Results for aliovalent doping of CeBr3 with Ca2+Paul Guss, Michael E. Foster, Bryan M. Wong, F. Patrick Doty, Kanai Shah, Michael R. Squillante, Urmila

Shirwadkar, Rastgo Hawrami, Joshua Tower, and Ding Yuan Citation: Journal of Applied Physics 115, 034908 (2014); doi: 10.1063/1.4861647 View online: http://dx.doi.org/10.1063/1.4861647 View Table of Contents: http://scitation.aip.org/content/aip/journal/jap/115/3?ver=pdfcov Published by the AIP Publishing

[This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to ] IP: 128.31.3.8

On: Fri, 17 Jan 2014 19:13:52

Results for aliovalent doping of CeBr3 with Ca21

Paul Guss,1,a) Michael E. Foster,2 Bryan M. Wong,2 F. Patrick Doty,2 Kanai Shah,3

Michael R. Squillante,3 Urmila Shirwadkar,3 Rastgo Hawrami,3 Joshua Tower,3

and Ding Yuan4

1Remote Sensing Laboratory – Nellis, P. O. Box 98521, Las Vegas, Nevada 89193-8521, USA2Materials Chemistry Department, Sandia National Laboratories, California, P. O. Box 969, Livermore,California 94551-0969, USA3Radiation Monitoring Devices, Inc., 44 Hunt Street, Watertown, Massachusetts 02472, USA4National Security Technologies, LLC, Los Alamos Operations, P. O. Box 809, Los Alamos,New Mexico 87544-0809, USA

(Received 20 August 2013; accepted 23 December 2013; published online 17 January 2014)

Despite the outstanding scintillation performance characteristics of cerium tribromide (CeBr3) and

cerium-activated lanthanum tribromide, their commercial availability and application are limited

due to the difficulties of growing large, crack-free single crystals from these fragile materials. This

investigation employed aliovalent doping to increase crystal strength while maintaining the optical

properties of the crystal. One divalent dopant (Ca2þ) was used as a dopant to strengthen CeBr3

without negatively impacting scintillation performance. Ingots containing nominal concentrations

of 1.9% of the Ca2þ dopant were grown, i.e., 1.9% of the CeBr3 molecules were replaced by CaBr2

molecules, to match our target replacement of 1 out of 54 cerium atoms be replaced by a calcium

atom. Precisely the mixture was composed of 2.26 g of CaBr2 added to 222.14 g of CeBr3.

Preliminary scintillation measurements are presented for this aliovalently doped scintillator.

Ca2þ-doped CeBr3 exhibited little or no change in the peak fluorescence emission for 371 nm

optical excitation for CeBr3. The structural, electronic, and optical properties of CeBr3 crystals

were studied using the density functional theory within the generalized gradient approximation.

Calculated lattice parameters are in agreement with the experimental data. The energy band

structures and density of states were obtained. The optical properties of CeBr3, including the

dielectric function, were calculated. VC 2014 AIP Publishing LLC.

[http://dx.doi.org/10.1063/1.4861647]

I. INTRODUCTION

Cerium-activated and cerium-based crystals appear to

be particularly promising fast scintillators.1,2 The discovery

of cerium-doped lanthanum tribromide (LaBr3:Ce) and

cerium-doped lanthanum trichloride (LaCl3:Ce)3–5 and the

characterization of their outstanding scintillation performance

sparked an interest in the lanthanide trihalide family of scin-

tillators as candidates for replacing thallium-activated sodium

iodide (NaI:Tl), the industry standard since its discovery in

1948.6,7 Since then, numerous derivatives of this family of

scintillators have been explored, including undoped,

self-activated cerium tribromide (CeBr3).8 Although the

scintillation performance of these cerium-doped and

cerium-based crystals is outstanding, their mechanical prop-

erties are not generally conducive to growth of large single

crystals.2,9,10

The CeBr3 material was chosen here, as previous work8

showed that it compared well with LaBr3:Ce, but without the

self-activity associated with the natural radioactivity of138La. The energy resolution was good for CeBr3, so this ma-

terial was chosen for this study. The intent here was to estab-

lish methods to model and fabricate the material for better

fracture strength in a way that degrades the energy resolution

as little as possible. However, our resulting crystal, based on

specifications calculated using a methodical density func-

tional theory (DFT) approach to be described, had a superior

energy resolution, for this cerium bromide material, of 3.2%.

The intent of this work was to report not only this final result

but also the methods that led us to this result, and the

approach used.

Cerium trihalides possess excellent scintillation proper-

ties, as do most rare-earth trihalide scintillators, such as high

light output, fast decay time, and excellent energy resolu-

tion.8 Therefore, they are widely used in various fields such

as high-energy physics and positron emission tomography,

as well as some chemical processes involved in the nuclear

industry.11 The structural, electronic, and optical properties

of many lanthanum trihalides, as well as cerium trifluoride

(CeF3), have been extensively investigated using both exper-

imental and theoretical methods.12–21 CeBr3 shows prefera-

ble optical and scintillation properties; for example, it has a

higher light yield than NaI:Tl as well as many other inor-

ganic scintillators.8,22

CeBr3 is a self-activated lanthanide scintillator that has

received considerable recent attention8 due to its linearity of

light output with photon energy; also, its high energy resolu-

tion for gamma spectroscopy is far superior to that of NaI:Tl.

Because the material possesses no intrinsic radioactivity,

CeBr3 has high potential to outperform scintillators such as

LaBr3:Ce or lanthanum-based Elpasolites,23 making it

an excellent candidate for gamma-ray spectroscopy.24,25a)Electronic mail: gusspp@nv.doe.gov.

0021-8979/2014/115(3)/034908/10/$30.00 VC 2014 AIP Publishing LLC115, 034908-1

JOURNAL OF APPLIED PHYSICS 115, 034908 (2014)

[This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to ] IP: 128.31.3.8

On: Fri, 17 Jan 2014 19:13:52

However, due to its hexagonal crystal structure (space group

P63/m, prototype structure UCl3), pure CeBr3 can fracture

during crystal growth, detector fabrication, and subsequent

field use.

Because the crystal fractures easily, manufacturing yield

is low, and the reliability of large crystals is questionable.26

Therefore, significant gains in the practical scale for CeBr3

scintillators will be realized by increasing fracture toughness

of the crystals.2 Aliovalent substitution in which a host ion is

replaced with an ion of different valence (e.g., Ca2þ for

Ce3þ in CeBr3) is a more potent method of strengthening

compared to isovalent substitution (i.e., replacing a fraction

of ions with like-valence ions). In this approach, the forma-

tion of intrinsic defects necessary to maintain charge neutral-

ity results in complexes with long-range interactions in the

crystal. The resulting increase in hardening rate can be

explained in terms of elastic interaction with dislocations.27

Because CeBr3 already exhibits superior scintillation

characteristics,8 the alloying element(s) used to strengthen

the crystal must not degrade the scintillation properties.

Aliovalent alloying provides more strengthening than isova-

lent alloying. Using a basic approximation, solid solution

strengthening based on lattice distortions due to some small

concentration of dopant can be estimated as2

s ¼ c � Gc12; (1)

where G is the shear modulus, c is the concentration of solute

in atomic fraction, and c is a proportionality constant.2,28 For

spherically symmetric distortions, such as those found in iso-

valent alloying, c typically takes on values that are signifi-

cantly smaller than unity, on the order of 10�4 to 10�6. For

tetragonal lattice distortions, such as those created from sol-

ute atoms of a different valence, c can be nearly unity.

Therefore, aliovalent alloying is more effective for a given

concentration of solute.2

Electronic valence effects exert a powerful influence on

material and optical properties. In particular, aliovalent sub-

stitution (making dilute alloys by substituting atoms of a dif-

ferent valence) has improved a variety of technological

applications such as gadolinium titanate (Gd2Ti2O7) ion con-

ductors for solid fuel cells;29–32 gallium indium oxide

(GaInO3) transparent conductors for solar cells;33,34

rechargeable batteries;35,36 thermoelectric devices;37 lead

zirconate titanate (Pb[Zr(x)Ti(1–x)]O3) or piezoelectric,

ceramics;38–41 and cubic zirconium dioxide (ZrO2)

crystals.42–44 These improvements, however, are far from

optimal because the fundamental aliovalent mechanisms are

largely unknown. Within the arena of scintillating materials,

CeBr3 single crystals have shown superior scintillating prop-

erties, and large CeBr3 crystals critically enable improved

identification of radioisotopes. Unfortunately, CeBr3 crystals

are very prone to fracture. Fracture during device testing and

implementation leads to catastrophic failures. Fracture dur-

ing synthesis complicates the growth of CeBr3 crystals in

sizes large enough to be used to identify radioisotopes, and it

accounts for the high production cost that is preventing

large-scale deployment required by homeland security

efforts. Aliovalent substitution was found to strengthen the

material without significantly degrading the scintillating

properties.2 An optimized material design, however, has not

been achieved due to a large number of possible aliovalent

combinations. Most empirical atomistic simulations can

approximately account for charge variation but not electronic

valence effects. These simplistic simulations cannot cor-

rectly reveal aliovalent phenomena such as the valence-

compensating interstitials or vacancies.

From a theoretical point of view, one would like to

explore a computational design beforehand to predict the

effect of aliovalent doping on the mechanical and optical

properties of these materials before they are used in experi-

ments. Indeed, in previous theoretical studies,45,46 it was

shown that first-principles-based approaches were necessary

to accurately predict effects due to defects and radiation.

Although there has been great progress in resolution-of-the-

identity techniques for computationally intensive coupled

cluster methods (described briefly in the Theoretical

Background section, below), these wavefunction-based

approaches are still computationally much more demanding

than DFT methods. Our DFT results for both the pure CeBr3

primitive cell and the (divalent) calcium-doped “supercell,”

which includes 200 atoms and well over 1000 electrons,

were based on spin-polarized, generalized gradient approxi-

mation (GGA) methods47–49 using Projector-Augmented-

Wave (PAW) pseudopotentials.50,51 During this study, it was

discovered that the inclusion of a Hubbard U DFT correc-

tion,52,53 often denoted as DFTþU, was essential for obtain-

ing accurate band structures and band gaps. Within the

Hubbard DFT approach, an onsite penalty function is placed

on the f electrons for each of the Ce atoms to prevent spuri-

ous over-delocalization of the electrons.

II. THEORETICAL BACKGROUND

In this work, state-of-the-art DFTþU approaches,

beyond those of standard DFT, were carried out for obtaining

electronic and optical properties for a Ca2þ-doped CeBr3

bulk material. The intent is not to resolve the many open

questions regarding how best to formulate exchange-

correlation functionals or provide universal benchmarks for

universal material systems. Rather, the calculations are pre-

sented to demonstrate that the DFTþU formalism offers an

efficient and accurate approach for describing the electronic

and optical properties in these novel scintillating materials.

In this work, aliovalent strengthening and the optical

properties of Ca2þ-doped CeBr3 using state-of-the-art, large-

scale, quantum mechanical simulations were explored. These

simulations go far beyond the current state of the art by incor-

porating valence and variable charge effects for both the cohe-

sive energies and optical properties of this doped compound.

The work leads to an understanding that affects a wide spec-

trum of materials of U.S. Department of Homeland Security

(DHS) and the U.S. Department of Energy National Nuclear

Security Administration interests, as mentioned above. In par-

ticular, it leads to a systematic design criterion that can opti-

mize the radiation detection materials explored by DHS.

Over the last few decades, DFT has made tremendous

progress in the accurate description of electronic and optical

034908-2 Guss et al. J. Appl. Phys. 115, 034908 (2014)

[This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to ] IP: 128.31.3.8

On: Fri, 17 Jan 2014 19:13:52

properties of molecules and materials. Based on the

Hohenberg–Kohm theorem,54 which relates electron den-

sities with the external electrostatic potential, DFT can, in

principle, be applied to any quantum-mechanical situation.

Despite the overwhelming success of DFT for predicting mo-

lecular and bulk properties, it is well-known that an accurate

description of electronic band structure and optical proper-

ties provides a significant challenge for the DFT formalism.

This shortcoming is not a failure of DFT itself nor is it a

breakdown of the Hohenberg–Kohm theorem (which is for-

mally an exact theory). As is the case for all DFT formal-

isms, i.e., for both orbital-free DFT and the Kohn–Sham

orbital-dependent approach used here,55,56 this limitation

arises from approximations to the (still unknown) exact

exchange-correlation functional. As shown by several groups

including ours, the use of conventional exchange-correlation

functionals results in severely underestimated band gaps and

incorrect asymptotic potential energy surfaces resulting from

electron-transfer self-interaction.22,46

An alternative to DFT methods is the use of wave-

function based methods which directly calculate the 3N-

dimensional wavefunction (as opposed to the 3-dimensional

electron density). The conventional approach for handling

these systems (at least for molecules) is to use Møller-

Plesset perturbative theory (MP2)57 in conjunction with

highly correlated coupled-cluster singles and doubles plus

perturbative triples methods (CCSD(T)) to estimate elec-

tronic energies. However, the computational cost of

CCSD(T) methods58 is too high for routine application to

molecules larger than about 30 atoms. DFT, on the other

hand, scales efficiently with size and is the method of choice

for studying large material systems. Unfortunately, the accu-

rate description of band gaps has been a traditional failure of

most current density functionals, and the development of

new methods for properly treating these electronic properties

is still a topic of active research.

Recognizing the shortcomings of conventional function-

als, major methodological progress has been made in DFT

techniques, which incorporate a strong intra-atomic interac-

tion in a (screened) Hartree–Fock like manner, as an on-site

replacement of the GGA.47–49 This approach is commonly

known as GGAþU or Hubbard DFT methods.52,53,59

The DFT results for both the pure CeBr3 primitive cell

and the (divalent) Ca2þ-doped supercell were based on

spin-polarized, GGA methods using PAW pseudopoten-

tials50 calculated by the Vienna Ab-initio Simulation

Package (VASP).60 It was found that the inclusion of a

Hubbard DFT correction, DFTþU, was essential for obtain-

ing accurate band structures and band gaps.61 This novel

DFTþU method was used to support experimental efforts in

understanding both the structural and electronic properties of

aliovalently doped CeBr3 materials. This formalism is essen-

tial for the Ce-based materials studied here, as conventional

GGAs fail to describe systems with localized (strongly corre-

lated) d and f electrons, which manifests itself in the form

of unrealistic one-electron energies due to spurious

self-interaction. This approach has been further modified and

applied in many forms by Liechtenstein62 and Dudarev.63 In

this work, the rotationally invariant approach by Dudarev is

used because the effective on-site Coulomb and exchange

parameters (U and J) do not enter separately, and only their

difference (U� J) is required. The Dudarev approach is

denoted here as just “DFTþU” unless otherwise specified.

III. CALCULATIONS

DFT calculations were performed to discern the optical

properties of the Ca2þ-doped CeBr3 compound. Introducing

the Ca2þ dopant (Figure 1(a)) resulted in calculations that

provided valuable insight to the increase in crystal strength.

These calculations may be used to compare the differences

of doping with different dopants or their concentrations and

to calculate the optical properties of the resultant crystals.

One must account for the electron densities over the space of

the crystal. To do this accurately, the calculations were made

for a “supercell” of 200 atoms with well over 1000 electrons

(Figure 1). In addition, the electron density of states was cal-

culated. If there is a crystal defect, or an atomic substitution,

then the electron density of states will change, leading to dif-

ferent results. The presence of a Ca2þ dopant can signifi-

cantly affect electronic properties by distorting the electronic

charge density near the Ca2þ atom (Figure 2).

Our DFT results for both the pure CeBr3 primitive cell

and the (divalent) calcium-doped supercell were based on the

FIG. 1. (a) Ca dopant shows the (divalent) Ca2þ-substituted atom (near the lower left corner). (b) A CeBr3 cross section showing the CeBr3 supercell that was

optimized and calculated using the Hubbard-U correction (i.e., DFTþU). White and red dots represent cerium and bromine atoms, respectively, and the green

dot represents the Ca2þ-substituted atom.

034908-3 Guss et al. J. Appl. Phys. 115, 034908 (2014)

[This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to ] IP: 128.31.3.8

On: Fri, 17 Jan 2014 19:13:52

Perdew-Burke-Ernzerhof (PBE) GGA functional using PAW

pseudopotentials. The PBE GGA48–51 is used for the exchange

and correlation correction. During our study, we found

that the inclusion of a Hubbard DFTþU correction for the

f-electrons on Ce was essential for obtaining accurate band

structures and band gaps. Within the Hubbard DFT approach,

an on-site penalty function is placed on the f-electrons for

each of the cerium (Ce) atoms to prevent spurious over-

delocalization of the electrons. Based on benchmarks for Ce3þ

ions, the Hubbard value (U) for our DFT studies was set at

4.5 eV. For the initial study of the primitive cell, a very high

cutoff energy of 400 eV was used for the plane-wave basis set,

and the Brillouin zone64 was sampled using a dense 8� 8� 8

gamma-centered Monkhorst–Pack grid.65 In addition to

spin-polarization and dispersion effects, a relativistic spin-

orbit coupling treatment was used for the valence electrons in

the bulk-lattice calculations. Geometry optimizations of both

the ions and the unit cell were carried out.

Since the Ca-defect calculations require the use of larger

supercells and significantly more atoms (>200 atoms), a

smaller 300 eV cutoff energy is currently used for these cal-

culations. For this same reason, we did not include spin-orbit

effects in these large supercell systems, as we found this

effect to be minimal in our initial calculations for the primi-

tive cell described above. In the point-defect calculations, a

large 3� 3� 3 supercell was used and, therefore, a smaller

2� 2� 2 Gamma-centered Monkhorst–Pack grid was used.

Due to the extremely large size of these systems, these calcu-

lations were carried out on a new massively parallelized

supercomputer (located at Sandia National Laboratories

headquarters in Albuquerque, NM) consisting of state-of-

the-art Intel Sandy Bridge processors, QLogic Infiniband

communication hardware, and a customized Linux software

environment. This unique computing resource is the newest

computing cluster at Sandia for high-performance capacity

computing work and is necessary for solving the self-

consistent, coupled Kohn–Sham equations with a large

plane-wave basis. Unconstrained geometry optimizations of

both the ions and the unit cell were carried out. Upon conver-

gence, we found that the Ca atom was slightly displaced

within the CeBr3 crystal lattice (Figure 1). However, the

Ca2þ atom substitutes for Ce3þ atom and basically stays at

the Ce3þ site when it is incorporated in the matrix. The

results of the DFT modeling performed at Sandia National

Laboratories suggested that a 1.9% substitution would result

in a suitable combination of crystalline structural properties,

increased strengthening of the lattice, and metrics for optical

properties and emission spectra. The Fermi levels that result

at different points in the lattice are also different because of

the aliovalent substitution.

Using the converged geometry, we then evaluated the

cohesive energy of the Ca-doped CeBr3 material. The cohe-

sive energy is defined as the energy required to break the

atoms of the solid into isolated atomic species, a measure of

the strength of the material. More specifically, the cohesive

energy Ecoh is given by

Ecoh ¼ Esolid �X

i

Eatomi ; (2)

where the summation index, i, represents all the different

atoms that constitute the solid. For the Ca-doped 3� 3� 3

CeBr3 supercell, we obtained a cohesive energy of

�3.66 eV/atom. In order to further understand the electronic

and structural properties due to the dopant, we also visual-

ized the electron density of the entire supercell.

Interestingly, we found that the presence of a Ca-dopant can

significantly affect electronic properties by distorting the

electronic charge density near the Ca atom (see Figure 2).

The distortions in charge density associated with Ca-dopant

site are in part from variations in energy transfer dynamics,

electronic transitions around the dopant atom, and the effects

of chemical modification on optical emission.

The electronic structures and optical properties of

both the undoped and the Ca-doped CeBr3 crystal may be

derived using a k-point grid generated according to the

Monkhorst–Pack scheme,65 for the sampling of the Brillouin

zone. The optical properties of CeBr3 can be determined

by the frequency-dependent dielectric function e(x) ¼ e1(x)

þ i e2(x) that is mainly connected with the electronic struc-

tures. The imaginary part e2(x) of the dielectric function

e(x) may be calculated from the momentum matrix elements

between the occupied and unoccupied states within selection

rules and given by22

FIG. 2. (a) Side view of electron charge density (red lobes) in a Ca2þ-doped CeBr3 scintillating material. The electron density is significantly modified and

accumulates near the Ca atom (blue). (b) Top-down view of electron charge density (red lobes) in a Ca-doped CeBr3 scintillating material. The calcium (blue)

lies underneath a Ce atom in the left corner of the primitive unit cell.

034908-4 Guss et al. J. Appl. Phys. 115, 034908 (2014)

[This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to ] IP: 128.31.3.8

On: Fri, 17 Jan 2014 19:13:52

e2 xð Þ¼ ½2e2p=ðXe0Þ�X

k;v;cjhwc

kju � rjhwvkij

2 �d Eck�Ev

k�E� �

;

(3)

where x is the light frequency and e is the electronic charge.

wck and wv

k are the conduction and valence band (VB) wave-

functions at k, respectively. The real part e1(x) of the dielec-

tric function e(x) can be derived from the imaginary part

e2(x) using the Kramers–Kronig dispersion equation. All

other optical constants can be derived from e1(x) and

e2(x).22,66

IV. STRUCTURE DETERMINATIONAND ELECTRONIC PROPERTIES

CeBr3 has a hexagonal structure with space group

P63/m, where the cerium and the halide occupy 2(a) in (1/3,

2/3, 1/4) and 6(h) in (0.375, 0.292, 1/4) sites, respec-

tively.22,67,68 Our structures are in reasonable agreement

with experiments.67–70 Figure 3 shows the Brillouin zone.64

The electronic structures of CeBr3 at their equilibrium struc-

tures were calculated using this Brillouin zone formulation

for context. We have carried out a series of DFT calculations

taking into account both ferromagnetic and antiferromag-

netic behaviors. We also performed a high-level calculation

(using the HSE06 hybrid functional71,72) to understand and

predict the band gap of this material. Our high-level calcula-

tions allow an accurate assessment of other CeBr3 dopants in

larger systems, which we will study in the near future.

The energy band structure of CeBr3 calculated at equi-

librium is shown in Figure 4. The energy band structure is

calculated along the path that contains the highest number of

high-symmetry points of the Brillouin zone, namely, C! A

! H! K! C! M! L! H. The zero energy is arbitra-

rily taken at the Fermi level.73 The lowest bands around

�36 eV consist of 5s states of Ce. The Ce 5p states are

located at around �19 eV. The bands around �16 eV are

derived from the 4s states of bromine (Br). The energy bands

in the range from �6 to �2 eV correspond to the 4p states of

Br that decide the top of the valence band. These results are

in agreement with CeCl3 experimental data.22 Above the

Fermi level, the conduction band (CB) consists of the 4f and

5d states of Ce. The 4f states have a sharp peak due to the

strong localization character of the 4f states. If the 4f electron

is kept in the core, the calculated result agrees with other

experiments.22,74

In addition, a partly filled f band situated right at the

Fermi level was found. This discrepancy from an experiment

by Park11,70 may arise from the different treatment method

of the 4f states of Ce. If the 4f electron is kept in the core, the

calculated result was found to be in agreement with the ex-

perimental results.74 Otherwise, because of the localized

character of the f states, the unpaired electron cannot lead to

metallic behavior, with its effective mass tending toward in-

finity.75 Therefore, CeBr3 is an insulator.22

In order to understand how the defect influences the

electronic density of states (DOS), we also evaluated the

DOS for both the pure, undoped CeBr3 material and

the Ca-doped supercell (Figures 5 and 6). Comparing

Figures 5 and 6, we notice that several peaks near 2 and

�18 eV disappear for the Ca-doped compound. These differ-

ences are undoubtedly due to the presence of the Ca atom

and its distortion of the electronic environment around this

dopant. Finally, using the converged DOS results, we com-

puted the optical properties of these compounds by evaluat-

ing the real parts of the frequency-dependent dielectric

tensor (Figures 7 and 8).

FIG. 3. CeBr3 Brillouin zone. Hex path: C–M–K–C–A–L–H–A–jL–Mj–K–H.

FIG. 4. Band structure of CeBr3 in a hexagonal crystal lattice. The band

structure corresponds to the Brillouin zone for the CeBr3 hexagonal crystal

lattice. The Fermi energy level is shifted to the valence band maximum at

zero.

FIG. 5. The energy band structure and electronic partial density of states

around the Fermi level for the 3� 3� 3 pure CeBr3 supercell are displayed.

The lowest bands around �36 eV consist of 5s states of Ce. The Ce 5p states

are located at around �18 eV. The bands around �16 eV are derived from

the 4s states of Br. The energy bands in the range from �6 to �2 eV corre-

spond to the 4p states of Br that decides the top of the valence band. There

are more peaks near 2 and �18 eV for the pure CeBr3 supercell.

034908-5 Guss et al. J. Appl. Phys. 115, 034908 (2014)

[This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to ] IP: 128.31.3.8

On: Fri, 17 Jan 2014 19:13:52

Although the optical response in Figures 7 and 8 are

similar, there are noticeable differences at 5 and 20 eV with

a much stronger variation near 5 eV for the pure CeBr3 mate-

rial. We note, in passing, that these effects would probably

be more significant if a higher Ca-doping level were used

(i.e., 1 Ca atom for every 4 Ce atoms). Furthermore, under-

standing how to modulate this distortion in electron density

(which results in other spectroscopic observables) with other

types of dopants would enable us to maximize their theoreti-

cal energy resolution for advanced spectroscopy applica-

tions, using a rational experimental approach guided by

predictive modeling. Future studies will explore this factor.

To summarize, these new simulations demonstrate a

capability for predicting properties of doped CeBr3 materials

that is unavailable elsewhere but is critically needed to study

the property-limiting valence phenomena in ionic com-

pounds. The model may also begin to enable realistic simula-

tions of chemical reactions where the electronic valence

effects of different atoms may dynamically vary (e.g., differ-

ent oxidation states or even different dopant atoms). It can

be applied to improve properties and perform life-cycle anal-

ysis of ionic materials and enhance our ability to respond to

emerging mission critical problems. Examples of the appli-

cations that are directly relevant include radiation detection,

photovoltaics, thermoelectrics, solid fuel cells, rechargeable

batteries, super-capacitors, piezoelectric ceramics, micro-

electronics, photonics, micro-electro-mechanical systems

(MEMS), and nano-electro-mechanical systems (NEMS).

V. RESULTS AND ANALYSIS

A. Detectors

In 2012, in support of a Nevada National Security Site

(NNSS) Directed Research and Development (SDRD) pro-

ject, Dynasil Radiation Monitoring Devices, Inc. (RMD) was

commissioned by the Remote Sensing Laboratory (RSL) to

build a CeBr3 crystal alloy detector with 2% substitution

using the Ca2þ cation. RMD grew a 1-in.-diameter CeBr3

crystal with 1.9% calcium dibromide (CaBr2), as requested.

The vertical Bridgman was used as the crystal growth tech-

nique. Materials were purified prior to crystal growth. The

charge composition was 222.14 g of CeBr3 and 2.26 g of

CaBr2. The resulting crystal is polycrystalline white but

FIG. 6. The energy band structure and electronic partial density of states

around the Fermi level for the 3� 3� 3 Ca2þ-doped CeBr3 supercell are dis-

played. The lowest bands around �36 eV consist of 5s states of Ce. The Ce

5p states are located at around �18 eV. The bands around �16 eV are

derived from the 4s states of Br. The energy bands in the range from �6 to

�2 eV correspond to the 4p states of Br that decides the top of the valence

band. There are fewer peaks near 2 and �18 eV for the Ca2þ-doped CeBr3

supercell.

FIG. 7. Real parts e1(x) of the dielectric tensor for the 3� 3� 3 pure CeBr3

supercell, where e(x) ¼ e1(x) þ i e2(x) and x is the light frequency. The

dielectric function of CeBr3 is calculated based on its electronic structure.

e1(x) is a function of the photon energy. The imaginary part e2(x) of the

dielectric function is connected with the energy band structure but correlates

with the real part e1(x) of the dielectric function. The first peak at 1 eV

observed for the pure CeBr3 supercell is not evident for the Ca2þ-doped

CeBr3 supercell. The peak at 3 eV corresponds to the transition 4f1! 5d0

of Ce. The main peaks of about 4 eV may be ascribed to the transition from

the Br 4p VB to the Ce 5d CB. The peaks near 20 eV correspond mainly to

the transition of inner electron excitation from Ce 5p VB to CB.22

FIG. 8. Real parts e1(x) of the dielectric tensor for the 3� 3� 3

Ca2þ-doped CeBr3 supercell, where e(x) ¼ e1(x) þ i e2(x) and x is the light

frequency. The dielectric function of Ca2þ-doped CeBr3 is calculated based

on its electronic structure. e1(x) is a function of the photon energy. The

imaginary part e2(x) of the dielectric function is connected with the energy

band structure but correlates with the real part e1(x) of the dielectric func-

tion. The first peak at 1 eV observed for the pure CeBr3 supercell is not evi-

dent for the Ca2þ-doped CeBr3 supercell. The peak at 3 eV corresponds to

the transition 4f1 ! 5d0 of Ce. The main peaks of about 4 eV may be

ascribed to the transition from the Br 4p VB to the Ce 5d CB. The peaks

near 20 eV correspond mainly to the transition of inner electron excitation

from Ce 5p VB to CB.22

034908-6 Guss et al. J. Appl. Phys. 115, 034908 (2014)

[This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to ] IP: 128.31.3.8

On: Fri, 17 Jan 2014 19:13:52

transparent. The crystal structure is columnar or filament

like. When RMD cut the crystal, it fell apart. Smaller crystal-

lites were easily separated from the bulk. Samples were then

harvested from each ingot as available and tested for fluores-

cence and scintillation. The detector crystals produced are

shown in Figure 9. The small crystallites are clear and trans-

parent, but the large sections of crystal are translucent or

whitish-opaque. Small crystallites are typically �2 to 3 mm

wide and 10 mm or more in length. Adding 500 ppm calcium

in a prior run (in 2009) did not appear to have a significant

impact on the performance; however, there was no cracking

of that ingot. In neither run has the calcium concentration in

the final ingot been measured, though we do not believe the

intended 1.9% doping concentration used in our calculations

was actually realized in the new crystal.

This investigation involved measurement of the light out-

put, the emission spectrum, and the fluorescent decay time of

the crystals. Preliminary results for a select set of samples

from the as-grown ingots were obtained. Fluorescence emis-

sion curves were recorded to determine the effect of the sol-

ute cation on the emission wavelengths of the base compound

CeBr3. Initial scintillation characterization was completed as

well. The small crystallites were also tested for scintillation

properties.

B. Pulse shape

The time profiles of the CeBr3:Ca2þ crystals grown at

RMD were characterized. A CAEN digitizer DT5720 was

employed to collect raw anode pulses from the Hamamatsu

H6610 photomultiplier tube (PMT), as opposed to the previ-

ously used delayed coincidence method.76 During measure-

ments, samples were coupled to a PMT and irradiated with

gamma rays. Figure 10 shows the decay-time spectrum

recorded for a CeBr3:Ca2þ sample along with a fit (to an ex-

ponential rise and decay time plus a constant background) to

the data. The pulse shape has two components (�21 ns and

�38 ns) as seen in Figure 10. A possible reason for a faster

and slower time component was provided by Drozdowski,77

which he ascribes to self-absorption phenomena. The second

pulse in Figure 10 is an artifact; it is an after pulse from the

PMT. The decay constant for the fast component is �21 ns.

The 21 ns decay component can be attributed to optical emis-

sion arising from direct capture of electron-hole pairs at the

Ce3þ sites.8 The rise time of the scintillation pulse from

CeBr3:Ca2þ is estimated to be �0.1 ns using the data shown

in Figure 10. The initial photon intensity, a figure of merit for

timing applications, is estimated to be �4000 photons/

(ns-MeV) for CeBr3:Ca2þ, which is higher compared to all

common inorganic scintillators [including barium difluoride

(BaF2)], a benchmark for timing applications).

C. Fluorescence measurements and gamma-ray totallight yield

The emission spectrum of the CeBr3:Ca2þ scintillator

was measured. Fluorimetry and preliminary scintillation are

presented here for this aliovalently doped scintillator. The

CeBr3:Ca2þ samples were excited with radiation from a

Philips x-ray tube having a Cu target, with power settings of

40 kVp and 20 mA. The scintillation light was passed

through a McPherson monochromator and detected by a

Hamamatsu R2059 PMT with a quartz window. The system

was calibrated with a standard light source to enable correc-

tion for sensitivity variations as a function of wavelength.

A normalized emission spectrum for the CeBr3:Ca2þ

(2% Ca2þ) sample is shown in Figure 11. The observed

bands, peaking at 360 and 380 nm, are characteristic of Ce3þ

luminescence. The Ca2þ exhibited little or no change in the

peak fluorescence emission for 370 nm excitation.

Fluorescence spectra indicate that the Ca2þ dopant is suitable

for strengthening CeBr3. The emission wavelength spectrum

peaks at �372 nm (Figure 11), indicating a nearly impercep-

tible red shift from the 370 nm already documented for

CeBr3.8 The peak emission wavelength for the CeBr3 sample

is at �370 nm and this emission is anticipated to be due to

5d!4f transition of Ce3þ. The peak emission wavelength ofFIG. 9. Crystals grown by RMD: (a) CeBr3 ampoule after growth; note the

visible cracking; (b) CeBr3 crystals with calcium doping.

FIG. 10. Pulse shape recorded for 1.9% nominally Ca2þ-doped CeBr3. Red

lines depict best fits to the data with a decay time of t1 � 21 ns for the fast

component.

034908-7 Guss et al. J. Appl. Phys. 115, 034908 (2014)

[This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to ] IP: 128.31.3.8

On: Fri, 17 Jan 2014 19:13:52

372 nm for CeBr3:Ca2þ is attractive for gamma-ray spectros-

copy because it matches well with the spectral response of

the PMTs, as well as the newest generation of silicon

photodiodes.

Light output (or luminosity) of the CeBr3:Ca2þ and sim-

ilarly sized NaI:Tl crystals have been measured at RMD.

Total light yield was compared to NaI:Tl by first recording a137Cs spectrum with a CeBr3:Ca2þ sample, then replacing

the CeBr3:Ca2þ with NaI:Tl and recording a new 137Cs spec-

trum with the same system settings. Spectra were collected

for both the doped CeBr3:Ca2þ crystal and a NaI:Tl crystal.

Samples of doped CeBr3:Ca2þ were submerged in oil and/or

coupled to the photocathode of a PMT. Then the detector

system was irradiated with 662 keV photons (137Cs source)

to record pulse height spectra using standard Nuclear

Instrumentation Module (NIM) electronics. A similar experi-

ment was then performed with a calibrated NaI:Tl crystal

under the same operating conditions. The NaI:Tl crystal was

wrapped with a Teflon tape coupled to a PMT. Comparison

of the 662 keV gamma-ray peak position recorded with

CeBr3:Ca2þ and the NaI:Tl scintillators (Figure 12) provided

an estimate of the light output of CeBr3:Ca2þ. Compared to

NaI:Tl, the doped CeBr3:Ca2þ ingots produced a full energy

peak 1.7 times higher. Light output is estimated by compari-

son to the known NaI:Tl to be approximately �58 000

photons/MeV. The light output of NaI:Tl used for compari-

son was assumed to be 38 000 photons/MeV. Comparison of

the measured spectra is shown in Figure 12. Scintillation

spectroscopy has thus far indicated the doped crystals still

excel over conventional NaI:Tl, despite the fact that total

light yield analysis appears to indicate a loss in light produc-

tion in the measured samples.

D. Gamma-ray scintillation spectroscopy

Gamma-ray spectroscopy has also been performed with

CeBr3:Ca2þ crystals grown at RMD by coupling them to a

PMT. The scintillator (2% Ca2þ) was irradiated with a 137Cs

source (662 keV photons) and the resulting spectrum is

shown in Figure 12. Referring to Figure 12, the energy reso-

lution of the 662 keV photopeak was measured to be 3.2%

FWHM at room temperature. Such high energy resolution

has never before been achieved with any of the established

inorganic CeBr3 scintillators (even in small sizes) at room

temperature. Figure 12 also shows a 137Cs spectrum recorded

with NaI:Tl. Energy resolution of NaI:Tl crystal was meas-

ured to be 6.7% (FWHM). Thus, the energy resolution of

CeBr3:Ca2þ is twice as good as that for typical NaI:Tl detec-

tors. Higher light output and proportionality of CeBr3:Ca2þ

(compared to NaI:Tl and other inorganic scintillators) are re-

sponsible for high energy resolution obtained with those

crystals. It is important to note that energy resolution of

CeBr3:Ca2þ at 662 keV gamma-ray energy is starting to

approach that of room-temperature semiconductor detectors

such as cadmium telluride (CdTe) and cadmium zinc tellu-

ride (CdZnTe) of similar size. The energy resolution is not

presently understood. The energy resolution at 662 keV

appears to be much better than that typical of CeBr3.23

E. Scintillation light output proportionality

Proportionality of response for a CeBr3:Ca2þ sample

(with 2% Ca2þ) as a function of gamma-ray energy was

measured. Relative light yield proportionality was measured

for both doped and undoped samples of CeBr3 to ensure that

no loss in performance was incurred in aliovalently doping

the crystal. Over the energy range from 100 to 1275 keV, the

nonproportionality of CeBr3:Ca2þ was measured to be 5%,

which is substantially better than that for many established

scintillators. For example, over the same energy range, the

nonproportionality is about 35% for cerium-doped lutetium

oxyorthosilcate (Lu2Si05:Ce or LSO) and about 20% for

NaI:Tl and CsI:Tl.78 Proportionality of light yield is one area

of performance in which Ce-doped and Ce-based lanthanide

halides excel. Maintaining proportionality is important to

producing a strengthened, high-performance scintillator.

FIG. 11. Emission spectra are shown for both 2% (red) and 500 ppm Ca2þ

(blue). The emission spectra peak at �372 nm. There is essentially no

change by increasing the Ca2þ content.

FIG. 12. 137Cs spectra collected with NaI:Tl and with 1.9% Ca2þ-doped

CeBr3 crystals coupled to a PMT. The energy resolution of the 662 keV

peak for NaI:Tl (at channel 351) is �6.7% (FWHM) and for CeBr3:Ca (at

channel 591) is �3.2% (FWHM). The green curves represent Gaussian fits

to the CeBr3:Ca2þ photopeak at 662 keV and to the low energy tail for this

same peak. The reported peak resolution of 3.2% (FWHM) was derived

from this parametric fit to the data.

034908-8 Guss et al. J. Appl. Phys. 115, 034908 (2014)

[This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to ] IP: 128.31.3.8

On: Fri, 17 Jan 2014 19:13:52

There was a reduced yield at low energy. Measurements are

ongoing and indications are that CeBr3:Ca has better propor-

tionality than standard CeBr3. Preliminary work supports

indications that linearity is maintained as CeBr3 is aliova-

lently doped with Ca2þ. We anticipate that these new mea-

surement results will be reported at upcoming IEEE-NSS

meeting.

VI. CONCLUSION

Ca2þ-doped CeBr3 exhibited little or no change in its

fluorescence spectrum, yet produced slightly less light. The

dopant may also have introduced an absorption center,

which produces an apparent red shift by adding new energy

states into the CeBr3 band gap without actually altering the

emission spectra. Ca2þ may have produced self-absorption

in the materials as members of the crystal lattice or perhaps

the dopants were present in quantities above their respective

solubility limits, creating secondary phase particles within

the crystals that absorbed scintillation light. Finally,

as-grown crystals may have had a large amount of internal

strain. Unseen grain boundaries and other discontinuities

due to internal strain would necessarily alter light propaga-

tion within the crystals, potentially decreasing the observed

light yield.

In summary, the crystal structure, band structure, density

of states, and optical properties of Ca2þ-doped were calcu-

lated by means of DFT within the generalized gradient

approximation. Our structural parameters are in good agree-

ment with the experimental results. A 3.2% energy resolu-

tion (Figure 12) was measured for Ca2þ-doped CeBr3.

Energy resolution at 662 keV appears to be better than the

undoped CeBr3. This result compares favorably with

LaBr3:Ce, considering that LaBr3:Ce also has considerable

internal self-activity. A lesson learned is that 1.9% calcium

is too much to incorporate into CeBr3. Many dopants are yet

to be fully tested, let alone used at differing concentrations.

Future work will focus on completing measurements of the

as-grown doped ingots and determining the solid solubility

limit for each dopant investigated.

ACKNOWLEDGMENTS

This manuscript has been authored by National Security

Technologies, LLC, under Contract No. DE-AC52-

06NA25946 with the U.S. Department of Energy and sup-

ported by the Site-Directed Research and Development

Program. The United States Government retains and the pub-

lisher, by accepting the article for publication, acknowledges

a non-exclusive, paid-up, irrevocable, worldwide license to

publish or reproduce the published form of this manuscript, or

allow others to do so, for United States Government purposes.

The authors acknowledge the professional staff of

RMD, Watertown, Massachusetts, for the production of the

detectors, for providing these detectors to the Remote

Sensing Laboratory, and for their support and advice. The

authors thank Michael Lukens and Sanjoy Mukhopadhyay

for their contributions.

1P. Rodnyi, Physical Processes in Inorganic Scintillators (CRC Press, New

York, 1997), pp. 1–160.2M. J. Harrison, P. Ugorowski, C. Linnick, S. Brinton, D. S. McGregor, F.

P. Doty, S. Kilpatrick, and D. F. Bahr, “Aliovalent doping of CeBr3,”

Proc. SPIE 7806, 78060M-1–78060M-14 (2010).3E. V. D. van Loef, P. Dorenbos, C. W. E. van Eijk, K. Kr€amer, and H. U.

G€udel, “High-energy-resolution scintillator: Ce3þ activated LaCl3,” Appl.

Phys. Lett. 77, 1467–1468 (2000).4E. V. D. van Loef, P. Dorenbos, C. W. E. van Eijk, K. Kr€amer, and H. U.

G€udel, “High-energy-resolution scintillator: Ce3þ activated LaBr3,” Appl.

Phys. Lett. 79, 1573 (2001).5P. Dorenbos, “Light output and energy resolution of Ce3þ-doped

scintillators,” Nucl. Instrum. Methods Phys. Res. A 486, 208–213 (2002).6R. Hofstadter, “Alkali halide scintillation counters,” Phys. Rev. 74,

100–101 (1948).7R. Hofstadter, “Properties of scintillation materials,” Nucleonics 6(5),

70–72 (1950).8K. S. Shah, J. Glodo, W. Higgins, E. V. D. van Loef, W. W. Moses, S. E.

Derenzo, and M. J. Weber, “CeBr3 scintillators for gamma-ray

spectroscopy,” IEEE Trans. Nucl. Sci. 52(6), 3157–3159 (2005).9M. J. Harrison, C. Linnick, B. Montag, S. Brinton, M. McCreary, F. P.

Doty, and D. S. McGregor, “Radioluminescence and scintillation results

of horizontal gradient freeze grown aliovalently-doped CeBr3,” in IEEENuclear Science Symposium Conference Record, edited by P. Sellin

(Institute of Electrical and Electronics Engineers, 2008), pp. 2850–2855.10M. J. Harrison, C. Linnick, B. Montag, S. Brinton, M. McCreary, F. P.

Doty, and D. S. McGregor, “Scintillation performance of aliovalently-

doped CeBr3,” IEEE Trans. Nucl. Sci. 56(3), 1661–1665 (2009).11V. Vetere, C. Adamo, and P. Maldivi, “Performance of the ‘parameter

free’ PBE0 functional for the modeling of molecular properties of heavy

metals,” Chem. Phys. Lett. 325, 99–105 (2000).12M. J. Harrison and F. P. Doty, “Initial investigation of strengthening agents

for lanthanide halide scintillators,” Proc. SPIE 6707, 67070B (2007).13F. P. Doty, D. McGregor, M. Harrison, K. Findley, and R. Polichar,

“Structure and properties of lanthanide halides,” Proc. SPIE 6707, 670705

(2007).14F. P. Doty, “Fracture resistant lanthanide scintillators,” US patent

7,683,572 B1 (2011).15X. W. Zhou, F. P. Doty, and P. Yang, “Atomistic simulations of mechani-

cal properties of LaBr3 single crystals,” J. Appl. Phys. 107(12), 123509

(2010).16X. W. Zhou and F. P. Doty “Embedded-ion method: An analytical energy-

conserving charge-transfer interatomic potential and its application to the

La-Br system,” Phys. Rev. B 78(22), 224307 (2008).17C. G. Olson, M. Piacentini, and D. W. Lynch, “Optical properties of single

crystals of some rare-earth trifluorides, 5-34 eV,” Phys. Rev. B 18,

5740–5749 (1978).18L. Joubert, G. Picard, and J. J. Legendre, “Structural and thermochemical

ab initio studies of lanthanide trihalide molecules with pseudopotentials,”

Inorg. Chem. 37, 1984–1991 (1998).19T. Tsuchiya, T. Taketsugu, H. Nakano, and K. Hirao, “Theoretical study

of electronic and geometric structures of a series of lanthanide trihalides

LnX3 (Ln ¼ La-Lu; X ¼ Cl, F),” J. Mol. Struct.: THEOCHEM 461–462,

203–222 (1999).20R. D. Wesley and C. W. DeKock, “Geometry and infrared spectra of

matrix-isolated rare-earth halides. I. LaF3, CeF3, PrF3, NdF3, SmF3, and

EuF3,” J. Chem. Phys. 55, 3866 (1971).21C. Shi, G. Zhang, Y. Wei, Z. Han, J. Shi, G. Hu, M. Kirm, and G.

Zimmerer, “The dynamics properties on luminescence of CeF3 crystals,”

Surf. Rev. Lett. 9(1), 371–374 (2002).22C. Li, B. Wang, R. Wang, and H. Wang, “First-principles studies on the

electronic and optical properties of CeCl3 and CeBr3,” Solid. State

Commun. 144, 220–224 (2007).23P. P. Guss, M. Reed, D. Yuan, A. Reed, and S. Mukhopadhyay, “CeBr3 as

a high-resolution gamma-ray detector,” Nucl. Instrum. Methods Phys.

Res., Sect. A 608(2), 297–304 (2009).24K. S. Shah, J. Glodo, M. Klugerman, W. M. Higgins, T. Gupta, and P.

Wong, “High energy resolution scintillation spectrometers,” IEEE Trans.

Nucl. Sci. 51(5), 2395–2399 (2004).25P. P. Guss, M. Reed, D. Yuan, M. Cutler, C. Contreras, and D. Beller,

“Comparison of CeBr3 with LaBr3:Ce, LaCl3:Ce, and NaI:Tl detectors,”

Proc. SPIE 7805, 78050L (2010).26F. P. Doty, D. S. McGregor, M. Harrison, K. Findley, R. Polichar, and P.

Yang, in Basic Materials Studies of Lanthanide Halide Scintillators, edited

034908-9 Guss et al. J. Appl. Phys. 115, 034908 (2014)

[This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to ] IP: 128.31.3.8

On: Fri, 17 Jan 2014 19:13:52

by D. L. Perry, A. Burger, L. Franks, and M. Schieber (Mater. Res. Soc.

Symp. Proc., 2007), Vol. 1038, pp. 1038-O01-03.27B. J. Pletka, T. E. Mitchell, and A. H. Heuer, “Solid solution hardening of

sapphire (a-Al2O3),” Phys. Status Solidi 39, 301–311 (1977).28T. H. Courtney, Mechanical Behavior of Materials, 2nd ed. (Waveland,

Long Grove, Illinois, 2000), p. 232.29H. L. Tuller, S. A. Kramer, M. A. Spears, and U. B. Pal, “Method for mak-

ing an electrochemical cell,” U.S. patent 5,509,189 (23 April 1996).30S. A. Kramer and H. L. Tuller, “A novel titanate-based oxygen ion con-

ductor: Gd2Ti2O7,” Solid State Ionics 82(1-2), 15–23 (1995).31S. A. Kramer, “Mixed ionic conduction in rate earth titanate/zirconate

pyrochlore compounds,” Ph.D. thesis (Massachusetts Institute of

Technology, 1994).32O. H. B. Hassan, Oxides with Polyatomic Anions Considered as New

Electrolyte Materials for Solid Oxide Fuel Cells (SOFC) (Reihe Energie

and Umwelt, Band 68 Schriften des Forschungszentrum J€ulich GmbH,

Germany, 1965), pp. 17–59.33R. J. Cava, J. M. Phillips, J. Kwo, G. A. Thomas, R. B. van Dover, S. A.

Carter, J. J. Krajeewski, W. F. Peck, Jr., J. H. Marshall, and D. H.

Rapkine, “GaInO3: A new transparent conducting oxide,” Appl. Phys.

Lett. 64(16), 2071–2072 (1994).34R. J. Cava, “Transparent conductors comprising gallium-indium-oxide,”

U.S. patent 5,407,602 (18 April 1995).35N. Meethong, Y.-H. Kao, S. A. Speakman, and Y.-M. Chiang, “Aliovalent

substitutions in olivine lithium iron phosphate and impact on structure and

properties,” Adv. Funct. Mater. 19(7), 1060–1070 (2009).36C. Zhang, E. Staunton, Y. G. Andreev, and P. G. Bruce, “Raising the con-

ductivity of crystalline polymer electrolytes by aliovalent doping,” J. Am.

Chem. Soc. 127(51), 18305–18308 (2005).37E. Erunal, P. Jakes, S. Korbel, J. Acker, H. Kungl, C. Elsasser, M. J.

Hoffmann, and R. A. Eichel, “CuO-doped NaNbO3 antiferroelectrics:

Impact of aliovalent doping and nonstoichiometry on the defect structure

and formation of secondary phases,” Phys. Rev. B 84(18), 184113 (2011).38A. K. Zak, A. Jalalian, S. M. Hosseini, A. Kompany, and T. Shekofteh

Narm, “Effect of Y3þ and Nb5þ co-doping on dielectric and piezoelectric

properties of PZT ceramics,” Mater. Sci. (Poland) 28(3), 703–708 (2010).39E. Erdem1, R.-A. Eichel1, H. Kungl, M. J. Hoffmann, A. Ozarowski, H.

van Tol, and L. C. Brunel, “Local symmetry-reduction in tetragonal

(La,Fe)-codoped Pb[Zr0.4Ti0.6]O3, piezoelectric ceramics,” Phys. Scr.

T129, 12–16 (2007).40S.-H. Lee, C.-B. Yoon, S.-B. Seo, and H.-E. Kim, “Effect of lanthanum on

the piezoelectric properties of lead zirconate titanate–lead zinc niobate

ceramics,” J. Mater. Res. 18(8), 1765–1770 (2003).41M. R. Benam, “Investigation the dielectrical and electromechanical prop-

erties of PZT thin films,” Int. J. Eng. Res. Ind. Appl. 3(2), 680–682 (2013).42A. Boricevic, C. Wolverton, G. M. Crosbie, and E. B. Stechel, “Defect

ordering in aliovalently doped cubic zirconia from first principles,” Phys.

Rev. B 64, 014106 (2001).43P. D. Edmondson, W. J. Weber, F. Namavarc, and Y. Zhang, “Lattice dis-

tortions and oxygen vacancies produced in Auþ-irradiated nanocrystalline

cubic zirconia,” Scr. Mater. 65, 675–678 (2011).44J.-M. Costantini, F. Beuneu, and W. J. Weber “Radiation damage in cubic-

stabilized zirconia,” J. Nucl. Mater. 440, 508–514 (2013).45B. M. Wong, “Optoelectronic properties of carbon nanorings: Excitonic

effects from time-dependent density functional theory,” J. Phys. Chem. C

113, 21921–21927 (2009).46B. M. Wong and T. H. Hsieh, “Optoelectronic and excitonic properties of oli-

goacenes: Substantial improvements from range-separated time-dependent

density functional theory,” J. Chem. Theory Comput. 6, 3704–3712 (2010).47X. Hua, X. Chen, and W. A. Goddard, “Generalized gradient approxima-

tion: An improved density-functional theory for accurate orbital

eigenvalues,” Phys. Rev. B 55, 16103–16109 (1997).48J. P. Perdew, K. Burke, and M. Ernzerhof, “Generalized gradient approxi-

mation made simple,” Phys. Rev. Lett. 77, 3865–3868 (1996).49J. P. Perdew, K. Burke, and M. Ernzerhof, “Erratum: Generalized gradient

approximation made simple [Phys. Rev. Lett. 77, 3865 (1996)],” Phys.

Rev. Lett. 78, 1396 (1997).50P. E. Bl€ochl, “Projector augmented-wave method,” Phys. Rev. B 50(24),

17953 (1994).51P. E. Bl€ochl, C. J. F€orst, and J Schimpl, “Projector augmented wave

method: Ab-initio MD with full wave functions,” Bull. Mater. Sci. 26(1),

33–41 (2003).

52V. I. Anisimov, J. Zaanen, and O. K. Andersen, “Band theory and Mott

insulators: Hubbard U instead of Stoner I,” Phys. Rev. B 44, 943–954

(1991).53V. I. Anisimov, F. Aryasetiawan, and A. I. Lichtenstein, “First-principles

calculations of the electronic structure and spectra of strongly correlated

systems: The LDAþU method,” J. Phys.: Condens. Matter 9, 767 (1997).54P. Hohenberg and W. Kohn, “Inhomogeneous electron gas,” Phys. Rev.

136(3B), B864–B871 (1964).55W. Kohn and L. J. Sham, “Self-consistent equations including exchange

and correlation effects,” Phys. Rev. 140, A1133 (1965).56S. K€ummel and L. Kronik, “Orbital-dependent density functionals: Theory

and applications,” Rev. Mod. Phys. 80, 3–60 (2008).57C. Møller and M. S. Plesset, “Note on an approximation treatment for

many-electron systems (abstract),” Phys. Rev. 46(7), 618–622 (1934).58I. D. Mackie and G. A. Dilabio, “Approximations to complete basis set-

extrapolated, highly correlated non-covalent interaction energies,”

J. Chem. Phys. 135(13), 134318-1–134318-10 (2011).59H. J. Kulik and N. Marzari, “A self-consistent Hubbard U density-

functional theory approach to the addition-elimination reactions of hydro-

carbons on bare FeOþ,” J. Chem. Phys. 129(13), 134314 (2008).60See http://cmp.univie.ac.at/vasp/, for VASP, Vienna Ab-initio Simulation

Package, What is VASP? (last accessed 30 April 2012).61P. Guss, B. Wong, M. Foster, F. P. Doty, K. Shah, M. Squillante, J. Glodo,

and D. Yuan, “Low-cost cerium bromide alloys,” Nevada National Security

Site–Directed Research and Development, FY 2012 Report, National

Security Technologies, LLC, Las Vegas, Nevada, 2013, pp. 11–19.62A. I. Liechtenstein, V. I. Anisimov, and J. Zaanen, “Density-functional

theory and strong interactions: Orbital ordering in Mott-Hubbard insu-

lators,” Phys. Rev. B 52, R5467–R5470 (1995).63S. L. Dudarev, G. A. Botton, S. Y. Savrasov, C. J. Humphreys, and A. P.

Sutton, “Electron-energy-loss spectra and the structural stability of nickel

oxide: An LSDAþU study,” Phys. Rev. B 57, 1505–1509 (1998).64L. Brillouin, “Les �electrons dans les m�etaux et le classement des ondes de

De Broglie correspondantes,” C. R. Hebd. S�eances Acad. Sci. 191, 292

(1930).65H. J. Monkhorst and J. K. Pack, “Special points for Brillouin-zone integra-

tions,” Phys. Rev. B 13(12), 5188–5192 (1976).66S. Saha and T. P. Sinha, “Electronic structure, chemical bonding, and

optical properties of paraelectric BaTiO3,” Phys. Rev. B 62, 8828–8834

(2000).67W. H. Zachariasen, “Crystal chemical studies of the 5f-series of elements.

I. New structure types,” Acta Crystallogr. 1, 265–268 (1948).68W. H. Zachariasen, “The UCl3 type of crystal structure,” J. Chem. Phys.

16, 254 (1948).69S. Sato, “Optical absorption and x-ray photoemission spectra of lanthanum

and cerium halides,” J. Phys. Soc. Jpn. 41, 913–920 (1976).70K. H. Park and S. J. Oh, “Electron-spectroscopy study of rare-earth

trihalides,” Phys. Rev. B 48, 14833 (1993).71J. Heyd, G. E. Scuseria, and M. Ernzerhof, “Hybrid functionals based on a

screened Coulomb potential,” J. Chem. Phys. 118, 8207 (2003).72A. V. Krukau, O. A. Vydrov, A. F. Izmaylov, and G. E. Scuseria,

“Influence of the exchange screening parameter on the performance of

screened hybrid functional,” J. Chem. Phys. 125, 224106 (2006).73C. Kittel and H. Kroemer, Thermal Physics, 2nd ed. (W. H. Freeman, New

York, 1980), p. 357.74N. V. Skorodumova, R. Ahuja, S. I. Simak, A. Abrikosov, B. Johansson,

and B. I. Lundqvist, “Electronic, bonding, and optical properties of CeO2

and Ce2O3 from first principles,” Phys. Rev. B 64, 115108 (2001).75F. Goubin, X. Rocquefelte, D. Pauwels, A. Tressaud, A. Demourgues, S.

Jobic, and Y. Montardi, “The dielectric function of LnSF rare-earth fluoro-

sulfides (Ln ¼ La, Ce): Experiment and theory,” J. Solid State Chem. 177,

2833 (2004).76L. M. Bollinger and G. E. Thomas, “Measurement of the time dependence

of scintillation intensity by a delayed-coincidence method,” Rev. Sci.

Instrum. 32, 1044 (1961).77W. Drozdowski, P. Dorenbos, A. J. J. Bos, G. Bizarri, A. Owens, and F. G.

A. Quarati, “CeBr3 scintillator development for possible use in space mis-

sions,” IEEE Trans. Nucl. Sci. 55(3), 1391–1396 (2008).78O. Guillot-No€el, J. C. van’t Spijker, J. T. M. de Haas, P. Dorenbos, C. W.

E. van Eijk, K. Kr€amer, and H. U. G€udel, “Scintillation properties of

RbGd Br:Ce advantages and limitations,” IEEE Trans. Nucl. Sci. 46,

1274–1284 (1999).

034908-10 Guss et al. J. Appl. Phys. 115, 034908 (2014)

[This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to ] IP: 128.31.3.8

On: Fri, 17 Jan 2014 19:13:52