Lifetime of the Sm++ 5D0 → 7F0 emission in alkali halides

Solid State Communications,

Vol. 8, pp. 1697—1701, 1970. Pergamon Press.

5D LIFETIME OF THE Sm~~ 0

Printed in Great Britain

‘F0 EMISSION IN ALKALI HALIDES

-,

G~Baldini, M. Cartoceti and M. Guzzi Istituto di Fisica, Université degli Studi, Milano and Gruppo Nazionale di Struttura della Materia del C.N.R., Milano. (Received 21 July 1970 by C.W. ~1cCombie) 5D The lifetime of the Sm~luminescence ( 0 ‘F0) has been investigated in KC1, KBr and RbC1 from liquid helium5D to room temperature. It has been found that the lifetime of the 0 excited state (4f ~ configuration), a few ~isec at room temperature, upon cooling to temperatures below 20°K is increased by a factor greater than iO~.From a mechanism proposed in order 5D to describe the results, it appears that a level slightly above the 0 ~onfiguration is responsible of the strong thermal dependence of the lifetime. This level, which had not been detected neither 5DO in undabsorption is attributed nor to in the 4f55s is or lovated 4f5Sp configuration. also shown here that the emission, a few 100 cm-’It is above thermal behavior of the decay lifetime describes correctly the apparently anomalous oscillator strength of the zero phonon emission discussed in a previous paper. .-~

INTRODUCTION

the expected theoreticallaw but it appeared that

THE OPTICAL spectra of Sm~~ in alkali halides and other ionic crystals which have been examined by a number of workers have suggested the following picture for its transitions.’ The most

some other mechanism was reducing its intensity.2 It was then proposed the existence of an energy level, not directly observed, which, by means of thermal activation from the 5D 0 state, could subtract oscillator strength from high. the transition the temperature w~ssufficiently In the if present paper we report on lifetime measurements which provethe thatdecay such time a level does exist and it at can modify of the luminescence

prominent absorption of light by a Sm~ion 6) conraises the to optic electron fromone theof‘F0 figuration states in which the(4f 4f electron is excited into levels such as 5d; 6s etc. In most cases no light emission corresponding to the downward transitions is observed but rather, as in the crystals examined here, only the intraconfigurational 5D 6) ‘F (4~)transitions, 0 (4f with 0, 1. 6, are detected in emission. These appear as a sequence of groups of sharp lines when the crystal temperature is of the order of a few absolute degrees. In a previous

high temperature (T> 15°K),thereby giving experimental support to the original assumption. EXPERIMENTAL TECHNIQUE AND RESULTS

-.

=

.

.

.

Lifetime measurements have been performed on the different zero-phonon transitions of the 5D 0 ‘F~ multiplets with the high resolution, 4 A,, provided by a double grating monochromator. Because of the extremely low emission intensities, a photon counting apparatus was -~

paper on the thermal behavior of the Sm ~ luminescence intensity it was shown that the strength of the zero phonon line did not follow the 1697

LIFETIME OF THE Sm~4 5D0

1698

.-~

~

1 ~ I0

RbCL ~ x702

70’•

•

I

KC(

~—&___

KBr

‘F0 EMISSION IN ALKALI HALIDES

be easily by measured. be absorpachieved shining Excitation light in the could different tian bands of Sm~through the use of suitable filters. However, no measurable difference was observed in the lifetime values when varying the exciting wavelength. A filter, opaque to exciting light with A> 6200 A, was employed in order to reduce the light scattered by the sample in the wavelength range of the emission. All the samples had been previously quenched at 600°C in air, in order to dissolve Sm~~ precipitates.

-

~N’

DISCUSSION

-

The curves of Fig. 1 clearly show the presence of an exponential process which can be justified according to the scheme of Fig. 2. The excitation produced by the flash raises the system from the ‘F 0 configuration (level 1) to an excited state which relaxes in a very short time

\

-

70

-

-

.

-

to the D to state (level 2). process No radiation corresponding this relaxation is observed.

t \ 70 I

0.05 1/

~

From level 2 radiative emission is possible with the transition 2 1 whose rate is proportional to the reciprocal of TR, the lifetime for radiative emission. If there exists a level 3 with energy E2 + L~E, a Boltzmann distribution can be established between the levels 2 and 3 such that the total probability of transition from level 2 i~ -+

i.,~

070

0.15

Vol.8, No.21

0 .

5D FIG. 1. Lifetime for the 0 ‘F0 transition of Sm ion in KC1, KBr, RbC1; solid curves have been obtained from equation (1), and the parameters of Table 1, part a. -~

=

1 T

developed so that photons could be resolved in time by means of a gate circuit operated with variable delay times. In Fig. 1 are reported the decay times of the Sm~ ~ ‘F0 emission line (X 6890 A, ~w 1.8eV) for KC1, K8r and RbCl at temperatures between 6 and 300°K. 5DMeasurements in the phonon sideband of the 0 ‘F0 transition and 5D in the other zero-phonon multiplets ( 0 ‘F’,) have yielded values of the lifetime which appear to be the same as that of the 5D 0 ‘F0 line within the experimental uncertainty of by ±10 per cent.a The data havebulb been obtained employing xenon flash operated

=

1 TR

+

!.

(1)

~

TJ

1,’-r being proportional to the transition rate of 2 3 process. If the ion is excited with light pulses which are short compared with T, then the radiative decay. of level 2 is exponential with time constant T.

-,

—

-.

-.

The analysis of the data reported in Fig. 1 according to equation (1) yield the values of the parameters which are reported in Table 1. Level E ~ does not contribute to any line in the emission spectrum. The only lines observed 5D in emission are those corresponding to the 0 multiplets which have 3energies below of the No transition that is found ‘F0 transition. in emission with energy greater than 5D 7F 0 0 (3 1 transition). In order to correlate level 3 with levels some configuration, we note that the 5D, the 5D, configuration levelamong lies at -.

...

-.

at frequencies from 10 Hz to 1000 Hz. The decay time of the exciting source was 1.6~.tsecso that luminescence lifetimes > 2 ~sec could be

*

Vol.8, No.21 LIFETIME OF THE Sm~ 5D0

‘F0 EMISSION IN ALKALI HALIDES

1699

Table 1.

(a)

(b)

Tj~(m sec)

T @sec)

AE/k(°K)

~E(cnr’)

KC1

11.5 ±1.0

1.14 ±0.10

186.2 ±3.3

130.1 ±2.5

KBr RbCl

10.0 ±1.0 10.7 ±1.0

1.29 ±0.14 1.26 ±0.07

150.8 ±3.5 371.0 ±3.9

105.0 ±2.5 257.9 ±3.0

191.8 ±2.0 128.4 ±1.7

133.3 ±1.4 89.2 ±1.2

397~2±4.9

276.0 ±3.5

KCI KBr

—

1.31 ±0.07 4.60 ±0.4

RbCI

—

1.08 ±0.07

—

______________

\ ~\ 1,

E,

—

‘0,

~

______________

I E

corresponding to a configuration different from 4f are very close to the ~ (4f°) level; in particular, in the case of CaF2, these 5D 6 levels are lower in energy than the 0 level. Preliminary data on NaBr and NaCl show that the lifetime does not change substantially when cooling the crystals to temperatures of the order of that of liquid helium. ~ This might indicate that in Na halides level 3 is closer to ‘D 0 than in K halides. It is to be noted also that the emission from NaBr and NaC1 doped with Sm~shows stronger electron phonon coupling (larger Huang—Rhys factors) than that from K halides. In .Nal Feofilov and Kaplyanski have not fo~andthe ‘D0 ‘F, (4f’ 4f’) emission, which may indicate that 8D level 3 in this crystal lies below 0. -.

-.

1‘D FIG. 2. Decay scheme for the Sm~ 0 ‘F0 transition in alkali halides (energies not in scale). -~

5D energies thecrystals,’ order of ~l000cm’ above3 is0 toin different of host so that level be sought in configurations such as 4f’5s, 4f’5p. Furthermore the energies of the 5D, levels should not depend much upon the host matrix, due to the extreme localization of the 4f orbitals. We observe instead that the energy difference i~E,as given in Table 1 (part a), 130 (KC1), 105cm’ (KBr) and 258cm’ (Rod), depends appreciably upon the crystal. A similar dependence is found for the lowest absorption band which involves interconfigurational transitions (4f Sd).5 It is likely then that level 3 is due either to 4f5s or to 4f5p configurations which are not connected to the ground state by an allowed optical transition. It is also observed that in alkaline earth fluorides the Sm~levels

The results discussed here are useful for a 2 further analysis of the oscillator strength data of the ‘D0 zero phonon line. ..

We may say that the efficiency TJR of the radiative process which refer to the luminescencc of Sm + + in alkali halides can be written as V

1 T

=

/

1

+

~4!.

-~E/kT

.~.

(2)

e

where 1/r is the total transition rate, whereas l/TR is the rate for the purely radiative process.

The expression for the oscillator strength of the zero phonon transition is then rewritten as

1700

LIFETIME OF THE Sm~ bD0

‘F0 EMISSION IN ALKALI HALIDES

T(’K) 70

20

f0(T)

50 700

Vol.8, No.21 *

=

f~(O)77R(T) exp~—S0(coth~.—1)

Rb Ct

__~

f0(O)m~(T)

KCL

=

f0(O)

1..

(3)

TR

70’

KBr

-

~

where S0 is of the order of unity and T*

-

150°K

Since the exponential factor changes only by 1 per cent in the temperature range 0—50°K, ~ 10’

o

-

it can be taken as a constant. Therefore the strength of the zero phonon line follows a thermal behaviour given by 77R(T) which is directly

-

-

.

t 70

-

proportional to the lifetime T, In Fig. 3 we report the thermal dependence of the zero phonon emission oscillator strength in the crystals examined; solid lines are obtained from equation (3) with a suitable choice of the

~-

-

I 0.75

I

0.70

0.05 l/T (‘K’)

5D FIG. 3. Oscillator strength of the 0 ‘F0 transition. The solid curves are obtained employing equation (3) and the parameters of Table 1, part b.

0..

parameters. These values are reported in Table 1, part b~We can note a good agreement between the values obtained from the mean life and the oscillator strength data; this confirms our original assumption about the5D mechanism for the thermal de-excitation of 0 level of Sm~ion in 2 the alkali halides.

Acknowledgements We are grateful to Prof. P.F. Manfredi for useful suggestions during the design of the electronics and to Mr. C. Ricci for designing most of the circuitry used in this paper. —

2.

REFERENCES See, for example, DIEKE G.H., Spectra and Energy Levels of Rare Earth Ions in Crystals, Interscience (1968) and references quoted therein. BALDINI G. and GUZZI M., Phys. Status. Solidi. 30, 601 (1968).

3.

BRON W.E. and HELLER W.R

4. 5.

ALAM M. and DI BARTOLO B., Phys. Rev. Lett. 19, 1030 (1967), and J. Chem. Phys. 47, 3790 (1967). KAPLYANSKH A.A. and FEOFILOV P.P., Soy. Phys. Optics and SpectrQscopy, 16, 144 (1964).

6. 7.

FEOFILOV P.P. and KAPLYANSKII A.A., Soy. Phys. Optics and Spectroscopy, 12, 272 (1962). BALDINI G. and GUZZI M., to be published.

1.

,

Phys. Rev. 136, A1433 (1964).

2

Vol.8, No.21 LIFETIME OF THE Sm~‘D0

-,

‘F0 EMISSION IN ALKALI HALIDES

On a mésuré la durée de vie de la luminescence de Ia transition 5 7 D0 F0 de I ,.ion Sm++ dans KC1, KBr and RbC1 a des temperatures comprises entre 6 et 300°K.On observe que la durée de vie, dc l’order de quelque ~sec a tempe~ratureordinaire, augmente d’une facteur iO~en réfroidissant l’échantillon a la temperature de l’heiium liquide. Selon-le modéle propose pour expliquer ces résultats, on fait l’hypothèese 5D d’un niveau, quelques lOOcm’ au~dessusde la configuration 0, qui cause la forte variation de la durée de ‘vie avec Ce niveau-ládea laCté assigné 55slaoutemperature. 41 55p. La dependence durée de viea la de configuration la temperature 4f explique complètement la variation avec in temperature de l’intensité de la raie s;~nsphonons observée en emission, que nous avions disc utée dans une publication préce dente. -

-~

1701