FERROMAGNETIC RESONANCE

By C. KITTEL,Bell Telephone Laboratories, Murray Hill, New Jersey.

Sommaire. - Cette communication passe en revue les travaux exprimentaux et thoriques dansle domaine de labsorption de rsonance ferromagntique. Leffet lui-mme a t dcouvert expri-mentalement par Griffiths en 1946. On montre quil est prvu par les quations de rsonance clas-sique, puis on en esquisse une interprtation par la Mcanique quantique. On discute les vrificationsexprimentales des conditions de rsonance pour des corps de diffrentes formes. On tudie la liaisonentre les expriences de rsonance en ondes trs courtes et les mesures ordinaires du coefficient gyro-magntique. On tablit lquation donnant le point dantirsonance. On discute diverses influencessur la largeur des raies. On rsume les rcentes tudes exprimentales de Bl0153mbergen, Bickford etYager et Galt. Le mmoire se termine par une thorie de linfluence de la frquence sur la permabilitinitiale des ferrites qui est en accord avec les rsultats exprimentaux rcents.

LE JOURNAL DE PHYSIQUE ET LE RADIUM. TOME J9 MARS 1951, PAGE 291.

1. - Historical Review.

Ferromagnetic resonance absorption was discussedtheoretically (Gans and Loyarte [1]; Dorfmann [2])before its experimental discovery by Griffiths [3]in the Clarendon Laboratory in Oxford in 1946.In particular the theoretical paper by Landauand Lifshitz [4] written in Kharkov in 1935 providesa foundation from which, after certain essentialmodifications, many of the later results can bederived. After the first observation of ferromagneticresonance the development of the subject wasperhaps more strongly affected by contemporarywork in the fields of nuclear and electronic para-magnetic resonance, than by the earlier thinkingon the subject.Resonance Equations. - The pioneer paper

by Griffiths gives the results of resonance absorp-tion experiments at several microwave frequencieson thin films of iron, cobalt and nickel. If hisresults are interpreted in terms of the Larmorequation connecting resonance frequency and appliedfield :

we are led to values of the spectroscopic splittingfactor 9 which are considerably too high, sometimesas high as 12, while the g-value for electron spinis 2.00.The derivation of Equation (1) is based on the

equation of motion

where J is the angular momentum per unit volumeand M is the magnetization. We suppose that

where y is the magnetomechanical ratio and isequal to 2 for electron spin. If we set H equal to

me

the applied field H= we have

these equations may be solved for Mr, giving

so that the resonance frequency is given by

This means that the energy absorption from anincident r

- f field of frequency Wo will go througha maximum when the static field intensity is equalto w0. We have neglected demagnetizing effects

yin reaching the above result.

In the actual experimental arrangement employedby Griffiths and most subsequent workers thereare appreciable demagnetizing effects which act toshift the resonance frequency. We consider aferromagnetic sheet whose surface is the plane y = o,and which is subjected to a strong static field H~.The essential point is that we must explicitly take intoaccount the demagnetizing field H~ =-4 7rMy whichresults when in the course of the precessional motionof the spin system there occurs a magnetizationcomponent normal to the plane of sheet. Theequations of motion become

giving [5, 6]

When this equation is used to interpret the results

Article published online by EDP Sciences and available at http://dx.doi.org/10.1051/jphysrad:01951001203029100

292

of Griffiths we find g-values often in the vicinityof 2,1 to 2.2, which is somewhat closer to our originalexpectation. The results are exhibited in figure 1.A typical resonance curve is shown in figure 2, fromthe work of Yager and Bozorth [7] on Supermalloy.

Fig. i. - Summary of data of Griffiths on ferromagneticresonance in plane specimens : comparison of effective g

1

values as calculated using H and (BH)*4 in the resonance1

equation. Notice that (BH).4 gives g values independent, of frequency and close to the free electron value 2.

The resonance frequency depends in general onthe form of the specimen, and we have a simpleresult only when the specimen is in the form of ageneral ellipsoid, and here only when the dimensionsof the specimen are small enough in relation to theskin depth and to the wavelength in the materialso that the r

- f field penetrates uniformly and sothat the phase of the r - f field is substantiallythe same throughout the specimen. In the caseof a plane specimen these requirements may berelaxed, however, for geometrical reasons. Theresonance frequency in an ellipsoid with demagne-tizing factors Nx, N" N= with the static magneticfield in the z-direction is given by

This relation reduces in the case of a sphere to

which was first verified experimentally by Hewitt [8].The predictions of Equation (10) regarding paralleland perpendicular orientations of a plane specimenwere verified by Kittel, Yager and Merritt [9].Other experiments with spherical specimens offerrites have been performed by Beljers [20] andYager, Galt, Merritt, Wood and 3Iatthias [26].The resonance frequency in a ferromagnetic

single cristal will depend on the angle the magneti-zation makes with the crystal axes, as a result of

the magnetocrystalline anisotropy energy. Inthe (001) plane of a cubic crystal we have

2. - Resonance curve for Supermalloy, according toYager and Bozorth; the apparent permability is plottedon a logarithmic scale.

where 0 is the angle between the z-axis and the [100]direction. In the (011) plane we have, accordingto Bickford [10]

where 0 is the angle between the z-axis and a [100]direction. It must be emphasized that the aboveequations are valid only when the static field is suffi-ciently large that the magnetization may be consi-dered to be oriented approximately in the directionof the static field; that is, we must have - MzThe anisotropy effect in a single crystal was first

verified experimentally by Kip and Arnold [11],who worked in the (001) plane of a Si - Fe crystal.They did experiments at two wavelengths, near cmand 3 cm. At I cm H K 1 and the predictionsM

293

Kof the theory were confirmed; but at 3 cm H Kand the simple theory is no longer applicable.At 3 cm two resonance peaks were found for certainorientations, and this result was explained satisfac-torily by considering the deviations of the magne-tization from the direction of the static field.Thus far our discussion has been based on classical

theory. One expects classical theory to be appli-cable as the quantum numbers of the entire systemare extremely large. It is, however, reassuringthat quantum-mechanical calculations have beencarried out which lead to the same resonance condi-tions. A most exhaustive investigation of thequantum-mechanical problem has recently beenmade by Van Vleck [12], who discusses the shapeand anisotropy effects as well as considering linewidth and g-values. This treatment is of particularvalue in exhibiting the approximations which mustbe made in order that the quantum theory resultsbe equal to the classical results. Other quantum-mechanical treatments have been given by Pol-der [13], Richardson [14], and Luttinger andKittel [15].The latter paper gives perhaps the simplest method

by which one can see that the Zeeman eigenvaluesare in fact displaced by demagnetizing effects inthe manner predicted by classical theory, as inEquation (10). The classical macroscopic Hamil-tonian of the system is

where V is the volume of the specimen. Thisequation may be rewritten to give the quantum-mechanical Hamiltonian

where the Js are angular momentum operators.We let H operate on the function

where the utn are eigenfunctions of the operators J2,J,,. We are thus led to a difference equation onthe ps; the difference equation is then approxi-mated by a differential equation. The differentialequation is essentially that for a harmonic oscil-lator, and gives equally spaced eigenvalues withthe separation

which is just the classical result.Theory of g and g. - It was a surprising and

puzzling result of the ferromagnetic resonance

experiments that they led to g values appreciablyhigher than the values obtained from gyromagneticexperiments, such as the Barnett and Einstein-deHaas experiments. The situation now appearssomewhat improved, as it has been shown thatthe resonance experiments do not in principlemeasure precisely the same quantity as the gyro-magnetic experiments. The theoretical situationis not yet entirely satisfactory, and on the expe-rimental side there is an unfortunate lack of agree-ment in the results of gyromagnetic experimentscarried out by different observers. It would beof considerable importance to resolve the existinggyromagnetic discrepancies.We now look more deeply into the philosophy

underlying the two types of experiment. In agyromagnetic experiment we determine the magne-tomechanical f actor g by the equation "

where At is the magnetization and J the angularmomentum per unit volume. For conveniencewe shall imagine that we are always dealing withunit volume of material. Now the total AJ iszero by conservation of angular momentum, but weactually observe the angular momentum ofthe lattice, while

the terms Jspm and J~,~.;, being the angular momentaassociated with the spin and orbital motion of theelectrons. Now we may similarly decompose themagnetization :

but Mlat may be neglected as the angular velocityof the lattice is always negligible in comparisonwith the electronic angular velocity. We havethen the result that g is determined by the equation

Now we know that

so that if we imagine

where -- ~ l, we have the result

In a resonance experiment we determine thespectroscopic splitting f actor g by the equation

294

Now in a magnetic dipole transition

while for first-order approximation wave functionsincluding spin-orbit interaction

so that we must have

Therefore

which should be compared with Equation (28)for g. We have further

so or

as given independently by Van Vleck ( 12~ in consi-derable generality and by the present author [16]for a particular model.

Equation (33) is approximately verified by expe-riment to the extent that g is usually greater than 2while g is usually less than 2, but there remainsthe unsolved problem that g - 2 is usually greaterthan 2 - g. Numerical results are given in Table I.

TABLE l.

Conipaiison of 7-ep7-esentative experimental values ofspectroscopic splitting factor (g) and magnetontecha-nical ratio ( g ) for jerrornagnetic suhstccnces ( 1 ).

Aticrowave GyromagneticResonance Experiments

(1) The values of ~ are taken from the SUn1l11(lI, v bpb. J. BARNETT, Pro:;.

-4cad., 1944? 75, iog. The values3f g are based largely on J. H. E. Griffiths, reference [3] andunpublished data kindly communicated by Dr. Griffiths; alsoon various papers and unpublished measurements by W. A. Yagerand his collaborators, to which detailed reference is made inthe text. The value of g for magnetite is from L. R. BICKFORD,PAys. Rev., , 19)y, 76, 137. For Permalloy, g is from Kip andArnold, unpublished.

We have made above a statement about theproperties of the first order wave functions for astate with no orbital degeneracy, but includingspin-orbit coupling as a perturbation. For a proof

of the statement the reader is referred to the papersjust cited, and also to the paper by Gorter andKahn [17]. The original suggestion that g mightdiffer from g in the ferromagnetic case was madeby Polder [13] and Van Vleck [18].The quantity c- which determines the amount

of orbital momentum mixed in with the spin isgiven by an expression of the form

where C is a constant of the order of unity, A is thespin-orbit coupling constant, and 3 is an energylevel separation of the order of the separation inthe solid of energy levels which were coincidentin the free ion. The coupling constant a is positivefor an electronic shell less than half full, and negativefor a shell more than half full. We then expects --,to be positive, and g therefore > 2, for paramagne-tic salts containing Fe++, Co~-~, Ni++1 and Cu++ ions,and this is generally confirmed by experiment;s is also positive for the ferromagnetic metals Fe,Co and Ni, and the alloys thereof which have beeninvestigated. In Heusler alloy and this is what would be expected on the naivepicture that ions are here the principalcarriers of the magnetic moment, as Mn+- has ahalf-filled 3 d shell.

Anti-resonance o point. - It was pointedout by Yager [23] that when the results of a reso-nance experiment are plotted as log vs H a well-defined minimum is frequently exhibited at lowfield strengths, as well as the usual maximum athigher fields. For a plane specimen with the staticfield parallel to the plane, Yager gives in the limitof small damping

which may be compared with the familiar expressionfor the point of maximum I-Lit :

The point of minimum IJ-R corresponds physicallyfo 4i7M, being equal to H n but 180 out of phase,so that the microwave induction B,, - o; in otherwords, we have a field but no flux. The expressiontor the minimum is of considerable use in situationswhere it is not convenient to determine the satura-tion magnetization by ordinary methods; one maythen solve Equations (35) and (36) for both M,and y, as was done for example by Yager andMerritt [19].

It is easy to derive the general condition for theminimum. The r

- f susceptibility is, neglectingdamping,

295

where given by Equation (10) and

now will be zero when Zr = -1, > so that the, ,

i ;condition for o becomes

or

Special cases follow :Plane, II parallel

plane :Plane, II perpen-

dicular plane :

Sphere:The minimum in the plane parallel case is illus-

trated by figure 3, which was obtained with Heusleralloy by Yager and Merritt ( I 9~.

Fig. 3. - Ferromagnetic resonance in Heusler alloy, accordingto Yager and Merritt; note the minimum in the apparentpermeability.

Line width. - The first task in the investigationof line width is to set up a suitable form of a relaxa-tion expression to be included in the equations ofmotion, a form which will give an adequate repre-sentation of the experimental results. The firstattempt was to write

but this was shown by Yager (r g4g) to fit the expe-

rimental results at one frequency as H is variedonly if X is taken proportional to H. It is moresatisfactory to write , _

it is easily verified that the relaxation term is alwaysperpendicular to M, so that the absolute magnitudeof the total magnetization does not vary. Theconstant ~ has the dimensions of a frequency andwill be called the relaxation frequency. The relationbetween h and ~ is readily seen to be given by

Fig. 3 a. - Line width in nickel and Supermalloy,as a function of temperature (BIoembergen).

In the interpretation of experimental results inferromagnetic resonance one must always bearin mind that the actual resonance curves may bebroadened not only by genuine relaxation effects,but also by crystalline anisotropy effects in poly-crystalline specimens, and by local and surfacestrains and inhomogeneities in both polycrystallineand single crystal specimens; in alloys there mayalso be unnatural a3 sources of broadening asso-ciated with variations in the order-disorder structure.

Experimental values of i determined by Yagerare given in Table II. Studies by Bloembergenwill be referred to below.The progress which has been made in the theore-

tical explanation of line width consists in showingthat the several mechanisms for which calculationshave been made fail entirely to account for theobserved width. Akhieser [32] has calculated spin-spin and spin-lattice relaxation effects on the basisof magnetic dipole interactions between electrons;

296

this calculation leads to relaxation frequencies muchlower than observed; Van Vleck [12] has confirmedthis tendency with a calculation which bears rathermore directly on the experimental situation thandoes Akhiesers calculation, which has been criti-cized by Polder [13].

TABLE It.

Experimental values of Relaxation Frequency i,,from JleasttrC71?,eitts Qy A. Yager.

It is not surprising that magnetic dipole inter-actions do not account for the observed line-width,as it is well known that such interactions fail toaccount for the observed magnetostriction, whichalso results from interaction between magnetizationdirection and elastic strain.

Exchange interaction. - The effect of exchangeforces on ferromagnetic resonance has recentlybeen treated [24]. In a uniformly magnetizedspecimen exchange forces have no effect on theresonance frequency, as a consequence of the factthat the magnetization operator commutes

i

-

with the exchange operator E Si S j ; in classicali, j

language, the Weiss molecular field H,,, = isalways parallel to the magnetization M, and hencethe torque MxqM must vanish. But in metallicspecimens the microwave field penetrates onlyto a depth of 0-5 to cm, so that the r - fcomponent of the magnetization is non-uniform,and the exchange energy will play a role in deter-mining the resonance frequency and line width.The effects will in general be of importance onlyin pure metals at low temperatures, where the skindepth is considerably smaller than at room tempe-ratures.We write the exchange energy density as [25]

h A 9 JS2 f .

where A = 20132013 =2.0X1 ergs : cm for iron.a

The shift for the plane parallel arrangement is foundto be

where is the classical depth for permeability 2 (~.2;here ia == ?1 - j !J-2 There are contributions to

the line width of the same order of magnitudeas the field shift. At room temperature the effectin pure iron is only of the order of I o cerstedsat 2 4oo mc : s, but may be of the order of 100to 1000 oersteds at liquid hydrogen temperature.

II. - Recent Experimental Results.

Line width and g-values at elevated tempe-ratures. - N. Bloembergen [22] has recently inves-tigated the temperature dependance of the g-valueand the line width in nickel and supermalloy fromroom temperature to above the Curie point. This

Fig. 4. - Resonance curves vs. temperature for Supermalloy,according to Bloembergen.

marks the first observation of resonance in theparamagnetic region above the Curie point ofnormal ferromagnetic materials. Bloembergen foundthat the g-value is independent of temperaturewithin the limit of error (2 per 100), with g = 2 . ~ ~for supermalloy and 2.20 for nickel.The results of the line width determinations

at 24 4oo mc : s are plotted in figure 3 a. Thewidth Ago, which is equal to 2013 in Bloembergensnotation, is half the width in frequency betweenthe points where the imaginary part of the suscep-tibility has dropped to half its maximum value. Theabsorption is found to be continuous across the Curie

297

point in both materials. The Curie point is 3580 Cin nickel and about ![ooo C in supermalloy. Theresonance curves in supermalloy at various tempe-ratures are plotted in figure /{. The plane parallelarrangement was used.

_ Magnetite crystal at low temperatures. -L. R. Bickford [10] has studied ferromagneticresonance in an artificial crystal of magnetite(FeS04) at low temperatures, and from microwave

Fig. 5. - Anisotropy constant Ki of magnetite vs. tempe-rature, from microwave measurements by Bickford.

measurements alone he has determined the crys-talline anisotropy constant as a function of tempe-rature ; he has also contributed to theun ders-

Fig. 6. - Dependence of resonance field on crystal orientation,for magnetite at a fixed frequency (Bickford).

tanding of the transition which magnetite undergoesat 2013 160 C (i 130 K), characterized by anomalouseffects in specific heat, magnetization and electricalconductivity.At room temperature, values of g = 2.12

and 1 05 ergs : e cc were ohtained,

It was found that below about - goo C the absolutevalue of K1 decreases with decreasing temperature,as shown in figure 5, reaching zero at about - 1 fy3 C.Between - 1 130 C and the transition point Kl ispositive and increases with decreasing temperature.The g-value was found to decrease gradually withtemperature. A typical plot of the angular variationof the resonance field is given in figure 6.The behavior of magnetite in the resonance

experiments below the transition indicates thatthe magnetic symmetry is uniaxial in this region.Below the transition the magnetic axis is the [100]direction most nearly parallel to a strong magneticfield applied to the crystal as it is cooled throughthe transition. At temperatures not far below thetransition it was found to be possible to change themagnetic axis from one [100] direction to anotherby means of a strong magnetic field.

Nickel ferrite crystals. - Yager, Galt andtheir collaborators [26] have recently carried outresonance experiments on small spheres of nickelferrite NiO-Fe,03 cut out from single crystals. The

Fig. 7. - Resonance curve for nickel ferrite single crystalin form of sphere o, i o cm diameter (Yager, Galt, Merritt,Wood and Matthias).

frequency was about 2I ooo mc : s, and the obser-vations were made at room temperature. Theanisotropy constant was determined from thevariation of resonance field as a function ofcrystallographic orientation. The easy directionwas found to be [111], as in magnetite, andK, = - 6.o x 104 ergs : cc.The first sphere measured was o. I o cm in diameter,

and it was found here that for every orientationof the crystal the resonance consisted of a main lineand one satellite on either side, figure 7. Thesatellites are explained as caused by cavity-typeelectromagnetic resonances in the sphere. Theidea is that as we go up the permeability resonance

1curve we pass through values of sufficiently

298

large so that the half-wavelength of the radiationin the material is of the order of the diameter ofthe specimen, and near this point we expect a cavitytype resonance. It is not possible to verify thisexplanation directly, as the dielectric constant Eis not known for nickel ferrite, but the order of

Fig. 8. - Same as figure 7, except that sphere is o.o38 cmin diameter. Difference between curves is ascribed toelectromagnetic cavity-type resonance effects in the largersphere.

magnitude of the required wavelength in the materialappears to be reasonable. The explanation can betested, however, by measurements on a muchsmaller sphere, as for a sufficiently small sphere atthe same frequency one would not expect cavityresonances to be possible. Measurements on asphere about o.038 cm in diameter showed in factonly a single resonance peak, figure 8.

III. - frequency Dependence

of Initial Permeability.

Our recent understanding of the ferromagneticresonance effect enables us to explain the principalaspects of the frequency dependence of the permea-biliy of ferromagnetic materials under conditionsof normal usage, in which there is no static or biasingfield present. The discussion is particularly simplein the case of high-resistivity ferromagnetic subs-tances, such as the ferrites, as here we may neglectentirely eddy current effects, even on a microscopicscale. We first discuss frequency dependence insuch substances; it is also assumed that the mea-surements are interpreted in such a way that thetrue magnetic permeability is obtained, quiteseparate from dielectric effects, especially geome-trical resonance effects.

The average permeability of a polycristallinematerial may be written as

where is the average susceptibility associatedwith rotation of magnetization within the domains,and is the average susceptibility associatedwith displacement of domain boundaries. At verylow frequencies both and will contributeto the permeability; the two terms will, however,depend on frequency in different ways. In somecases it is possible experimentally to separate thetwo types of dispersion.The rotational susceptibility is determined largely

by the crystalline anisotropy energy, or in somecases by strain energy and the magnetostrictiveconstants. In either situation there will be aneffective anisotropy energy density K which deter-mines the magnitude of the rotational suscepti-bility, according to the classical relation

where M, is the saturation magnetization. Theremay be in special cases demagnetizing effects whichwill change the susceptibility appreciably, but weshall suppose that the average rotational suscepti-bility at zero frequency in a randomly-orientedpolycrystalline specimen is given by

This definition corresponds to the use of 2 h i asM,the effective anisotropy field.The frequency variation is given, using Equa-

tions (44) and neglecting all demagnetizing effects, by

where ~, is an average rotational susceptibility

approximately equal to and the resonancefrequency in the anisotropy field is

The sharpness of the resonance depends on the valueof the ratio

that is, if the resonance frequency in the field IVT,is much greater than the relaxation frequency ?, theresonance will be well defined, and vice versa. Inactual materials intermediate cases occur mostcommonly, but sometimes the rotational dispersion

299

r esembles a resonance process. In the two limitingcases the Equation (51) assumes the f orm :

. -;

~/ l1~;; resonance

), > ~ ~~T, ; relaxation

The displacement susceptibility at low frequenciesis determined by the restoring force tending toreturn the boundary wall to its equilibrium position,and by the number of walls intercepting a line ofunit length. We suppose that the restoring forceper unit area of wall is - q3x; the magnetizingpressure is 2A1 (3H), so that

The magnetization change is

where -r, is the average number of boundaries perunit length. Thus

It was first shown by Landau and Lifshitz [4]that there is a frictional term in the equation ofmotion of a boundary, of such a nature that in theabsence of a restoring force the wall velocity isgiven by

1

where (A )2 is the usual wall-thickness parameter;Kit is near the order of magnitude of the wall thickness,but somewhat smaller [25, 30]. A simple deriva-tion of Equation (59) has been given Kittel [38].The complete equation of boundary motion is

where in Equation (59)

We have then

or

further

This result [31] shows that the displacementsusceptibility has a frequency dependence of therelaxation type, with relaxation frequency

We estimate the order of magnitude of B bytaking 10% r, ~ ~ 10.3, ZSi-.P rV 102; and arough estimate gives G ~ 10.) cm : s : oersted, sothat B N cycles : s, although in high permea-bility materials -2k may be very much less thanone 3Ic : s.

Fig. 9. ~--~ Frequency dependence of permeability in magne-sium ferrite, according to Welch, Nicks, Fairweather andRoberts.

The most complete measurements available tothe author on frequency dependence in ferritesare these of Welch, Nicks, Fairweather andRoberts [29] on magnesium ferrite. Their resultsare shown in figure g. One may also refer to workby Birks [21].

Relationship between frequency and permea-bility. --- We have from Equation (52)

where is the rotational resonance frequencyand y is the magnetomechanical ratio. Thisrelation was first stated by Snoek [28], althoughin incomplete form, as he does not point out thatit is not the total susceptibility, but only the rota-tional susceptibility, with which the equation isconcerned. Now at high frequencies we are oftenconcerned only with the rotational part of thesusceptibility, as the displacement contributionoften drops out first. Equation [66] therefore

300...

conveys the interesting and technically importantresult that the product of the maximum usablefrequency by the effective high-frequency per-meability of a material is essentially a constant,dependent for practical purposes only on the satu-ration magnetization. We have approximately

(67)This relation is plotted in figure I o, and there areincluded several experimental data, which are seento be in fair agreement with the theory.

Fig. o. - Plot for several materials of maximum usablefrequency vs. effective high frequency permeability.

We pause to note that for y ~,,

and using Equation (65),

.

f k .

1 I -v~f. 3 if we take as a numerical example = 3, fj = 103.

which suggests that 2~ may often be less than ~.;further, for the same numerical values,

so that the displacement susceptibility may dropout with increasing frequency before the rotationalsusceptibility.

---

We note further that Equation (65) gives a rela-tion involving which is parallel in functionto. the relation (66) involving but appre-ciably more complicated and thus less useful.Rado, Wright and Emerson [37] have suggested

independently in connection with their measure-ments on a magnesium ferrite " Ferramic A " thatthe permeability may be broken down into rotationaland displacement contributions. In their materialthe wall damping is unusually low, and they accor-dingly find definite effects of the wall inertia termcalculated by D6ring [27]. This marks the firsttime that the inertial effects of wall motion havebeen observed. The condition for the resolutionof the inertial term is given as, approximately,

Dispersion in conducting materials. - Inconducting materials the wall movements aredamped out for the most part at lower frequen-cies [33] than in the ferrites. This is the resultof eddy currents in the material. At high fre-quencies the magnetization is expected to be attri-buable principally to domain rotation. One maycite the measurements of Johnson, Rado andMaloof [34], for example, who find that in iron200 mc : s magnetization by rotation is the dominanteff ect. Another confirmation is off ered by thereasonably good agreement of Equation [67] withthe data on conducting materials.

It seems probable then that the higher frequencyrelaxation effects are associated with rotationalrelaxation, rather than with eddy current effectspreviously suggested by several workers [35],including the present author [36]. It is also unlikelythat effects of the inertia of domain boundaries [27]are important here, as the frictional terms in theequation of motion of a Bloch wall are dominantin conducting materials and serve to eliminatewall motion at relatively low frequencies.

_

Acknowledgments. - I wish to express myindebtedness to my colleague W. A. Yager for hiskindness in making available various unpublishedresults, and thanks are also due to Professor J. H.Van Vleck and Dr N. Bloembergen for friendlydiscussions of the ferromagnetic resonance problem.The writing of this paper has been facilitated bythe constant assistance of J. G. Walker and MissPhyllis Brown. The drawings were kindly producedby the Research Drafting Department under E. H.Hasbrouck.

Remarque de M. Bates. - I only wish to say thatDr Standley at Nottingham has found a preliminaryvalue of g for MnAs by the method of ferromagnetic,

301

resonance of about 2.6; it is certainly greaterthan 2.5.

Remarque de M. Casimir. - Je propose dadopterles notations de M. Kittel en d6signant par g lavaleur trouv6e dans les exp6riences de resonancegyromagn6tique et par g les valeurs donn6es parles exp6riences Einstein de-Haas ou Barnett.

Remarque de M..Roberfs. -- I would like to makethree remarks on Dr Kittels contribution : io Regar-ding g values. I have obtained in some magnesiumand other ferrites, in powder form, with samplesof a few mm3 dispersed across the wave guide

cross-section, two resonance peaks as the transversestatic field is increased. These two peaks areusually unequal in height and very broad, andsuggest a splitting of energy levels of the magneticions in these ferrites. The direct interpretationof the results of high frequency measurementsin terms of the g-factor alone can therefore lead toerroneous values. 20 Regarding the experimentalline widths obtained in nickel and supermalloy,I wonder whether this could be accounted forin terms of a magnetic zero point energy such asProf. Stoner has found, in effect, to be operativein preventing the complete parallel alignementof all spins at oo K. 30 I should like to add newresults to those shown in Dr Kittels figure g.

Fig. A.

These are shown in my figure A and refer to a rangeof Mg ferrites in which Fe is gradually replacedby Al. It is found in all cases that these curvescan be matched by a relaxation term plus twodamped resonance terms like Frenkels dampedresonance, these two terms being nearly equaland opposite in low frequency magnitude anddiffering slightly in resonance and damping fre-quencies. To complete the picture, a third Frenkelterm should be added, to correspond with thosespins which give the resultant low frequency magne-tisation. The other two Frenkel terms then corres-pond with the two groups of Fe+++ spins whichmutually cancel at low frequencies.

Remarque de M. Stoner. - I should like to reportvery briefly on some work at Leeds by Millership,Webster and others on the measurement of initialpermeability of metallic ferromagnetics in the formof wires over a frequency range down to a wavelength of 3 cm, It is essential to measure both the

" resistive " and inductive " permeabilities fromwhich thee real " and 11 imaginary " componentsof the complex permeability may be determined.For iron the curves fit roughly, but not well, onthe theoretical curves obtained by Becker. Thefit is no better when a further theoretical term corres-ponding to resonance is introduced. For nickel,there are clear indications of resonance effects, nodoubt because the anisotropy field frequency isin the range examined, while for iron it is too high.The work on some five ferromagnetic metals andalloys is nearing completion.

Remarque de M. Casimir. - There have beenmade some measurements at Delft by Miss vanLeeuwen and pupils in cooperation with Kronig.Preliminary results have been published in " Phy-sica " and one may look for further results.

Remarque de M. Went. - I should like to saya few words on the role of the rotational processes

302

and wall displacements in the magnetization processif sintered ferrites. It is well known that in metallicferromagnetics the major contribution to the per-meability at low amplitudes is due to reversiblewall displacements. We have found (Went andWijn, to be published) that in sintered ferrites therotational processes are mainly active in the lowamplitude region. The observations are the fol-lowing. The initial permeability is virtually constantup to the frequency where precessional resonancesets in. While at larger amplitudes and low fre-quencies the permeability is usually considerablyhigher as a result of (irreversible) wall displace-ments, we find that, in many materials, at largeamplitudes and frequencies above roughly 100 Kc : sthe permeability decreases to a value exactly equalto the initial permeability. We have found thatthe gradual disappareance of the contribution ofthe wall displacements as a function of frequencycan be described by a single relaxation time. Furth-ermore we have found that the occurence of thedamping of the wall dispacements is closely conn6c-ted with the existence of volume magnetostriction.

Remarque de M. 1 I should like toadd a few comments to Wents remarks. He hasparticularly stressed that he has only done measu-

rements on sintered specimens and it seems thateven there the walls contribution of initial suscep-tibility is negligible only for certain ferrites and itseems that those ferrites have always considerablevolume magnetostriction. This suggest a mecha-nism of freezing in connected with volume magneto-striction. 2 In connection with Bozorths remarkthat the size of a specimen in a resonant cavityhas to be small compared to the wavelenght Ishould like to point out that it has to be smallcompared to the wavelenght inside the specimen;this wavelength may be considerably shorterthan that in vacuum. The problem of calculatingthe corrections when this is not the case is ratherdifficult to solve when the susceptibility is aniso-tropic, as in the case for gyromagnetic resonance;so far we have not been able to find a solutionand not even the value of the first coefficients in a

series development in powers of -8

Remarque de M. Shockley. - Dr Roberts hasremarked that values of g as high 2.6 to 3.o havebeen obtained in Mg ferrite powder. It may bepossible that these large values are due in part tomagnetostatic effects which tend to act as anincreased restoring force.

REFERENCES.

[1] GANS R. and LOYARTE R. G. 2014 Ann. Physik, 1921,(4), 64, 209.

[2] DORFMANN J. 2014 Z. Physik, 1923, 17, 98.[3] GRIFFITHS J. H. E. 2014 Nature, 1946, 158, 670.[4] LANDAU L. and LIFSHITZ E. 2014 Physik. Z., Sowjetunion,

1935, 8, 153-169.[5] KITTEL C. - Phys. Rev., 1947, 71, 270.[6] KITTEL C. - Phys. Rev., 1948, 73, 155.[7] YAGER W. A. and BOZORTH R. M. - Phys. Rev., 1947,

72, 80.[8] HEWITT W. H. Jr. 2014 Phys. Rev., 1948, 73, 1118.[9] KITTEL C., YAGER W. A. and MERRITT F. R. - Physica,

1949, 15, 256.[10] BICKFORD L. R. 2014 Phys. Rev., 1950, 78, 449.[11] KIP A. F. and ARNOLD R. D. 2014 Phys. Rev., 1949, 75,

1556.[12] VAN VLECK J. H. - Phys. Rev., 1950, 78, 266.[13] POLDER D. - Phil. Mag., 1949, 40, 99.[14] RICHARDSON J. M. - Phys. Rev., 1949, 75, 1630.[15] LUTTINGER J. M. and KITTEL C. 2014 Helv. Phys. Acta,

1948, 21, 480.[16] KITTEL C. - Phys. Rev., 1949, 76, 743.

1940, 7, 753.[17] GORTER C. J. and KAHN B. 2014 Physica, 1940, 7, 753.[18] VAN VLECK J. H. 2014 Physica, 1949, 15, 197.[19] YAGER W. A. and MERRITT F. R. - Phys. Rev., 1949,

75, 318.

[20] BELJERS H. G. 2014 Physica, 1949, 14, 629.[21] BIRKS J. B. 2014 Proc. Phys. Soc., 1950, 63 b, 65.[22] BLOEMBERGEN N. 2014 Phys. Rev., 1950, 78, 572.[23] YAGER W. A. 2014 Phys. Rev., 1949, 75, 316.[24] KITTEL C. and HERRING C. - Phys. Rev., 1950, 77, 725.[25] KITTEL C. 2014 Rev. Mod. Physics, 1949, 21, 541.[26] YAGER, GALT, MERRITT and WOOD 2014 Phys. Rev., 1950,

80, 744.[27] DRING W. - Z. Naturforschung, 1948, 3, 373.[28] SNOEK J. L. - Physica, 1948, 14, 207. [29] WELCH, NICKS, FAIRWEATHER and ROBERTS. 2014 Phys.

Rev., 1950, 77, 403.[30] NEL L. - Cahiers de Physique, 1944, 25, 1-20.[31] KITTEL C. - Bull. Amer. Phys. Soc., Washington mee-

ting, 1950.[32] AKHEISER A.2014 J. Physics. U. S. S. R., 1946, 10, 217.[33] WILLIAMS SHOCKLEY and KITTEL. 2014 Phys. Rev., 1950,

80, 1090. [34] JOHNSON, RADO and MALCOOF. 2014 Phys. Rev., 1947,

71, 322.[35] BECKER R. - Physik. 8 m Z., 1938, 39, 856.[36] KITTEL C. - Phys. Rev., 1946, 70, 281.[37] RADO G. T., WRIGHT R. W. and EMERSON W. H. 2014

Phys. Rev. 2014 1950, 80, 273.[58] KITTEL C. 2014 Phys. Rev. 1950, 80, 918.