μSR Investigations of Cerium Hydride*

(1)

Physikalische

Folge,

164,

(1989)

©

by

R.

Oldenbourg

Verlag,

Miinchen 1989

-0044-3336/89

$3.00+0.00

jjSR Investigations

of Cerium

_Hydride*

P.

_Birrer,

F.N.

_Gygax,

B.

_Hitti,

E.

_Lippelt

and A. Schenck

InstituteforIntermediate

_{Energy Physics}

_(IMP),

ETHZ,

CH-5234

_Villigen,

Switzerland

L.

_Schlapbach

Institut de

_Physique,

Universite de

_Fribourg,

CH-1700

_Fribourg,

Switzerland

The

_Knight

shift and transversefield

_spin

relaxation rateof

_positive

muons

(/x+),

implanted

in

_LaH2.75-,

CeH2.7o-

and

_{CeH2.95-powder samples,}

have been measured

as afunction of

temperature

andin the

Cerl^-samples

also as afunction offield.

These data are used to extract informationon the

/i+-site, ^-diffusion

and the

magnetic properties

of the

_{CeH^-samples.}

Introduction

Cerium

_hydrides

CeHj.

areknowntoexhibit

fascinating

properties

regarding

their

phase

diagram j\j.

Their electronic structure

_/2/

_changes

from metallicat x =2

to

_insulating

(or

_{semiconducting)}

at x = 3 and their

magnetic properties

include

both

_{ferromagnetic}

order

_(for

2.3 < x <

2.7)

and

antiferromagnetic

order

(for

x < 1.95 and x >

2.75) /3/.

In addition measurements ofthe electronic

specific

heat ina

CeH2

₆₅

sample yielded

evidence for

heavy

fermion behavior

/4/.

Hydro-gen

occupies

up to x = 2the available tetrahedral sites in the

Ce

fee lattice.

On

(2)

1048

P. Birreretal.

The indication for

_heavy

fermion behavior and the

_{complex magnetic phase}

dia-gram in these

systems

motivated us to

investigate

them further

by

means of the

fiSR. technique

/5/

which has proven to be of considerablepower in the

study

of

heavy

fermion

_systems

_(see

_e.g.

_/6/).

In addition we wereinterested in the

diffu-sional behaviorof the

_fi+,

inits site of residence

_(tetrahedral

or

octahedral)

and

its interaction with the

_neighboring

_protons.

In this contribution we

report

on first results from a

systematic

/xSR

study

of

cerium

_hydrides

_/7/.

Two

_powder

_samples

with nominal values ofx = 2.70 and x =2.95 which bracketed well the metal tosemiconductor transitionwere

investi-gated.

The

_sample

with x =2.95 shouldshow

antiferromagnetism

with

Tjv

—

4K while the other

_sample

was

expected

to be either

ferromagnetic

with

Tc

around

~ 2 K or tofall within the narrow

paramagnetic region

abovex = 2.70

/3/.

For

comparison

also results froma

nonmagnetic

LaH2

₇₅

sample

areincluded here. The

experiments

were

performed

at the Paul Scherrer Institute

(formerly SIN).

Resultsand discussion

Generally

the

_{/xSR signal}

in the

CeHj.

compounds

consisted oftwo

_components,

irrespective

of the

_applied

field. The

_amplitudes

ofthe two

components

showed

a weak

temperature

dependence

and a

_stronger

field

dependence,

however their sumremained

independent

of

temperature

or

applied

field. This

implies

that the

muons do not

probe

two different

phases

in the

samples

but that the two

com-ponents

are

probably

originating

from

fi+

in different

magnetic

environments the

extent of which is controlled

_by

_applied

field and

_temperature.

The two compo-nentsaredistinct

by

their relaxationratesand

_frequencies.

At afield of0.02 Tthe

minority

component

(called

the fast

_component)

_displayed

a rather

_temperature

independent

relaxation rate of~ 1.25

fis'1

for both

CeHx

samples.

This

compo-nent

_disappeared

close to room

temperature.

At low

temperatures

the

minority

component

(~ 25%)

became the

_majority

_component

_(~

_75%)

for

_applied

fields above ~ 0.08 T. Above ~ 0.1 T the relaxation _rate increased

linearly

with the

field. The other

_component

_(called

the slow

_component)

_{displayed always}

amuch

smaller relaxationratewitha

pronounced

temperature

dependence

butnofield

de-pendence.

This

_signal

resembles the

_single

_component

found in the

_LaH2.75

_sample.

(3)

;jSR

Investigations

Hydride

The

_magnitude

of the relaxationrateof the fast componentcan

only

_be

_explained

if static electronic moments are invoked. Since _we _have _no _{clear idea} _{about the}

origin

ofthis

_component,

wewill discusshere

only

the slow

_component

which_has

manyfeatures incommon with proton NMR

signals

in

LaHj.

and

CeHx

/8,9,10/.

Fig.

lashows the

_temperature

_dependence

ofthe

_fi+

relaxation rate orlinewidth orin

LaH2.75,

displaying

the onset of motional

_narrowing

around 50 K

and,

more

pronounced,

above ~ 150 K. Around the latter

_temperature

hydrogen

diffusion_is

knownto setin

_/8/.

Similarcurves wereobtainedin

CeH2.7o

(Fig. lb)

and

CeH2.95

(Fig. lc).

The low

_temperature

_plateau

values of or can be used to learn about the

_/x+

position

in thelattice. In Tab. 1 wecomparethe

_experimental

valueswith calculated

_predictions

for the twodifferent

_possible

sites,

assuming

forthe lattice constant a„ =5.57

A.

Wesee thatin all cases the

experimental

values are

signifi-cantly larger

than thecalculatedones. Toobtain

agreement

wecan scaledown the

lattice constants

_(<r

oc

a~3).

Foranoctahedralsite

assignment

this would

imply

a

surprisingly

large

near

neighbor

latticecontractionof

—8.5%

in

CeH2.70,

—5.5%

in

CeH2.95

and

—3.2%

in

_LaH2

_76. Thesevalueswould beeven

larger

foratetrahedral

site

_assignment

which seems to ruleout the latter one. In

Cerl*

the low

temper-ature <t_may be enhanced

by magnetic

effects,

nevertheless _it _appears _as _{if there}

is a

_strong

_attractiveinteraction between_the

_fi+

_at _an _{octahedral interstitial} _site

and the nearest

_neighbor

_protons

at tetrahedral interstitial sites.

<texp(/is

')

<rcal(a0

=5.57

A)

a„,e//

(A)

aocxp

A

An/a

(%)

octh. tetr. octh. octh.

LaH2

75 0.26 0.244 0.220 5.450 5.63

/8/

3.2

CeH2.70

0.31 0.233 0.196 5.064 5.54

/ll/

-8.5

CeH2.9B

0.28 0.235 0.212 5.254 5.54

_/ll/

-5.5 Tab. 1.

_Compilation

of resultson cr and

_comparison

withcalculations. The

_temperature

_dependence

of<r at

higher

temperatures

wasused to extract the

fi+

diffusionrate ormore

precise,

acorrelationratev= ₊_{up which}_contains

alsothe

_proton

_jump

rate_{up. It is} found that

_v(T)

followedanArrhenius law for

temperatures

above about 180 K. Tab. 2 containsthe

fitted

activation

_energies

and

_{preexponential}

factors

_together

with

_{corresponding}

data

froma

proton

NMR

(4)

1050

P. Birreretal. La H

SS

0.20 s oj & 0.10 0.00 L-LlJ-1_L_ 10u 10' Temperature(K ) CeH,05 0.40

101

Temperature(K)

Fig.

1.

Temperature dependence

of (t^out m

(a)

LaH2.7o,

(b)

CeH2.7o

and

_(c)

CeH2.95-

Thedatawere_obtained

in an

applied

field of_{0.02 T.}

study.

It is seen that

the

present

activation

ener-gies

and

_{preexponential}

fac-tors are

significantly

smaller

than in the

_proton

case. A

similar observation was also

made in

PdH*

/12/.

This is hard to

_understand,

be-cause one

_expects

always

v >

up. A way out would be

a

highly

correlated

fi+-proton

diffusion or a kind of

immo-bilizationofthe

_tightly

bound

ft+

—

nn

_proton

complex.

An-other

_possibility

is a

stoichio-metric

CeH3

environment in the

_fi+

_vicinity

which should also inhibit

_strongly

uncorre-lated

_jump

events.

The

_Knight

shift of the

_p+

Larmor

_frequency

in

LaH2.75

was too small to be detected in the

_present

_study.

In

contrast the

_Knight

shift in the

_{CeHj.-samples}

was

quite

sizeable and showed remark-able

temperature

dependen-cies

_(see

_Fig.

_2a,b).

In

CeH2.7o

the

_Knight

shift

_changes

from about

—0.2%

at room

tem-perature

to about

_+0.25%

at

~ 5

K,

displaying

_a shallow

maximum there. In contrast

(5)

uSR

Investigations

Cerium

Hydride

fi+

E(kcal/Mol)

f„(a_1)

LaH2

75 - 2.28 1.2•

10*

CeH2.7o

~0.50 4.6 •

106 CeH2

95 ~1.63 3.2

-107

P

LaH2.86

~3.50 ~

10"

/8/

Tab. 2. Diffusion

_parameter

for T > 180 K.

CeH 2.778 2.772 2.766 2.760 10 10' Temperature(K) CeH,0c 2.778 2.772 2.766 2.760 tt-1-r 10" 101 Temperature IK I —1_i_.1_i_l

Fig.

2.

Temperature dependence

of

/z+-Larmor

frequency

of the slow

_component

in

_(a)

CeH2.7o

and

_(b)

CeH2.95.

The dashed line

corresponds

to the external field

_(~

0.02

_T).

close to zero both at room

temperature

and below 5 K and has a shallow minimum

of ~

-0.2%

at 30 K. A

few

_proton

_Knight

shift data for

CeHj.

with x = 2.85 can

be found in

_/13,9/

for

tem-peratures

between 6 K and 180 A".

_They

_agree

_relatively

well with the

_present

results in

_CeH2.70.

The

_magnitude

of the

_Knight

shift in

_Fig.

2

clearly

points

to a

_strong

in-fluence of the Ce4f-moments.

The

_negative

shift in

CeH2.70

at room

_temperature

seems

to

_suggest

an

antiferromag-netic

_coupling

ofthe conduc-tion electrons

_(or

what ever screens the

/i+-charge)

to the 4f-electrons. The

_change

of

sign

at ~ 20 K is for now

a

_mystery

as is the difference

between the results in

_Figs.

2a and 2b.

(6)

1052

P. Birreretal. Conclusions

Positive muons

implanted

in

CeH,,.

(x

=

2.70,

2.95)

and

LaH2.76

_seem to occupy

the octahedral interstitial sites. The nearest

_neighbor

_protons

at tetrahedral in-terstitial sites arecontracted towards the

/i+

_by

_afew

%.

Uncorrelated diffusional

motion ofthe

_fi+

and its

proton

neighbors

is slowed down

_{considerably compared}

to

_proton

diffusion in the bulk ofthe material. The

_/i+

_Knight

shift shows

_quite

a different andunusual

temperature

dependence

for the two

CeHj.

samples.

It is

suggested

that the

_hydrogen

concentration via the

_density

of free carriers has a

pronounced

influenceonthe

(exchange)

coupling

of the electrons

screening

the

fi+

(mostly

conduction

_electrons?)

and the 4f-electrons. Whether this is correlated with the metal-semiconductortransition remains tobeseen. The low