J. Phys. Chem.

A 2005, 109, 11521-11525 11521

Electronic Structure Differences in ZrO2 vs HfO2†

Weijun Zheng and Kit H. Bowen, Jr.*

Departments of Chemistry and Materials Science, Johns Hopkins UniVersity, Baltimore, Maryland 21218

Jun Li,‡ Iwona Da̧bkowska,§ and Maciej Gutowski*,§,#

W. R. Wiley EnVironmental Molecular Sciences Laboratory, Pacific Northwest National Laboratory,
Richland, Washington 99352, Department of Chemistry, UniVersity of Gdańsk, 80-952 Gdańsk, Poland, and
Chemical Sciences DiVision, Pacific Northwest National Laboratory, Richland, Washington 99352
ReceiVed: June 30, 2005; In Final Form: October 18, 2005

Although ZrO2 and HfO2 are, for the most part, quite similar chemically, subtle differences in their electronic
structures appear to be responsible for differing MO2/Si (M ) Zr, Hf) interface stabilities. To shed light on
the electronic structure differences between ZrO2 and HfO2, we have conducted joint experimental and
theoretical studies. Because molecular electron affinities are a sensitive probe of electronic structure, we
have measured them by conducting photoelectron spectroscopic experiments on ZrO2- and HfO2-. The adiabatic
electron affinity of HfO2 was determined to be 2.14 ( 0.03 eV, and that of ZrO2 was determined to be 1.64
( 0.03 eV. Concurrently, advanced electronic structure calculations were conducted to determine electron
affinities, vibrational frequencies, and geometries of these systems. The calculated CCSD(T) electron affinities
of HfO2 and ZrO2 were found to be 2.05 and 1.62 eV, respectively. The molecular results confirm earlier
predictions from solid state calculations that HfO2 is more ionic than ZrO2. The excess electron in MO2-
occupies an sd-type hybrid orbital localized on the M atom (M ) Zr, Hf). The structural parameters of ZrO2
and HfO2 and their vibrational frequencies were found to be very similar. Upon the excess electron attachment,
the M-O bond length increases by ca. 0.04 Å, the OMO angle increases by 2-4°, and frequencies of all
vibrational modes become smaller, with the stretching modes being shifted by 30-50 cm-1 and the bending
mode by 15-25 cm-1. Together, these studies unveil significant differences in the electronic structures of
ZrO2 and HfO2 but not in their structural or vibrational characteristics.

Introduction Nevertheless, in the solid state, differences emerge between

ZrO2 and HfO2. For example, significant differences in elastic
Zirconia (ZrO2) and hafnia (HfO2) are important materials behavior and transition pressures were suggested for ZrO2 and
due to their present and potential future applications in HfO2,16 and the temperature-induced monoclinic to tetragonal
microelectronics,1-4 catalysis,5-7 and ceramics.8-11 Zirconium transition is about 500 K higher in HfO2 than in ZrO2.17
([Kr]4d25s2) and hafnium ([Xe]4f145d26s2) reside in the same Moreover, recent theoretical and experimental work18 found that
group of the periodic table, with the most obvious electronic the thermodynamic stabilities of ZrO2 and HfO2 films are quite
structure difference between them being that hafnium possesses different when they are in contact with silicon. In fact, of
a closed subshell of f-electrons and zirconium has no f-electrons. substantial practical interest, the same study indicates that the
Also, as a result of ensuing lanthanide contraction,12 both the HfO2/Si interface is stable with respect to silicide formation,
atomic and the ionic radii of these atoms are nearly the same. whereas the ZrO2/Si interface is not, making HfO2 a contender
As a result of these similarities, Cotton et al.12 state “The as a replacement for SiO2 as a high-k gate dielectric.18 Because
chemistries of hafnium and zirconium are more nearly identical thermodynamic stability is a critical property for microelectronic
than for any other two congeneric elements.” Indeed, infrared devices, it is important to better understand the source of the
spectra,13 microwave spectra,14,15 and theoretical calculations13-15
difference. Given that most structural properties of ZrO2 and
show that the vibrational frequencies of molecular ZrO2 and
HfO2 are very similar, differences between their solid state
HfO2 are quite close in value, and their bond lengths and bond
properties must trace back to subtle differences in electronic
angles are very similar. This is a manifestation of the well-
structure. The difference between adiabatic electron affinities
known chemical similarity of Zr and Hf, which is exemplified
of otherwise similar molecules provides a sensitive measure of
in their geochemistry: Hf is found in all zirconium minerals
their electronic structure differences. To determine these values
and the separation of the two elements is difficult.12
experimentally, we measured the gas phase, anion photoelectron
spectra of ZrO2- and HfO2-, whereas to obtain them theoreti-
† Part of the special issue “Jack Simons Festschrift”.
* Corresponding authors. E-mail: K.H.B., kbowen@[Link]; M.G.,
cally, we conducted electronic structure calculations on MO2
[Link]@[Link]. and MO2- (M ) Zr, Hf) using density functional theory (DFT)
‡ W. R. Wiley Environmental Molecular Sciences Laboratory, Pacific
and wave function-theory-based ab initio methods.
Northwest National Laboratory.
§ University of Gdańsk. Molecular ZrO2 and HfO2 were previously reported as polar
# Chemical Sciences Division, Pacific Northwest National Laboratory. structures.13-15,19 The results of a matrix isolation study indicated
10.1021/jp053593e CCC: $30.25 © 2005 American Chemical Society
Published on Web 11/16/2005
11522 J. Phys. Chem. A, Vol. 109, No. 50, 2005 Zheng et al.

TABLE 1: Experimental Electron Affinities and Vibration Frequencies for TiO2, ZrO2, and HfO2a
adiabatic electron affinity vib freq from PES vib freq from matrix IRb
TiO2 1.59 ( 0.03 eV c
940 ( 40 cm (V1)
-1 c
946.9 cm-1 (V1), 917.1 cm-1 (V3)
ZrO2 1.64 ( 0.03 eVd 887 ( 40 cm-1 (V1)d 884.3 cm-1 (V1), 818.0 cm-1 (V3)
HfO2 2.14 ( 0.03 eVd 887 ( 40 cm-1 (V1)d 883.4 cm-1 (V1), 814.0 cm-1 (V3)
a
V1 symmetric stretching, V3 asymmetrical stretching, V2 bending, V2 is not shown here. b Reference 13. c Reference 38. d This PES experiment.

that the MO2 (M ) Zr, Hf) is bent, and the upper limit for
∠OMO was assigned to 113 ( 5° and 115 ( 5° for M ) Zr
and Hf, respectively.13 The stretching frequencies of these C2V
structures are strikingly similar: V1(a1) ) 884.3 and 883.4 cm-1,
V3(b2) ) 818 and 814 cm-1, for ZrO2 and HfO2, respectively;
see also Table 1.13 More accurate structures were determined
from rotational spectra: the Zr-O bond length is 1.7710 (
0.0007 Å, ∠OZrO ) 108.11 ( 0.08°,14 the Hf-O bond length
is 1.7764 ( 0.0004 Å, ∠OHfO ) 107.51 ( 0.01°.15 The electric
dipole moments determined from Stark-effect measurements are
7.80 and 7.92 D for ZrO2 and HfO2, respectively.13,14 Finally,
the frequency of the bending vibration of ZrO2 was estimated
to be 290 cm-1, on the basis of its inertial defect.14

Methods
Experimental Methods. Anion photoelectron spectroscopy
is conducted by crossing a mass-selected beam of negative ions
with a fixed-frequency laser beam and energy-analyzing the
resultant photodetached electrons. The technique is governed
by the following energy-conserving relationship:

hV ) EKE + EBE (1)

where hV is the photon energy, EKE is the measured electron

kinetic energy, and EBE is the electron binding energy. The Figure 1. Photoelectron spectra of ZrO2- and HfO2- recorded with
experiment was conducted on an apparatus consisting of a Nd: 3.493 eV photons.
YAG laser vaporization source, a linear time-of-flight mass
spectrometer, a mass gate, a momentum decelerator, a second for Zr and Hf.26 As the excess electron in the anions are localized
Nd:YAG laser, and a magnetic bottle energy analyzer. The at the metal centers, the (8s7p6d)/[6s5p3d] basis sets of Zr and
resolution of our magnetic bottle electron energy analyzer was Hf were further expanded with polarization and diffuse functions
∼35 meV at an EKE of ∼1 eV. The apparatus has been as described below. A small basis set (denoted as S) on M (M
described in the detail elsewhere.20 ) Zr, Hf) included a set of diffuse d functions ú(M,d) ) 0.01
The anions, ZrO2- and HfO2-, were generated by a laser and a large basis set (denoted as L) included two f and one g
vaporization source by focusing the pulsed (10 Hz), second polarization functions27 and two diffuse s and two diffuse d
harmonic (532 nm) beam of a Nd:YAG laser onto a zirconium functions. The exponents of s and d diffuse functions were
rod (6.2 mm diameter, 99%, Alfa AESAR #10443) in the former obtained from even-tempered progressions with a constant of
case and onto a hafnium rod (5.0 mm diameter, 97%, Good- 0.5, initiated from the most diffuse s and d functions in the
fellow HF007910) in the latter case. In both cases, the target original (8s7p6d)/[6s5p3d] basis set. The scalar relativistic
rod was continuously rotated and translated so that the laser effects were taken into account via the relativistic pseudopo-
struck a different spot each time it was fired. The carrier gas tentials obtained from Wood-Boring all-electron calculations
used in the laser vaporization source was highly purified helium, on the atoms.26 Spin-orbit coupling effects might be important
which issued through a pulsed valve, having a backing pressure for the hafnium compounds but they were neglected in the
of ∼4 atm. Oxygen was not added to the carrier gas, because current approach. The CCSD(T) forces and curvatures were
the oxygen on the surface of the zirconium and hafnium rods calculated numerically. Excited electronic states of neutral ZrO2
was found to be adequate for making the metal oxide anions of and HfO2 were calculated using time-dependent density func-
interest. tional theory (TDDFT) with the L basis set and the B3LYP
Franck-Condon analyses were conducted on both the ZrO2/ exchange-correlation functional. The DFT calculations were
ZrO2- and the HfO2/HfO2- systems using the PESCAL performed with Gaussian9828 and CCSD(T) calculations with
program.21 MOLPRO.29 Molden was used for the visualization of singly
Theoretical Methods. Initial calculations were performed occupied molecular orbitals.30
with relativistic SBKJ pseudopotentials and basis sets22 and the
B3LYP exchange-correlation functional.23 Advanced ab initio Results
calculations were further carried out using the coupled cluster
method with single, double, and noniterative triple excitations Experimental Results. The photoelectron spectra of ZrO2-
(CCSD(T)).24 All electrons were treated on oxygens with aug- and HfO2- are presented in Figure 1. They were both recorded
cc-pVTZ basis sets.25 Stuttgart energy-consistent, small-core with 3.493 eV photons, the third harmonic of Nd:YAG laser.
pseudopotentials and the (8s7p6d)/[6s5p3d] basis sets were used Because Zr and Hf each have several isotopes, we recorded these
Electronic Structure Differences in ZrO2 vs HfO2 J. Phys. Chem. A, Vol. 109, No. 50, 2005 11523

spectra on mass peaks, 90ZrO2- and 176HfO2- to avoid the TABLE 2: Calculated M-O Bond Lengths (Å), OMO Bond
possibility of contamination from nearby ZrO2H- and HfO2H-. Angles (deg), Adiabatic Electron Affinities (eV), and
The electron binding energies (EBE) in both spectra were Vibrational Frequencies (cm-1) for ZrO2 and HfO2*
calibrated against the atomic lines of the copper anion, and the system method M-O ∠OMO AEA V1 V2 V3
uncertainty of the calibration is (0.03 eV. The spectrum of ZrO2 CCSD(T)/L 1.797 109.6 1.62 887 281 835
ZrO2- is consistent with our former low resolution result.31 The CCSD(T)/S 1.816 109.4 1.67 876 272 842
spectrum of ZrO2- exhibits a main peak at EBE ) 1.64 eV, B3LYP/SBKJ 1.806 108.0 1.85 906 295 854
corresponding to the transition from V ) 0 of the ground state ZrO2- CCSD(T)/L 1.833 111.7 - 839 266 785
CCSD(T)/S 1.853 111.9 - 829 255 790
of ZrO2- anion to V ) 0 of the ground state of ZrO2 neutral. B3LYP/SBKJ 1.843 110.3 - 850 272 799
This implies that the adiabatic electron affinity (AEA) of ZrO2 HfO2 CCSD(T)/L 1.815 109.9 2.05 869 266 801
is 1.64 ( 0.03 eV. The peaks at EBE ) 1.75 and 1.86 eV CCSD(T)/S 1.826 109.8 2.07 859 254 805
correspond to excited vibrations (V ) 1 and 2, respectively) in B3LYP/SBKJ 1.797 107.3 2.32 898 293 821
the electronic ground state of neutral ZrO2. The energy spacings HfO2- CCSD(T)/L 1.855 114.3 - 819 242 752
CCSD(T)/S 1.867 115.1 - 808 224 753
between adjacent peaks are both 0.11 ( 0.005 eV, corresponding B3LYP/SBKJ 1.833 111.6 - 844 255 772
to a vibrational frequency of 887 ( 40 cm-1 for ZrO2. This a
value is consistent with that of 884.3 cm-1 obtained for the The values of AEA include vibrational zero-point energy correc-
tions. V1 is the symmetric stretching mode, V2 is the bending mode,
symmetrical stretching mode (V1) of ZrO2 in matrix infrared and V3 is the asymmetric stretching mode. “L” and “S” stand for the
spectroscopy experiments;13 see Table 1. In analogous fashion, “large” and “small” basis sets, see text. For “SBKJ”, see ref 22.
the photoelectron spectrum of HfO2- exhibits a main peak at
EBE ) 2.14 eV, corresponding to the transition from V ) 0 of
the ground state of HfO2- anion to V ) 0 of the ground state of
HfO2 neutral. This implies that the AEA of HfO2 is 2.14 (
0.03 eV. The peaks at EBE ) 2.25 and 2.36 eV correspond to
excited vibrations (V ) 1 and 2, respectively) in the electronic
ground state of neutral HfO2. Again, the energy spacings
between adjacent peaks are both 0.11 ( 0.005 eV, corresponding
to a vibrational frequency of 887 ( 40 cm-1 for HfO2. This
value is also consistent with that of 883.4 cm-1 obtained for
the symmetrical stretching mode (V1) of HfO2 in matrix infrared
spectroscopy experiments;13 see Table 1. We should mention
that the uncertainties of V1 frequencies obtained from the Figure 2. Excess electron in MO2- (M ) Zr, Hf) occupying an a1-
photoelectron spectra are larger than those from the matrix symmetry orbital dominated by the M’s ns and (n - 1)d orbitals.
infrared spectra.13
experiment for the Hf systems. The B3LYP values of AEA are
Franck-Condon analyses were conducted on both the ZrO2/ overestimated with respect to the CCSD(T) results by 0.2-0.3
ZrO2- and the HfO2/HfO2- systems and the results fitted to eV, which is typical for this exchange-correlation functional.32,33
their respective photoelectron spectra. For the ZrO2- photo- Our calculations confirm that the neutral ZrO2 and HfO2 have
electron spectrum, our best fit was obtained when the symmetric C2V symmetry and a very similar bond length and ∠OMO.14,15
stretch (V1) of neutral ZrO2 equaled 871 cm-1 and when the The M-O distances calculated at the CCSD(T)/L level are 1.797
AEA of ZrO2 equaled 1.65 eV. Both of these values are Å for ZrO2 and 1.815 Å for HfO2, thus ca. 0.03-0.04 Å longer
consistent with our a priori assignment of the ZrO2- photo- than the microwave spectroscopy predictions.14,15 The bond
electron spectrum. For the HfO2- photoelectron spectrum, our lengths are sensitive to the selection of one-electron basis set
best fit was obtained when the symmetric stretch (V1) of neutral (L vs S) and the computational method (CCSD(T) vs B3LYP).
HfO2 equaled 872 cm-1 and when the AEA of HfO2 equaled The CCSD(T)/L values of ∠OMO of 109.6° and 109.9° are in
2.15 eV. Both of these values are again consistent with our a excellent agreement with the previous findings of 108.11 (
priori assignment of the HfO2- photoelectron spectrum. For both 0.08° and 107.51 ( 0.01° for ZrO214 and HfO2,15 respectively.
systems, the measured values of AEA and the V1 frequencies The calculated, harmonic, CCSD(T)/L frequencies for the
are in excellent agreement with the calculated values, vide infra symmetrical stretching mode of 887 cm-1 (ZrO2) and 869 cm-1
and Table 2. (HfO2) are consistent with the 887 ( 40 cm-1 value resulting
Theoretical Results. The most stable structures of ZrO2 and from the photoelectron spectra. However, the more accurate
HfO2 in the ground electronic states have C2V symmetry. These results from matrix infrared measurements13 of 884.3 cm-1
are closed-shell systems; thus the symmetry is 1A1. The (ZrO2) and 883.4 cm-1 (HfO2) reveal that the calculated values
structures of anionic ZrO2 and HfO2 maintain the C2V symmetry might be inaccurate by ca. 15-20 cm-1. The calculated
and the excess electron occupies a fully symmetric orbital (a1); harmonic CCSD(T)/L frequencies for the asymmetrical stretch-
see Figure 2. Thus the symmetry of the open-shell electronic ing mode are also inaccurate by ca. 20 cm-1. The calculated
anionic states is 2A1. The calculated bond lengths, OMO bond frequency for the bending mode of ZrO2 (281 cm-1) agrees well
angles, adiabatic electron affinities, and harmonic vibration with the 290 cm-1 value determined from the inertial defect.14
frequencies for ZrO2 and HfO2 are listed in Table 2. The The B3LYP vibrational force constants are systematically larger
reported values AEA include vibrational zero-point energy for ZrO2 than for HfO2 and so are the reduced masses resulting
contributions. The CCSD(T)/L values of AEA are 1.62 and 2.05 in very similar values of vibrational frequencies for the two
eV for ZrO2 and HfO2, respectively, thus in excellent agreement molecules.
with the corresponding measured values of 1.64 ( 0.03 and Upon the excess electron attachment the M-O bond lengths
2.14 ( 0.03 eV. There are only small differences in the values increase by ca. 0.04 Å and ∠OMO’s increase by 2-4°; see
of AEA obtained with the S and L basis sets. Including spin- Table 2. The differences in geometries between the neutral and
orbit coupling effects may further improve the agreement with the anion are along the symmetrical stretching mode, V1, and
11524 J. Phys. Chem. A, Vol. 109, No. 50, 2005 Zheng et al.

the bending mode, V2. The frequencies of all vibrational modes phase. This result is surprising because usually for congeneric
decrease upon the excess electron attachment with a larger shift elements an increase in the atomic number leads to an increase
(30-50 cm-1) for the stretching modes and a smaller shift of of the atomic or ionic radius. For typical elements from the
15-25 cm-1 for the bending mode. second and third transition metal rows such an increase is
ZrO2 and HfO2 are polar molecules with the experimentally reduced by the electron occupation of the 4f and relativistic
measured dipole moments of 7.8014 and 7.9215 D, respectively. effects, whereas for Zr and Hf the relative sizes are apparently
The calculated dipole moments are consistent with experiment reversed, as in the case of monovalent coin metals.37 Indeed,
in predicting a larger polarity of HfO2 than of ZrO2, but the the recommended ionic radius for “Hf(+4)” is smaller by 0.01
absolute values of dipole moments are overestimated. Indeed, Å than that for “Zr(+4)”,12 which qualitatively confirms the
the corresponding B3LYP/S values are 8.00 and 8.06 D. The structural predictions for bulk phases. For the molecular systems,
overestimation is even larger at the CCSD/S level with the however, the measured M-O distance is larger for ZrO2 (1.7710
corresponding values being 8.40 and 8.60 D. The effective ( 0.0007 Å)14 than for HfO2 (1.7764 ( 0.0004 Å).15 Clearly,
charge on the metal, determined from the Mulliken population the parallelism between structural properties of bulk phases and
analysis with the B3LYP/SBKJ wave functions, is +1.15 and molecular systems is limited.
+1.10 e, for HfO2 and ZrO2, respectively. The excess electron The neutral ZrO2 and HfO2 are highly polar and the excess
is found to localize on the metal atom, where the positive pole electron is described by a sd-type hybrid orbital localized on
of the dipole is (Figure 2). An excess electron bound by a polar the metal atom. In the case of HfO2- small amplitudes can be
molecule, which does not contain a transition metal atom, is also identified on oxygens (Figure 2). The availability of
typically described by a fully symmetric sp-type hybrid orbital; (n - 1)d orbitals makes the charge distribution of an excess
see ref 34 for numerous examples. In the case of ZrO2 and HfO2, electron different than in anions of polar molecules that do not
however, the ns and (n - 1)d orbitals are available on the metal contain a transition metal atom. One might only speculate what
atom, whereas the np orbitals are higher in energy. The excess the charge distribution of an excess electron would be in polar
electron is then described by a fully symmetric sd-type hybrid molecules with available (n - 2)f orbitals, such as CeO2 or
orbital with a1 symmetry (Figure 2). BaO anions.
The low lying electronic states of the neutral ZrO2 and HfO2 Because Ti, Zr, and Hf are all in the same group of the
were characterized at the TDDFT/B3LYP/S level of theory. For periodic table, it is also interesting to compare the electron
both systems, the lowest excited singlet and triplet states are of affinities and vibrational frequencies of ZrO2 and HfO2 with
B1 symmetry, with the b1 orbital localized on the metal atom. those of TiO2. Wu and Wang38 reported the photoelectron
The more stable 3B1 states are at 2.15 and 1.55 eV with respect spectrum of TiO2-. They found the adiabatic electron affinity
to the ground singlet state for ZrO2 and HfO2, respectively. The of TiO2 to be 1.59 ( 0.03 eV and the frequency of V1 to be 940
related 1B1 states are less stable by 0.07-0.08 eV. Thus the ( 40 cm-1. These properties of TiO2, HfO2, and ZrO2 are
excited electronic states of the neutral do not contribute to the compared in Table 1. Note that ZrO2 and HfO2 have essentially
photoelectron spectra recorded with the 3.493 eV photons. identical vibrational frequencies, whereas the V1 vibrational
frequency of TiO2 differs from them. This indicates differences
Discussion in bonding strength and reduced masses between TiO2 and ZrO2
(HfO2). Consistent with this, the atomic and ionic radii of
The significant electron affinity difference between ZrO2 and zirconium and hafnium are the same to within 0.01 Å, whereas
HfO2, combined with the lack of a measurable difference in the ionic radius of Ti(+4) is smaller by 0.1 Å than those of
their V1 vibrational frequencies and very similar molecular Zr(+4) and Hf(+4).12 The comparison reverses, however, when
geometries, appears to indicate electronic structure rather than adiabatic electron affinities are compared. The electron affinities
size or bonding effects as the primary basis for reactivity and of TiO2 and ZrO2 are rather similar, differing by only 0.05 eV,
stability differences between ZrO2 and HfO2. The difference in whereas the electron affinities of HfO2 and ZrO2 differ
stability of the ZrO2/Si and HfO2/Si interfaces with respect to substantially, i.e., by 0.50 eV. These comparisons reinforce the
formation of metal silicides was interpreted in terms of a larger implication that differences in the electronic structures of Hf
stability of (i) HfO2 than ZrO2 and (ii) zirconium rather than and Zr are at the root of chemical differences between ZrO2
hafnium silicides.18 The current results shed new light on and HfO2. This points to the nonnegligible role of f-electrons
differences in electronic structure of ZrO2 and HfO2. and relativistic effects in the chemistry of hafnium compounds.
Recent computational results on various polymorphs of HfO2 It may be useful to mention that there are other examples
and ZrO2 indicate that the former is more ionic than the latter.35 among atomic metals, metal-ligand complexes, and metal dimer
The former also has a larger band gap across all polymorphs. and cluster anions, in which the excess electron occupies a metal
These results are consistent with the finding that the calculated s orbital, which likewise shows an increased electron affinity
heat of formation for the monoclinic phase is larger by 0.60 for the third transition series congener, mirroring the trend
eV/formula unit for HfO2 than for ZrO2, with the experimental observed here.39 This pattern has been ascribed to relativistic
difference being 0.49 eV.17 The difference in ionicity is also effects, resulting in more strongly bound s electrons in the third
confirmed by dipole moments of HfO2 and ZrO2, 7.92 and transition series. Because the orbital occupied by the extra
7.80 D, respectively.15,14 A difference in the dipole moment electron in ZrO2- and HfO2- has a significant metal s character
might contribute to the different values of AEA for HfO2 and (Figure 2), the reported values of AEA are consistent with this
ZrO2; see Tables 1 and 2 and Figure 1. All these findings are trend. This observation supports our conclusions that the
consistent with a small difference in the electronegativity of Zr presence of f electrons in the third transition series affects
and Hf: 1.4 and 1.3, respectively, according to Pauling’s scale.36 valence electronic structure controlled by the 5d and 6s orbitals.
How important is the lanthanide contraction for Hf? It has
recently been demonstrated for various polymorphs of MO2 that Acknowledgment. We acknowledge the reviewers for their
the volume per formula unit is smaller for HfO2 than for ZrO2, insightful and constructive comments. K.B.’s portion of this
as if the ionic radius was smaller for “Hf(+4)” than for “(Zr- work was supported by The Division of Materials Science and
(+4)”.35 The difference was 3.9% for the most stable monoclinic Engineering, Office of Basic Energy Sciences, U.S. Department
Electronic Structure Differences in ZrO2 vs HfO2 J. Phys. Chem. A, Vol. 109, No. 50, 2005 11525

of Energy under Grant No. DE-FG0295ER45538. Acknowledg- (21) Ervin, K. M.; Ramond, T. M.; Davico, G. E.; Schwartz, R. L.;
ment is also made to the donors of the Petroleum Research Fund, Casey, S. M.; Lineberger, W. C. J. Phys. Chem. A 2001, 105, 10822. K.
M. Ervin, PESCAL, Fortran program, 2004.
administered by the American Chemical Society, for partial (22) Stevens, W. J.; Basch, H.; Krauss, M. J. Chem. Phys. 1984, 81,
support of this research under grant No. 28452-AC6. M.G.’s 6026. Stevens, W. J.; Krauss, M.; Basch, H.; Jasien, P. G. Can. J. Chem.
and J.L.’s portion of this work was supported by DOE Basic 1992, 70, 612.
Energy Sciences, Division of Chemical Sciences. I.D. is a holder (23) Becke, A. D. Phys. ReV. A 1988, 38, 3098. Becke, A. D. J. Chem.
Phys. 1993, 98, 5648. Lee, C.; Yang, W.; Paar, R. G. Phys. ReV. B 1988,
of a Foundation for Polish Science (FNP) Award. Computing 37, 785.
resources were available through (i) a Computational Grand (24) Taylor, P. R. In Lecture Notes in Quantum Chemistry II; Roos, B.
Challenge Application grant from the Molecular Sciences O., Ed.; Springer-Verlag: Berlin, 1994.
Computing Facility (MSCF) in the Environmental Molecular (25) Kendall, R. A.; Dunning, T. H., Jr.; Harrison, R. J. J. Chem. Phys.
1992, 96, 6796.
Sciences Laboratory, a national scientific user facility sponsored
(26) Andrae, D.; Haeussermann, U.; Dolg, M.; Stoll, H.; Preuss, H.
by the U.S. DOE, OBER and located at PNNL, and (ii) the Theor. Chim. Acta 1990, 77, 123.
National Energy Research Scientific Computing Center (NER- (27) Martin, J. M. L.; Sundermann, A. J. Chem. Phys. 2001, 114, 3408.
SC). PNNL is operated by Battelle for the U.S. DOE under (28) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb,
Contract DE-AC06-76RLO 1830. M. A.; Cheeseman, J. R.; Zakrzewski, V. G.; Montgomery, J. A., Jr.;
Stratmann, R. E.; Burant, J. C.; Dapprich, S.; Millam, J. M.; Daniels, A.
D.; Kudin, K. N.; Strain, M. C.; Farkas, O.; Tomasi, J.; Barone, V.; Cossi,
References and Notes M.; Cammi, R.; Mennucci, B.; Pomelli, C.; Adamo, C.; Clifford, S.;
Ochterski, J.; Petersson, G. A.; Ayala, P. Y.; Cui, Q.; Morokuma, K.; Malick,
(1) Copel, M.; Gribelyuk, M.; Gusev, E. Appl. Phys. Lett. 2000, 76,
D. K.; Rabuck, A. D.; Raghavachari, K.; Foresman, J. B.; Cioslowski, J.;
436.
Ortiz, J. V.; Stefanov, B. B.; Liu, G.; Liashenko, A.; Piskorz, P.; Komaromi,
(2) Wilk, G. D.; Wallace, R. M.; Anthony, J. M. J. Appl. Phys. 2001,
I.; Gomperts, R.; Martin, R. L.; Fox, D. J.; Keith, T.; Al-Laham, M. A.;
89, 5243 and references therein.
Peng, C. Y.; Nanayakkara, A.; Gonzalez, C.; Challacombe, M.; Gill, P. M.
(3) Jeon, T. S.; White, J. M.; Kwong, D. L. Appl. Phys. Lett. 2001,
W.; Johnson, B. G.; Chen, W.; Wong, M. W.; Andres, J. L.; Head-Gordon,
78, 368.
M.; Replogle, E. S.; Pople, J. A. Gaussian 98, revision A.11; Gaussian,
(4) Robertson, J. J. Vac. Sci. Technol. B 2000, 18, 1785.
Inc.: Pittsburgh, PA, 1998.
(5) Yang, X.; Jentoft, F. C.; Jentoft, R. E.; Girgsdies, F.; Ressler, T.
Catal. Lett. 2002, 81 (1-2), 25. (29) MOLPRO is a package of ab initio programs written by H.-J.
(6) Rossi, S. De.; Ferraris, G.; Valigi, M.; Gazzoli, D. Appl. Catal. Werner, P. J. Knowles, M. Schütz, R. Lindh, P. Celani, T. Korona, G.
A-Gen. 2002, 231 (1-2), 173. Rauhut, F. R. Manby, R. D. Amos, A. Bernhardsson, A. Berning, D. L.
(7) Mulas, G.; Deledda, S.; Monagheddu, M.; Cocco, G.; Cutrufello, Cooper, M. J. O. Deegan, A. J. Dobbyn, F. Eckert, C. Hampel, G. Hetzer,
M. G.; Ferino, I.; Solinas, V. Metastable, Mechanically Alloyed and A. W. Lloyd, S. J. McNicholas, W. Meyer, M. E. Mura, A. Nicklass, P.
Nanocrystalline Materials; 2000; Vol. 343, p 889. Palmieri, R. Pitzer, U. Schumann, H. Stoll, A. J. Stone R. Tarroni, and T.
(8) Andrievsky, E. R.; Redko, V. P.; Lopato, L. M. Euro Ceramics Thorsteinsson.
VII, 2002, 206 (2), 1113. (30) Schaftenaar, G.; Noordik, J. H. J. Comput.-Aided Mol. Design 2000,
(9) Hastie, J. W.; Bonnell, D. W.; Schenck, P. K. J. Nucl. Mater. 2001, 14, 123.
294 (1-2), 175. (31) Thomas, O. C.; Xu, S.-J.; Lippa, T. P.; Bowen, K. H. J. Cluster
(10) Sasakim, H. J. Eur. Ceram. Soc. 1995, 15 (4), 329. Sci. 1999, 10, 525.
(11) Andrews, M. J.; Ferber, M. K.; Lara-Curzio, E. J. Eur. Ceram. (32) Rak, J.; Skurski, P.; Gutowski, M. J. Chem. Phys. 2001, 114, 10673.
Soc. 2002, 22 (14-15), 2633. (33) Gutowski, M.; Dabkowska, I.; Rak, J.; Xu, S.; Nilles, J. M.; Radisic,
(12) Cotton, F. A.; Wilkinson, G.; Murillo, C. A. Bochmann, M. D. K.; Bowen, H. Eur. Phys. J. D 2002, 20, 431.
AdVanced Inorganic Chemistry; John Wiley & Sons: New York, 2000. (34) Gutowski, M.; Skurski, P.; Boldyrev, A. I.; Simons, J.; Jordan, K.
(13) Chertihin, G. V.; Andrews, L. J. Phys. Chem. 1995, 99, 6356. D. Phys. ReV. A 1996, 54, 1906. Gutowski, M.; Jordan, K. D.; Skurski, P.
(14) Brugh, D. J.; Suenram, R. D.; Stevens, W. J. J. Chem. Phys. 1999, J. Phys. Chem. A 1998, 102, 2624. Gutowski, M.; Hall, C.; Adamowicz,
111, 3526. L.; Hendricks, J. H.; de Clercq, H. L.; Lyapustina, S. A.; Nilles, J. M.; Xu,
(15) Lesarri, A.; Suenram, R. D.; Brugh, D. J. Chem. Phys. 2002, 117, S. J.; Bowen, K. H., Jr. Phys. ReV. Lett. 2002, 88, 143001. Gutowski, M.;
9651. Skurski, P. Chem. Phys. Lett. 1999, 303, 65.
(16) Lowther, J. E. Mater. Res. Bull. 2003, 28, 189. (35) Jaffe, J. E.; Bachorz, R. A.; Gutowski, M. Phys. ReV. B 2005, 72,
(17) Barin, I. Thermochemical Data of Pure Substances, 3rd ed.; VCH: 144107.
Wienheim, 1995; Vol. I and Vol. II.
(18) Gutowski, M.; Jaffe, J. E.; Liu, C.-L.; Stoker, M.; Hegde, R. I.; (36) Pauling, L. The Nature of the Chemical Bond, 3rd ed.; Cornell
Rai, R. S.; Tobin, P. J. Appl. Phys. Lett. 2002, 80, 1897. University Press: Ithaca, NY, 1960.
(19) Kaufman, M.; Muenter, J.; Klemperer, W. J. Chem. Phys. 1967, (37) Liao, M. S.; Schwarz, W. H. E. Acta Crystallogr. 1994, B50, 9.
47, 3365. (38) Wu, H. B.; Wang, L.-S. J. Chem. Phys. 1997, 107, 8221.
(20) Thomas, O. C.; Zheng, W.; Bowen, K. H. J. Chem. Phys. 2001, (39) Bengali, A. A.; Casey, S. M.; Cheng, C-L.; Dick, J. P.; Fenn, P.
114, 551. T.; Villalta, P. W.; Leopold, D. G. J. Am. Chem. Soc. 1992, 114, 5257.