Magnetic saturation intensity and some other related measurements

HAL Id: jpa-00234402

https://hal.archives-ouvertes.fr/jpa-00234402

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 saturation intensity and some other related

measurements

W. Sucksmith

To cite this version:

MAGNETIC SATURATION INTENSITY AND SOME OTHER RELATED MEASUREMENTS

By

W. SUCKSMITH. Sommaire. - Dans la

première Partie, l’auteur attire l’attention sur le _{fait que des} mesures

exactes de l’aimantation spontanée n’ont été effectuées que dans un très petit nombre de cas. Les résultats pour le _{cobalt, qui} possède un réseau _{hexagonal jusqu’au point} de Curie et un réseau _cubique centré au-dessus, sont donnés et discutés. Une _augmentation de 1,5 pour 100 se _produit à la tempé-rature de transition.

Dans la deuxième Partie, la variation de l’aimantation à saturation et le _degré de l’ordre dans

quelques alliages binaires sont donnés. Dans la dernière Partie, on montre l’importance des mesures

de l’aimantation à saturation pour étudier la détermination de l’équilibre des phases. L’auteur cite

l’application heureuse de cette méthode au cas du système fer-carbone, caractérisé par quelques phases qui ne _{figurent pas à l’état d’équilibre.}

LE JOURNAL DE PHYSIQUE ET LE RADIUM. 12,

1~~~?

I. Saturation

_Intensity

Measurements and

the Iaaur of

_{Corresponding}

States. - Of the three

_{ferromagnetic}

elements,

the variation of the

spontaneous

magnetization

with

_temperature,

has

been most

_thoroughly

_investigated

in the case of

nickel,

the most exhaustive and accurate

measu-rements

_having

been carried out

_by

Weiss and Forrer

_[1].

For

_iron,

the lower

_temperatures

were

covered

_by

the work of the same two authors

_[21

whilst the

_{investigation by}

Potter

_[3]

deals with the range from room

_temperature

to the Curie

point.

Cobalt,

which was

_{investigated by}

Bloch

_[4]

presents

additional

difficulties,

chiefly

on account

of the

_large

_crystalline

_anisotropy.

Bloch’s

mea-surements in fields _{up to} about i3ooo Oe have been hitherto

_regarded

as

_representing

the

tempe-rature variation of

_spontaneous

_{magnetisation,}

a conclusion which must be

_regarded

as _open to

considerable criticism on the

following

_grounds.

In the first

_place

the

_{abnormally high crystalline}

anisotropy

makes the

_approach

to saturation of the

polycrystalline

material _{very slow.} This rate of

approach

decreases as the

_temperature

is lowered

owing

to the increased

_difficulty

of

_attaining

satu-ration in the basal

_plane,

i.e. at -

i goo C only

80 per 100 saturation is reached in an effective field of io ooo Oe. In addition there is the

diffi-culty

that at

_{ordinary temperatures polycrystalline}

cobalt exists as a mixture of face-centred cubic and close

_{packed hexagonal}

structures, the

pro-portions

of which vary

considerably according

to

the method of

_preparation

and is

_subsequent

thermal and mechanical treatment and is not

_fully

control-lable.

Further,

there is a

_phase

transition

_beginning

at about

_’700

C,

above which the metal is

_wholly

face-centred cubic.

_Lastly,

it is not

_impossible

that the wide variations in the recorded

_density

of cobalt exercise some

_effect,

if

_only

in

limi-ting

accuracy of measurement. The

only

method of

_securing

more

_precise

data on the variation of

_spontaneous

_{magnetisation}

with

_temperature

appears to be to use

_{single crystals}

of the

_metal,

and to make measurements

_along

_{the easy}

direc-tion of

_{magnetisation}

_up to the transition

tempe-rature. Above the _{transition range,} the

_wholly

face-centred metal has

_anisotropy

constants similar

to the other cubic

_{ferromagnetics.}

Myers

and Sucksmith have

_just

_completed

this

investigation

with the

_following

results.

_By

slow

cooling

of the molten metal

_(99.9

_per Ioo

purity)

through

the

_{melting point}

and the solid metal

thus

_{produced through}

the transition _range,

_single

crystals

of

_hexagonal

cobalt were

_produced.

The variation of the saturation

_{intensity along}

its easy direction of

_{magnetisation}

shows the usual trend until a

_temperature

of

4ooo

C is reached when the

phase

change

takes

_{place. Up}

to this

_temperature

quantitative comparison

of the results with other

ferromagnetic

materials is not

_possible,

other than

on the somewhat

_{specious assumption}

of a Curie

temperature

for the

_hexagonal

material. On the

other

hand,

because of the close

_agreement

in

_part

of _{this range between} our results and those of Guillaud

_{(see below),}

it is not without interest to

extrapolate

such values as we have. The reduced

magnetisation

decreases from

_unity

to a value

of

_o.942

at

₆₇₃₀

abs. If we

_extrapolate

these

values

_{according to j}

=

2 ,

the Curie

_temperature

2

is about 1150

abs.,

_considerably

lower than the

Curie

_temperature

for the cubic cobalt of

1404°

abs.

Assuming

a behaviour identical with that of nickel

431

(see

below),

leads to a Curie

_temperature

at

about

1420°

abs. Whilst this close

_agreement

is

_wholly

fortuitous,

it can be taken to

_point

to

the conclusion that the curve for

_hexagonal

cobalt lies below

_{the j}

_curve

and close to that for the face centred nickel.

At the transition

_temperature

the

_{magnetisation}

rises

_{discontinuously}

at the transition

_by

about

1,5

per ioo

(see

_{f g.}

_{I ),}

and

_thereupon

Fig. I.

follows a trend _up to the Curie

_temperature

identical

with that of nickel when the reduced

magneti-sation

G

is

_plotted

_against

the reduced

tempe-0

rature ( n )’

This means that in this upper

tempe-0

rature

_region

the values are

_greater

than for the theoretical curve

_{for j}

= §.

_Subsequent

_heating

and

_{cooling through}

the transition zone for both the

single

crystal specimens

and the

_{polycrystalline}

material showed similar behaviour of a rather remarkable nature. With this thermal treatment

the

_intensity

both above and below the transition increased

_by

small amounts _{which decreased}

pro-gressively

until the cobalt was 11 stabilized " after about 10

_repetitions

of

_the

thermal

_cycle.

The

total increase amounts to a

_change

at 3000

_(i.

e. below the transition

_region)

of about u =

0.4.

The

only possible

explanation

of this minor

_anomaly

appears to be to

_postulate

11

lattice mistakes "

in the structure.

The value of 6 for the

_{hexagonal phase}

is found

to be 162.5 with pn =

1.72, and a Curie

tempe-rature of i 13 1

± 3° C,

the

density

at room

tempe-rature

_being

_p =

8.84!t

± o.oo5 in all cases. These

results are in fair

_agreement

with the recent

measu-rements on a

_{single crystal}

of cobalt of Guillaud and Roux

_[5],

who obtained from measurements at and below room

_temperature

c -=

_16o.9

with

1.70. Their value for the ratio of the

intensity

at room

_temperature

to that at absolute zero is about 1.005 and _{compares well} with ours of

The best value for 7 for the cubic cobalt we estimate

to be

16~1.5

(p~ _ ~

this

_being

obtained

_by

extrapolation

from above the _{transition range} and

consequently higher

than values obtained

_by

others.

Comparison

of all the

_{ferromagnetic}

elements

with

_{simple theory gives}

the best accord

_{for j}

_{== ~}

7

with minor deviations in

_general

common to all

threee

elements,

in that the measured

_spontaneous

magnetisation

lies

_slightly

below the theoretical

T

curves for small values of

(j’

, but at

temperatures

approaching

the Curie

_temperature,

_experimental

results exceed those demanded

_{by theory.}

If one considers the

application

to

_alloys,

in

particular

intermetallic

_compounds,

the work of Guillaud

_[6]

has added

_considerably

to the

expe-rimental data available. For the three

_compounds

for which data are available up to the Curie

tem-perature,

two of

these,

MnP and

_{Cr02’ approximate}

more

_{closely to j}

= ~

with

_departures

of the same

type

as those

_experienced

_{in the pure metals. On} the other hand. Cr - Te

_approximates

_closely

to the

_{curve j}

= 3.

The inherent

_difficulty

in the utilization of all these measurements of saturation

_intensity

lies in the determination of the

_spontaneous

_{magnetisation.}

At low

_{temperatures,}

the law of

_approach

is of

secondary

importance

and

_{extrapolation}

to infinite field

_gives

an

_adequate

method in which the error

cannot be

_large.

For

_higher

_{temperatures,}

however,

where the

_change

oaf " saturation "

_{magnetisation}

with

_temperature

is

_greater,

and the

_approach

to

saturation is more

_{gradual, extrapolation}

becomes

increasingly

difficult and uncertain. The

best,

method for

_determining

the

_spontaneous

magne-tisation comes from the

_{magnetocaloric}

effect,

the rise of

_temperature

on

_{magnetisation}

is

_directly

related to the _{variation of the square of} the spon-taneous

_{magnetisation}

with

_temperature,

so that

T

only

for values

_{of 6}

above about

o.6,

where there is a

sufflciently

_{great change}

of cr with

T,

is it pos-sible to obtain measurable results.

Weiss,

however,

has made use of curves

_giving

the field as a function of

_temperature

for constant values _{of the}

magne-tisation. These curves are linear for the

_higher

values of the

field,

and

_{extrapolation}

to zero field

gives

the

_temperature

which

_corresponds

to the

particular

value of 6

employed.

The two methods have been found to a _gree for metal over the limited range of

temperature

where both measurements

are

_possible,

but the _accuracy is far from

_adequate

to enable

_{satisfactory comparison}

with

_theory.

There is therefore a need for considerable extension of

_{magnetocaloric}

measurements on

_{ferromagnetics.}

Accurate measurement of the small

_temperature

which accurate

_knowledge

of the values of the

spontaneous

magnetisation

to be

_{considerably,}

increased. As

_yet,

no

_attempts

have been made to

extend the measurements to any of the

important

alloys

which are the

_subject

of

investigation

to

_day.

In the absence of these

data,

one is

_compelled

to

assume that the saturation

_intensity

is an

_adequate

measure of the

_spontaneous

_{magnetisation,}

this

assumption becoming increasingly

hazardous with the slower

_approach

to saturation characteristic of

_nearly

all

_{ferromagnetic alloys.}

For

_example,

the reduced curves

G,

T)

_{change shape}

markedly

in solid solutions of a

ferromagnetic

and a

non-ferromagnetic

element. It is not

_yet

_definitely

known whether the

_{spontaneous magnetisation}

curve is of the same

_shape,

and more extensive

measurements of the

_{magnetocaloric}

effect as well as

comprehensive

determinations of

.I~

curves over the whole

_temperature

_range are

requi-red. The writer is

_initiating

_experimental

work in the

_hope

of

_extending

the _range of information of this nature.

II. Saturation

_Intensity

Measurements and Order-Disorder. - The

writer,

in

_{investigations}

on the Fe-Ni-Al

_alloy

_system,

has

_previously

noted differences in the variation of saturation

inten-sity

of ordered and disordered

_alloys.

At least three diff erent

_examples

are

_present

in this

_system,

two of which have been the

_subject

of further

inves-tigations

and will be dealt with below. The

third,

which is the

_prototype

of the

_{permanent magnet}

precipitation alloys,

showed saturation

_intensity

data which fit in well with the now

_generally

accep-ted _{view of the processes involved. The}

_{(g, T)}

curve for the annealed

_{alloy (two phase),}

shows

behaviour characteristic of the

_{single phase type}

with a Curie

_temperature

approaching

closely

to _{that of pure iron and} a saturation

_intensity

of about 105

units,

i. e.

_slightly

less than half that

of

iron,

thus

_justifying

the view that the two

_phases

are Fe and the

_nonmagnetic

Ni Al. The

_quenched

material, however,

has a

_higher

saturation

_intensity

than the annealed

_alloy,

and after

_tending

to a lower Curie

_temperature

than pure iron up to

_{4ooo C,}

gradually

shows the _{break up of the}

_{single phase by}

an arrest in the fall of the

_{magnetisation,}

and

_finally

at about 6ooO C reaches the

_equilibrium

_intensity

of the two

_{phase alloy.}

Of other work in this

field,

probably

the most

important

is that on the 5o _per 10o

Fe,

5o _per I oo Co

superlattice.

Ellis and Greiner

_[7]

had shown evidence of the existence of the

_{superlattice by}

X-rays, supported

by

the additional evidence of the thermal arrest method.

Later,

Goldman and Smoluchowski

_[8]

in

_dealing

with

_experimental

and theoretical data on saturation

_{magnetostriction}

and

order,

show that the

_{magnetostriction}

at

satu-ration increases

_by

4o

per 10o _upon

ordering.

In a

later

contribution,

the theoretical

_prediction

that

the saturation moment should be about

4

per 100

higher

than in the random

_alloy

has been confirmed. One

_contributory

factor

_leading

to the

_high

permea-bility

of this

_{alloy (known}

as

Permendur)

_{may be}

due to the establishment of order

_{by appropriate}

heat treatment. Some

_support

of this view is

given by

work on the

_Fe,Al

_{superlattice by}

Ben-nett

_[9].

The

_{(6, T)}

curve for this

_alloy,

both in the

_quenched

and annealed states and for others

near this

_composition,

show the usual

trend,

but the

intensities at low

_temperatures

show notable features. Whilst the

_intensity

of the ordered

_alloys

falls off

smoothly

with increased aluminium content at a

gradually

increasing

rate,

the disordered

_alloy

is more

_magnetic

_up to about 25 _per i oo

Al,

but

_drops

sharply

to a lower value than the ordered for

_higher

aluminium content. This effect

_persists

_up to

about 3 o _{per 100. The} reason for this is

_presumably

due to the formation of

_non-magnetic

Fe Al. In addition to this there is a

_slight

_{discontinuity}

in the

_{intensity-temperature}

curve in the

neighbour-hood of the

_ordering

_{temperature, though}

the Curie

temperatures

of both

_quenched

and annealed

specimens

do not differ

_appreciably.

The effect

on the low field

properties

at room

_temperature

was

investigated,

but the increase in

permeability

due to

_ordering,

e. _g. from about 2 ooo to 10 _000,

was not marked. The

_{coercivity, irrespective}

of the

_degree

of order was less than i Oe in all cases. On the other

hand,

measurements of the

_{(~, H)}

curve at

_higher

_temperatures

show

_quite

different

results,

the

_region

of

_partial

order

_showing

the

most marked effects. Here the

_equilibrium

per-meability

fell from the

_high

values of the order of some thousands to a value of about

4oo

just

below 5ooo C

_{rising again}

with

_increasing

tempe-rature. The

_coercivity

reached a maximum of about 20 Oe in the same

_region.

The difference in behaviour between

_high

and low

_temperature

measurements can be attributed to the time

_required

for the establishement of order. A tentative

suggestion

for the

_higher

coercivities in the transition

range

might

be made on the lines of the

_existing

theories,

by postulating

small

_regions

of order in a metrix of disordered

material,

with the

_probable

accompaniment

of strains in the

_{boundary regions}

between order and disorder.

An

_{investigation}

into the

_magnetic

_properties

of the Fe-Cr

_system

has been made

_by

Ardron

_[10].

Around the

_{composition Fe3Cr}

there is

_satisfactory

evidence of the existence of the

_superlattice

both from

_magnetic

saturation

_intensity

and resistance

measurements.

_Alloys

_containing

20 and _{27 per} 100

chromium have low

_temperature

_{intensities up}

to

2,5

per 10o more for the ordered state than for

433

to a Curie

_temperature

of about

_8oo-gooo

C,

the Curie

temperature being

about

_65~°

C for the 25 per 100

_alloy.

The break down of order is

around

600° C,

so that the ordered

_alloy

does not

reach its true Curie

_temperature.

_The

permea-bilities do not accord with the views

_expressed

above,

the ordered

_{alloys showing}

_{very low}

permea-bilities,

but it has been

_{satisfactorily}

shown that this is due to the

_{precipitation}

of the

_non-magnetic

sigma phase

which is

_responsible

for the

_high

coer-civity

with the

_accompanying

low

_{permeability.}

The writer has

_previously

drawn attention to the

magnetic

evidence for the existence of the super-lattice

_{Ni3Fe [11].}

The small

_dependence

in the

ato-mic numbers of the two constituents makes

_X-ray

evidence

difficult

but the

_magnetic

saturation

intensity

measurements show a similar trend to those

_already

referred to in the case of

Fe3Cr.

Fig. 2. - Variation of (c, T)

for ~5 atomic per 100 Ni, after cooling at 23~ : h.

Wakelin

_{(in progress)}

has

_recently

extended these

observations,

of which the evidence for

the

_Ni3Fe

_superlattice

is shewn in

_figure.

_The spe-cimens were cooled from about 600~ C to

4000 C

at 0.23° _{per hour} to

_produce

order,

after which more

_{rapid cooling}

to room

_temperature

was pos-sible.

_Magnetic

measurements were then taken from low

_temperatures

_upwards.

Above 5000 C the order is

_destroyed

too

_rapidly

for the

_high

Curie

temperature

of the

_superlattice

to be

reached,

and the curve falls

_rapidly

to meet the lower curve for the disordered

_alloy.

Similar measurements

have been carried out on a _{range of}

_{alloys containing}

between

₄₅

and 85 _{atomic per 10oo iron. At both}

these limits no difference in

_magnetic

behaviour for

_cooling

rates between

_quenching

and the above slow rate were

detected,

but at 55 and 80 _per I o0

quite

measurable differences of the

_type

shown

in

_figure

2 were

_observed,

thus

_indicating

some

ordering.

Resistance

_temperature

measurements of the well-known

_shape

confirmed the

_magnetic

observations,

but

_only

for the 25 per i oo Iron

alloy

was it

_possible

to show the existence of the

super-lattice

_by

direct

_X-ray

measurements. The maxi-mum increase of v on

ordering

was about

₄

units in I 15 for

_Ni3Fe,

this

_decreasing

on both sides of this

composition.

The

_extrapolated

Curie

_temperature

fort his ordered

_alloy

is

_evidently

about 2oo-3ooO

higher

than that for the disordered material. On the basis of such evidence as is available at

present,

one _{may conclude that the ordered}

_binary

alloys

have in

_general,

intensities around _per 100

greater,

with Curie

_temperatures

up to 200-300° C

higher

than for the disordered

_alloys.

III. Saturation

_Intensity

Measurements and

Magnetic

Phase

_Analysis.

- At _a conference

at the Institute of

_Physics

in the

_University

of

Strasbourg

in

_1939,

the writer

_presented

a _paper

dealing

with an

_apparatus

for the convenient

measu-rement of saturation intensities at all

_{temperatures,}

irrespective

of the

_shape

of the

_{specimen, together}

with results which had been obtained on a suitable

ternary

alloy

system,

Much

infor-mation can be derived from

_analyses

of this

_type.

Fig. 3.

A recent and useful summary of the measurements

and the information to be derived has been made

by Guillaud

[12].

In

_particular,

he has made a

study

of the

_alloys

of _manganese with

_arsenic,

bismuth,

antimony

and silicon. Two

_parameters,

the variation of saturation

_intensity,

and the

tem-perature

at which

_{ferromagnetism}

_disappears,

can be utilized to

_provide

information. In a solid solution both saturation

_intensity

and Curie

tempe-rature

_usually

decrease with addition of a

non-ferromagnetic

diluent,

initially linearly,

but with

increasing proportions

at a more

_rapid

rate. At a

temperature appearing

if the second

_phase

is

ferro-magnetic.

The saturation intensities at _any

tem-perature

are

_proportional

to the

_products

of the

intensity

per unit mass and the

_intensity

at that

temperature,

thus

_providing

_quantitative

infor-mation.

A detailed

_example

of this

_procedure

has been

given

in the

_{magnetic analysis}

of the Iron-Silicon

System by Guggenheimer,

Heitler and Hoselitz

_[13].

In the

_{non-equilibrium}

condition,

no less than four

phases,

three

_magnetic

and one

_{non-magnetic,}

have

_{been’satisfactorily}

isolated

_{(see f g. 3).}

In

_addition,

however, it has been found

_possible

to follow out rates of reaction of

_{physical changes}

bv means of continuous observations of the

intensity

changes.

By

suitable heat treatment,

i. e.

_quenching

from

_appropriate

_{temperatures,}

the

_approach

to

_equilibrium

can be followed

_by

means of the

_magnetic

observations.

Fig. 4.

One

_example

is

_given

in

_{figure 11,}

which shows

the effect of

_heating

at 65oo C an

_alloy

which was

quenched

from

I 100° C,

after

_being

maintained

at that

_temperature

a sufficient time to

_produce

equilibrium.

The

_proportions

of the

_phases

«" and E are each about o.5

_(the

E

_phase

is

_non-magnetic

and the amount

_present

determined

_by

difference. Continuous measurements of the

_intensity

show the break down of the all

_phase

into « and _n.

Pro-longed heating

over a

_period

of weeks showed the

gradual disappearance

of the n

_{phase (which}

is

_only

stable between about 8500 C and io5oO

_C),

with the

accompanying

increase of a and s. ’The whole

of the

_{ferromagnetics regions}

of this

_alloy

_system

. has been examined in this

way. The

accompanying

table

_{gives typical}

_examples

of the

_weight

propor-tions at different

_stages

in the reaction. The

calculated amount of Si is shown in the last column

and _agrees well with the chemical

_analysis.

The most recent

_development

in this field has

been carried out

_by

_Crangle

and Sucksmith

(in

progress)

in an

_{investigation}

into _pure Iron-Carbon

alloys.

Here an

_{interpretation}

of

_equilibrium

condi-tions

_presents

little

_{difficulty .}

For low carbon

contents, up to about 1.2 _per 100

by weight,

the

components

are ferrite

(a

saturated solution of

carbon in

_body

centred cubic iron which contains

~ o.oo5

per Ioo C at room

_temperature

and is therefore

_{magnetically indistinguishable}

from pure

iron),

and cementite

_Fe,C,

for which a- = ~ 66 with a Curie

_temperature

of about 215° C. The dotted curve in

_figure

shows

_equilibrium

conditions and the amounts of cementite and ferrite can be

deter-mined from the

_{intensity-temperature}

magneti-sation curve, the results

_agreeing

well with the known constitution of the

_alloys.

The examination of

_quenched

material shows

_clearly

and

quanti-tatively

the three known

_stages

in the

_tempering

process.

Measurements of the first

_change

_up to about 1 50°C

gives (Ci,

T)

curves of the

_type

marked ~. in

_{figure 5}

Fig. 5.

a, decomposition of martensite; b, extrapolation of reversible

part of _graph; c, curve obtained on _cooling from This graph is steeper than the _equivalent diluted iron

graph, showing that the _precipitate is _{ferromagnetic;} d, start of the austenite transformation; e, transformation at constant _temperature of _non-magnetic retained

aus-tenite ; f , that there is here no marked inflection of the

curve shows that the precipitate is mainly something other than cementite; g, decomposition of the first precipitate;

h, region where there appears to be a _precipitated mixture of Fe.,C and Fe20 C9. The _approach of the " _quenched "

and " annealed "

graphs depends on _temperature but not

sharply on time.

which is known to be due to the breakdown of the

martensite solid solution. The identification of the

breakdown as distinct from a Curie

_temperature

is indicated

_by

the

_{irreversibility,}

_{whilst the}

pro-duction of a

_{ferromagnetic component}

with a lower

Curie

_point

than iron is shown

_by

the

_cooling

curve c.

being

steeper

than for iron mixed with a

non-ferromagnetic

material.

_{Subsequent heating gives}

reversible values

_along

c. The Curie

_temperature

of the new constituent is not less than about 3ooo C,

and it is

_possible

_{that it may be} identified with

the s-carbide

_Fe,C,

which

_though

not

_yet

_prepared

435

has a Curie

point

of 380~ C. It is also

_noteworthy

that Jack has obtained

_X-ray

reflections of this carbide from a

_high

carbon steel

_tempered

at I20~ C.

The second

_stage

shows the transformation of the

non-magnetic

austenite

_{(face-centred}

cubic

_iron)

into a

_magnetic

form thus

_giving

the increase

indi-cated at e. After this transformation is

_completed

the

_{(c, T)}

curve is reversible

_{along f .}

This is

_steeper

than that measured before the commencement

of the austenite breakdown and contains an inflexion

in the

_slope

at about 2 150

C,

i. e. at the Curie

tempe-rature of cementite. The

_implication

is that some cementite is

_{produced, though}

whether this is due to the austenite or the lower

_temperature

martensite transformation is not

_yet

clear.

The third

_tempering

_stage

is characterised

_by

a

sharp

fall in

_{magnetisation,}

and becomes

_fairly

rapid

above 26oo C. The

_change

is

_again

irrever-sible and is

_presumably

due to the breakdown of a carbide to a

_phase,

the

magnetic

intensity

of which is low at this

_temperature.

At 3ooO C the

_change

_{becomes very}

_slow,

but the

_magnetic

intensity

remains

_higher

than the annealed state

until a

_temperature

of about 5oo-55oo C is attained. The

_{interpretation}

of this

_{stage presents}

considerable

difficulty.

The trend of the

_{(o-, T)}

curve _up to about 3ooo C can be

_{reproduced by assuming}

that

the three constituents are

_ferrite,

cementite and

iron

_{percarbide (Fe2oC9).}

This latter has been

isola-ted and measured

_{magnetically.}

The Curie

tempe-rature is

_{2 ~o°}

C _{which agrees well with the values} obtained

_by

Pichler and Merkel

_{[14], though}

sligh-tly higher

than the value of

₂₄₇₀

C

_{given by}

Hofer,

Cohn and Peebles

_[15].

Iron

_percarbide

however,

decomposes

into cementite above

_45oo

C,

and thus the

_{only ferromagnetic}

constituent available in the

temperature

range 300°-500°C is the

ferrite,

the

quantity being

in excess of the

_equilibrium

percen-tage.

Further work is in _progress to

_try

and elucidate these

_points.

TABLE.

Results

_of

_{Quantitative Magnetic}

_{Analysis of}

Some I’hree-Phase

_{cr-1Y Curçes,}

and

_Cornparison

with 1-rotal Silicon Content.

w = Relative amounts of each phase; £w - i.

It is

_hoped

that the

_examples

considered above

indicate the

_importance

of

_magnetic

saturation

intensity

measurements both in the field of

_magnetic

phase analysis

and that of the

_physical

interpre-’

tation of

_{magnetic phenomena.}

Remarque

de M. Goldman. - 11 existe _un

pr6-cedent pour

l’hypothèse

de Sucksmith relative à la coexistence des

_phases

ordonn6e et

desordonnee

dans

_Fe3Al.

Smoluchowski et Newkirk ont trouv6 que dans CoPt les

phases

ordonn6e et

desordonnee

pouvaient

exister en même

_temps.

Dans un tel

syst6me,

le durcissement

_{magn6tique peut provenir}

de la

_{precipitation}

d’une

_phase

ordonn6e

_t6tragonale

dans une matrice de

_phase

_cubique

desordonnee.

Remarque

de 11,1. Hoselitz. -

During

the

_magnetic

studies of iron nickel

_alloys

of

_composition

approxi-matively Fe3Ni,

I found that in the

_non-magnetic

_y state there existed a

_superlattice

which can be detected

_by

_magnetic

measurement

_owing

to the

fact that a _{25 per} 100 nickel

alloy undergoes

irre

versible

_{phase change}

to the

_magnetic

oc form when cooled to the

_temperature

of

_liquid

air. Thus

an

_alloy

with 23 _per 100 at.

nickel,

when annealed

at a

_temperature

of 5250 C and

_subsequently

cooled

to the

_temperature

of

_liquid

air,

showed three

magnetic

constituents. The two

_equilibrium

concen-trations at 5250

C,

i. e.

_containing

about _g

and _{27 per} 100 nickel

respectively,

and a 25 _per 100

nickel constituent which must have been formed

during

the

_annealing

treatment

_prior

to

_cooling

in

_liquid

air

_(HOSELITZ,

J.

_{o f}

Iron a Steel

Inst.,

I9!l4,

Part

I,

p.

193).

Remarque

de M.

_Shockley.

- In connection with the iron aluminium

_alloy

discussed

_by

Prof. Sucks-mith it may be

appropriate

to remark that the in

immediate

_neigborhood

of the critical

_temperature

for order, there _may be ordered and disordered

phases

present

of different

_{compositions.}

The

observation

_by

Prof. Smoluchowski of ordered and

by

Prof. Goodman and the

_theory

of

_phase

_diagrams

requires

in

_general

that the

_compositions

be diff

e-rent. The difference in

_composition

_{may lead} to stresses and other discontinuities

’which

contribute

to the coercive force.

-Remarque

de iV. Sfoner. - 11 faut

souligner

l’im-portance

de ces mesures

_pr6cises

de la variation

de 1’aimantation

_spontan6e

avec la

_temperature.

Comme

_je

1’ai

_deja

dit,

les resultats ne sont _pas

en accord avec la theorie

_classique

_{pour j}

2

ni avec la th6orie des electrons collectifs.

Mais,

en

prin-cipe

on

_peut

d6duire des d6saccords comment les interactions

_d’6change

varient avec

_{1’aimantation,}

ce

_qui

est tr6s

_important

du

_point

de vue

_th6orique.

Les mesures

_{magn6to-caloriques}

_que M. Sucksmith a commene6es a faire sur les

_alliages

_pourront

aussi donner des

_{renseignements pr6cieux.}

_L’analyse

des r6sultats

_{exp6rimentaux pr6sente beaucoup}

de difficultes mais nous _pensons

_{pouvoir d6velopper}

a Leeds des m6thodes

_convenables;

des mesures

pr6cises

comme celles de M. Sucksmith sont

indis-pensables

pour cela.

Remarque

de M. Néel. - Les désaccords

signal6s

par M. Stoner

peuvent

provenir

non seulement

de la variation des interactions

_d’6change

avec 1’aimantation mais aussi de la variation avec la

temperature.

Remarque

de ll4I. Van Vleck. - 11 faut

souligner

le fait

_qu’a

basse

_temperature

le calcul de 1’aiman-tation a saturation est très difficile. On

_peut

com-prendre

pourquoi,

a basse

_temperature,

la dimi-nution de _{1’aimantation par}

_rapport

a sa valeur pour T = o est

plus grande

exp6rimentalement

que la diminution calcul6c avec le mod6le du

_champ

mol6culaire;

on doit

_plut6t

utiliser la m6thode des

ondes de

_{spin qui}

donne une loi

d’approche

de la

forme a

= o~o ~

Aux basses

_{temp6ratures,}

il

_s’agit

de fluctuations li6es aux ondes de

_spin

et non

_comprises

dans la th6orie du

_champ

mole-culaire. Les difficultés relatives aux basses

temp6-ratures se

_pr6sentent

d’une

_{façon f rappante}

dans la

m6thode de Bethe-Peierls

_d6velopp6e

_{pour le}

magn6-tisme par Peter Weiss. Anderson a observe

qu’avec

cette m6thode il existe une

_temperature

au-dessous

de

_laquelle

il ne devrait pas exister de

_f

erromagne-tisme. On doit attribuer ce r6sultat absurde au fait

qu’a

basse

_temperature

il existe des corr6lations

et des

_longues

ondes

_qu’on

ne

_peut

_pas traiter

facilement _{par la m6thode de} Bethe-Peierls.

REFERENCES.

[1] WEISS and FORRER. 2014 Ann.

Physique, 1926, 5, 153.

[2] WEISS and FORRER. 2014 Ann.

Physique, 1929, 12, 279.

[3] POTTER. - Proc.

Roy. Soc., 1934, 146, 362.

[4] BLOCH. 2014 _Z.

Physik, 1930, 64, 817, and 1933, 81, 790.

[5] GUIELLAUD and Roux. - C. R. Acad. Sc., 1949, 229,

1062,

[6] GUILLAUD. - C. R. Acad. Sc., 1946, 222, _1110.

[7] ELLIS and GREINER. - Bell

Telephone Tech. Publ., B 1257.

[8] GOLDMAN and SMOLUCHOWSKI. 2014 _Phys. Rev., 1949, 75,

140, 310.

[9] BENNETT. 2014 Thesis. Sheffield, 1949

(unpublished).

[10] ARDRON. 2014 Thesis. Sheffield, 1949 (unpublished).

[11] SUCKSMITH. - Proc.

Roy. Soc., 1939, 171, 525.

[12] GUILLAUD. - Rev. Métall., 1949,

46, 453.

[13] GUGGENHEIMER, HEITLER and HOSELITZ. - J. Iron and

Steel Inst., 1948, p. 192.

[14] PICHLER and MERKEL. 2014 U. S. Bureau

of Mines Techn,

Paper 718, 1949.

[15] HOFER, COHN and PEEBLES. 2014 J. Amer. Chem. Soc.,