Magnetic domain patterns

HAL Id: jpa-00234388

https://hal.archives-ouvertes.fr/jpa-00234388

Submitted on 1 Jan 1951

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.

Magnetic domain patterns

R.M. Bozorth

To cite this version:

MAGNETIC DOMAIN PATTERNS

By R.

M. BOZORTH,

Bell

_Telephone

_{Laboratories, Murray} Hill

_{(New Jersey).}

Sommaire. 2014 La

technique et l’interprétation des diagrammes de _{poudres magnétiques} est briè-vement _passée en revue, d’un point de vue _historique. Les diagrammes les plus simples observés sont ensuite décrits et _expliqués dans la mesure du possible. Dans la troisième Partie, on décrit et discute

de nouveaux _diagrammes relatifs : a. à un monocristal dont la direction _{(1 11)} est celle de facile aiman-tation ₍₆₀ pour 100 _{Co, 40 pour} 100 _Ni); b. à un monocristal de cobalt; c. à un alliage polycristallin

fer-silicium et d. à un _{alliage polycristallin pour aimants permanents (Alnico 5).}

JOURNAL PHYSIQUE 12,

19~1,

Brief revievsr. ---

For many decades iron

_filings

have been used to

_portray

the directions of lines of

_magnetic

force in air and to detect flaws or

inho-mogeneitics

in

_magnetic

materials. In

₁₉₃

1 it

occurred to von Hamos and Thiessen

_{[ 1]}

to use

magnetic powder

to detect the local

_{inhomogeneities}

irr

_{magnetization}

that the domain

_theory

_predicts.

Independently

Bitter

_{[2] applied}

a

_suspension

of siderac

_{(Fe,O,), having particles}

about 10-4 cm in

diameter,

to a

_{polished magnetized}

surface and

observed under the

_microscope

that the

_powder

formed

_parallel

lines

_{regularly spaced}

about o. mm

apart

and

_{approximately perpendicular}

to the direction of

_{magnetization.}

The

_technique

and

_{interpretation}

of such

_patterns

.was then the

_subject

of

study

of a number of workers

_[3].

The

_preparation

of colloid for these studies has been described in some detail

_by

Elmore

_[4]

who recommends a

_suspension

of

magne-tite,

ground

to colloidal

dimensions,

peptized

with

hydrochloric

acid and

_{protected by}

one _per cent of soap; an

_improvement

on his

_technique

has

_recently

been

_developed

and will be

_published

soon.

Elec-trolytic polishing

[4]

overcomes the

_{objectionable}

mechanical

_polishing

which disturbs the surfaces. A notable advance was made

_by

McKeehan and Elmore

_[5]

who first observed a well-defined

pattern

on a

_{demagnetized single crystal. Figure}

i

shows such a

_{pattern (b)}

and also those

patterns

observed when the

_{magnetization}

is directed

_(a)

into,

or

_(c)

out

_of,

the same

_portion

of the surface as that shown in

_(b).

The

_suspension

used for the

experiments

was a true colloid of

_{Fe,O, particles}

small

_enough

to show Brownian movement, and a

change

in

_{magnetization}

of the

_{magnetic specimen}

was

_accompanied

_by

a movement of the lines

immediately

visible to the _eye.

More recent

_work,

_reported

in various articles

_by

Williams,

Bozorth and

_{Shockley [6],}

has made visible for the first time the domain boundaries characteristic of unstrained

iron,

and has

_improved

considerably

our

_knowledge

of the processes of

magnetization. They

used

_{single crystals}

contai-ning

3.8

_weight

_per cent silicon and

_having

surfaces cut

_{nearly parallel}

to

_{crystallographic}

_planes.

The

specimens

were annealed and

_polished

_carefully,

first

_mechanically

and then

_{electrolytically.}

After mechanical

_polishing

the

_powder

_pattern

on a surface almost

_parallel

to

₍₁₀₀₎

is the cc maze »

pattern

of

_{figure 2 (a),}

similar to that of

_figure

i.

Fig. I. - n Maze » _pattern observed on _polished surface of

single crystal of _{iron; (b) demagnetized, (a)} and (c)

magne-tized in _opposite directions.

After

_{electrolytically polishing}

and

_reapplying

the

powder

to the same area the result is the « _tree »

pattern

of

_figure

2

(b).

It is evident from this

and other

_experiments

that the maze

_pattern

is characteristic of a strained surface and that the tree

_pattern

shows the domain boundaries of strain-free material.

The directions of

_{magnetization}

in the domains can be determined in _{several ways,}

_{using techniques}

described in the

_original

_{paper. The result for} a

portion

of one tree

_pattern

is shown in

_figure

2

(c).

The local

_{magnetization}

in

_unmagnetized

material

is

_{always parallel}

to one of the

_crystal

axes, and the

boundaries

_separate

domains

_magnetized

at _goo or at 180° to each other.

Fig. 2. - Maze and a tree » _{patterns, (a)} and _(b), observed on the same _portion of a _{single crystal} after mechanical

and eliectrolytic polishing, respectively. Directions of _{magnetization} in the tree _pattern are shown at _(c).

Fig. 3. - Effect of increasing tension (a) to (d), _on the tree

pattern. In _(fi tension has been released.

Visible

movement of domain

boundaries

takes

place

upon

application

of field or stress. The

effect of uniform tension is shown in

_figure

3. In this material tension increases the

_{magnetization}

in the direction of the

tension,

and the mechanism

by

which this is

_accomplished

is here

_{apparent :}

domains oriented

_parallel

to the axis of tension are

enlarged

by displacement

of domain

_boundaries,

at _{the expense}

of

domains oriented at

_right

_angles,

so that the latter domains

_disappear

almost

_entirely

when the tension is

_{sufficiently large.}

With release of tension the

_original

kind of tree

_pattern

forms,

cause the return of the

_powder

lines to the same

_places

after

_they

have been disturbed

_{temporarily by}

field or uniform stress.

When the surfaces are not

_parallel

or

_nearly

parallel

to

_{simple crystallographic}

_planes,

the

patterns

are

_likely

to be more

_{complicated. Figure}

4

Fig. 4. -

Complicated patterns observed on _{(1 10)} and (i I i) planes.

Fig. 5. -° Patterns on cobalt surfaces cut parallel and perpendicular to the _hexagonal axis.

shows two

_examples

of such

_patterns.

_Although

the

_simple

_patterns

are well

_understood,

it has not

yet

been

_possible

to understand in detail the more elaborate ones. It is

_{believed, however,}

that the basic

_principles

that

_apply

to the

_simple

ones are also

_applicable

to the more

_complex

ones. These

principles

are discussed below.

Experiments

on cobalt have also been instructive.

Bitter

_[7]

observed two

_types

of

_patterns

on

poly-crystalline

material and Elmore

_{[8], working}

with

single crystals,

found the

_hexagonal

lace-like

_patterns

on surfaces

_parallel

to the

_hexagonal

_{planes (o01 )}

perpendicular

to the

_crystal

axis,

and the

_straight

line

_patterns

on

_{prism planes,}

as shown in the

properties

of

_cobalt,

known to have a direction of easy

magnetization parallel

to the

crystal

axis. The domains are then

_expected

to be

_long

in the

direction of the axis and

_{packed together}

like a bundle of needles

_{(or sheets).}

The boundaries of such domains thus

_correspond

to the

_patterns.

Moving pictures

of the

_patterns

taken with

_slowly

changing

field

_strength

show sudden

_{displacements}

of the boundaries

_{corresponding}

to

_jumps

much

larger

than those

_usually

attributed to the Bark-hausen effect.

Germer

_[9]

has measured the

_strength

of

_magnetic

fields close to the surface of an

_unmagnetized

cobalt

_crystal

and found that near a

_hexagonal

face it is of the order of 104 Oe and falls off with distance from the surface so that it is

_relatively

weak at mm. The fields near

_prism

faces

are weaker and fall off more

slowly

with

distance,

in the way that

they

would

_expect

if the domains are needlelike as assumed.

The domain structure around cavities and inclu-sions was

_{investigated theoretically by}

Néel

_[10]

before _any direct observations were made. Obser-vations of a number of

_crystal

surfaces under the

microscope

showed the presence of an occasional hole that had formed

_{accidentally during}

_freezing

or

_etching

or

_polishing

of the

_crystal.

The

_patterns

around two holes in

₍₁₀₀₎

surfaces are shown in

_(a)

and

_(b)

of

_figure

6. The structure observed was almost identical with that

_predicted

_by

_N6el,

and can be

_interpreted

with the

_help

of

_figure

6

_(c).

Briefly,

the _energy is lowered

_by

the formation of

«

spikes o

which

_help

the

magnetic poles

to

_spread

out over a

_larger

area.

Fig. 6. - N6el spikes a around holes in a _{crystal surface,} and their _{interpretation.}

The

_{interpretation}

of the various structures can

be carried out in terms of

_energies

of associated

with domain

walls,

_magnetic

_{poles (magnetostatic}

energy), crystal

anisotropy,

strain and the inter-action of the

_{magnetization}

with the _{field if any} be

_present.

The

_theory

has been summarized

recently

by

Kittel

_[11].

In the next section the

simpler

types

of structure will be discussed on this basis. The

_following

section describes the

appli-cation of the

_powder

_pattern

_technique

to various

problems,

and the new conclusions that can be drawn from the various

_experiments

will be

_pointed

out.

Interpretation

of

_simple

_patterns.

-- The

simple

domain

_patterns

that were first understood may be listed as follows :

Plate

_pattern,

_{(100) plane;}

Tree

_pattern

on surfaces

_nearly

_parallel

to

_(1oo);

Neel "

_{spikes "}

around

cavities;

Line

_pattern

on cobalt

_parallel

to

_{[oo .}

r

] axis.

Others,

interpreted

more

_recently,

are referred to in the third

_part

of this _paper.

Plate Pattern. - A

typical

«

plate»

_pattern,

with domains of

closure,

is shown in

_figure

7. This is a stable

_{configuration}

in zero

_applied

field,

for reasons illustrated in

_figure

8. In the

_possible

single

domain

_(a),

the

_poles

at the end

_give

rise to

magnetostatic

energy

equal

to In

_(b)

this is reduced

_{considerably by}

the introduction of the

wall,

with which is associated some _energy. In

_(c)

the

_poles

are eliminated

_completely

but there is some strain energy associated with the magne-tostriction of the domains of

closure,

the material

of which does not fit into _{the space it would occupy} if

_{unmagnetized.}

In

_(d)

the _energy of

magnetostric-tive strain is reduced

further,

and the wall _energy is increased.

The various

_energies

may be evaluated and the

Fig. 7. -- ~ _Plate » _{pattern with domains of closure,} on the (100) surface of an iron-silicon _crystal.

Fig. 8. - Theoretical domain

structures illustrating the

energies associated with _{magnetic poles,} domain walls and

magnetostriction.

strain

_energy

is

ergs per cm8 of the volume of the domains of closure.

Young’s

modulus

_[12]

in the direction

is x 1012 ergs :

cm2,

the saturation

magnetostriction [13]

in this direction is

and the _{domain wall energy about} i _{erg per cm2}

of wall. For one cm3 of a

_crystal

slab

_composed

of domains L cm

_long

and D cm wide the volume of domains of closure is

D

and the wall

( 2 L ) area is

The sum of the

_energies

is then

per cm3 of

crystal,

which is a minimum for

or 0.08 for silicon iron. In the

_crystal

of

_figure

7, L = 0.22 cm and the calculated and observed values of D are o.03 and

o.o5cm,

respectively.

The

agreement

is

_good,

_considering

the

_{approximations}

made.

When a field is

_applied

the thickness of the domains diminishes. The

_theory

has been worked out

_by

Néel

_[14],

and confirmed

_by

the data of Bates and

Neale [15J.

This will not be considered further here. Tree Pattern. - This _was the first of the

complicated

patterns

shown to be in accord with our

_present

ideas of domain

theory

[6].

This

_pattern

is observed when the surface of an iron

_crystal

(positive anisotropy constant)

is

_slightly

inclined

(at

angle 0)

to the

_{(100) crystallographic planes,}

and its

_explanation

can be understood

_{qualitatively}

by

reference to

_figure

_9. In the

_plate-like

domains

Fig, 9. - Explanation of the tree pattern.

that _compose most of the

_crystal

the

_{magnetization}

is

_parallel

to the

_{crystal planes,}

and therefore the

lines of

_magnetic

flux will cut the surface at an

angle

and

_produce

a

_density

of

_{magnetic poles}

of ±

1s

sin 0 on alternate

_strips

of width W. There will then be a

_{magnetostatic}

_energy

_proportional

to W _{sin2 0 per unit} area. This energy is reduced

by

formation of the tree

_patterns

even

_though

the wall energy is added. The branches

_transport

flux accross the « trunks » of the treees and as

they taper

this flux is distributed as

_magnetic

poles

over the domain wall

_separating

the branch from the

_underlying

domain. A minimum _energy

theory

has been worked out

_[6],

and this

_explains

the dimensions of the branches and their variation with 0 within a factor of about 2. When the

angle

0 becomes

_larger

the branches lie closer

together,

as shown in

_flgure

I o, and then

overlap

so that the trunk of the tree becomes

_completely

hidden.

«

Spikes

». - N6el’s theoretical

investi-gations

of the domain structure around cavities

and

_conclusions,

as well as the

_previous

work of Kersten

_{[16], prompted}

an

_{investigation}

of the

powder

patterns

in areas where visible cavities occured in

_crystal

surfaces. Two

_patterns

observed in

_(IOO)

surfaces have

_already

been shown in

_figure

6 and

_they

have the form

_predicted

_by

Néel on

_purely

theoretical

_grounds.

The

_energetics

of this kind of

_pattern,

as

_already

reported [6]

is as follows.

The

_{magnetostatic (or}

_{demagnetization)}

_energy

associated with a hole

_[see

₍₁₎

of

_{f g.}

6

_(c)]

around

which there is no domain _structure, is

No

being

the

_{demagnetizing}

factor,

"Yo

the volume of the hole and

Is

the saturation

_{magnetization}

of the material around the hole. If domains are formed as in

_{(2), poles}

are not

_present

at the

_edges

of the

_cavity

but are distributed

_along

the domain boundaries as indicated. In this case there will be

Fig. I o. -

Dependence of the tree _pattern on the angle between

the surface and the _{(1oo) planes.}

energies

associated both with the

_{demagnetization}

(a

volume

_effect)

and with the domain walls

_(area

effect) :

N

_being

the

_{demagnetizing}

factor of the volume V enclosed

_by

the domain

boundaries,

the _energy per unit area of wall and A the total wall area. Assume a

_spherical

hole of diameter d and a domain

of this

_ellipse

is

and the volume and surface are

_easily

calculable. The

_{magnetostatic}

_{energy is} then

a factor

of ’

_being

included to take account of

25 °

the

_permeability

of the domains

themselves,

and the wall energy is

Using

the

_appropriate

numerical values d=o.oo1 cm,

Is

=

_1580,

ew = 1.5

_{ergs :}

cm2,

the calculated

value of I for which

Ed

-~- is a minimum is o.1 o cm and the ratio

of j is

then

_{approximately}

100. The

observed

ratio

for the domain of

_figure

6

_{b)

is about 50, smaller

_by

a factor of 2. This is a

satis-factory

agreement

in view of the

_{simplifications}

used.

Moving

Boundaries. -

Simple

geometry

has been observed in at least two

_experiments

on the move-ment of domain boundaries

_accompanying

a

_change

of

_{magnetization}

with

changing

applied

field.

Williams and

_Shockley

_[6]

observed a

_simple

struc-ture in a hollow

_rectangle

with sides

parallel

to

[100]

directions,

and

_they

have described

_[17]

some

experiments

on the movement of such a

_boundary

with slow and

_{rapid changes}

in

_magnetic

field.

Fig. I I. - Effect of Néel «"spikes » on the movement of a Bloch wall.

Spikes

such as those

just

described have an

impor-tant effect on the movement of

large

domain

walls,

as noted

by

Williams and

Shockley [ 6]

and illustrated

Fig. 2. - Movement of a well _{during magnetization parallel} to a _{[on] ]} direction in a _{(100) plane.}

is consumed in

_forming

the additional wall. This

mechanism seems to be

_important

in the

inter-pretation

of

_hysteresis

loss and coercive force and is discussed

_by

_Shockley

and Williams in an

accom-panying

paper.

A

_{moving boundary composed}

of a number of

segments

of

_straight

lines is observed in a

₍₁₀₀₎

plane

when

_{magnetized parallel}

to a

_[001]

direction

(fig. 12).

The boundaries that move are between domains

_{magnetized antiparallel}

to each other

( 1 8 0 ° wall).

Patterns. - Cobalt-nickel

crystal.

- The

geometry

of

_patterns

on the materials of cubic

symmetry

already

examined is

_closely

connected with the fact that their directions of easy

magnetiza-tion are

_[100].

Heretofore no

_simple

_patterns

have been

_reported

on

_crystals,

like

nickel,

in which the directions of easy

magnetization

are

_[111].

It seemed

_probable

that the failure to observe

_patterns

in such

_crystals

was connected with the fact their

_{crystal anisotropy}

was too small

_(for

nickel the

_anisotropy

constant K is 60 ooo as

_compared

Fig. 14. - Effect of

magnetization on the pattern on the cobalt-nickel _crystal.

with 280 ooo for iron

containing

several per cent

of

_{silicon). Consequently}

a

_crystal

of a cobalt-nickel

_{alloy containing}

60 _per cent cobalt and

_having

a constant of about 200 ooo

_according

to Shih

_[18],

has been

_{prepared by}

slow

_cooling

of the melt as described

_by

Walker,

Williams and Bozorth

_[19].

The surface of the

_crystal

was cut

_parallel

to a

(0 1 1)

plane

so that

₄

of the 8 directions of _easy magne-tization were

_parallel

to the surface. The

_pattern

and its

_{interpretation}

are shown in

_figure

13.

All of the

_{theoretically expected angles}

between

adjacent

domains -

180°,

109°, 71°

-

are observed. It is also noticed that the domain structure is smaller than that observed on iron-silicon

_crystals.

This difference in _{size may} be due to the more

compli-cated

_pattern

that _may be

_expected

in a structure

having

8 instead of 6 directions of _easy

magnetiza-tion and

_having

no

_goo

boundaries.

Fig. I5. --- Patterns _on the (roo) surface of cobalt; a. positive

field, b.

zerolfield, c. _{negative field, parallel} to the axis.

Fig. 16. - Pattern

on cobalt _{(100) surface,} with small field applied normal to the surface.

structure,

_originally

of

_complicated

_geometry,

resolves itself in

_high

fields to a series of lines at

_{right angles}

to the direction of the

_applied

field,

the direction of

_{magnetization}

_alternating

in

_adjacent

domains between the two

_[111]

_]

directions oriented most

nearly

to the direction of the field. The boundaries

are between domains in which the directions of

magnetization

differ

_by

_{i ogo.}

Cobalt. -

are

_reproduced

here

₁₅₎

because

_interesting

regularities

are observed in the

_pattern.

The

specimen,

cut with a surface

_parallel

to a

_{(I 00)}

_plane,

is

_magnetized

_parallel

to the surface in the direction of the

_hexagonal

axis

_[001].

In

_figure

15, b is

unmagnetized

while a and c are

_magnetized

in

opposite

directions. The

_{displacements}

of alternate lines in

_opposite

directions shows that the boun-daries move so that more material is

_magnetized

parallel

to the field and less

_antiparallel

to it. Careful

_comparison

of a and c shows also that the lines which move

_upwards

in one move down-wards in the other so that

_{imperfections}

that _appear in a thinner domain in a are in a thicker domain in c.

In

_figure

16 the contrast between

_neighboring

domains is enhanced

_by

_applying

a small normal field. The

_specimen

surface is

_slightly

inclined to the

_hexagonal

axis, so the

_{magnetic poles}

on the

surface are

_alternately

north and south

_poles

in successive domains. The verticle field enhances the

_pole

_strength

on half of the domains and

neu-tralizes the

_poles

on the other domains, so that the colloid is attracted in the one kind and not in the other.

Polycrystalline

material. - In _some commercial

silicon-iron sheet material used for transformer cores the

_separate

_crystals

are

_aligned

with their

_{[ 100]}

axes

_{approximately parallel}

to the

_long

dimension of the sheet. A

_{powder photograph}

of such material

1 ~),

taken

_by

H. J.

Williams,

shows that domain

Fig. 17. - Pattern _on

polycristalline iron-silicon alloy, showing that domains sometimes cross _grain boundaries.

boundaries often cross

_crystal

boundaries. This occurs when the

[100]

directions in

adjacent crystals

are almost

parallel,

and one believes that the

_crystals

must also be

_{nearly aligned}

_{in 3-dimensional space} so that the

_platelike

domains of the two

_crystals

will

_join

_together

_{along planes}

of contact that _go below the surface observed. The

_alignment

in

some of the

_crystals

is

_obviously

too _{poor for domains}

to cross

_crystal

boundaries.

Alnicos. - In a recent

_study

of the mechanism in Alnico 5

_{(5 1}

per cent

Fe, 2l~

per cent

Co, 14

per cent Ni 8 _per cent

_Al,

_{3 per cent}

_Cu)

heat treated in a

magnetic

field,

Nesbitt has observed

_{powder pattern}

that

_help

in

_{understanding}

the nature of this maetrial. In a series of

_experiments

a

_specimen

was heated

to 1300° C and cooled at the « normal » rate

of 20 C : s to 8ooo C and then

_quenched

in oil. A

magnetic

field was

_present

from

_gooo

C to room

temperature.

Examination

_by

the

_{powder technique}

shows the existence of

_long

domains,

about 0.02 mm in

width,

_{aligned approximately parallel}

to the field

_present

_during

the heat treatment.

_They

cut across

_crystal

boundaries with no substantial

change

in direction

_{(see fig.}

₁₈₎

and show that the

magnetization

is

_{everywhere parallel}

or

_antiparallel

to the field used

_during

heat treatment. In this material the

_{magnetization}

is

_obviously

not

_parallel

to the _easy

_{crystallographic}

direction nearest to the

Fig. 18. - Pattern

on Alnico V : domains _{crossing crystal} boundaries.

Fig. 19. - Movement of domain boundaries during magnetization of Alnico V.

When a field is

applied

for measurement

parallel

to the

_{heat-treating}

field,

domain boundaries are observed to move

_Ig).

The

_{hysteresis loop}

Fig. 20. - Rotation of

magnetization of domains in Alnico V when magnetized at right angles to field present during heat treatment.

field is

_applied

at

_{right angles}

to the

_{heat-treating}

field the

_powder

_patterns

show domain rotation without

_{boundary displacement (see}

_{f g. 20).}

This is in

_agreement

with the form of the

_{magnetization}

curve, which rises almost

_linearly

to saturation at about

4oo

to

₄₅₀

Oe.

These

_experiments

show that

_large

domains

having

a

_{strongly preferred}

orientation exist when the material is

_prepared

in the manner described. When the material is cooled at the normal rate to

temperatures

lower than 8ooo C the domains are much

smaller,

and when the material is

_aged

at 6ooo C no domain has so far been detected in the

unmagnetized

condition.

_Recently

Nesbitt has observed their formation upon the

application

of a field to material

_given

the usual Alnico 5

treat-ment-cooling

in a field and

aging.

The coercive force of the

_{quenched specimen having large}

domains is 15

Oe,

that of the

_{specimen having}

smaller domains is 35o, while that of the material

_given

the usual Alnico 5 treatment is 600.

The manner in which the

_preferred

direction of

magnetization

is fixed in the structure is not

_yet

known,

but it is

_possible

that it is connected with

the atomic

_ordering

that exists in these

_alloys,

as shown

_{by X-rays.}

I am indebted to Messrs H. J. Williams and E. A. Nesbitt for several of the

_photographs

repro-duced in this

article,

and to Mr J. C. Walker for assistance with the

_experimental

work,

especially

in the

_preparation

of the

_{single crystals}

of the cobalt-nickel

_alloy.

Remarque

de M. Bates. - I

congratulate

Dr Bozorth on his beautiful

_pictures

on alnico V. I

myself

have tried on and off _{for four years} to obtain

tem, but without success. Was the

_crystal

sur-face

_prepared

in a

_special

_way or was the

_{dry powder}

technique

used ?

Question

de M.

_Epetboin.

- Le monocristal de

cobalt at-t-il subi au

_pr6alable

un

_polissage

elec-trolytique ?

Au

_sujet

des difficult6s rencontr6es par M. le Prof.

Bates,

je

me souvients que dans notre laboratoire M. Amine a obtenu deux

_aspects

diff6-rents de la surface

_polie

d’un

_alnico,

_{suivant que l’on} utilisait un bain a base d’acide

_phosphorique

à chaud

_{(8oo C)}

ou d’acide

_{perchlorique-anhydride}

Réponse

de M. Bates. -- We have done

some

experiments

on

_electrolytic

_polishing

and have devised a

_bridge

circuit for

_controling

it. This is to be described in a

forthcoming

_{paper in Journ.} Scient. Instr.

Remarque

de M. Sucksmith. - I would like to

obtain a three dimensional

_picture

of a domain in

cobalt,

which should be

_simple

on account of the

_single

easy direction of the

magnetization.

What is the reason for

_assuming

that

_they

are « needles » or « sheets » since the

right

hand

picture

of the

_figure

5

_suggests

that

_they

are rods whose cross section is that of a

_hexagonal

cluster

Remarque

de M.

_Shockley.

-- It

may be

appro-priate

to

_point

out that in

_figure

₁₇

_{presented by}

Dr _{Bozorth there appear domains of} closure which can be understood in terms of the

_relationship

between the

_{crystal grain}

boundary

and

_crystal

axes. The

_diagram

pre-sented here

_{(fig. 21 )}

repre-sents a

_{grain boundary}

ABC which is

_symmetrical

between the two

_crystals

for the

_segment

BC. For Fig. 21. - Domain structure the

segment

BC,

domains

near a grain boundary. can form

along

easy di-rections with no free

poles generated

on the

boundary.

Between

_points

A and

B, however,

there will be

_{magnetic poles.}

The

_{magnetostatic}

energy of these

poles

may be reduced

by

introducing

reentrant

_spike

domain which reduce the

_density

on the

_{grain boundary.}

The effect of one such

spike

is shown in the

_diagram;

in

_{figure 17}

of Bozorth’s

aticle,

a number of such

spikes

are

_present.

The

_generation

of such

_spikes

are

_frequently

observed to be discontinuous and

_consequently

irreversible. This

_suggests

that in

_{polycristalline}

material an

appreciable

contribution to the coercive force and

hysteresis

may be made

by

discontinuities in the

pattern

of domain

_walls,

a _process

_quite

similar to that discussed

_by

N6el.

A une

_question

de M.

Guillaud,

M. Bates

_répond.

-The field

_{applied perpendicular}

to the surface was not measured

_exactly;

but it was of the order of 13o Oe.

Remarque

de M. Hoselitz. - The

powder

patterns

shown

_by

Dr Bozorth to occur on Alnico V when

quenched

from 8ooo C can be

_explained

without any additional

assumptions

from our views

(HOSELITZ

and

MCCAIG,

Nature,

1949,

164,

p.

581;

Proc.

_Phys.

Soc.,

I go9,

B,

62,

p.

163). Magnetostriction

and other measurements on this material in the

_fully

heat treated condition have shown that the

magneti-sation _energy can be

_{represented by}

a cubic term and a uniaxial term

_{probably acting}

in one of the

_[100]

directions. The order of

_magnitude

of these _energy are _up to I o _{ergs :} cc. In the

_quenched

material described

_by

Dr

Bozorth,

these _{values may} be

_considerably

smaller and I cannot comment on this

_particular

case.

However,

as mentioned

_by

Dr

_Bozorth,

similar

powder

patterns

are observed

_by

Nesbitt in the

fully

heat treated

material,

but

_only

in

_relatively

high

fields.

If a field of about 5oo Oe is

_applied

_{in the} pre-ferred direction of

_{magnetisation,}

the field _energy term

HJS

becomes of the same order of

_magnitude

as the

magnetic

anisotropy

term estimated in our

experiments,

and it is

_consequently

_likely

that at fields

of this

_magnitude,

the actual field direction will become the direction of the

_{magnetisation.}

Hence it can be understood that no unidirectional domains are observed

_{by powder}

_patterns

until fields of about the coercive force are

_applied,

when most domains will be very

hearly aligned

with the field direction and

_consequently

_{many domains boundaries} will have

_disappeared.

Thus

_large

unidirectional domains boundaries will exist in this condition.

Demande de M. Bauer. - Serait-il

possible

de faire ces

_exp6riences

_{pour des corps} dont le

_point

de Curie est voisin de la

_temperature

ordinaire et de voir comment les domaines

_changent

avec la

_{temperature ?}

Reponse n6gative

de MM. Bozorth et Bates. REFERENCES.

[1] HAMOS L. v. and THIESSEN P. A. - Z. Physik, 1931,

71, 442.

[2] BITTER F. 2014 Phys. Rev., 1931, 38, 1903.

[3] KAYA S.2014 Z. Physik, 1934, 89, 796, and _{1934, 90,} 551.

- SIXTUS K. J. -

Phys. Rev., 1937, 51, 870. -SOLLER T. 2014 Z. Physik, 1937, 106, 485. - ELMORE W. C. and Mc KEEHAN L. W. - Tr. Amer. Inst. Min. Met. Engrs, 1936, 120, 236.

[4] ELMORE W. C. 2014 Phys.

Rev., 1938, 54, 1092, and 1942, 62, 486.

[5] Mc KEEHAN L. W. and ELMORE W. C. - Phys. Rev., 1934, 46, 226.

[6] WILLIAMS H. J. -

Phys. Rev., 1946, 70, 106 and

subse-quent brief _reports. - _{WILLIAMS H.}

J., BOZORTH

R. M. and SHOCKLEY W.- _Phys. Rev., 1949, 75, 155. -WILLIAMS H. J. and SHOCKLEY W. - _Phys. Rev., 1949, 75, 178.

[7] BITTER F. 2014 Phys. Rev., 1932, 41, 507.

[8] ELMORE W. C. 2014 Phys. Rev., 1938, 53, 757.

[9]

GERMER L. H. 2014 _Phys. Rev., 1942, 62, 295. [10] NÉEL L. - Cahiers de Physique, 1944, 25, _21.

[11]

KITTEL C. 2014 Rev. Mod.

Physics, 1949, 21, 541.

[12]

BENFORD F. G. - Metal Progr., 1946, 46, 94.

[13] CARR W. J. - Private Communication.

[14] NÉEL L. - J. Physique Rad., 1944, (8), 5, 241.

[15] BATES L. F. and NEALE F. E. -

Physica, 1949, 15, 220. [16] KERSTEN M. - Physik. Z., 1943, 44, 63.

[17] WILLIAMS H. J. and SHOCKLEY W. - _{Phys. Rev.,} 1950, 80, ooo.

[18] SHIH J. W. - Phys. Rev., 1936, 50, 376.

[19] WALKER J. G., WILLIAMS H. J. and BOZORTH R. M. 2014

Rev. Sci. Instruments, 1949, 20, 947. [20] HOSFLITZ K. and Mc CAIG M. -