Optical absorptions and rotations in the ferrimagnetic garnets

HAL Id: jpa-00236052

https://hal.archives-ouvertes.fr/jpa-00236052

Submitted on 1 Jan 1959

HAL is a multi-disciplinary open access archive for the deposit and dissemination of sci- entific research documents, whether they are pub- lished or not. The documents may come from teaching and research institutions in France or abroad, or from public or private research centers.

L’archive ouverte pluridisciplinaire HAL, est destinée au dépôt et à la diffusion de documents scientifiques de niveau recherche, publiés ou non, émanant des établissements d’enseignement et de recherche français ou étrangers, des laboratoires publics ou privés.

Optical absorptions and rotations in the ferrimagnetic garnets

J.F. Dillon

To cite this version:

J.F. Dillon. Optical absorptions and rotations in the ferrimagnetic garnets. J. Phys. Radium, 1959,

20 (2-3), pp.374-377. �10.1051/jphysrad:01959002002-3037400�. �jpa-00236052�

OPTICAL ABSORPTIONS AND ROTATIONS IN THE FERRIMAGNETIC GARNETS

By ^{J. F.} DILLON, Jr.,

Bell Telephone Laboratories, Incorporated, Murray ^{Hill, New} Jersey, U. S. A.

Résumé. 2014 Les propriétés optiques ^de plusieurs grenats ferrimagnétiques ^ont été observées.

Il y a plusieurs ^maxima d’absorption au-dessous de la limite d’absorption qui ^{se trouve à envi-}

ron 5 200 Å. La lumière qui passe à travers le cristal subit une rotation non-réciproque ^{de son} plan ^de polarisation. La ^structure de la courbe représentant la rotation en fonction de l’énergie

reflète celle de la courbe d’absorption. ^{On observe un} dichroisme magnétique circulaire. Ces résul- tats nous permettent ^d’étudier les niveaux d’énergie électroniques ^{dans ces matériaux} magnétiques.

Des propriétés ^comme la structure des domaines peuvent être facilement vues et étudiées _par la lumière transmise.

Abstract.

The optical properties ^of several of the ferrimagnetic garnets ^have been measured.

There are several maxima in the absorption ^{below an} absorption edge ^{at about 5 200} Å. Light passing through ^the crystals undergoes ^a nonreciprocal rotation of its plane ^of polarization. ^The

structure in the plot of rotation versus photon energy reflects that in the absorption ^{curve. A} magnetic circular dichroism is observed. These data allow us to study the electronic energy levels in these magnetic materials. The properties ^are such that domain structure can easily ^be

seen and studied by transmitted light.

PHBSIQUE 20, 1959,

Introduction.

The properties ^{of the} recently

discovered ferrimagnetic garnets [1], [2] (FMG)

have been the subject ^{of a} very considerable study.

The interest in these compounds ^{has been} greatly augmented by ^the availability ^{of sound} single crystals [3]. ^These have, ^for instance, ^{become the} preferred experimental ^material for fundamental studies of ferrimagnetic resonance. Recently ^the

author has found that sections cut from these crys- tals are sensibly transparent ^{to visible} light. ^This transparency ^{has been} briefly reported elsewhere, [4], [5], and summaries have been given ^{of thé} optieal. properties ^{as well as} the results of domain observations [6]. This paper will deal with ^the

optical properties of several of the FMG.

Sampl(s.

^The crystals ^{from which our} samples

were eut were grown by J. W. Nielsen from lead monoxide flux [3]. The actual specimens ^{for the}

transmission measurements were made by grinding

and polishing thin sections. The most suitable

samples ^were those of about . 0025 cm in thickness.

A detailed description ^{of the} preparation proce- dures is given ^elsewhere [6].

Abscrption.

^{The results} reported in the pre- sent work are restricted to the visible spectrum.

D. Wood [7] ^{of these} laboratories and Porter et al. [8] have measured the absorption of several of the FMG in the near infrared. Figure ^{1 shows the}

overall course of the absorption ^{from the near}

infrared to the visible. Starting ^{at a} few hundred

wave numbers the curves show vibrational absorp-

tions up ^to about 1 200 cm-i, ^{then a} long spectral

interval in which the crystals ^are quite transpa-

rent. Just below the visible there is a peak ^{in the} absorption ^at about 11 000 cm-1.

In figure ^{2 is} plotted ^the absorption (as ^{in the}

FIG. 1.

A plot ^of optical absorption ^{versus wave number}

from the ^near infrared into the visible. This curve is a

composite ^of data taken by ^D. Wood and the author.

Below 1 000 cm-’ there are a number of high sharp peaks corresponding ^to the fundamental lattice vibra- tions.

Beer-Lambert law I

Io exp(- at), ⁱⁿ ern-1) against ^wave number in that part ^{of the visible}

Article published online by EDP Sciences and available at http://dx.doi.org/10.1051/jphysrad:01959002002-3037400

375

spectruin ^{in which} measurements have so far been

possible. ^{This curve} applies ^to yttrium iron gar- net. At energies higher than about 19 000 wave

numbers, ^no detectable light ^is transmitted through

Fie. 2.

Plots of absorption against ^{wave number in the} visible portion ^{of the} spectrum. ^The absolute and rela- tive vertical positions ^{of the various} curves are only approximate since with these _very small samples ^{it was}

found difficult to exactly normalize the curves.

specimens ^of the thicknesses used. The absorp-

tion was measured at several temperatures ^down

to that of liquid helium. As the temperature ^is lowered the structure in the absorption ^becomes

somewhat sharper. ^The peak ^at 16 000 cm-’ nar- rows by ^{about a factor of two in} going ^{from 300,DK}

down to 4,2 OK. The low temperature ^curves

seem to indicate ^more clearly ^an absorption peak

at about 14 000 cm-1. The values plotted ^{for the} absorption ^{have not} been corrected for reflection.

The measurements were, made with ^a single ^beam

instrument in which the light transmitted through

a small hole (~ 0. 5 x 1.2 mm) ^was compared

with the light transmitted through another small

supposedly ^identical hole which was covered by

the sample. ^{Thus the exact vertical} positions ^of

the curves in figure ^{2 are not} significant ^and they

have in fact been slightly adjusted ^{so the curves}

would not overlap.

The gadolinium ^iron garnet absorption ^{curve has}

been measured also. Any of ^{the curves in} figure ⁴

may be taken as representative ^{of it at room} tempe-

rature. Note the similarity ^to figure ^{2. The}

overall shape ^{of the} absorption ^{curve for erbium}

iron garnet ^is again ^the same, however there is

one important différence in detail. Superimposed

on this curve are a number of relatively sharp absorption lines, ^{as are often seen in the} absorption spectra ^of compounds containing certain of the rare

earths. The most prô minent group of lines lies

near 6 525 A. Lines near 5 206 À have been seen.

These results will be given ^{in a later} publication.

Rotation.

If plane polarized light is incident

on the sample, it is observed to undergo ^{a non-} reciprocal ^{rotation on} transmission. For a given wavelength ^and sample ^{thickness the amount of}

the rotation is proportional ^{to the component of}

the magnetization along ^the optical path. ^{In the}

case of yttrium ^iron garnet ^{the rotation at two} temperatures ^is given ⁱⁿ figure 3. These differ

FIG. 3.

Specifie ^rotation plotted against ^{wave number}

in the visible portion ^{of the} spectrum. ^{These curves} apply ^to yttrium ^iron garnet. They ^were taken with a

field of 2 500 Oe along ^{the line of} sight.

rather little in net rotation, though ^{the lower tem-} perature ^curve does show ^{an extra} inflection at

about 17.500 cm-1 as compared ^{with the 300 OK}

curve. At 300,OK the rotation in the gadolinium

and erbium compounds ^is again essentially ^the

same, except that in the erbium there are anomalies in the rotation corresponding ^{to the sharp} absorp-

tion lines mentioned above. In the ^case of the

gadolinium compound, ^{the rotation} changes sign

when the sample is cooled through the compensa- tion point ^{near 283,DK. The} rotation is obviously

tied to the component ^{of one} of the sublattice

magnetizations along the line of sight, rather than of the net magnetization ^{of the} ferrimagnet.

Gadolinium iron garnet ^{at the} compensation point ^{with no} applied ^{field is an} antiferromagnet.

Yet by examination of the sense of rotation of

light passing through i1., ^{we can} directly ^détermine

tlie orientation of the iron sublattice magnetization.

Comparing figures ^{2 and} 3, ^the absorption ^maxi-

mum at about 16,000 ^{cm-1 is} clearly ^associated

with the rotational anomaly ^{at the same} energy.

Similarly, ^the general upward ^{trend of the rotation}

is clearly connected with the absorption edge ^at

say 19,000 ^{crn-1. If we were to subtract out this} upward trend, ^{it would be} clear that the part ^of

the rotation which is associated with the

16,000 ^cnrl peak ^has opposite sign ^on opposite

sides of the absorption.

This contrasts with the ordinary Faraday ^rota-

tion associated with the inverse Zeeman effect.

In that casse the rotation has the ^same sign ^{on both}

sides of the anomalous region, though ^{it reverses its} sign within that region. The variation of rota- tion with wavelength found here is similar to that observed in paramagnetic materials. However

Clogston’s ^{treatment of the} question indicates that the origin of the rotation in these ferrimagnets ^is fundamentally différent from that now accepted

for the paramagnetic rotation. Note that the rota- tions observed here are quite large, ^{2 000} o/cm ^is 5°/ .0025 cm, ^a typical specimen ^thickness.

Circular dichroism.

In view of the large ^rota-

tions observed, ^{an obvious} experiment ^{was to}

FIG. 4. - With the gadolinium ^iron garnet sample ^satu-

rated along the line of sight absorption measurements

were made for the two senses of circularly polarized light.

Curves (a) ^{were taken a few} degrees above the compen- sation point, ^curves (b) ^{a few} degrees ^below.

measure the absorption separately ^{for the two}

xrnsos of circularly polarized ^{light,. When this}

was donc the absorptions ^{were found to} bye diffe- rent. The solid and the dotted curves in figure ^4a apply ^{to the two senses of} ciroularly polarized light. In both the cause of the yttrium ^{and the} gadolinium ^iron garnets ^{at 300 °K the curves are}

like figure ^{4a. This’} particular plot applies ^{to the} gadolinium ^{case. When the} sample ^{was cooled} through ^the compensation point ^{the two curves}

cross as shown in figure ^4b.

In view of the difficulty ^of obtaining ^{a measure}

of the absolute absorption, further studies of this

phenomenon might ^{best be made with a double}

beam apparatus ⁱⁿ which the transmission of right

and left circularly polarized light could be compar- ed directly.

Preliminary measurements indicate that the difference between the two absorptions disappears

as the temperature is lowered. At 120 °Kit was

barely observable.

Conclusion.

Its is desirable to identify ^the

transitions which have been observed in the absorp-

tion spectrum ^{of the} ferrimagnetic garnets. ^The absorption data show strong ^{maxima at} 12,000

and 16,000 ^{cm-1 and a weaker one at about} 14,000 cm-1. Low temperature ^{rotation measu-}

rements suggest ^{an .} anomaly ^at perhaps 17,500 cm-1. The similarity ^{of the} absorption

curves for the yttrium, gadolinium, ^{and erbium} garnets ^{and the} fact of rotational anomalies implies

that the transitions with which we are concerned

are between energy levels of the magnetic ^elec-

trons of the Fe+++ ion. A number of authors have treated the effect of a cubic crystal ^{field on the}

energy levels of the 3d5 configuration appropriate

to Fe+++. Perhaps the correlation diagram given by Tanabe and Sugano ^{[9] is the most} complete.

Consideration of it leads to the following ^tentative

identifications. The 12,000 ^and 16,000 ^cm-1 absorptions may be the parity ^and spin ^forbidden (for ^{the free} ion) transitions 3d5 685/2 ^{- 3d5} 4 Tx (4 G) ^{and 3d5} 685/2 ^-+ 4T2 ^{(4G) of the Fe+++}

ions in octahedral sites. We ^{do not} expect ^the corresponding transitions for tetrahedrally ^coordi-

nated ions to be depressed ^{into the} visible, though

these might ^{account for the} 14,000 ^and 17,500 cm-1 anomalies. On the other hand the

14,000 ^cm-l peak may well represent ^{the transi-}

tion 3d5 68 -> 3d5 2T2 (2I). Though highly forbidden, this transition would borrow intensity by spin ^orbit coupling ^{from the} nearby transitions to the triplets arising ^{from the 4G.}

We have seen that the study ^{of the} optical pro-

perties ^{of these} transparent magnetic ^materials gives ^{us access to} information on their electronic energy levels. Clogston [10] discusses the strength

of the absorptions ^{and the} origin ^{of the} ferrimagne-

tic rotation ^{in an} accompanying paper. ^{As has}

been reported elsewhere, ^the absorption ^{and rota-}

377

tion properties ^{of these} crystals ^make possible

the direct observation of magnetic domains. In all of the FMG, domains may be studied under a

wide variety ^of field and temperature conditions.

The only requirement ^{is that we} prepare samples sufficiently ^{thin to transmit a} reasonable ^amount of light. ^It has, ^for instance, ^been possible ^to

obtain _very striking moving pictures showing ^the

effect of external fields ^on domain structures and of the effects on those structures of large changes

in the spontaneous magnetization. ^The veryfact

of transparency ^{enables us to} examine for internal

perfection many of the specimens ^{on which we do}

other types ^of experiments, ferrimagnetic ^resonance

for example. Finally, ^{it seems that the same com-}

bination of rotation and absorption properties

which allows us to see the domains, will allow us

to observe changes ^of Mz ⁱⁿ ferrimagnetic ^resonance experiments.

Aeknowledgements.

The author wishes ^to thank C. J. Ballhausen, ^{A. M.} Clogston ^and

A. D. Liehr for _many helpful discussions. He is

grateful ^{to D. Wood for} permission ^to quote ^his unpublished ^data.

REFERENCES

[1] ^BERTAUT (F.) ^{and FORRAT} F.), ^{C. R. Acad.} Sc., 1956, 141, 382.

[2] ^GELLER (S.) ^{and GILLEO} (M. A.), ^Acta Cryst., ^1957, 10, 239.

[3] ^NIELSEN (J. W.) and DEARBORN (E. F.), ^{Growth of} Single Crystals ^of Magnetic. GARNETS, ^J. Phys.

Chem. Sol., 1958, 5, ^202-207.

[4] ^DILLON (J. F., Jr.), ^{Bull. Amôr.} Phys. Soc., 1957, 2, 238.

[5] ^DILLON (J. F., Jr.), Optical Properties of Several

Ferrimagnetic Garnets. J. Appl. Physics, 1958, ^29,

539-541.

[6] ^DILLON (J. F., Jr.), Observation of Domains in the

Ferrimagnetic ^Garnets by Transmitted Light.

J. Appl. Physics, Sept. 1958, ^29.

[7] ^WOOD (D.), Private communication.

[8] ^PORTER (C. S.), ^SPENCER (E. G.) ^{and LECRAW} (R. C.), Transparent Ferromagnetic Light ^Modulator Using

Yttrium Iron Garnet. J. Appl. Physics, 1958, 29,

495-496.

[9] ^TANABE (Y.) and SUGANO (S.), J. Phys. ^Soc. Jap., 1954, 9, ^766-779.

[10] ^CLOGSTON (A. M.) (Proceedings ^{of this} conference, ^J.

Physique ^Rad., 1959, 2, 151.) DISCUSSION

Mr. Porter.

The change ^{of sense of rotation on} cooling through ^the compensation point, ^{as describ-}

ed for gadolinium ^iron garnet, ^is strong ^evidencc

for the relation of ^the rotation to a mechanism within one of the sublattiees, perhaps ^{in the way} Clogston has described. However, ^{it should be}

borne in mind that one encounters extremely high coercivity ^{and an} appreciable susceptibility ^{in the} neighborhood ^{of the} compensation point. ^The

text does not indicate clearly ^{that the} applied

field was sufficiently large ^{to cause a reversal of} magnetization ^{within a small} temperature change.

What value of fiehd was used ?

Mr. Dillon. - The magnetic ^field used in all the measurements of specific ^rotation, absorption,

and circular dichroism was 2 500 Oe along ^the

line of sight. ^{Data were taken} sufficiently ^away

from the compensation point ^{so that} the magne- tization was clearly ^reversed in that field.

Mr. Kornetzki.

Which was the direction of the magnetic ^{field in} your pictures ?

We have seen that thé .elementary ^{domains are}

in the direction of the field in the point ^{of rema-}

nence, if the magnetic ^{field was in the} direction of the plane.

Mr. Dillon.

The domain pictures ^{shown were}

of domains in gadolinium ^iron garnet. ^{In these} particular photographe ^the applied ^{fields were nor-}

mal to the major surfaces, ^i.e. along the line of

sight. Crystals ^{which are dominated} by ^surface

strain have, ^{at zero} applied field, ^{a structure con-} sisting ^{of domains normal to} the surface. If the surface layers ^are removed, ^the magnetization ^will

revert to any easy directions in the plane. ^{We will}

show this in subsequent publications.

Mr. Bates.

How was the surface layers ^{of the} garnets, ^{which must have been} strained by grind- ing ^and polishing, ^{removed ? I ask this} question

because the tubular patterns ^{remind me of the}

Bitter figures which Craik has obtained by ^his

electron microscope technique ^{on cobalt} crystals.

Mr. Dillon.

At the time this work was carried

on we did not have techniques ^for removing ^the

strained layers. However, ^{the effect of the sur-}

face strain is to create an easy axis normal ^to the

plane. ^This generally overrides the magneto-

crystalline anisotropy. ^{Thus we can saturate the}

crystal ^{normal to} the surface by applying ^{a field}

larger ^than 4nMs.