOH, 0.700 eV, is also lower than the EA of EuH, 0.771 eV, while the calculations on their respective neutral 9Σ− states predict that the Eu center in EuOH is significantly more positively charged (+0.53 in EuOH versus +0.17 in EuH). It would be interesting to determine whether this trend holds for all lanthanidebased small molecules, and whether the effect points to greater delocalization of the “extra” electron beyond metal-localized orbitals. The 6s-like orbitals associated with the detachment transitions (Figs. 3 and 5) are indeed diffuse.

FIG. 6. Schematic of the transitions contributing to the PES collected from unresolved EuOH− and EuO− anions [Fig. 4(a)]. The different colored arrows represent transitions associated with the simulated spectra [Figs. 4(b) and 4(c)] of the same color.

DFT calculations on lanthanide-containing molecules have their limitations and must be treated with appropriate caution. However, results of DFT calculations have proven to be helpful in providing the groundwork for understanding the lowest energy orbital occupancies of EuH anion and neutral, along with CeO and Ce(OH)2 anions and neutrals.15 Based on the calculations done in the current study, the 4f 6σ6s 2 7Σ− and 4f 7 σ6s 9Σ− states of EuO− are very close in energy, which, in light of the ligand-field predicted stability of the 4f 6σ6s 8Σ− neutral state26 is not unexpected, though no Eu+ state with 4f 6 6s2 occupancy is identified in the NIST Atomic Spectra Database (the lowest-energy Eu+ state with 4f 6 . . . occupancy is the 4f 65d 6s 9D2 state, Te = 3.74 eV).21 The computational results also reflect how shifting the relative nuclear shielding associated with the two occupancies: Promotion of a 4f electron to the σ6s orbital increases the effective nuclear charge on Eu, decreasing the Eu—O bondlength, and increasing the vibrational frequency. An additional effect is that the relative energies of the 4f like orbitals in the two different anion states change relative to each other as well as relative to the bonding orbitals. In particular, the shorter bondlength in the 4f 6σ6s 2 7Σ− state results in overlap between the Eu 4fπ orbitals and the O 2p⊥ orbitals [see Fig. 5(b)], resulting in their being stabilized relative to the other 4f -like orbitals. The overall energy interval spanned by the 4f -like orbitals is broader in the 4f 6σ6s 2 7Σ− state (0.74 eV) than the 4f 7 σ6s 9Σ− state (0.58 eV), and the hybridization of the 4fπ orbitals accompanied by the bondlength decrease is evocative of volume collapse and mixed-valencies observed in high-pressure studies on bulk EuO.37 V. CONCLUSIONS

The anion PE spectrum of EuH− bears out the ionic nature of this lanthanide metal hydride in that the Eu+ 9S − 7S splitting is largely preserved. The vibrational frequency of the neutral 9 − Σ and 7Σ− states arising from the 4f 7σ6s orbital occupancy is 1250 ± 100 cm−1, reflecting the weak Eu—H bond. The neutral electron affinity is 0.771 ± 0.009 eV. Calculations on the relative energies of the 8Σ− anions ground state and the neutral 9Σ− and 7Σ− states are within 0.15 eV of the observed transition energies. The calculations underestimate the change in Eu—H bondlength upon photodetachment of the anion and may also significantly overestimate the EuH− vibrational frequency. Analysis of the PE spectrum of the overlapping EuOH− and EuO− anions relied more heavily on computational results. The electronic structure of EuOH was determined to be largely analogous to the electronic structure of EuH, with neutral 9 − Σ and 7Σ− states accessed via photodetachment of the 8Σ− anion ground state. The EA of EuOH was determined to be 0.700 ± 0.011 eV, which is lower than the EA of EuH, in spite of the Eu center in EuOH being more positively charged than in EuH. Again, agreement between calculated and observed transition energies was good, though calculations underestimated the Eu—OH bondlength change upon photodetachment of EuOH−. Analysis of spectroscopic features not attributed to EuOH− photodetachment was more complicated because of

034305-9

Kafader, Ray, and Jarrold

the likelihood of two energetically competitive electronic states of EuO− populating the ion beam. However, based on the calculated relative energies of the close-lying anion states arising from the 4f 7σ6s and 4f 6σ6s2 configurations and the relative energies of the one-electron accessible 4f 7 and 4f 6σ6s neutral states based on ligand-field theory,26 the remaining features were assigned and are consistent with the 4f 6σ6s2 7Σ− and 4f 7 σ6s 9Σ− anion states lying very close in energy (the former was calculated to be 0.15 eV lower in energy than the latter). The true anion ground state and neutral EA could not be established unambiguously. Calculations on the various EuO anion and neutral states suggest 4f -orbital overlap with 2p orbitals in species with 4f 6 occupancy.

ACKNOWLEDGMENTS

The authors gratefully acknowledge generous support for this work from the National Science Foundation, Grant No. CHE-1265991. 1D. DiMarzio, M. Croft, N. Sakai, and M. W. Shafer, Phys. Rev. B 35, 8891R

(1987). 2N. M. Souza-Neto, D. Haskel, Y.-C. Tseng, and G. Lapertot, Phys. Rev. Lett.

102, 057206 (2009). Schmehl, V. Vaithyanathan, A. Herrnberger, S. Thiel, C. Richter, M. Liberati, T. Heeg, M. Röckerath, L. Fitting Kourkoutis, S. Mühlbauer, P. Böni, D. A. Muller, Y. Barash, J. Schubert, Y. Idzerda, J. Mannhart, and D. G. Schlom, Nat. Mat. 6, 882 (2007). 4H. Ott, S. J. Heise, R. Sutarto, Z. Hu, C. F. Chang, H. H. Hsieh, H.-J. Lin, C. T. Chen, and L. H. Tjeng, Phys. Rev. B 73, 094407 (2006). 5T. J. Konno, N. Ogawa, K. Wakoh, K. Sumiyama, and K. Suzuki, Jpn. J. Appl. Phys. 35, 6052 (1996). 6Y. Shapira, S. Foner, and T. B. Reed, Phys. Rev. B 8, 2299 (1973). 7A. G. Swartz, J. Ciraldo, J. J. I. Wong, Y. Li, W. Han, T. Lin, S. Mack, J. Shi, D. D. Awschalom, and R. K. Kawakami, Appl. Phys. Lett. 97, 112509 (2010). 8R. L. Zanowick and W. E. Wallance, Phys. Rev. 126, 537 (1962). 9A. Mustachi, J. Phys. Chem. Solids 35, 1447 (1974). 10M. Hirscher, Handbook of Hydrogen Storage: New Materials for Future Energy Storage (WILEY-VCH, Germany, 2010). 11H. Saitoha, A. Machidaa, T. Matsuokab, and K. Aokic, Solid State Commun. 205, 24 (2015). 12I. Infante, L. Gagliardi, X. Wang, and L. Andrews, J. Phys. Chem. A 113, 2446 (2009). 13R. W. Field, Ber. Bunsenges. Phys. Chem. 86, 771 (1982). 14L. A. Kaledin, M. C. Heaven, R. W. Field, and L. A. Kaledin, J. Mol. Spectrosc. 179, 310 (1996). 3A.

J. Chem. Phys. 143, 034305 (2015) 15M. Ray, J. A. Felton, J. O. Kafader, J. E. Topolski, and C. C. Jarrold, J. Chem.

Phys. 142, 064305 (2015). D. Moravec and C. C. Jarrold, J. Chem. Phys. 108, 1804 (1998). 17S. E. Waller, J. E. Mann, and C. C. Jarrold, J. Phys. Chem. A 117, 1765 (2013). 18M. J. Frisch, G. W. Trucks, H. B. Schlegel, G. E. Scuseria, M. A. Robb, J. R. Cheeseman, G. Scalmani, V. Barone, B. Mennucci, G. A. Petersson, H. Nakatsuji, M. Caricato, X. Li, H. P. Hratchian, A. F. Izmaylov, J. Bloino, G. Zheng, J. L. Sonnenberg, M. Hada, M. Ehara, K. Toyota, R. Fukuda, J. Hasegawa, M. Ishida, T. Nakajima, Y. Honda, O. Kitao, H. Nakai, T. Vreven, J. A. Montgomery, Jr., J. E. Peralta, F. Ogliaro, M. Bearpark, J. J. Heyd, E. Brothers, K. N. Kudin, V. N. Staroverov, R. Kobayashi, J. Normand, K. Raghavachari, A. Rendell, J. C. Burant, S. S. Iyengar, J. Tomasi, M. Cossi, N. Rega, J. M. Millam, M. Klene, J. E. Knox, J. B. Cross, V. Bakken, C. Adamo, J. Jaramillo, R. Gomperts, R. E. Stratmann, O. Yazyev, A. J. Austin, R. Cammi, C. Pomelli, J. W. Ochterski, R. L. Martin, K. Morokuma, V. G. Zakrzewski, G. A. Voth, P. Salvador, J. J. Dannenberg, S. Dapprich, A. D. Daniels, Ö. Farkas, J. B. Foresman, J. V. Ortiz, J. Cioslowski, and D. J. Fox,  09, Revision D.01, Gaussian, Inc., Wallingford CT, 2009. 19X. Cao and M. Dolg, J. Chem. Phys. 115, 7348 (2001). 20J. A. Felton, M. Ray, and C. C. Jarrold, Phys. Rev. A 89, 033407 (2014). 21A. Kramida, Yu. Ralchenko, and J. Reader, NIST ASD Team (2014). NIST Atomic Spectra Database (ver. 5.2), [Online]. Available: http://physics.nist. gov/asd, National Institute of Standards and Technology, Gaithersburg, MD, 15 April 2015. 22S. Yamamoto, H. Tatewaki, and H. Moriyama, Theor. Chem. Acc. 131, 1230 (2012). 23S. P. Willson and L. Andrews, J. Phys. Chem. A 104, 1640 (2000). 24M. Dolg and H. Stoll, Theor. Chim. Acta 75, 369 (1989). 25H. Heiberg, O. Gropen, J. K. Laerdahl, O. Swang, and U. Wahlgren, Theor. Chem. Acc. 110, 118 (2003). 26M. Dulick, E. Murad, and R. F. Barrow, J. Chem. Phys. 85, 385 (1986). 27S. P. Willson and L. Andrews, J. Phys. Chem. A 103, 3171 (1999). 28C. Linton, M. Dulick, R. W. Field, P. Carette, and R. F. Barrow, J. Chem. Phys. 74, 189 (1981). 29S. M. O’Malley and D. R. Beck, Phys. Rev. A 74, 042509 (2006). 30V. T. Davis and J. S. Thompson, J. Phys. B: At., Mol. Opt. Phys. 37, 2004 (1961). 31S. M. O’Malley and D. R. Beck, J. Phys. B: At., Mol. Opt. Phys. 38, 2656 (2005). 32D. R. Lide, Handbook of Chemistry and Physics, Spectroscopic Constants of Diatomic Molecules, 95th ed. (CRC Press, Boca Raton, FL, 2014-2015), pp. 9–106. 33S. E. Waller, M. Ray, B. L. Yoder, and C. C. Jarrold, J. Phys. Chem. A 117, 13919 (2013). 34H. Wu, X. Li, X. Wang, C. Ding, and L. Wang, J. Chem. Phys. 109, 449 (1998). 35R. Klingeler, G. Lüttgens, N. Pontius, R. Rochow, P. S. Bechthold, M. Neeb, and W. Eberhardt, Euro. Phys. J. D 9, 263 (1999). 36J. O. Kafader, M. Ray, J. E. Topolski, and C. C. Jarrold, “Anion photoelectron spectroscopy of cerium oxide clusters” (unpublished). 37N. M. Souza-Neto, J. Zhao, E. E. Alp, G. Shen, S. V. Sinogeikin, G. Lapertot, and D. Haskel, Phys. Rev. Lett. 109, 026503 (2012). 16V.

Low-lying electronic structure of EuH, EuOH, and EuO neutrals and anions determined by anion photoelectron spectroscopy and DFT calculations.

The anion photoelectron (PE) spectra of EuH(-) and the PE spectrum of overlapping EuOH(-) and EuO(-) anions are presented and analyzed with supporting...
1KB Sizes 0 Downloads 6 Views