ANTIFERROMAGNETISM OF IRON IN FACE-CENTERED CRYSTALLINE LATTICE AND THE CAUSES OF ANOMALIES IN INVAR PHYSICAL PROPERTIES

• Document Type: CREST
• Collection: General CIA Records
• Document Number (FOIA) /ESDN (CREST): CIA-RDP80T00246A009700440001-6
• Release Decision: RIPPUB
• Original Classification: K
• Document Page Count: 17
• Document Creation Date: December 22, 2016
• Document Release Date: August 6, 2009
• Sequence Number: 1
• Publication Date: March 1, 1960
• Content Type: REPORT

Approved For Release 2009/08/06: CIA-RDP80T00246AO09700440001-6
Approved For Release 2009/08/06: CIA-RDP80T00246AO09700440001-6
 Approved For Release 2009/08/06: CIA-RDP80T00246AO09700440001-6
The anomalies of electrical and magnetic properties in invars at low
temperatures and the fact that these anomalies are observed in iron-nickel
alloys with the nickel content ranging from 30 to 4+0 percent are explained
on the grounds of an assumption that the exchange integral of electrons of
neighboring ions of iron in a face-centred lattice is negative which entails
a "latent" antiferromagnetism in the invars. This assumption is founded on
experimental data obtained by authors and proving that in an iron alloy with
a face-center l2attice, stable at low temperature thanks to the presence of
chrominium and nickel alloying elements, an antiferromagnetic transformation
occurs. The paper presents results of investigation of the influence of
pressure upon the magnetic saturation
and the residual resistivity
.f: of iron, nickel and iron-nickel alloys at low temperatures. It is hown
that the values - and in invar alloys at low temperatures
&p 'o 4 p
and other iron-nickel alloys; the ratios Kul are approximately equal to the
ratios which is in agreement with the conclusions of the suggested
theory.
Fig I a
Approved For Release 2009/08/06: CIA-RDP80T00246AO09700440001-6
 Approved For Release 2009/08/06: CIA-RDP80T00246A009700440001-6
ANTIFERROMAGNETISM OF IRON IN FACE-CENTERED
CRYSTALLINE LATTICE AND THE CAUSES OF ANOMALIES
IN INVAR PHYSICAL PROPERTIES
by E. I. Kondorsky, V. L. Sedov
The Academy of Sciences of the USSR,
State University of Moscow, Moscow, USSR
The physical properties of 36 percent Ni invar alloys have a number of
anomalies (I): the low heat expansion coefficient, a substantial value of
electrical resistivity, comperatively big values of volume magnetostriction,
magnetic susceptibility in the magnetic saturation region and the coefficient
?~ describing the ,pressure influence upon the specific magnetic
It follows from the thermodynamic equality: ( p/ p '(4)H
saturation
that at big values of big values of volume magnetostriction should -
be observed in invars and as a sequence the heat expansion coefficient anomalies.
The comparatively big values of 1 L.- in invars have been explained heretofore
(2-4) by the assumption that the exchange interaction energy between the
electrons of neighboring ions in these alloys changes radically under the
expansion or compression of the crystalline lattice. It was believed that the
point corresponding to the mean exchange integral for ions in invars is on the
steep part of the positive section of the Bete-Slaiter curve, and it has been
considered that the changes in the spontaneous magnetization of these alloys
caused by pressure come as a sequence of the Curie point displacement which in
its turn is related to the change in the exchange energy which the distance
between the lattice ions shrinks or increases. Fig.I-a shows how the
spontaneous magnetization should change under the influence of pressure if
the latter caused only the Curie point shift.
Approved For Release 2009/08/06: CIA-RDP80T00246A009700440001-6
 Approved For Release 2009/08/06: CIA-RDP80T00246AO09700440001-6
This explanation of the p4 anomaly and of the related, peculiarities
of physical properties in invar alloys can hardly be treated as a satisfactory
one. It doesn't supply a direct an answer as to why these anomalies are observed
precisely in the iron-nickel alloys with nickel concentration ranging from 30
to 45 percent or in iron-platinum alloys at definite concentrations, and is
not observed in these alloys with some other contents of nickel or platinum.
There are also doubts as to the validity of comparing the dependence curve of
Curie temperature upon the nickel concentration in iron-nickel alloys with the
Bete-Slaiter curve. Finally, from the point of view of the above said concepts
it is impossible to explain the high electrical resistivity and a number of
anomalies in magnetic properties that have been recently found (5-7) in invars
at low temperatures. In their paper (7) the authors of the present work have
shown that the limit value of magnetic saturation c= 0j. and the residual
resistivity 8 , change with hydrostatic compression, and that the coef-
ficients ( ti) 440 and o do not turn into zero. Thus, the actual
shape of the curve for the invar alloys at 13;o is not as is shown in Fig I-a
outline, but follows the ,pattern of Fig. Ib. Apparently, the changes of e
due to the ,pressure in the low temperature region (the left part of the curve in
Fig. Ib)cannot be explained by the Curie point displacement due to pressure.
In the paper (7) we have indicated the possible causes of pressure
influence upon 6~ and S. U1 particular that the changes of G, and
P? can be a result of the influence of pressure upon the values of the d-d
exchange integrals provided certain exchange integrals are negative and there
is a non-compensated antiferromagnetism in the alloy. The present work attempts
to explain the anomalies of electrical and magnetic properties in invars at low
temperatures and that these anomalies are manifest precisely at a nickel content
Approved For Release 2009/08/06: CIA-RDP80T00246AO09700440001-6
 Approved For Release 2009/08/06: CIA-RDP80T00246AO09700440001-6
-3-
ranging from 30-40 percent, preceeding from the assumption that the exchange
integral of the electrons of the neighbouring iron ions in a face-centered
lattice is negative.
I. Experimental Grounds for Supposing that Iron is Antiferromagnetic
in a Face-centered Lattice
It has been shown in paper (8) that for the r -phase (face-centered
cubic lattice) of iron in the temperature interval from 910-1400?C when the
phase is stable, the Curie-Weiss'law is valid but with other values of the
and ? parameters than for the '' phase (body-centered cubic
lattice). In keeping with the paper (8) the G parameter for the r -iron
phase is a negative value ( y --1340?K), which gives us grounds to suppose
the absence of the ferromagnetic trans-formation in this phase up to the
absolute zero. In order to determine whether the '-phase of iron is
paramagnetic or antiferromagnetic at low temperatures we should prevent the
Y'-, d. transition by introducing alloying elements and by appropriate heat
treatment. The paper (9) has a brief reference to, the fact that according to
the neutronographic analysis data the Fe -Y alloys with a face-centered cubic
lattice containing over 12 percent of manganese are antiferromagnetic. However,
thusfar no works have been published where these data would have been described.
We have investigated the temperature dependence of the magnetic susceptibility
of an alloy with a face-centered lattice, containing 73 percent of iron, 18 percent
of chromium and 9 percent of nickel. No precipitate of the ferromagnetic
- -phase at low temperature has been observed in the alloys of this content
that have been aged at room temperature following their preparation (10). The
specific susceptibility `L was measured by means of the procedure described
by me in paper (7), The results are presented in Fig 2. The anti?erromagnetic
Approved For Release 2009/08/06: CIA-RDP80T00246AO09700440001-6
 Approved For Release 2009/08/06: CIA-RDP80T00246AO09700440001-6
transformation takes place at 40?K. The value of the paramagnetic Curie
point proves to be -28830K
The obtained results give grounds to suppose that the exchange inter-
actions in.a face-centered iron lattice would at low temperatures result in
antiferromagnetism similarly to the neighbouring elements manganese (11-12)
and chromium (13), and the exchange integral of the d electrons of the neigh-
bouring ions of iron.
2. Anomalies in invar alloys magnetic and electrical properties
at low temperatures
The magnetic saturation Is of iron-nickel alloys at hydrogen temperatures
has been studied in the work of Kondorsky and Fedotov (14). In paper (15) and in
our paper (7) we have studied the residual electrical resistivity of these
alloys. Fig. 3 presents the curves describing the dependence and of iron-nickel alloys upon nickel content. In the invar alloy region there
takes place a decrease of I. and a radical growth of .f* at the increase of the
iron content.
In order to determine how far the k :_L and /
differ for invar alloys from the same values in the alloys of other content we
have studied the influence of pressure upon G. and d, of iron, nickel,
and iron-nickel alloys of different concentration.
A pressure of about 150 atm. was created by gaseous helium which was fed
into a beryllium-bronze bomb from a gas tank through a capillar valve. The bomb
with the specimen was in a field of a water-cooled sollemoid. The measuring
coils, the basic and the compensating one were placed outside the bomb and were
in the circuit of the photoelectrical fluxmeter, with the sensitivity of 5 Narwell/
dag. The error in measuring the value 16 -1 was +0.15.10 Gauss/atm. The
value was detertined, by the procedure described in papers (16-17). In this case
Approved For Release 2009/08/06: CIA-RDP80T00246AO09700440001-6
 Approved For Release 2009/08/06: CIA-RDP80T00246AO09700440001-6
-5-
the pressure was created by freezing a 3,percent aqueousL.solution of alcohol.
It turned out that within the accuracy limits of our measurements the
values) G, of the studied metals and alloys are not dependent on temperature
in the range from 4.2 to 20.I?K, while in the same interval 1`? n values slowly
diminish with temperature.
Fig. 4 presents the curves describing the dependence of the and 11~
limit values obtained by extrapolation to absolute zero, from the nickel content
in the alloys. In the region of invar alloys (30-40%) there takes place a
radical increase of ft 0,, and a peak fig, . Thus, the f6? anomaly in the
invars takes place both at high and low temperatures.
3. Ferromagnetism-Antiferromagnetism Transformation in Iron-Nickel
Alloys With Face-centered Lattice as the Cause of Anomalies in
Magnetic and Electrical Properties of Invars
The quantum-mechanical calculation of the magnetic saturation of disordered
alloys involves profound mathematical difficulties and thusfar has not been
carried out. A calculation of this type for the ordered AB and AB3 alloy has
been performed by Kondorsky and Pakhomov (18) but the results cannot be applied
to disordered alloys like invars. Therefore, proceeding fromthe quasi-classical
model we shall try to make a rough estimate of the least nickel concentration in
iron-nickel alloys at which these alloys can be ferromagnetic, provided that the
exchange integral of the d., electrons of the neighbouring ions of iron is
negative. 04, ,a and $, are exchange integrals d electrons of
neighbouring ions of iron, iron and nickel and nickel, respectively: Cl and C2 -
are the concentrations of iron and nickel in the alloy ( Cl - C2 =1 ). The
concentrations of ions of iron and nickel with the right and left spins shall be
denoted as C1 , C2 , C} , and C2 respectively.
As has been indicated for the first time by Vonsovsky (19) it is impossible
to consider the spontaneous magnetization of parts of the lattice occupied by ions
Approved For Release 2009/08/06: CIA-RDP80T00246AO09700440001-6
 Approved For Release 2009/08/06: CIA-RDP80T00246AO09700440001-6
-6-
of nickel separately from the parts of the lattice occupied byjthe ions of iron
but rather only the full numbers of and C' of right and left spins
respectively. Therefore we should take C,4- C
2
2. ''L a
At j, where Jet -Z j_e
~- ti )CI, GA . ?1- dC
Then considering the mean exchange energies of iron and nickel ions with right
and left spins surrounded by Z neighbours, (Z=12) and taking into account only
the interaction between the neighbouring ions it is not difficult to obtain for
the exchange energy Wa alloy from N ions.
(I)
which coincides with the formula obtained for the first time in the work (19).
It follows from (I) that at low temperatures the antiferromagnetism is
possible provided there.are negative exchange integrals. The pure nickel is
ferromagnetic, within an increase of the iron content up to 60 percent the magnetic
moment of iron-nickel alloys increases too Hence 3-,," 0, In view of
the experimental data that have been mentioned earlier we may take
Then the state with i. e., a ferromagnetic state, will correspond
to the smallest values of the exchange energy Wa when the trinominal in the second
brackets (I) is positive. When this trinominal is negative the least value of
Wa is obtained at which corresponds to antiferromagnetism. Hence, if we
have the CA 'PC,,concentration, the ferromagnetism is more advantageous from the
energy point of view at low temperatures, if C2 '- C RI then under the same
condition the antiferromagnetism is more advantageous. Taking the trinominal
in the formula (I) as being equivalent to zero we find that
(2)
in a special case when ~' a o
a quite satisfactory result. 3-0
which, apparently, is
The random distribution of ions with anti-parallel spins is more likely
Approved For Release 2009/08/06: CIA-RDP80T00246AO09700440001-6
 Approved For Release 2009/08/06: CIA-RDP80T00246AO09700440001-6
- T -
at high temperatures then at low ones when there should take place the
groupings of spins (the closest magnetic order). Therefore between the fully
ferromagnetic and the antiferromagnetic there should be alloys with some of the
ions having the anti-parallel spins. This type of "latent" antoferromagnetism
should take place when the mean energy of the ion of iron with the left spin
W ? .,z ( " ')((; C'r f'"r a C g-) is negative, i.e. at a nickel concentration
of Ca a Hence at
C. ACS ,`~~. While Co the
I%e
ferromagnetic moment of iron-nickel alloys grows linearly with the increase in
the iron content, at ?'C2. = (, appears a deviation from linearity and at C g close
to C,,t there sets in a radical drop of I., thanks to an increase in the number
of the anti-parallel spins. As it is seen from Fig. 3, the change in the I. with
the growth of iron content in the iron-nickel alloys with the face-centered lattice
occurs precisely in this way. Let us note that the deviation from the linearity
in the curve I. occurs in the interval of nickel content ranging from 50 to IU)
,percent, while with further decrease in nickel content, beginning with h+0 ,percent
a radical drop of I. is observed.
We believe that the latent antiferromagnetism should be regarded as the
main cause of the anomalies in invar physical properties. In fact, the presence
of the anti-parallel spins in these alloys explains the comparatively big values
of the volume magnetistriction and the susceptibility of the ,paraprocess in
strong magnetic fields. Further, as it is'seen.from formula (I) the transitional
concentration e., grows with the increase in the absolute value , which
most likely grows at hydrostatic compressionx). An increase in. Cz entails a
shift of the steep part of the curve I. towards the nickel ani which should be
accompanied by radical decrease in I. and (, precisely of the invars which are
in the transition region. As it is seen. from Fig. 4 the experimental data
corroborate this conclusion.
x) Let us note that to explain. the C & "displacement" there is no necessity of
Approved For Release 2009/08/06: CIA-RDP80T00246AO09700440001-6
 Approved For Release 2009/08/06: CIA-RDP80T00246AO09700440001-6
-8-
assuming that any of exchange integrals correspond to the point on the steep
part of the Bete-Slaiter curve
and are more sensitive to the changes of the
inter-atomic distances that others.
For instance from (2) at ?, _,,
and ad~ it follows that
if ~.e i' id
ei -_% !' tL P Gd
P
q(R- . Then
~.
In a similar way it is possible to explain the anomaly of magnetic
properties observed in iron-platinum alloys at ,platinum content of 30-40 percent.
The exchange integral of d-electrons of neighboring ions of platinum
should be taken as equivalent to zero, since up to helium temperatures platinum
remains paramagnetic. Since iron-platinum alloys are ferromagnetic, 0
Taking just as in the iron nickel alloys we shall obtain from (2) for
concentration C-rC of iron-platinum alloys ?~. that at-9t 7 gives
e 7, Cv 1ITRa.-70
Cq< 0.73
If platinum concentration\the alloys are ferromagnetic.
Substituting in (2) the sign minus before the radical for plus at the same values
of the exchange integrals we shall find that ferromagnetism in iron-,platinum
alloys disappear at iron concentration close to zero.
An indirect confirmation of "latent" antoferromagnetism in invars is their
great electrical resistivity, which we believe is due to the conductivity
electron scattering on the magnetic moment heterogeneities. With an increase in
the iron content from 50 percent the electrical resistivity begins to grow and
particularly vigorously in the region of the magnetic moment drop. As a value
describing the magnetic moment of anti-parallel spins we may regard Io= Io - Ic
where I - is the spontaneous magnetization value corresponding to the ordinate
c
obtained by extrapolating the linear part of the Io curve on the iron and
c - are the spontaneous magnetization values of alloys with x concentra
(Fig-5). I
tions in real or hypothetical case when all the spins are parallel. If we
determine the .?.X electrical resistivity incriments and the Al spontaneous
Approved For Release 2009/08/06: CIA-RDP80T00246AO09700440001-6
 Approved For Release 2009/08/06: CIA-RDP80T00246AO09700440001-6
-9-
magnetization., which correspond to the increase in the iron concentration on
~. , compose the ratios ~~ 4Z and compare them with the ratios
6Gand N~ / , for the alloys of the same concentrations, it
? X68
turns out that all these ratios are approximately the same. Table I presents
the values of these ratios for the alloys with the nickel content of 34.7 percent
and 38 percent.
The value gl~ AY. in strong fields is negative in ferromagnetics and
ordinarily this has been associated with the fact that when the true magnetization
grows in the strong fields the scattering of electrons on spin waves decreases.
The "latent" antiferromagnetism considered by us in invar alloys is reduced to
the existence of spin waves at absolute zero. An approximate confidence of the
mentioned ratios may be regarded therefore as a confirmation of the assumption
on the "latent" antiferromagnetism.
Approved For Release 2009/08/06: CIA-RDP80T00246AO09700440001-6
 Approved For Release 2009/08/06: CIA-RDP80T00246AO09700440001-6
References
1. R. M. Bozorth, Ferromagnetism Van Nostrand N.Y. 1951.
2. M. Kornetzky ZS Phys.98.289 (1935).
3. U. Dehlinger Z. Metallunde 28.194 (1936).
4. K. P. Belov JETP 19.1032 (1949)-
5. D. Smit Physics 17. 612 (1951).
6. K. P. Belov A. V. Pedko, Dokaldy of high school (USSR), Nb.2,
214 (1958).
7. E. I. Kondorsky, V. L. Sedov JETP 35.845 (1958).
8. P. Weiss G Foex Journ. Phys. et Rad. (5), 1.744 (1911).
9. N. Mott K. Stevens Phil. Mag. 2,1364 (1957).
10. T. Colembier Hochmann, Aciers Inoxi Dobles, Aciers
Refractuires, Paris (1955).
11. T. B. NcGuire, C. T. Kriesmann Phys. Rev. 85, 452, 9, I (1952).
12. C. B. Shu11:.N. K. Wilkinson Bull. Amer. Phis. Soc. 27,24 (1952).
13. C. B. Shull N. K. Wilkinson Rev. Med. Phys. 25,100 (1953).
14. E. I. Kondorsky, L. N. Fedotov Iswestija Academy of Science of
USSR (Phys) 16,432 (1952).
15. I. Shirakava Sci. Rep. Tohoku Irnp. Univ. 27,485 (1939).
16. B. G. Lazarev L. S. Kan JETP 14, 439 (1944).
17. N. B. Brandt A. K. Tomaskuk J.P.T.E. (Russ)2,113, (1958).
18. E. I. Kondorsky A. S. Pahomov JETT 32.323 (1957).
19. S. V. Vonsovsky Doklady Academy of S. USSR 26,564 (1949).
Approved For Release 2009/08/06: CIA-RDP80T00246AO09700440001-6
 Approved For Release 2009/08/06: CIA-RDP80T00246AO09700440001-6
Approved For Release 2009/08/06: CIA-RDP80T00246AO09700440001-6
CC-r"Z o
 Approved For Release 2009/08/06: CIA-RDP80T00246AO09700440001-6
-g2
Capttowo
0
0d0 D dr Of Piio?urQ influence upon spout?neouo
ntiBnti?~0
ae opaitcmcotto c otiQatlaz nitoro only duo
to GU.5o 0oflat CM
b) SEIa Q,> Mtt e 1cj ~Q$iocat~4i?~ Ql4aeo duo to Curl e,
?ir~t c f,, d ~2mo t? tbo obci o in t ho evoPege
n c6lo Dc osi@ =Olnto 0
Pig020 `r-o tc io a dopoZQoncQ of tho a ratia augcepti.bYii -,r
and. it invorcjca valmo in Qr tonito otoo1. 71 th I6" G.
Al and 9% PPecttuzo pog nto la4ieatiz tho antifor?romag-
natio
Pig-3., UQgalg ? Qnft ration of abooluto zoy~o point and residual
P001otiv5lt r of t a5L?)ro1. a.ll.oyo voa a n.ic11e1 content
imXogn
Unordorod elloyo (Data of Authora
t pmper? (R)
t rc -- ? los?Gd all@yo ( tea of paper (6)
/ 040 VQ1LlOO 0Q0 j cnq DerIZUo
f
o
nicho1 oOntCIMt $2Vn-'dUcio1 Q31oyo. The Purvae ettia1bit,
tho r ialUoo @f
cad 4 in i rivaro rogioo 0
0- P
Ig. 5 o Di.c )P ilimotgnt?o too dotominrat1o3n of Z' aid .I-,
Approved For Release 2009/08/06: CIA-RDP80T00246AO09700440001-6
 Approved For Release 2009/08/06: CIA-RDP80T00246AO09700440001-6
~, lI
~
i rl k
Approved For Release 2009/08/06: CIA-RDP80T00246AO09700440001-6
 Approved For Release 2009/08/06: CIA-RDP80T00246AO09700440001-6
Approved For Release 2009/08/06: CIA-RDP80T00246AO09700440001-6
 Approved For Release 2009/08/06: CIA-RDP80T00246AO09700440001-6
C3
z
Approved For Release 2009/08/06: CIA-RDP80T00246AO09700440001-6