Valence and magnetism in EuPd 3S4and Y LaxEu1xPd3S4 D. H. Ryan Physics Department and Centre for the Physics of Materials McGill University

2025-05-06 0 0 1.34MB 16 页 10玖币

侵权投诉

Valence and magnetism in EuPd3S4and (Y,La)xEu1−xPd3S4

D. H. Ryan

Physics Department and Centre for the Physics of Materials, McGill University,

3600 University Street, Montreal, Quebec, H3A 2T8, Canada

Sergey L. Bud’ko, Brinda Kuthanazhi, and Paul C. Canﬁeld

Ames National Laboratory, and Department of Physics and Astronomy, Iowa State University, Ames, Iowa 50011, USA

(Dated: October 11, 2022)

151Eu M¨ossbauer spectroscopy shows that yttrium substitution in mixed-valent EuPd3S4drives

the initial 50:50 mix of Eu3+ and Eu2+ towards pure Eu2+, whereas lanthanum substitution has the

opposite eﬀect, but only for substitution levels above 50%. We ﬁnd that total valence electron count

and chemical pressure eﬀects cannot account for the observed behaviour, however conserving the

cell volume provides a consistent description of the changes in the Eu2+:Eu3+ ratio. Remarkably,

lanthanum substitution also leads to a clear transition from static mixed-valent behavior at lower

temperatures to dynamic mixed valent behavior at higher temperatures, with the onset temperature

monotonically increasing with Eu content and extrapolating to a value of ∼340 K for the pure

EuPd3S4compound. Magnetic order persists at least as far as x=0.875 in both series, despite the

drastic reduction in the amount of moment-carrying Eu2+ ions.

I. INTRODUCTION

The rare earth palladium sulphides RPd3S4have been

reported for the majority of the rare earths, including

yttrium1–4. They all crystallise in the cubic NaPt3O4

structure (P m3n#223) with the rare earth occupying

the 2asite forming a bcc sublattice, the palladium on

the 6dsite and the sulphur on the 8esite. Remark-

ably, although the RPd3S4phases exist for the trivalent

rare earths, but apparently not for the divalent alkaline

earths (Ca and Sr), when prepared with europium5or

ytterbium4a roughly 50:50 mix of divalent and trivalent

rare earth is found.

Here we will use chemical substitution of yttrium and

lanthanum for europium to investigate the stability of the

valence distribution and its eﬀects on magnetic ordering.

Although both 170Yb and 151Eu M¨ossbauer spectroscopy

can generally be used to identify the valence of their re-

spective target ions, for 170Yb M¨ossbauer spectroscopy,

the isomer shift between the two valence states is ex-

tremely small so the technique is almost totally depen-

dent on the presence of an electric ﬁeld gradient (efg)

at the Yb3+ ions to identify trivalent ytterbium. Un-

fortunately the high symmetry of the 2asite makes the

efg contribution eﬀectively zero and the presence of the

Yb3+ ions is only apparent in the magnetically ordered

state well below TN∼2 K4. We will therefore only study

EuPd3S4using 151Eu M¨ossbauer spectroscopy where the

two valence states are clearly isolated by a large diﬀer-

ence in isomer shift, even at ambient temperatures.

We ﬁnd that, by substituting Y for Eu, the remaining

Eu sites become more and more divalent. In contrast,

by substituting La for Eu we ﬁnd that, initially, the re-

maining Eu sites stay roughly a 50:50 mixture of di- and

tri-valent Eu, but for higher La substitution levels the re-

maining Eu rapidly becomes more trivalent. La substitu-

tion also leads to a transition from statically mixed valent

behavior at lower temperatures to dynamically mixed

valent behavior at higher temperatures with the onset

temperature (Tonset) monotonically increasing with Eu

content and passing through room temperature as pure

EuPd3S4is approached. Despite the decreasing fraction

of moment-carrying Eu2+ ions, both YxEu1−xPd3S4and

LaxEu1−xPd3S4continue to exhibit some form of mag-

netic order at least as far as x=0.875, with transition

temperatures of ∼3 K (Y) and ∼6 K (La).

II. EXPERIMENTAL METHODS

Polycrystalline samples of EuPd3S4and

(Y,La)xEu1−xPd3S4were prepared from stoichio-

metric mixtures of EuS (99.9% – American Elements)

Y2S3(99.9%), La2S3(99%), Pd (99.95%) and S (99.5%),

all from Alfa-Aesar. The powders were mixed and

then pressed to form a dense pellet. This was loaded

into an alumina crucible and sealed under a partial

pressure of helium in a fused silica tube. The sample

was heated to 650◦C over three hours, held for an hour

and then taken to 900◦C over a further three hours.

After 75 hours at 900◦C the sample was furnace cooled

and removed once it reached ambient temperature. In

most cases this single thermal cycle was found to yield

a single-phased product, however when an impurity

was found (typically PdS seen by x-ray diﬀraction or

EuS seen in susceptibility vs. temperature) the sample

was crushed, pressed and subjected to a second 75 hr

annealing cycle to 900◦C to remove the impurity.

X-ray diﬀraction measurements were made on a

Rigaku Miniﬂex-II diﬀractometer using a Cu-Kαsource.

The instrument calibration was checked using NIST 676a

Al2O3and found to be consistent within ﬁtted uncertain-

ties. Full Rietveld reﬁnement of the diﬀraction pattens

was carried out using the GSAS/EXPGUI packages6,7.

As all three species occupy special sites in the P m3n

structure, no positional parameters were adjusted dur-

arXiv:2210.03860v1 [cond-mat.mtrl-sci] 8 Oct 2022

ing the ﬁts. For the yttrium and lanthanum substituted

samples only random occupation of the Eu(2a) site was

considered.

151Eu M¨ossbauer spectroscopy measurements were car-

ried out using a 4 GBq 151SmF3source, driven in sinu-

soidal mode. The drive motion was calibrated using a

standard 57CoRh/α-Fe foil. Isomer shifts are quoted rel-

ative to EuF3at ambient temperature. The 21.6 keV

gamma rays were recorded using a thin NaI scintillation

detector. For temperatures above 5 K, the samples were

cooled in a vibration-isolated closed-cycle helium refrig-

erator with the sample in a helium exchange gas. Tem-

peratures below 5 K were achieved using a helium ﬂow

cryostat while pumping on the sample space and using

a needle valve to throttle the ﬂow. The spectra were

ﬁtted to a sum of Lorentzian lines with the positions

and intensities derived from a full solution to the nuclear

Hamiltonian8.

Temperature- and magnetic ﬁeld- dependent magne-

tization measurements (1.8 K ≤T≤300 K, 0 T ≤

µ0H≤7 T) were performed using a Quantum Design

MPMS-3 SQUID magnetometer. The sample was con-

ﬁned in a #4 gelatin capsule and a transparent drinking

straw was used as a sample holder. Low-temperature

heat capacity measurements were made using semiadi-

abatic thermal relaxation technique as implemented in

the heat capacity option of a Quantum Design Physi-

cal Property Measurement System (PPMS). For selected

samples, the 3He option was used to cool to ∼0.4 K.

Sintered samples of 20–70 mg mass with at least one

ﬂat surface were mounted on a micro-calorimeter plat-

form using Apiezon N grease. The addenda (platform +

grease) heat capacity was measured separately for each

sample and subtracted from the total heat capacity using

the PPMS software. Although the samples possibly had

reduced density, the measured sample coupling parame-

ter took reasonable values of more than 97%.

III. RESULTS

A. EuPd3S4

Fitting the x-ray diﬀraction pattern of the EuPd3S4

sample showed it to be single phased with the expected

cubic NaPt3O4structure2and a lattice parameter of

a= 6.6786(1) ˚

A (Fig. 1). The room temperature 151Eu

M¨ossbauer spectrum showed two distinct contributions

from Eu3+ and Eu2+ in the ratio 50.4(4):49.6(4), with the

linewidth of the Eu2+ component being slightly broader,

in complete agreement with previous reports5. On cool-

ing to 5 K the ratio becomes 46.2(5):53.8(5) (Fig. 2) as

the Debye temperature of the Eu3+ component is slightly

higher than that of the Eu2+ component. Fitting the

temperature dependence of the two component areas to

a simple Debye model, as shown in Fig. 3, yields Debye

temperatures of 227(3) K (Eu3+) and 204(3) K (Eu2+).

It is important to emphasise that any apparent changes in

40 50 60 70 80 90 100

2 (°)

Counts (×103)

10 20 30 40

2 (°)

EuPd3S4

FIG. 1. Cu-Kαx-ray diﬀraction pattern for EuPd3S4. Solid

line is a full Rietveld reﬁnement using the GSAS/EXPGIU

packages6,7 . The line below the data shows the residuals.

Tick marks between the data and residual lines show the cal-

culated positions of the Bragg peaks. The inset shows the

low-angle range. The diﬀraction pattern was collected in two

overlapping blocks with a longer counting time at higher an-

gles to compensate for the loss of intensity due to the eﬀects

of the x-ray form factor.

the Eu2+:Eu3+ ratio with temperature in Fig. 3 do not

reﬂect actual changes in the ratio, rather they are the

result of the diﬀerent temperature dependences of the

recoil-free-fractions (often denoted f) for the two species.

In order to minimise the impacts of this eﬀect, all valence

ratios will be taken from low-temperature spectra.

On cooling to 1.8 K, the Eu2+ component develops

a clear magnetic splitting whereas the Eu3+ component

is unchanged, consistent with trivalent europium being

non-magnetic (Fig. 2). In principle one might anticipate

a small transferred ﬁeld at the Eu3+ sites from the or-

dered Eu2+ moments, however none was observed and

no increase in the width of the Eu3+ component was

detected (Bhf 1 T), perhaps as a result of cancella-

tions arising from the antiferromagnetic ordering of the

Eu2+ moments. The hyperﬁne ﬁeld (Bhf ) for the Eu2+

component at 1.8 K is 29.9(2) T, typical for ordered

Eu2+, and ﬁtting the temperature dependence shown in

Fig. 4 to the expected J=7

2Brillouin function yields

a T=0 Bhf =37.9(6) T and an ordering temperature of

TN=2.90(1) K in good agreement with both previous

work5and our own susceptibility and Cpdata (Fig. 5).

The clear cusp in the susceptibility vs. temperature is

consistent with antiferromagnetic (AF) ordering and it is

accompanied by a sharp peak in the heat capacity. The

high-ﬁeld magnetisation curve taken at 1.8 K (inset to

Fig. 5) shows that the system is readily saturated despite

the AF order, consistent with the low ordering temper-

ature as well as the low anisotropy typically associated

with the Eu2+ ion. Furthermore, the maximum moment

observed in the applied ﬁeld of 7 T suggests a Eu2+ frac-

tion of 50% (assuming a moment of 7 µB/Eu2+) con-

sistent with the 53.8(5)% derived above from the 151Eu

M¨ossbauer spectrum at 5 K, and with earlier results5,9.

FIG. 2. 151Eu M¨ossbauer spectra of EuPd3S4at RT (top),

5 K (middle), and 1.8 K (bottom). Between RT and 5K the

primary changes are a signiﬁcant increase in the absorption

due to conventional thermal eﬀects and a slight change in the

2+:3+ area ratio due to diﬀerences in the Debye temperatures

of the two components. At 5 K the two valence contributions

are almost equal in area and fully resolved, with the Eu2+ at

−10.91(2) mm/s and Eu3+ at +0.07(1) mm/s. At 1.8 K only

the Eu2+ component is ordered and shows a hyperﬁne ﬁeld

(Bhf ) of 29.9(2) T, while the Eu3+ component is unchanged.

The solid red lines are ﬁts as desribed in the text. For the

1.8 K spectrum we also show the magnetically split Eu2+ and

unchanged Eu3+ components.

The unusual and apparently stable valence mix in

EuPd3S4leads to the question: “why?”. What makes

europium (and ytterbium) “special”? How robust is the

valence distribution? Can we change it?

As the Eu3+ ion is smaller than the Eu2+ ion,

one might expect hydrostatic pressure to promote

Eu2+ →Eu3+ conversion. Alternatively, if we force some

fraction of the R sites to be unambiguously trivalent, by

replacing some of the europium with a formally trivalent

ion, will this cause more of the remaining europium to

become divalent to preserve the average electron count?

Although the driving that can be achieved by chem-

ical substitution is not as clean as that generated by

direct hydrostatic pressure, it is much easier to make

direct measurements of the valence distribution, mag-

netisation and transition temperatures in doped sam-

ples at ambient pressures. We turn therefore to an in-

vestigation of the impacts of chemical substitution on

EuPd3S4using the non-moment bearing, trivalent La and

Y ions with rionic(Eu2+)&rionic(La3+)> rionic(Eu3+)

> rionic(Y3+). If, on the one hand, the total valence elec-

FIG. 3. Temperature dependence of the normalised (adjusted

for total counting time) area of both the total spectrum and

the two valence components in the 151Eu M¨ossbauer spectra

of EuPd3S4. Solid lines in each case are ﬁts to a simple Debye

model yielding Debye temperatures of 227(3) K (Eu3+) and

204(3) K (Eu2+). Fitting the total spectral area yields an

average Debye temperature of 212(3) K.

FIG. 4. Temperature dependence of hyperﬁne ﬁeld (Bhf ) for

EuPd3S4ﬁtted using the expected J=7

2Brillouin function

to obtain the ordering temperature of TN=2.90(1) K.

tron count is a dominant factor, then partially replacing

the europium with an unambiguously trivalent ion should

lead to a compensating increase in the Eu2+ fraction.

On the other hand, if preserving the cell volume is criti-

cal, then the lattice expansion that would be caused by

introducing the rionic(La3+).rionic(Eu2+) lanthanum

ion could be compensated by some Eu2+ →Eu3+ con-

version (the reverse being expected for the rionic(Y3+)

< rionic(Eu3+) yttrium ion). Alternatively, if the sub-

stitutions act as chemical pressures, then expanding

the cell using lanthanum substitution should lead to

M / H ( e m u / m o l )

E u P d 3S4

µ0H = 5 m T

0 2 4 6 8 1 0 1 2

1 0

Cp ( J / m o l K )

d ( T * M / H ) / d T ( m u m o l )

0 2 4 6

M ( µB/ m o l E u )

µ0H ( T )

T = 1 . 8 K

FIG. 5. Top: DC susceptibility vs. temperature for EuPd3S4

showing a cusp at TN∼3 K. Inset shows M vs. µ0H at 1.8 K

conﬁrming that half of the europium is divalent. Bottom:

Temperature dependence of the heat capacity (black points)

revealing the corresponding cusp associated with the AF tran-

sition, with d(T·M/H)/dT (red line) showing that Cp(T) and

the temperature derivative of T·M/H take the same form

around the transition.

Eu3+ →Eu2+ conversion (again, the reverse process be-

ing expected for yttrium substitution). As we show be-

low, we can clearly distinguish between these three op-

tions.

B. Valence impacts

Fitting the x-ray diﬀraction data for the yttrium and

lanthanum substituted compounds showed that they all

retained the expected NaPt3O4structure but with pro-

gressively smaller (Y) or larger (La) lattice parameters

(Fig. 6). There is a clear, and signiﬁcant, change in

lattice parameters for both Y and La substitution. We

detect no indication of phase separation in the powder

x-ray diﬀraction data, i.e. no broadening or splitting of

peaks that would suggest segregation of the samples into

Eu-richer and Eu-poorer phases.

For the Y-substituted series, the ﬁtted lattice param-

eters all lie visibly above the line connecting the Eu

and Y compounds, suggesting that a shift towards more

of the larger Eu2+ ion occurs, as the yttrium content

is increased. This valence shift is conﬁrmed directly

by the 5 K 151Eu M¨ossbauer spectra shown in Fig. 7,

where the line near 0 mm/s associated with the triva-

lent europium decreases rapidly in intensity as the level

FIG. 6. (top) Lattice parameters for (Y,La)xEu1−xPd3S4

showing the expansion associated with La-substitution and

the contraction when yttrium is substituted. We note that

whereas the values for the La-substituted series lie on the line

connecting the pure La and Eu compounds, those for the Y-

substituted samples lie signiﬁcantly above the corresponding

line between the pure Eu and Y compounds.

(bottom) The Eu2+ fractions in the pure and Y,La-

substituted compounds showing that Y-substitution leads to

a signiﬁcant shift towards more Eu2+, while La-substitution

initially appears to have no eﬀect, but for x > 0.5 there is a

marked shift towards Eu3+ . Open circles show values taken

from 151Eu M¨ossbauer spectroscopy at 5 K(Y-substituted) or

10 K(La-substituted); solid symbols show values derived from

bulk magnetisation data: high-temperature Curie-Weiss ﬁts

(red squares), saturation magnetisation (blue circles). The

deviation of the Curie-Weiss values for the La-rich compounds

reﬂects the development of dynamic eﬀects discussed in the

text.

of yttrium substitution increases. (All valence ratios for

the Y-substituted series were taken from 5 K spectra to

minimise the impacts of f-factor diﬀerences for the two

species, as noted above.) Both magnetisation measure-

ments in the ordered state at 1.8 K (Fig. 8) and Curie-

Weiss ﬁts to the temperature dependence of the suscepti-

bility above 10 K, further support these observations (see

below). The Eu2+ fractions derived from all three mea-

surements are summarised in the lower panel of Fig. 6.

By contrast with the yttrium-substituted series, the ef-

fects of lanthanum substitution are more complex and nu-

anced. The lattice parameters of the La-substituted com-

pounds lie much closer to the line connecting the two end

文档加载中……请稍候！
如果长时间未打开，您也可以点击刷新试试。

下载文档到电脑，查找使用更方便

10 玖币 0人已下载

立即下载

摘要：

ValenceandmagnetisminEuPd3S4and(Y;La)xEu1xPd3S4D.H.RyanPhysicsDepartmentandCentreforthePhysicsofMaterials,McGillUniversity,3600UniversityStreet,Montreal,Quebec,H3A2T8,CanadaSergeyL.Bud'ko,BrindaKuthanazhi,andPaulC.CaneldAmesNationalLaboratory,andDepartmentofPhysicsandAstronomy,IowaStateUniversity,A...

展开>> 收起<<

Valence and magnetism in EuPd 3S4and Y LaxEu1xPd3S4 D. H. Ryan Physics Department and Centre for the Physics of Materials McGill University.pdf

共16页,预览4页

还剩页未读，继续阅读

声明：本站为文档C2C交易模式，即用户上传的文档直接被用户下载，本站只是中间服务平台，本站所有文档下载所得的收益归上传人(含作者)所有。玖贝云文库仅提供信息存储空间，仅对用户上传内容的表现方式做保护处理，对上载内容本身不做任何修改或编辑。若文档所含内容侵犯了您的版权或隐私，请立即通知玖贝云文库，我们立即给予删除！

Valence and magnetism in EuPd 3S4and Y LaxEu1xPd3S4 D. H. Ryan Physics Department and Centre for the Physics of Materials McGill University

相关推荐

开通VIP享超值会员特权

作者详情

相关内容

热门标签

举报选择: