best available - defense technical information · pdf fileapproved for public release;...
TRANSCRIPT
![Page 1: Best Available - Defense Technical Information · PDF fileApproved for public release; distribution unlimited. Reproducpcl by NATIONAL TECHNICAL INFORMATION SERVICE U S Daparlmtnl](https://reader031.vdocuments.mx/reader031/viewer/2022030413/5a9f16157f8b9a7f178c57e4/html5/thumbnails/1.jpg)
' . '
I
' ! I
Best Available
Copy ·
...
![Page 2: Best Available - Defense Technical Information · PDF fileApproved for public release; distribution unlimited. Reproducpcl by NATIONAL TECHNICAL INFORMATION SERVICE U S Daparlmtnl](https://reader031.vdocuments.mx/reader031/viewer/2022030413/5a9f16157f8b9a7f178c57e4/html5/thumbnails/2.jpg)
^^^^■^■■pp^-»w^—• '■n1' ' i mmmn^mmm^^mm .nuii. aiu i i n iiim vnspa^pf
AD-760 777
MATERIALS PROCESSING OF RARE EARTH COBALT PERMANENT MAGNETS
Paul '. Jorgensen, et al
Stanford Research Institute
Prepared for:
Air Force Materials Laboratory
February 1973
DISTRIBUTED BY:
KFÜ National Technical Information Service U. S. DEPARTMENT OF COMMERCE 5285 Port Royal Road, Springfield Va. 22151
'ii'^ilMflM^fd^h^Miir' f
![Page 3: Best Available - Defense Technical Information · PDF fileApproved for public release; distribution unlimited. Reproducpcl by NATIONAL TECHNICAL INFORMATION SERVICE U S Daparlmtnl](https://reader031.vdocuments.mx/reader031/viewer/2022030413/5a9f16157f8b9a7f178c57e4/html5/thumbnails/3.jpg)
■MMMM^^^vrnp"*»
AFML-TR-73-60
MATERIALS PROCESSING OF RARE EARTH COBALT PERMANENT MAGNETS
o
<
P. J. Jorgensen R. W. Bartlett
Stanford Research Institute
TECHNICAL REPORT AFML-TR-73-60
February 1973
Approved for public release; distribution unlimited.
Reproducpcl by
NATIONAL TECHNICAL INFORMATION SERVICE
U S Daparlmtnl of Comirnr« Si)fing(i»l'l VA 771 "it
Air Force Materials Laboratory Air Force Systems Command
Wright-Patterson Air Force Base, Ohio 45433
riMii JIlllMlitilMMiaTMmtliiiiiM
^
— -■'■■ ■- '■-"
![Page 4: Best Available - Defense Technical Information · PDF fileApproved for public release; distribution unlimited. Reproducpcl by NATIONAL TECHNICAL INFORMATION SERVICE U S Daparlmtnl](https://reader031.vdocuments.mx/reader031/viewer/2022030413/5a9f16157f8b9a7f178c57e4/html5/thumbnails/4.jpg)
""-•^"■"■■OPW l"1"11 ii iinipini i u
1
NOTICE
When Government drawings, specilications. or other data arc used lor any purpose other than in connection with a definitely related Government procurement operation, the United States Government thereby incurs no responsibility nor any obligation whatsoever; and the tact that the govern- ment may have lormulated, lurnished, or in any way supplied the said drawings, specilications, or other data, is not to be regarded b" implication or otherwise as in any manner licensing the holder or any other person or corporation, or conveying any rights or permission to manufacture, use. or sell any patented invention that may in any way be related thereto.
Copies of this report should not be returned unless return is required by security considerations, contractual obligations, or notice on a specific document.
«CCRSIMJir
mi Mil1« '•«"•" "flf* DOC Bulf Stctlin a MMROailOIH □ .'tS'.li iwAHCH
11 DISUIlllTION/AmHIllIt MOO
~ Oi:,. (..»IU MJ/'i y.CI'l
y
I
I
Bi -'
.«mmmmmtmm mtmm
![Page 5: Best Available - Defense Technical Information · PDF fileApproved for public release; distribution unlimited. Reproducpcl by NATIONAL TECHNICAL INFORMATION SERVICE U S Daparlmtnl](https://reader031.vdocuments.mx/reader031/viewer/2022030413/5a9f16157f8b9a7f178c57e4/html5/thumbnails/5.jpg)
mmm mm
UNCLASSIFIED s*- .rit\ l KisMtu .**i'-n
DOCUMENT CONTROL DATA R 4 D .,,...,,.„. ..t,,;. .../. (..> ..../,..i ,,. ■ •■i...,/ .1 i ...I' -, '■■ -"■ '
Stanford Hosoarch Institute
33U Ravanswood Avenue Mcnlo Park, Callfurnla 94025
üNCa^ASSIFIED
MATERIALS PROCESSING OK KAItE EAHTH-COHALT PERMASEMT MAGNETS
ICtCRI^tlVI NOTf»(Ap«■,,' ftpcft .m.i i»u /u t\> l.ti-
• »u t,.oH.!.. iFlnl nmmr. rni./Jlr inrlKl. In«! namrl
Paul J, Jorgenaen Robert w. Bartlett
Mf IJO*' ' * ' '
February 1973 •• CONTMACT OR CtANT NO
F33615-7()-C-162 1 h PHOJ I C T NO
ARPA Order No, 1617
Program Code No, OD10
m i o • »L NO o> ''*-i 7A 24
PYU-H731
«^ o * »*i ** '** I'OM T NO
A1ML-TR-73-GO
'/l»iv ufhvr numbvn f/'wr nmy b* «aalgnvd
t' Dl»l »1 ÖU T ION S^ATEMtNT
Approved for public release; distribution unlimited,
.PPLLME». '*Ni . IP ON SOn IN 6 Mil 'A^*^ ACTIVITY
Air Force Materials Laboratory Wrißht Patterson Air Force Base Ohio 15433
* ÜST«, A C
A study of the oxidation kinetics of SmCo involving selective oxidation of
samarium was extended from high temperatures to lüo'C without a change in the diffud.01
controlled rate mechanism. The low temperature oxidation rates are consistent with
the amount of oxidation normally observed in SmCo powders and with the temperature
rise expected from oxidation of ground powders. The solubility of oxygen in SmCo5 at
1130OC (sintering temperature) in excess of the 800 C solubility was determined to be
3500 + 500 ppm. The source of oxygen during sintering is the oxide subscale. Sub-
micron oxide particles precipitate within the SmCor grains on cooling from the sinter-
ing temperature. Oxidation causes depletion of samarium and precipitation of Sm^Co
particles within grains. Both inclusions are postulated to be sources of domain wall
nucleatlon. The oxide inclusions can be removed at 800 C by an aging treatment, that
collects the oxide into a few large grains ouUlde the SmCo5 grains by a solution-
reprecipitation mechanism involving grain bounoary transport of samarium. The Sm2Co
17
inclusions are not affected by thermal aging, but they can be prevented from occurring
by including excess samarium in the sintered compact to replenish samarium lost by
oxidation.
The kinetics of solid phase sintering of SmCo have been studied as a function of The kinetics can be described by a grain time, temperature, and stoichlometry
md oy
boundary transport mechanism that Is independent allr"
the samarium cobalt ratio In llie
DD fr.,1473 S/N 0101.807.6801
(PAÜE 1) UNCLASSIFIED /a bPcurltv Cliissifit ütirn
rwMMMi --■- IHM ■maillHll
![Page 6: Best Available - Defense Technical Information · PDF fileApproved for public release; distribution unlimited. Reproducpcl by NATIONAL TECHNICAL INFORMATION SERVICE U S Daparlmtnl](https://reader031.vdocuments.mx/reader031/viewer/2022030413/5a9f16157f8b9a7f178c57e4/html5/thumbnails/6.jpg)
"'"■■, wm ^"^^^^m
UNCLASSIFIED Srcunty CUtudcition
Kfy MOHU« LINK B
Magnetic Particles
Rare Earth
Cobalt
Oxidation
Internal-Oxidation
Sintering
Diffusion
Magnetic-coercivity
DD .^..1473 'BACK, (PAGE 2) /£ UNCLASSIFIED
Security Classification
UM liH|j-||-((ilBMMII
![Page 7: Best Available - Defense Technical Information · PDF fileApproved for public release; distribution unlimited. Reproducpcl by NATIONAL TECHNICAL INFORMATION SERVICE U S Daparlmtnl](https://reader031.vdocuments.mx/reader031/viewer/2022030413/5a9f16157f8b9a7f178c57e4/html5/thumbnails/7.jpg)
~~w~. — ''■"• '!"< U.I«PP«II i i iwiiin M i >»i uamwnmH i L i IJI inn ii \wmmm^^^^mmm^ ■ ■ ' ^1
MATERIALS PROCESSING OF RARE EARTH COBALT PERMANENT MAGNETS
P. J. Jorgensen R. W. BartlfU
Approved for public release; distribution unlimited.
/fy
■IIÜHIII mmiläm^mmmmmä^tamM,
![Page 8: Best Available - Defense Technical Information · PDF fileApproved for public release; distribution unlimited. Reproducpcl by NATIONAL TECHNICAL INFORMATION SERVICE U S Daparlmtnl](https://reader031.vdocuments.mx/reader031/viewer/2022030413/5a9f16157f8b9a7f178c57e4/html5/thumbnails/8.jpg)
•w^Kii^wMmnHOTiiwii ^iipii" ■'-' ' ■
FOREWORD
This is the fifth semiannual interim technical report of the research program "Materials Processing of Rare Earth-Cobalt Permanent Magnets" under Contract F33615-70-C-162A. Stanford Research Institute project number is PYU-8731. This project is being conducted by the Materials Laboratory of Stanford Research Institute. Dr. Paul J. Jorgensen, Manager of the Ceramics Group, is the project supervisor. Dr. Robert W. Bartlett of Stanford University is project consultant. The research described in this report is part of the contractual research program of the Electromagnetic Materials Division, Air Force Materials Laboratory, Air Force Systems Command, Wright-Patterson AFB, Ohio. Mr. Harold J. Garrett (AFML/LPE) is the project monitor. It was sponsored by the Advanced Research Project Agencv, ARPA Order No. 1617, Program Code No. OD10.
This report covers research conducted between July 1 and December 31, 1972 and was submitted in February 1973 by the authors for publication.
This technical report has been reviewed and is approved.
/^/^«^v^^V
CHARLES H. ROBISON, Major, USAF Chief, Solid State Materials Branch Electromagnetic Materials Division Air Force Materials Laboratory
li
y^i ------ —■ - ■ - i iiiüiiMitiii miiiiainii
![Page 9: Best Available - Defense Technical Information · PDF fileApproved for public release; distribution unlimited. Reproducpcl by NATIONAL TECHNICAL INFORMATION SERVICE U S Daparlmtnl](https://reader031.vdocuments.mx/reader031/viewer/2022030413/5a9f16157f8b9a7f178c57e4/html5/thumbnails/9.jpg)
wmma^tm*. w« »"N*! HIIIPIHI mr~*i^m*v*^imimn^*wm .... .. . , - IB I wa^WnBH^
ABSTRACT
A study of the oxidation kinetics of SmCo Involving selective oxi- 5
dation of samarium was extended from high temperatures to 100 C without
a change in he diffusion-controlled rate mechanism. The low temperature
oxidation rates are consistent with the amount of oxidation normally
observed in SmCo powders, and with the temperature rise expected from
oxidation of ground powders. The solubility of oxygen in SmCo at 1130 C 5
(sintering temperature) in excess of the 800 C solubility was determined
to be 3500 ± 500 ppm. The source of oxygen during sintering is the oxide
subscale. Submicron oxide particles precipitate within the SmCo grains on 5
cooling from the sintering temperature. Oxidation causes depletion of
samarium and precipitation of Sm Co particles within grains. Both
inclusions are postulated to be sources of domain wall nucleation. The
oxide inclusions can be removed at 800 C by an aging treatment, that
collects the oxide into a few large grains outside the SmCo grains by a D
solution/reprecipitation mechanism involving grain boundary transport of
samarium. The Sm Co inclusions are r.ot. affected by thermal aging, but
they can be prevented from occurring by including excess samarium in the
sintered compact to replenish samarium lost by oxidation.
The kinetics of solid phase sintering of SmCo have been studied as 5
a function of time, temperature, and stoichiometry, The kinetics can be
described by a grain boundar' transport mechanism that is independent of
the samarium/cobalt ratio in the alloy.
lii
■- ^ . _..
![Page 10: Best Available - Defense Technical Information · PDF fileApproved for public release; distribution unlimited. Reproducpcl by NATIONAL TECHNICAL INFORMATION SERVICE U S Daparlmtnl](https://reader031.vdocuments.mx/reader031/viewer/2022030413/5a9f16157f8b9a7f178c57e4/html5/thumbnails/10.jpg)
-..,.....,. u>p-i*.>i • min i ... MI ■ muw diim n i wmmmmq^immi'* i...« «.u»!.». i ...»-.«1™^. ■■■■ i iiiiiiinaa '•"■^■■^
CONTENTS
I INTRODUCTION 1
II OXIDATION 0"' SmCc) 3 5
A. Oxidation, Subscale Structure, and Composition 3
B. Kinetics of Internal Oxidation of SmCo,. at Elevated 5
Temperatures 5
C. Kinetics of Oxidation at Low Temperatures . 8
D. Expected Temperature Rise Associated with Low Temperature
Oxidation 14
E. Oxidation Rate Mechanism 16
III VACUUM ANNEALING PARTIALLY OXIDIZED SmCo AT SINTERING
TEMPERATURES 25
IV OXYGEN SOLUBILITY DURING SINTERING AND SUBSEQUENT OXIDE
PRECIPITATION IN SmCo^ 31 5
V PRECIPITATION OF Sm Co IN SmCo 39
VI DISCUSSION OF Sm 0 AND Sm Co INCLUSIONS AND INTRINSIC
COERCIVITY IN SINTERED MAGNETS 43
A. Summary of Inclusions 43
B. Inclusions and Intrinsic Coercivity 43
VII SINTERING
A. Introduction 49
B. Experimental Procedure 51
C. Results and Discussion 52
REFERENCES . 63
Preceding page blank
- - - .^^mätmmmm
![Page 11: Best Available - Defense Technical Information · PDF fileApproved for public release; distribution unlimited. Reproducpcl by NATIONAL TECHNICAL INFORMATION SERVICE U S Daparlmtnl](https://reader031.vdocuments.mx/reader031/viewer/2022030413/5a9f16157f8b9a7f178c57e4/html5/thumbnails/11.jpg)
I I Hill I 11 I IIUII. « ^ J HMV^M^«
ILLUSTIlfVriüNS
1 Course Fibrous Structure of Samarium Oxide in the Oxide/
Cobalt Composite Subscale Grown at High Temperatures, 1150OC 4
2 X-Ray Diffraction Spectra lor the Oxide Subscales Grown in
Air and in Pure Oxygen 6
3 Arrhenius Plot for the Internal Oxidation Rate at High Temperatures 7
1 OxyRen Dissociation Pressures for CoO in a 90% CO + 10% CO 2
Mixture 9
5 Internal Oxidation Rate Dependence on Oxygen Pressure (Air
Equilibrium at Surface) 10
6 Internal Oxidation Rate Dependence on Oxygen Pressure
(CoO Equilibrium at Surface) 11
7 Thermogravimctric Data for Internal Oxidation of 25 ^m SmCo Particles 13
8 Arrhenius Plot for the Internal Oxidation Rate Extended to Lower Temperatures 15
9 Ad_abatic Temperature Rise Accompanying Partial Oxidation to Sm 0 17
6 10 Oxygen Self-Diffusion in B-Type Sm 0 (After Stone ) .... 19
11 Fibrous Samarium Oxide with a Crack in the Subscale Grown at 800OC 22
12 Oxide Depletion Zone Resulting from Internal Oxidation of
SmCo Followed by Vacuum Annealing, 125X 26
13 Oxide Depletion Zone at Higher Magnification (SEM) 27
14 Boundary of the Oxide Depleted Zone (Right Side) Within the Subscale 28
vi
aiiir ^iiMMimifc -II üfmummiHimtf^m^Mti, „u
![Page 12: Best Available - Defense Technical Information · PDF fileApproved for public release; distribution unlimited. Reproducpcl by NATIONAL TECHNICAL INFORMATION SERVICE U S Daparlmtnl](https://reader031.vdocuments.mx/reader031/viewer/2022030413/5a9f16157f8b9a7f178c57e4/html5/thumbnails/12.jpg)
!■ Ill HUM Ml 111 .III II mi^vQvtfiafi^wiiiai ii rn~ iiiiuau.ii > wm ^^^^^trnm^Brnmammmm* wm^mrm^^mm"
ILLUSTllATIONS (Concluded)
18 Boundary Between the Oxide Depleted Zone and the SmCo Core 30 o
16 Oxide Precipitates in the SmCop Grains after OxyRen Satura- tion at 11250C and Asin« at 8o80C 33
17 Quenched and Aged SmCo Grains 35 D
18 Samarium Oxide Particles in a Sintered SmCo,. Magnet Uncovered 5
by Etching 37
19 Scanning Electronmicrograph of Sm Co Precipitates in SmCo 40 ^ 17 O
20 Magnetic Domains Passing Through Sm Co Precipitates Within
SmCo. Grains 16 5
21 Linear Thermal Expansion of Polycrystalline Randomly
Oriented SmCo^ 53 5
22 Samarium-Cobalt Shrinkage Isotherms for an Alloy Contain ng
27.7 wt% Sm 51
23 Samarium-Cobalt Shrinkage Isotherms for an Alloy Containing
31.35 wt% Sm 55
24 Samarium-Cobalt Shrinkage Isothems for Alloys Containing
33.0 and 33.33 wt% Sm 56
25 Arrhenius Plot of the Shrinkage Rate Versus Reciprocal
Temperature 58
TABLES
1 Some Efects Involving Magnetic Coercivity in SmCo5 and
Their Possible Causes 'I7
2 Sintering Constants for Equation 4 50
3 Oxygen Analysis of SmCo Alloys 59
vil
MMM — - ■ ■■■fclllll ■ - ■ ■
![Page 13: Best Available - Defense Technical Information · PDF fileApproved for public release; distribution unlimited. Reproducpcl by NATIONAL TECHNICAL INFORMATION SERVICE U S Daparlmtnl](https://reader031.vdocuments.mx/reader031/viewer/2022030413/5a9f16157f8b9a7f178c57e4/html5/thumbnails/13.jpg)
'■' '"'■■" ........._.. ..-„..I.,.,,., *^^*mmmm^**im**'mpmm
I INTRODUCTION
The objectives of this program are to imestigate materials pro-
cessiriK methods with the goal of achieving optimum properties of SmCo 5
and other RECo compounds as permanent magnets. The primary problem
is achieving high magnetic coercivity.
Previous to this research period, .1 variety of unconventional pro-
cesses for generating RECoc powders were investigated but none was found
to be more effective in obtaining high coercivity materials than the
properly conducted comminution process of bull milling.
The sintering process is also a critical step in magnet production.
Studios to gain further knowledge of the liquid-phase sintering process
and magnetic evaluation of sintered alloys were carried out prior to this
research period, and studies of ehe final stage of sintering (solid-state
process) were conducted during this period and arr reported.
Selective internal oxidation of samarium invariably accompanies
magnet powder processing. A study of the kinetics of oxidation and the
morphology of the oxide subscale was completed during this period. Studies
on oxygen solubility in SmCor and the precipitation of oxide inclusions D
during quenching and aging are reported.
A study of Sm Co precipitation in samarium-depleted SmCo grains is j i # 5
included in this report.
—■-"- ■-- —
![Page 14: Best Available - Defense Technical Information · PDF fileApproved for public release; distribution unlimited. Reproducpcl by NATIONAL TECHNICAL INFORMATION SERVICE U S Daparlmtnl](https://reader031.vdocuments.mx/reader031/viewer/2022030413/5a9f16157f8b9a7f178c57e4/html5/thumbnails/14.jpg)
II OXIDATION OF SmCo 5
A. Oxidation, Subscule Structure, and Cumpositlon
Oxidation of SmCo5 is selective, resulting in a subscale consisting
of samarium oxide platelets ana libers surrounded by cobalt. The oxide
fibers are oriented in their growth direction, which is generally normal
to the alloy surface. Scanning electron micrographs on the subscale
structure in the direction of growth and transverse to the direction of
growth are shown in Figure 1. These specimens were oxidized at USt/'c
lor two minutes in a 90% C02 + 10% CO mixture, which prevents any oxidation
of cobalt. After motallographic mounting, the specimens were etched in
50% HN0:} + 30% H2S04 + 20% ^0 and vapor coated with a gold-platinum alloy.
The etching removes cobalt and leaves the oxide structure in relief.
The transverse section indicates that the oxide consists of fibers, narrow
platelets, and a loosely Interconnected network of joined platelets.
Previous x-ray diffraction determinations regarding the crystallo-
graphic structure of the composite subscale formed by internal oxidation
of SmCo5 have been somewhat confusing. More detailed x-ray diffraction
studies following oxidation have now been completed on both powder samples
and flat surfaces )f SmCo inpeots. These studies have been made in pure
oxygon, in air, and in the 90% C02 + 10% CO mixture. Determinations have
been nade on samples oxidized at temperatures from 320 to lOOo'V In all
casej the primary metal produced is fl-cobalt (cubic).
The samarium oxide phase that is obtained during oxidation is either
Sm203 0r an oxynitri(lG depending on the presence or absence of nitrogen.
B-type Sm^^ (monoclinic) is produced at all oxidation temperatures. Pure
Sm 0 exhibits the monoclinic structure above a50-900OC, and C-type Sm 0 * 0 2 3
Preceding page blank
imM—tlh MU^~ , .
![Page 15: Best Available - Defense Technical Information · PDF fileApproved for public release; distribution unlimited. Reproducpcl by NATIONAL TECHNICAL INFORMATION SERVICE U S Daparlmtnl](https://reader031.vdocuments.mx/reader031/viewer/2022030413/5a9f16157f8b9a7f178c57e4/html5/thumbnails/15.jpg)
(a) TRANSVERSE
IUHI M >. I
lb) LONGITUDINAL TA B731 87
FIGURE 1 COARSE FIBROUS STRUCTURE OF SAMARIUM OXIDE IN THE OXIDE/COBALT COMPOSITE SUBSCALE GROWN AT HIGH TEMPERATURES,
1150'C
IMMM^M^^* ■ ■ -'"•"'—~- . ... ,.
![Page 16: Best Available - Defense Technical Information · PDF fileApproved for public release; distribution unlimited. Reproducpcl by NATIONAL TECHNICAL INFORMATION SERVICE U S Daparlmtnl](https://reader031.vdocuments.mx/reader031/viewer/2022030413/5a9f16157f8b9a7f178c57e4/html5/thumbnails/16.jpg)
• ^^^^^^^^^^^^^^^" ^^^^^^^^^^^^^m^
(cubic) is usually observed at lower temperatures. X-ray diffraction
peak intensities lor the C-type sosquioxide are very weak or nonexistent.
No other samarium oxide phases are observed in the absence ol nitrogen.
After oxidation in air, a suboxynitrido phase, SmN 0 (NaCl structure) 1-x x '
is .ibservod and the Sm 0 peaks are weak. This samarium compound lias been 9
previously prepared and idcntilied by Kelmlee and EyriiiK." Electron-beam
microprobe (EDM) analyses on air-oxidized samples have indicated that the
actual oxygen content oi the subscale is somewhat less than that required
stoichiomotrically for Sm 0 and somewhat greater than that required
stoichiometrically for SmO, This finding is consistent with the presence
of both the sesquioxide and suboxynitrido. However, nitrogen was not
detected by EBM analysis.
X-ray diffraction spectra taken after oxidation of subscale at 1000OC
in air and oxygen are shown in Figure 2. The shift from the sesquioxide to
the suboxynitride caused by nitrogen in the air oxidation tests is clearly
evident.
B- Kinetics of Internal Oxidation of SmCo_ at Elevated Temperatures ' O ' .....
The oxide subscale thickness increases parabolically with time
indicating a diffusion-controlled rate-process. The kinetics of this
process for both SmCor and PrOo are identical and have been studied at
elevated temperatures using a metallographic technique on small sectioned
cast pieces.
The temperature dependence of the internal oxidation rate follows
the Arrhenius law. Further work at higher temperatures conducted during
this period has shown that there Is a break in the Arrhenius plot, with
a lower apparent activation energy for the oxidation rate at higher
temperatures. The break in the curve occurs at about 65ü(>C (see Figure 3).
The apparent activation energies are 14.0 kcal/mole at lower temperatures
tmm ■ I^MMM-
![Page 17: Best Available - Defense Technical Information · PDF fileApproved for public release; distribution unlimited. Reproducpcl by NATIONAL TECHNICAL INFORMATION SERVICE U S Daparlmtnl](https://reader031.vdocuments.mx/reader031/viewer/2022030413/5a9f16157f8b9a7f178c57e4/html5/thumbnails/17.jpg)
r ^^^^ ^^^^■^^^^
z $ o cc (9 </) LU
< o CQ 3 <j)
LU a X o
i F oc O LL
< z IT LU i- a u >
I- 0. Ü < H OC —
U. Q Ü z Q < > a:
er ^
(N
LU tr 3 a
taHM^^-^M
![Page 18: Best Available - Defense Technical Information · PDF fileApproved for public release; distribution unlimited. Reproducpcl by NATIONAL TECHNICAL INFORMATION SERVICE U S Daparlmtnl](https://reader031.vdocuments.mx/reader031/viewer/2022030413/5a9f16157f8b9a7f178c57e4/html5/thumbnails/18.jpg)
- < "«ippw*nmn^npm "■IMIIP-IIIRJIPU
00 i- I
r- (fl LL.
r- < h <
CE LU a.
ID ?
I
I
n H <
o n 1-
< •- Q
O X o
IN 1 _i
t < 2
o OC •" LU (-
LU
I
o QC
O
—l «)
z LU I cc
5 n
ac D
B/l inoy luoiDiuj - givb NOIiVaiXO 1VNH3iNI OnOBVdVd
— - mm
![Page 19: Best Available - Defense Technical Information · PDF fileApproved for public release; distribution unlimited. Reproducpcl by NATIONAL TECHNICAL INFORMATION SERVICE U S Daparlmtnl](https://reader031.vdocuments.mx/reader031/viewer/2022030413/5a9f16157f8b9a7f178c57e4/html5/thumbnails/19.jpg)
■PTW in im 1" *"»»*m.immm
and 7.7 kcal/mole at higher temperatures. Each dat"ni point shown in
Figure 3 represents a .riplicate set of runs.
With two exceptions, all the data shown in Figure L are from runs
conducted in one atmosphere of air or oxygen. The remaining two data
points were obtained from runs conducted in a flowing gas mixture of
mr CO + 101 CO. The chemical potential of oxygen in this gas mixture 2
is below that required for oxidation of cobalt; see Figure 4. Conse-
quently, the usual thin oxide scale containing cobalt oxide over the
subscalc is not formed.
The markedly reduced oxygen potential in the 90% C02 + 10% CO mixture o
caused a small decrease in the subscale oxidation kinetics at 840 C and
nt J25"c. In both cases the rates wen; reduced to about half the respective
rut.'s obtained at the same temperatures In air. As determined from these
limited data, the parabolic oxidation rate dependence on oxygen partial
pressure is summarized in Figures 5 and 6. If the exterior samarium oxide
in the subscale is in equilibrium with air at one atmosphere, then the
increase in the parabolic rate constant with oxygen pressure is shown by
Figure 5. However, if the exterior samarium oxide in the subscale formed
by air oxidation is assumed to be in equilibrium with CoO, the oxygen
pressure dependence is shown by Figure 6. In both cases the oxygen pressure
dependence is very weak. Generally for parabolic oxidation of metals, it
is related to the point-defect transport mechanism in the oxide. Although
not enough is known about the defect mechanism in rare earth sesquioxides
to draw firm conclusions, the observed oxygen pressure dependence is less
than that encountered for any of the usual defect mechanisms responsible
for ion diffusion in sesquioxides.
C. Kinetics of Oxidation at Low Temperatures
Oidinarily, the oxidation of magnet powders occurs during handling
at room temperature, or at slightly elevated temperatures resulting from
.. . _ Mumm ■ ■ ---■■■ ■
![Page 20: Best Available - Defense Technical Information · PDF fileApproved for public release; distribution unlimited. Reproducpcl by NATIONAL TECHNICAL INFORMATION SERVICE U S Daparlmtnl](https://reader031.vdocuments.mx/reader031/viewer/2022030413/5a9f16157f8b9a7f178c57e4/html5/thumbnails/20.jpg)
^T- ■-""^^ ■,■ w^WJ+ «,,^(p.^ivH, .,„,,MWA„ <Tpii. ^im-inHi^iiui «iiKMIIlMipiMlfflpiJiPAUlL-ll.liMi I i^I^^pHB^pjiPHIWiriwil^Ii»! IW"''!^»."^« * l^'Wl^WIBl.'^^HlHi,1
500 1000
TEMPERATURE
1500
TA-8731-79
FIGURE 4 OXYGEN DISSOCIATION PRESSURES FOR CoO IN A 90% C02 + 10% CO MIXTURE
^^-^ niiMiirihMlili il i i wmi
![Page 21: Best Available - Defense Technical Information · PDF fileApproved for public release; distribution unlimited. Reproducpcl by NATIONAL TECHNICAL INFORMATION SERVICE U S Daparlmtnl](https://reader031.vdocuments.mx/reader031/viewer/2022030413/5a9f16157f8b9a7f178c57e4/html5/thumbnails/21.jpg)
Vl^.^lU^i'l'MJBfii^^ I ■■'^ l ! rfllBUIID^Ät^T^ i.1i1.-üijipiiwjiiiiJ *jiiii^npppi9p^.^>i-iiin^.4if7^piBgpppp^iBP^s^miniii^i!piippiiiPii«ftippii< i n.\ . Liiji^vmtHttmMvmm^Bimrt
30
26
22
UJ t-- 4 cc
y is j o CO
< < 'L 1.4
! 0
90% C02
10% CO
1 1 1 1 1 1 1 i i
y R*nur. — ►
/■ r^r"^ AIR
525"C -
^ir
i I I I L -26 -24 -22 -20 -18 -10 -14 -12 -10 -8-6-4-2
log P - atm 2
SA 8731 86R
FIGURE 5 INTERNAL OXIDATION RATE DEPENDENCE ON OXYGEN PRESSURE (AIR EOUILIBRIUM AT SURFACE)
U)
- -■- - ■ ■-
.^^ - - »in
![Page 22: Best Available - Defense Technical Information · PDF fileApproved for public release; distribution unlimited. Reproducpcl by NATIONAL TECHNICAL INFORMATION SERVICE U S Daparlmtnl](https://reader031.vdocuments.mx/reader031/viewer/2022030413/5a9f16157f8b9a7f178c57e4/html5/thumbnails/22.jpg)
■^^?^^'■—»—■^^BlWtPf^WlH Ul -1 >l-1 11 Wip.llH Jj'I II LI »L1 iH l]| I UIJ l-WWI^w^ I I IIIHHPHWWIIIII I ■ 11 ^»PWPIWW^ 1 • i '■'■■ ' ■ ^m^mmnißigm
30
26
1 Ü
1—^—r
840 C
i I I I \ I \ \ \ L -26 -25 -24 -23 -22 -21 -20 -19 -18 -17 -16 -15 -14 -13
log P atm 2
TA 8731 80H
FIGURE 6 INTERNAL OXIDATION RATE DEPENDENCE ON OXYGEN PRESSURE (CoO EQUILIBRIUM AT SURFACE)
1 1
M* _- ^ i ■ ii
![Page 23: Best Available - Defense Technical Information · PDF fileApproved for public release; distribution unlimited. Reproducpcl by NATIONAL TECHNICAL INFORMATION SERVICE U S Daparlmtnl](https://reader031.vdocuments.mx/reader031/viewer/2022030413/5a9f16157f8b9a7f178c57e4/html5/thumbnails/23.jpg)
ii m mm ■*■■ i^^mimmm wmr ^««^Hn>Mimf1IW»>iia'l< IW .1 . IUUI
the exotliermic oxidation reaction. Therefore, it is particularly
important that kinetic data bo obtained at very low temperatures so
that the extent of oxidation as a function of particle size, time, and
temperature can be predicted. Previous attempts to measure small amounts
of oxidation using a zlrconiu electrochemical cell, which is very sensi-
tive to small amounts of oxygen consumption at low oxygen pressures, 3
have been reported but the results were not very satisfactory. A review
of those data after accumulating more knowledge about ^idation at higher
temperature indicates that the ma.jori'y of these runs represent measure-
ments in such small amounts of oxygen that adsorption and desorption
effects were probably dominant. As a consequence, additional oxidation
kinetic studies with the zlrconia coll apparatus have been conducted in
higher oxygen pressures, although the oxygen pressure is still sufficiently
low that differences in oxygen pressure can be measured and used to o
detemine the oxidation kinetics below 200 C.
An electronic microbalance has been used to measure the oxl'' .ion o
kinetics gruvimetrically in the temperature range from 150 to 250 C. The
tliei-mogravimetrie data are shown in Figure 7.
For both the gravimetric und oxygon consumption methods, the results
have boon obtained using a single batch of SmCor particles with an average 0
particle diameter of 25 micrometers. This diameter was determined by
photographing and measuring numerous particles and then averaging tho diame-
ter. In those experiments the extent of oxidation was limited to an oxide
scale thickness very much smaller than tho particle radius. Conseqoontly,
one dimensional planar diffusion was an adequate approximation in calcu-
lating tho diffusion-controlled oxidation rate.
Tho weight gain and oxygon consumption curves obtained from these
experiments wore parabolic. The results were converted into equivalent
oxide scale thickness rates for comparison with tho oxide scale thickness
rates determined by the sectioning technique used at higher temperatures,
12
![Page 24: Best Available - Defense Technical Information · PDF fileApproved for public release; distribution unlimited. Reproducpcl by NATIONAL TECHNICAL INFORMATION SERVICE U S Daparlmtnl](https://reader031.vdocuments.mx/reader031/viewer/2022030413/5a9f16157f8b9a7f178c57e4/html5/thumbnails/24.jpg)
■„im,„ —w.,, in,,,imnjunnpiiui^nnmiiT.njiiii.juuiiti ,inimnimp^pmM^Fwn^^mpailf« ,iil,»^>jlll,mill.ll!i.jlli|i i, iiiiililiJIIl., «IIIIJIII llull..l<lin^pi|Kippm(Uf|iM>nUi
14 i—
FIGURE 7 THERMOGRAVIMETRIC DATA FOR INTERNAL OXIDATION OF 25-//m SmCot PARTICLES
13
. i
![Page 25: Best Available - Defense Technical Information · PDF fileApproved for public release; distribution unlimited. Reproducpcl by NATIONAL TECHNICAL INFORMATION SERVICE U S Daparlmtnl](https://reader031.vdocuments.mx/reader031/viewer/2022030413/5a9f16157f8b9a7f178c57e4/html5/thumbnails/25.jpg)
JIT- tliM^IIIW.II.pilll.llJJIWI'll'I'l"1''11''■■»'■''""'''■lllllllll».«!».!»»!! llll.lllll.ll»»W)'J»W>P«'''■ -IIIIJI I limUpWflfPf.ll »I'" »TIMilli. l.llPil^w»»»'»!-»»'l"ll,.'l!W'IWI^TWIWWi,.l I1 . I,«1IIIH»I
The oarabolic oxidation rates for the low temperature experiments
are shown in Figure 8, The straight line is an extrapolation of the
Arrhenius plot from the high temperature data of Figure [i. These data
show that the same oxidation mechanism observed at higher temperatures
continues down to the lowest temperature at which rates can be measured,
100 C. The kinetics for the oxidation process shown in both Figures 3
and 8 span a range of oxidation rates that is seven orders of magnitude,
o o between 100 C and 1125 C.
U, Expected Temperature Rise Associated with Low Temperature
Oxidation
The adiabatic temperature rise caused by partial oxidation of SmCo 0
has been calculated. However, the assumptions for this calculation need
to be emphasized. First, samarium is selectively oxidized to samarium
sesquioxide and hence all of the measured oxygen in a sample can be
assigned to Sm 0 . Second, the sensible heat of oxidation is uniformly
distributed in the alloy particle instantaneously; rapid heat transfer
occurs in the metal and the particle surface temperature, wheve oxidation
takes place, is no greater than the interior temperature. This approxi-
mation is probably adequate for a moderately slow oxidation process and
small particles such as those Involved in preparing sintered magnets. The
third assumption is that no heat is lost from the oxidizing metal system;
the aggregate alloy powder is an adiabatic system. This approximation is
probably valid for a large volume of particles and oxidation occurring
fast enough that heat transfer out of the crucible or particle container
is not significant.
The temperature rise will be proportional to the amount of oxygen,
the specific heat of SmCo , and the heat of formation of Sm O , The 5-l -1 2 J
specific heat is 0.1 cal g deg , which was estimated from an assumption
-1 -1 that the heat capacity is 7 cal g-atom deg . The heat of formation of
o 4 Sm O, at 298 K is 433,900 cal/molc,
14
m*m mm ^MM
![Page 26: Best Available - Defense Technical Information · PDF fileApproved for public release; distribution unlimited. Reproducpcl by NATIONAL TECHNICAL INFORMATION SERVICE U S Daparlmtnl](https://reader031.vdocuments.mx/reader031/viewer/2022030413/5a9f16157f8b9a7f178c57e4/html5/thumbnails/26.jpg)
'i.'« "»■W*1»miWI|fPW.lM'"n"pWW!^WW((«*! mm iwinit ■ i uimpitpiH.aniii iiuiwu iiii.iiiwiawii •— .1. . . . iii u.iK-paa^ain
15
Experimental:
• ■ Microbalance
A - Oxygen Consumption —
FiGURE 8
10 4/T — V
ARRHENIUS PLOT FOR THE INTERNAL OXIDATION RATE EXTENDED TO LOWER TEMPERATURES
"-'••'-' '■;
![Page 27: Best Available - Defense Technical Information · PDF fileApproved for public release; distribution unlimited. Reproducpcl by NATIONAL TECHNICAL INFORMATION SERVICE U S Daparlmtnl](https://reader031.vdocuments.mx/reader031/viewer/2022030413/5a9f16157f8b9a7f178c57e4/html5/thumbnails/27.jpg)
I III lIHPILWWV^nV* - ",1" ll"1 ' '■ "' l-. — .-F»-™.^.-. ......... ... „.»«». ..... —r. mi^^tPWip^lH l||l| I ■■ l»l»»MW»»tT»'«l». I I I IIWII
The temperature rise caused by an increase in oxygen is shown in
Figure 9. An increase of 1000 ppm of oxygen causes an aciiabatic tomper-
o ature increase of nearly 100 C.
5 As previously discussed, oxygen analyses in samarium cobalt materials
after grinding generally range from a few hundred to a few thousand ppm
depending on the processing and particle size. This amount of oxygen is
sufficient to account for an appreciable adiabatic temperature rise
o o during oxidation. A 100 C rise in temperatux-e from 25 to 125 C is suf-
ficient to account for a subscale thickness of about 100 A in a few hours.
o Furthermore, a 100 A subscale thickness is equivalent to a 1000-ppm oxygen
* increase for a 3-micrometer-diameter SmCo,. particle. Hence, for the
5
particle sizes employed in sintered magnets, the observed increase in
oxygen is consistent with the particle size, the expected temperature rise,
the kinetics of oxidation, and the expected subscale thickness as a result
of the temperature rise. Although these relationships do not prove that
a temperature rise and the internal oxidation by the diffusion-controlled
mechanism have occurred to match the observed oxygen concentration, they
at least show that all of the observed results are compatible.
E, Oxidation Rate Mechanism
3 It was previously shown that the published data on the oxygen
solubility in cobalt coupled with the rapid rate of internal oxidation
required that the diffusion coefficient for oxygen In cobalt would have
-6 2 o to be 1 x 10 cm /sec at 420 C and even greater at higher temperatures.
In fact, at high temperatures the diffusion coefficient must be greater
-4 -6 2 than those typically encountered for liquid metals (10 to 10 cm /sec).
Consequently, it appeared that the transport of oxygen in the B-cobalt
matrix could not adequately explain the rapid rate of growth of the
subscale.
* See Figures 27 and 28 of Reference 3,
16
■—■ UMMi^ ^^mmrn^. ^^^^M^^^.
![Page 28: Best Available - Defense Technical Information · PDF fileApproved for public release; distribution unlimited. Reproducpcl by NATIONAL TECHNICAL INFORMATION SERVICE U S Daparlmtnl](https://reader031.vdocuments.mx/reader031/viewer/2022030413/5a9f16157f8b9a7f178c57e4/html5/thumbnails/28.jpg)
^™r"-""'FrTF<™w«»w™«»'!*'™''«'™"'!"»r™w<^
200
0 200 400 600 800 1000 1200 1400 1600 1800
OXYGEN — ppm ,. „„, nn rA-8731-e2
FIGURE 9 ADIABATIC TEMPERATURE RISE ACCOMPANYING PARTIAL OXIDATION TO Sm203
17
mmm^^M,, . tmf'---~~ - ■
![Page 29: Best Available - Defense Technical Information · PDF fileApproved for public release; distribution unlimited. Reproducpcl by NATIONAL TECHNICAL INFORMATION SERVICE U S Daparlmtnl](https://reader031.vdocuments.mx/reader031/viewer/2022030413/5a9f16157f8b9a7f178c57e4/html5/thumbnails/29.jpg)
^JfOIMMI in i iipiMia '■ >'»iii'i mi ■ IM--Ii' . i- IIIWIMILimil ivi ' ^^ W11P»WI«111I 11 ■■•■
A literature search to determine the available information on
diffusion of oxygen in B-type and C-type rare earth sesquioxides has
been made and it presently appears that oxygen anion transport within
the oxide fibers is also too slow to account for the rapid growth of
the subscale. Kofstad provides an adequate review of the existing
diffusion literature for the rare earth oxides.
A study of self-diffusion of oxygen in D-type Sm 0 has been made
6 o by Stone over a temperature range between 693 and 907 C using an isotopic
exchange experimental technique. The activation energy for diffusion
was about 23 kcal/mole, and the data are summarized in Figure 10. Wirkus,
7 8 Berard, and Wilder ' have rcoxidized partially reduced rare earth
sesquioxides, and although they have not studied Sm 0 , they have obtained
diffusion coefficients of the same order of magnitude in the same tempera-
ture ranges as the diffusion coefficients determined by Stone. Their
implicit assumption is that these oxides are semiconductors, whereas Tare
9 and Schmalzried find them to be ionic conductors at low oxygen pressures
o (C-type structure). At 800 C the diffusion coefficient of oxygen in
-10 2 -1 Eu 0 determined by Wirkus and Wilder is D = 2.27 x 10 cm sec .
2 3 0 The value obtained from Stone's data for oxygen diffusion in Sm 0 at
o -10 2 -1 800 C is D = 2.76 x 10 cm sec .
0
If the rate of internal oxidation of SmCor is determined by the flux 5
of oxygen anions through Sm 0, platelets, then one can deduce the product
D AC from the parabolic rate constant. It is instructive to make this
calculation on the assumption that oxygen transport through the oxide is
rate-controlling at 800 C, and compare the results with the diffusion data
at the same temperature. The derived value of D AC using this hypothesis -7 -1
is 1.67 x 10 g cm sec . Matching the value for the diffusion coef- -10 2 -1
ficient determined by Stone, namely D = 2.76 x 10 cm sec , would 3
require an oxygen concentration difference of about 1000 g/cm . This
conclusion is clearly absurd since the total oxygen concentration is only
18
MHMMM I^IIM itmiiiiiMi v i - -■--^-'liiiiir'rfilliii'üllinifi'ill'iil i
![Page 30: Best Available - Defense Technical Information · PDF fileApproved for public release; distribution unlimited. Reproducpcl by NATIONAL TECHNICAL INFORMATION SERVICE U S Daparlmtnl](https://reader031.vdocuments.mx/reader031/viewer/2022030413/5a9f16157f8b9a7f178c57e4/html5/thumbnails/30.jpg)
• ii w»"r^p^^p^B»pi^w^'i''■" ' ' l,lll" m*mi*\vmmn^mm^w*~>*'i^^mmimmmi*r***'si'*'^ii**imimm***mmimi^t^mi^mi^mmm'^^^
10 8
10'
io-10u-
E
Q0
10 -11
10 12
= 1 1 1 1 1 1 z
1\ ■ ■
I \\
- \ \ \\
= >v\
—
' \
k 99 999% Sm203 -
\ — 99.9% Sm203 I -
._ \
....
— W —
— —
w \\
— \\ — \\ — \\ — w — — -
\
1 1 1 1 1 1 0.6 0.8 1.0 1.2 1.4
103/T — V1
1,6 1.8 2.0
TA-8731-81
FIGURE 10 OXYGEN SELF-DIFFUSION IN B-TYPE Sm^ (AFTER STONE6)
19
JiMtiiiiiiiiHiMmiiiBii II MwyiMM
![Page 31: Best Available - Defense Technical Information · PDF fileApproved for public release; distribution unlimited. Reproducpcl by NATIONAL TECHNICAL INFORMATION SERVICE U S Daparlmtnl](https://reader031.vdocuments.mx/reader031/viewer/2022030413/5a9f16157f8b9a7f178c57e4/html5/thumbnails/31.jpg)
"*'■> ' ■ ■■ ■ »i 11 ii i in mmmmmmfmwmmm^i^r^*^^****'*n ' "vmmmmmrmnrmr'mm i i mmmtw^mmmmirmiimfrtm
1.15 g cm and thu actual driving force could only represent a concentra-
tion difference that is matched by an appropriate point-dofect concentra-
tion difference In the oxide that would be undoubtedly much smaller than
the total oxygen concentration in the oxide. It is quite clear from these
calculations that if oxygen transport through the oxide is the rate-limiting
step, then the diffusivity must be much faster than the diffusivity observed
by Stone and by other investigators involved In the diffusion of oxygen
or reoxidation of partially reduced B-type and C-type rare earth sesqui-
uxides.
Vacuum annealing a partially oxidized sample at a temperature higher
than the oxidation temperature evidently prevents further oxidation when
the sample recnters oxygen. As discussed in the next section, annealing
causes dissolution of tue aligned fibers in the subscale and reprecipitation
of larger, more equidimensional oxide particles. The samarium oxide par-
ticles remain as the dispersed phase in the subscale while 0-cobalt and,
additionally. Sm Co are the matrix phases as a result of high temperature
vacuum appealing.
If rapid oxygen transport through fl-cobalt were rate-controlling,
then continued oxidation would be expected after vacuum annealing and
reexposing the specimen to oxygen at elevated temperatures. Hence, oxygen
diffusion in the cobalt phase is eliminated as the rate-controlling step.
The dispersed oxide phase occupies approximately 40 percent of the
subscale volume and there is scanning microscopic evidence that considei'-
able linking (sintering) exists among the various oxide particles after
vacuum annealing the subscale. Hence, if oxygen anion transport through
the oxide were rate-controlling, it is expected that there would be some
continued oxidation rather than passivation.
There are no consistent phase changes in either cobalt or the
samarium oxides that correlate with the apparent change in activation
20
— - - -*-^-- ■ -■-■-"--"-■■'—
![Page 32: Best Available - Defense Technical Information · PDF fileApproved for public release; distribution unlimited. Reproducpcl by NATIONAL TECHNICAL INFORMATION SERVICE U S Daparlmtnl](https://reader031.vdocuments.mx/reader031/viewer/2022030413/5a9f16157f8b9a7f178c57e4/html5/thumbnails/32.jpg)
r~"—'— i. i nm^^^mm m unjii ii mmmmmpimmmsn***si. •w.'^m^immmm^i^w^nKmimwfmmmiK^^'maßfm^m^^f^rmmmrmrmmi^mmmmmmimfm
energy above Toi/'c. This supports the conclusion that bulk diffusion of
oxygen in neither cobalt nor Sm 0 is fast enough to account for the
internal oxidation rate.
Migration of samarium outward by a vacancy exchange mechanism would
cause the accumulation of internal vacancies and the precipitation of voids,
which are not observed.
The rapid oxygen transport required for internal oxidation may occur
along the interface between oxide fibers and ß-cobalt. Because of the
small dimensions of the composite microstructure of the subscale, a large
amount of interface area is present. This mechanism can account for the
apparent change in activation energy for the internal oxidation rate
at about 700OC. The fiber size and spacing begin to increase with
increasing temperature at about that temperature and the interfacial area
per unit volume of oxide subscale decreases correspondingly. This effect
is shown by comparing Figure 1 (1150OC) with the scanning electron micrograph
shown in Figure 11 for subscale oxidation at 800 C, If it is assumed that
the same transport mechanism is involved at all of the temperatures for
which there are experimental data, then a reduction in the interfacial area
would reduce both the number of effective conduits for oxygen transport and
the apparent activation energy, even though the true activation energy for
surface diffusion and the enthalpy for generating mobile species at the
samarium oxide-cobalt interface remain constant.
Other experimental techniques for elucidating the mechanism of oxygen
transport involved in internal oxidation of SmCo have been attempted
without much success. Ion microprobe analysis following oxidation of SmCo5
18 16
with 99% 0 was not successful because considerable amounts of 0 wore
also observed. The source of the O is not clearly known, but evidently
a considerable amount of oxygen and water vapor adsorb on the sample and
is subsequently desorbed during ion microprobe analysis. This contaminating
21
fa ,.■„,■,...: ,.....„ i II Vi'itilttlrtMntii i -- ------ '
![Page 33: Best Available - Defense Technical Information · PDF fileApproved for public release; distribution unlimited. Reproducpcl by NATIONAL TECHNICAL INFORMATION SERVICE U S Daparlmtnl](https://reader031.vdocuments.mx/reader031/viewer/2022030413/5a9f16157f8b9a7f178c57e4/html5/thumbnails/33.jpg)
- "' ■■'■■■■■ "-p"»i—■■ ■• mim mit^mn^^i in i •mm*mmm*mmmmmm I I I I IIHU IMMM^p
TA-8731-91
FIGURE 11 FIBROUS SAMARIUM OXIDE WITH CRACK IN THE SUBSCALE GROWN AT SOCfC
22
-"■>—*-**^-*— ■Mil I1M IHMTiMn mi
![Page 34: Best Available - Defense Technical Information · PDF fileApproved for public release; distribution unlimited. Reproducpcl by NATIONAL TECHNICAL INFORMATION SERVICE U S Daparlmtnl](https://reader031.vdocuments.mx/reader031/viewer/2022030413/5a9f16157f8b9a7f178c57e4/html5/thumbnails/34.jpg)
.« I. II|I.HI^ I fc,*"*-PT^ .+,.:.»«..—H' i -r— ' -—I^II m u ..1U1|1I||IUII1II- _ , ^PWH^^»«^»*^^'»-™»»"!"^ ^^»
oxygen is probably adsorbed during the transfer of the sample from the
0 furnace chamber to the ion microprobe. Adsorbed oxygen and water
vapor may continuously migrate to the ion impact area by surface diffusion
and then be discharged and picked up by the ion detector. Although
18 16 moderately high levels of 0 could also be detected with the 0 con-
tamination, it was not possible to do marker experiments unambigously,
18 18
using 0 and subsequently detect the location of the 0 in the subscale.
Auger spectroscopy was also employed to investigate the oxide films
formed by low temperature oxidation. These experiments also indicate that
a considerable amount of adsorbed species, some of which contain oxygen,
are migrating to the surface and affecting the results. In both ion probe
analyses and Auger analysis, the oxygen concentration pivfile obtained from
continued sputtering appears to extend deeper into the specimen than is
known to be true from kinetic and metalloeraphic data.
23
![Page 35: Best Available - Defense Technical Information · PDF fileApproved for public release; distribution unlimited. Reproducpcl by NATIONAL TECHNICAL INFORMATION SERVICE U S Daparlmtnl](https://reader031.vdocuments.mx/reader031/viewer/2022030413/5a9f16157f8b9a7f178c57e4/html5/thumbnails/35.jpg)
•~~rr~—~'~~~—~—~~^~—-r^~*mm^^*m^~'—~~~'^^~^~^m^w-m~^^m^mm^mmm^m
III VACUUM ANNEALING PAKTIALLY OXIDIZED SmCo AT SINTERING TBiraRATURES
An example of the result of high temperature annealing in the
absence of oxygen after partial internal oxidation of SmCo is shown 5
in Figure 12. This particular specimen was oxidized 15 minutes in air
o . o at 800 C, and was subsequently annealed 16 hours at 1125 C. A scanning
electron micrograph of the same section is shown in Figure 13. The
original subscale prior to vacuum annealing penetrated to the farthest
extent of the large oxide particles shown in Figure 12 and had the
fibrous microstructure shown in Figure 11. As a result of vacuum anneal-
ing, the fine samarium oxide fibers have recrystallized into large,
roughly equidimensional particles. The subscale contains a conjugate
zone partially depicted in samarium oxide adjacent to the original SmCo 5
core. The few oxide particles in the oxide depleted zone are consider-
ably larger than the oxide particles in the undepleted part of the
subscale. A higher magnification view of the reprecipitated oxide
particles Is shown in Figure 14.
The oxide particles and alloy phases have been identified by electron
beam microprobe analyses. The matrix of the depleted region consists of
Sm D) . This phase results because samarium is transported from the 2 17
SmCo core into the subscale region where it reacts with cobalt. The 5
SmCo core is transformed from a single phase region into an SmCo matrix 5 5
containing Sm Co precipitates as a result of samarium depletion from
the core. The solid state diffusion and precipitation reactions are:
25
Preceding page blank
■mi
![Page 36: Best Available - Defense Technical Information · PDF fileApproved for public release; distribution unlimited. Reproducpcl by NATIONAL TECHNICAL INFORMATION SERVICE U S Daparlmtnl](https://reader031.vdocuments.mx/reader031/viewer/2022030413/5a9f16157f8b9a7f178c57e4/html5/thumbnails/36.jpg)
-OXIIH ■
$ - Co-
SmCor > 5;
AFTER OXIDATION
ii - Co -V* SmjCo^ —
OXIDE ■
OXIDE DEPLETED ZONE
SmCo
(Sm 5 ) AFTER
2C())7i)t)il J ANNEALING
&' I >
TA 8731 85R
FIGURE 12 OXIDE DEPLETION ZONE RESULTING FROM INTERNAL OXIDATION OF SmCo5 FOLLOWED BY VACUUM
ANNEALING, 125X
LT.
![Page 37: Best Available - Defense Technical Information · PDF fileApproved for public release; distribution unlimited. Reproducpcl by NATIONAL TECHNICAL INFORMATION SERVICE U S Daparlmtnl](https://reader031.vdocuments.mx/reader031/viewer/2022030413/5a9f16157f8b9a7f178c57e4/html5/thumbnails/37.jpg)
_-,
FIGURE 13 OXIDE DEPLETION ZONE AT HIGHER MAGNIFICATION (SEM)
27
![Page 38: Best Available - Defense Technical Information · PDF fileApproved for public release; distribution unlimited. Reproducpcl by NATIONAL TECHNICAL INFORMATION SERVICE U S Daparlmtnl](https://reader031.vdocuments.mx/reader031/viewer/2022030413/5a9f16157f8b9a7f178c57e4/html5/thumbnails/38.jpg)
SA 8731 93
FIGURE 14 BOUNDARY OF THE OXIDE DEPLETED ZONE (RIGHT SIDE) WITHIN THE SUBSCALE
2*
![Page 39: Best Available - Defense Technical Information · PDF fileApproved for public release; distribution unlimited. Reproducpcl by NATIONAL TECHNICAL INFORMATION SERVICE U S Daparlmtnl](https://reader031.vdocuments.mx/reader031/viewer/2022030413/5a9f16157f8b9a7f178c57e4/html5/thumbnails/39.jpg)
(1) At vacuum annealing temperature, diffusion reaction:
core + subscale matrix - subscale matrix + core
Mfy + ^ ^ - f ■■a0oi7 + ^c-x^s <1)
where 6 ~ 1.
(2) During quenching, precipitation:
If 6 = 1, then
Smi-xC05 - ( ' " T1 ) SmC05 + ( ? X) Sm2C017 ' (2)
While samarium migrates out of the core, oxygen dissolves and
mlRrates Into the core, causing the reduction In population of oxide
particles In the depleted zone. This phenomenon Is a direct result of
the solubility of oxygen In SmCo at the higher vacuum annealing temper- o
o aturc, 1125 C, Although the vacuum annealing time In this case was
16 hours, Identical results have been obtained with a 3Ü-mlnute vacuum
anneal, which Is Identical to our standard magnet sintering time and
temperature.
The scanning electron micrograph shown in Figure 15 shows essentially
all of the features of this system. The original boundary separating the
subscale from the SmCor core Is shown. Uelow this boundary Is the region
containing a Sm Co matrix and large samarium oxide parllcles. A two-phase
alloy region consisting of a SmCo,. matrix and Sm Co precipitates exists 5 2 17
above the boundary. A relic of the original fibrous subscale is also shown
in this unusuaJ SEM micrograph. A small samarium oxide particle is also
shown in the core region above the boundary line. This particle is pro-
duced by oxygen dissolved in SmCo and subsequently precipitated on cooling. 0
29
____
![Page 40: Best Available - Defense Technical Information · PDF fileApproved for public release; distribution unlimited. Reproducpcl by NATIONAL TECHNICAL INFORMATION SERVICE U S Daparlmtnl](https://reader031.vdocuments.mx/reader031/viewer/2022030413/5a9f16157f8b9a7f178c57e4/html5/thumbnails/40.jpg)
IT i .11 IPf
SA 873! 94
FIGURE 15 BOUNDARY BETWEEN THE OXIDE DEPLETED ZONE AND THE SmCob
com
30
imaum ■ M^——■ I .
![Page 41: Best Available - Defense Technical Information · PDF fileApproved for public release; distribution unlimited. Reproducpcl by NATIONAL TECHNICAL INFORMATION SERVICE U S Daparlmtnl](https://reader031.vdocuments.mx/reader031/viewer/2022030413/5a9f16157f8b9a7f178c57e4/html5/thumbnails/41.jpg)
■ipiiyw^piiniliiuiini iiii.iiii .mimUM-tiLUWIW. ^.»PI'K^'^'^WBWWiPiPWWffWWI'll'lll WWliH'UWUWl.f ■II.'II.'WIWIPI mm i .•■.•••■•.•« li nai i 11 ii ii inn i mi ill II I ,I«|IBIP|
IV OXYGEN SOLUBILITY DURING SINTERING
AND SUBSEQUENT OXIDE PRE-
CIPITATION IN SmCc) 5
Attention is now turned to the problem of oxide dissolution in
SmCo at sintering temperatures and the subsequent redistribution of the 5
oxide on cooling.
The experimental procedure used in this study has been to introduce
considerable amounts of oxygen deliberately into the SmCo system using o
small polycrystalline cast pieces by first partially oxidizing the
relatively oxygen-free alloy. Because of their large grain size and
absence of porosity, the cast alloys are easier to examine metallograph-
ically by SEM and by EBM methods than sintered magnets. An oxidation
temperature of 800 C was chosen because it is substantially below the
sintering temperature, and the homogeneity range in SmCo is believed 5
to be extremely narrow at that temperature. Furthermore, it has been
shown that aging sintered SmCo. magnets at about this temperature (usually 5
900 C) can often be beneficial to the magnetic properties. In particular,
10 an increase in magnetic coercivity is often obtained. After oxidation,
o the specimens were vacuum annealed at 1120 to 1125 C for 30 minutes.
We were initially concerned about obtaining chemical equilibrium through
the entire ingot, which is typically 0.5 cm on edge, and therefore very
long annealing times were used. Subsequently, we found that the standard
sintering time, 30 minutes, is sufficient to provide a macroscopically
homogeneous distribution of oxygen in the SmCo core of such large alloy 5
pieces. The initial alloy had the stoichiometric SmCor composition, i.e., D
there was essentially no depletion of samarium.
The results will be described for specimens after vacuum annealing
at sintering conditions and also after a third heat treatment, i.e,
31
■'■■ :■•■■■ --.■■-i.-..-;n. ,■
uttm ihlini ■"-—-'"-•J-'; --- - ■ "'— --
![Page 42: Best Available - Defense Technical Information · PDF fileApproved for public release; distribution unlimited. Reproducpcl by NATIONAL TECHNICAL INFORMATION SERVICE U S Daparlmtnl](https://reader031.vdocuments.mx/reader031/viewer/2022030413/5a9f16157f8b9a7f178c57e4/html5/thumbnails/42.jpg)
|lliHjmM>umiiiiuw.iwii.|ii,iuiiw^ HI «iMjimnuwpjiiiM << ' ■«■■^^■^ > m« ^*mmmm
aging at 80Ü C for 45 minutes. For convenience wo will refer to the
two types of specimens as "quenched" and "aged," Quenched specimens
are produced by quickly removing the specimen from the furnace, but
cooling occurs in vacuo so that "normalized" would be a more appropriate
classical metallurgical term.
The amount 01 oxygen transferred from the oxide subscale into the
SmCo was crudely estimated by measuring the thickness and the area of 5
the depletion zone. This estimate is .lot very accurate because not all
of the oxide in the depletion zone Is dissolved. From inspection of
Figure 13 and similar photographs it was estimated that 75 percent of
the oxide in the depleted zone was dissolved. The calculation Indicated
o that the solubility of oxygen in SmCor at '125 C in excess of the solu-
bility at SOc/'c is _ 0,35 wt%. It is impossible to put error limits
on this estimate.
Next, the oxide content within the core of ar1 aged specimen was
calculated metallographically. Large oxide particles are produced as a
result of the aging treatment and these oxide particles appear black under
reflected light; see Figure 16. Consequently, a Quantimet image analysis
was performed on several sections and the ar;a percent of oxide was
determinrd, The average area of oxide particles and the limits were
determined to be 2.6 ± 0.4%, The area percent is equtl to the volume
percent for randomly distributed equidimensional particles in a dilute
suspension. Hence, from the volume of oxide in the core obtained after
aging, the excess solubility of oxygen at the sintering temperature was
determined. This calculation involves the implicit assumption that all
of the soluble oxygen in excess of the 800 C solubility limit was pre-
cipitated as measurably large oxide particles during the 45-minute aging o
treatment at 800 C. The precipitation and growth of extremely large
oxide particles during this period, as shown in Figure 16, indicates that
32
k'-A ^^^^ .
■M ----■■■-■-^- -:-.-. »w«..
![Page 43: Best Available - Defense Technical Information · PDF fileApproved for public release; distribution unlimited. Reproducpcl by NATIONAL TECHNICAL INFORMATION SERVICE U S Daparlmtnl](https://reader031.vdocuments.mx/reader031/viewer/2022030413/5a9f16157f8b9a7f178c57e4/html5/thumbnails/43.jpg)
-rr^MM^.^y.x-nmmm-lvmmmm ll i.il.il I ■ .1 i iw^^rwf«. i ■■lllwni^^n^n^P« mmmmmm w*m^mm wmmmmmmmnmQ
SA Rm 91
FIGURE 16 OXIDE PRECIPITATES IN THE SmCo5 GRAINS AFTER OXYGEN SATURATION AT 1125"C AND AGING AT 800 C
■Mi
kjt~am ■ l_llMaKaM —^^tmm jjt
![Page 44: Best Available - Defense Technical Information · PDF fileApproved for public release; distribution unlimited. Reproducpcl by NATIONAL TECHNICAL INFORMATION SERVICE U S Daparlmtnl](https://reader031.vdocuments.mx/reader031/viewer/2022030413/5a9f16157f8b9a7f178c57e4/html5/thumbnails/44.jpg)
--.■-IT'WVW; I,IIIUJIJ,I|IIIUI»IIMII' 'wi:f..iiiMijiiiu«!W1wimiii j I .lllinjFnmpmin>«W1(l,4)l>|i « IIPI|I(.1|J|IIJJ,P1W,I|WII| .UJlllW'fWmilll.HIHIWilllnlJIM wwm^l'aTO W-MFWI u I , •!, .WWBWBTOI
this assumption is valid. The metallographic determination of the o
excess solubility at 1125 C is 0.35 ± 0.05 Wt% (3500 + 500 ppm).
Although the oxide particles observed in the aged specimens were
sometimes larger than 100 micrometers, particles in the quenched specimens
are generally smaller than one micrometer. Any exceptions can usually
be traced back to larger oxide inclusions that were present alter melting.
The oxide particle population is greater in the quenched specimens than
in the aged specimens, as shown by cont.-asting the quenched and aged
sepcimens illustrated in Figure 17. Again, the specimens were castings
subsequently oxidized and annealed at 1125 before quenching. These
micrographs were obtained from surfaces polished using cerium oxide
but not etched. The background also shows a faint relief of larger
ym Co precipitates, which will be discussed in the next section.
However, a large amount of ^ygen known to be present in the specimen
is not accounted for xn the low particle population regions of the
quenched specimen, which is most of the specimen. This oxygen ray occur
as small particles below the resolution limit, as coherent oxygen-rich
zones, or as a supersaturated soluto. The aging process collects the
smallor oxide particles into a few large particles by a solution-
reprecipitation process.
The quenched specimen micrograph of Figure 17 shows a region of
low oxide particle population in the upper part of the figure and a
region of high particle population in the lower half of the figure.
High particle population regions were less common than low particle
population regions in the specimens examined. The boxed area in Figure
17 was picked at random within the high particle population region.
The average oxide particle size is 0.4 micrometer and tho volume frac-
tion of oxide, V , was determined from
Vf = J N d f 6 s p (3)
34
MM ^^Mfcl».! .. Mi . i
![Page 45: Best Available - Defense Technical Information · PDF fileApproved for public release; distribution unlimited. Reproducpcl by NATIONAL TECHNICAL INFORMATION SERVICE U S Daparlmtnl](https://reader031.vdocuments.mx/reader031/viewer/2022030413/5a9f16157f8b9a7f178c57e4/html5/thumbnails/45.jpg)
—■ ,.. ...,1.,.,. [ jiuiu^.iuji, uji>'n><*pf*iiiai<iOTnmmi|i.>iwiiiiii.ii«inw<ni«w''*>iuiii' i »i nimmawm^mmfimimmr^mmiwmmmil'^imimmiim'wnmKmifi "wm
v. -
» i
■
*
• *
.. ^v
■/ V
•
■.»..
* •
. F« * 20/J m
(iit QUENCHED FROM 1125 C
_ *
1
^ >r
, **"* ,
Ibl QUENCHED AND AGED AT 800 C SA HI \\ 90
FIGURE 17 QUENCHED AND AGED SmCo6 GRAINS
35
^^MMM.
![Page 46: Best Available - Defense Technical Information · PDF fileApproved for public release; distribution unlimited. Reproducpcl by NATIONAL TECHNICAL INFORMATION SERVICE U S Daparlmtnl](https://reader031.vdocuments.mx/reader031/viewer/2022030413/5a9f16157f8b9a7f178c57e4/html5/thumbnails/46.jpg)
wimi^Hmiiii. ■ i«HT^v^^iinp mwwwi^FtPH^Brwpwim^piw^ -»-^"^■MW
where N is the particle population per unit area, and d is the mean s P
particle diameter. The particle population in the boxed area was 7 2 3.0 x 10 particles/cm . The resulting volume fraction wus 0.025 (2.5%),
which is very similar to the volume fraction (2.6%) in aged specimens.
Hence, in the high population regions, most of the dissolved oxygen has
precipitated as submicron but visible particles.
The appearance of small oxide inclusions within the primary SmCor
grains of both experimental and commercial sintered magnets has also been
observed. These magnets did not receive an aging treatment. Small rxidc
particles generally appear within the SmCor grains rather than at grain
boundaries. Fracture surfaces have also been examined with a scanning
electron micrograph and oxide particles are rarely observed until the
specimen is etched. Numerous oxide particles which were present in the
SmCo grains are uncovered by etching. Exposed oxide particles in an 5
etched sintered magnet are shown in Figure 18. The dark areas in this
scanning electron micrograph are pores.
~m*m
![Page 47: Best Available - Defense Technical Information · PDF fileApproved for public release; distribution unlimited. Reproducpcl by NATIONAL TECHNICAL INFORMATION SERVICE U S Daparlmtnl](https://reader031.vdocuments.mx/reader031/viewer/2022030413/5a9f16157f8b9a7f178c57e4/html5/thumbnails/47.jpg)
"■■ — m1 »■
SA 871t 97
FIGUF^E 18 SAMAf^UM OXIDE PARTICLES IN A SINTERED SmCOg MAGNET UNCOVERED BY ETCHING
M
![Page 48: Best Available - Defense Technical Information · PDF fileApproved for public release; distribution unlimited. Reproducpcl by NATIONAL TECHNICAL INFORMATION SERVICE U S Daparlmtnl](https://reader031.vdocuments.mx/reader031/viewer/2022030413/5a9f16157f8b9a7f178c57e4/html5/thumbnails/48.jpg)
——-
V PRECIPITATION OF Sm Co IN SmCo 2 17 5
A study was made of the precipitation of the Sm Co phase In 2 17
samarium-depleted SmCor USIIIK cast pieces. There are two stages of
Sm2C"l7 l,ruciPitution ln SmCo as a result of partial oxidation followed
by sinterinj; or vacuum anneal in«. The primary staue results from samarium
depletion caused by reaction between the SmCo core and fi-cobalt In the 5
subscale. This reaction occurs durin« the sinter-anneal inn process de-
scribed by liquation (1). As a single phase alloy is quenched from 1125*^,
precipitates of Sm Co occur within the SmCo. «rain as described by
Equation (2). These precipitates are one micrometer wide and u few
micrometers Long. They often occur in clusters as shown by the ■canning
electron micrograph in Figure 19.
One small samarium oxide precipitate also appears in this mlcroKiaph,
The average SmCo5 «rain size, which is controlled by solidification, was
about 80 micrometers i- ')ur tcjt specimens. There wat no tendency 'or
Sm2Coi7 to PreclPltatc ut thc SmCo «rain boundaries during the primary
sta^e of Sm2Coi7 precipitation after vacuum annealing at sintering tempera-
tures. However, cobalt-rich specimens quenched from the melt exhibit
Sm Co at boundaries of SmCo grains. ^17 5
It should be emphasized that primary precipitation of Sm Co within
the SmCo5 grains occurs whenever excess cobalt is exsolved from a samarium-
depleted SmCo5 phase during cooling, whether the cause of samarium
depletion Is oxidation or not.
Primary Sn^Co^ precipitation is evidently complete on quenching
in vacuo. A series of aging tests for various times ut 800OC using pieces
from a samarium-depleted double molted ingot (not oxidized) showed no
increase in the amount of Sm Co or any apparent rearrangement of Sm Co tu 2 17
as a result of aging.
Preceding page blank
39
^*^k*m
![Page 49: Best Available - Defense Technical Information · PDF fileApproved for public release; distribution unlimited. Reproducpcl by NATIONAL TECHNICAL INFORMATION SERVICE U S Daparlmtnl](https://reader031.vdocuments.mx/reader031/viewer/2022030413/5a9f16157f8b9a7f178c57e4/html5/thumbnails/49.jpg)
SA 8731 98
FIGURE 19 SCANNING ELECTROMICROGRAPH OF Sm?Co,7 PRECIPITATE': IN SmCo&
Id
■HMMHH ^mm
![Page 50: Best Available - Defense Technical Information · PDF fileApproved for public release; distribution unlimited. Reproducpcl by NATIONAL TECHNICAL INFORMATION SERVICE U S Daparlmtnl](https://reader031.vdocuments.mx/reader031/viewer/2022030413/5a9f16157f8b9a7f178c57e4/html5/thumbnails/50.jpg)
1 ■■■II«
The second stage of Sm Co precipitation is caused by rejection
of oxygen dissolved in SmCo and formation of Ba 0 . As already mentioned,
this does not occur to any appreciuble extent during quenching, but It
does occur with aging. During the aging process, both samarium and oxygen
are collected at the large oxide particles and must diffuse over distances
relatively large when compared with the primary SmCor grain size. Further
depletion of samarium from primary grains and rejection of additional
Sm Co results. Micrographs indicate that this secondary precipitation 2 17
of Sm Co occurs at the SmCoE grain boundaries rather than within the 2 17 0
grain i. This is particularly evident in the aged specimcMi micrograph of
Figure 17.
The precipitation of large Sm 0 particles is probably controlled by
rapid grain boundary diffusion of samarium. This process causes bulk
depletion of samarium from the region adjacent to SniCor grain boundaries
and precipitation of Sm^fo in the samarium-depleted zone, which is about
2 microm^ers wide. riix« model is consistent with the r i c ol .-lid state
sintering of SmCo=, a process t mt is also controlled by grain boundary 5
diffusion.
Similar behavior has been observed in sintered magnets. However,
when the primary SmCo grain size is small, ~ 5 micrometers, the Sn^Co^
precipitates are no smaller than obsc.-'ed in cast pieces, and, hence,
nearly as large as the primary grains.
The precipitation of fine Sm Co particles in samarium-depleted
alloys contrasts sharply with the Sm Co phase in samarium-rich alloys.
In the latter case the primary SmCo grains are clean and Bm Co occurs
us separate grains outside the SmCo primary grains.
■11
^^^ • - -- ... ■ mm iaMu>^~aMM
![Page 51: Best Available - Defense Technical Information · PDF fileApproved for public release; distribution unlimited. Reproducpcl by NATIONAL TECHNICAL INFORMATION SERVICE U S Daparlmtnl](https://reader031.vdocuments.mx/reader031/viewer/2022030413/5a9f16157f8b9a7f178c57e4/html5/thumbnails/51.jpg)
v^rmmammm*
VI DISCUSSION OF Sm 0, A.M) Sm Q) INCLUSIONS ANL)
INTHINSIC COERC1VITY IN SINTEHKI) MAGNETS
A. Summary ol Inclusiona
It is now quite clour thut internal surfaces can occur in primary
Sm'"« Rrains of sintered magnets from precipitation of second phase 5
inclusions: Sm Co and SmrO,■ Although the oxygen solubility dotermina-
tions for cast SmCo arc not very precise, it is certain that large 5
amounts of oxygen can dissolve in SmCo. at sintering temperatures. For 5
careful processing operations, where oxygen contamination is kept below
3500 ppm, all of the oxygen will be dissolved in SmCo during sintering. 5
The normal quenching of sintered magnets is expected to leave approximately
one-micrometer oxide particles within the SmCor grains, which has been
verified in sintered magnets. Much of the oxygen in quenched SmCo is
uimccountod for and may bi prc^tnt as oupci turned oxygen, joheie.it
o;,ygen-rich zones, and unresolvablc oxide inclusions. Selective oxida-
tion of the alloy also affects the formation of Sm Co by depleting
samarium and fhifting the composition, and this occurs in two sequential
Bt age n.
Although oxidation cannot be entirely prevented, primary and secondary
formation of Sm Co precipitates can be prevented by having excess sama-
rium present in the powder compact to replace the samarium depleted from
SmCo grains. Oxide precipitates can be removed from the primary magnet 5
grains by an aging treatment that tends to collect the oxygen in a few
large oxide grains. Optimum aging conditions need to bo determined.
D. Inclusions and Intrinsic Coercivity
This section begins with a brief review of coercivity models for high
Preceding page blank 43
■
![Page 52: Best Available - Defense Technical Information · PDF fileApproved for public release; distribution unlimited. Reproducpcl by NATIONAL TECHNICAL INFORMATION SERVICE U S Daparlmtnl](https://reader031.vdocuments.mx/reader031/viewer/2022030413/5a9f16157f8b9a7f178c57e4/html5/thumbnails/52.jpg)
——-"•^PBP
magnetocrystalline unlsotropy magnet materials. For a detailed review,
see LlvitiKston. The highest published i-oom temperature intrinsic
12 coercivity in sintered SmCo. magnets is IL,. = 43 kOe. A sintered
5 1
SmCo magnet with a room temperature intrinsic coercivity of 43 kOe has 5 3
also been produced and measured at SRI. This is only about 15% of the
theoretical coercivity. 2 K/M , where K is the magnetocrystalllne anisot- s
ropy and M is the saturation magnetization. Furthermore, sintered SmCor s J
magnets usually have much lower coerclvitles and SmCor powders and
samarium-depleted sintered magnets usually have intrinsic coerclvitles
well below 20 kOe.
Because of the high mannutocrystalline anisotropy, coherent rotation
<JI domains and curling cannot occur at the comparatively low fields at
which magnetic reversal actually occurs.
Magnetic domain nucleation and domain-wall pinning arc factors that
may control coercivity. Inclusions and possibly other defects such as
stacking faults and coherent zones o! second phase material are expected
to impede domain-wall movement aid Increase OOorolvity« However, thes.e
same defects can act as nucleation sites for new domains and decrease
coercivity. The experimental results show conclusively that processing
operations that lead to Sm,0, and Sm Co inclusions within SmCo grain
boundaries always lower coercivity and convorsely processing operations
thai decrease these inclusions raise coercivity. Hence, nucleation of
domains rather than domain-wall pinning evidently controls coercivity.
Additional arguments against a domain-wall pinning model can be
made based on other e^perimontal information. Sintered rare earth cobalt
magnets are made by aligning the powder (easy axis crystallographlc
parallel orientation). During sintering the crystallographlc orienta-
tion is maintained but the specimen is thermally demagnetized. If
domain-wall motion controlled the coercivity, then magnetization of the
sintered specimen should require a high field similar to H . The ci
M
■^MM MM - . ,__
![Page 53: Best Available - Defense Technical Information · PDF fileApproved for public release; distribution unlimited. Reproducpcl by NATIONAL TECHNICAL INFORMATION SERVICE U S Daparlmtnl](https://reader031.vdocuments.mx/reader031/viewer/2022030413/5a9f16157f8b9a7f178c57e4/html5/thumbnails/53.jpg)
experimental evidence is that magnetization nearly equal to M can be
obtained in a thermally demagnetized specimen at fields much lower
than H . ci
Domains have been observed in a large number of SmCo^ grains in which 5
Sm Co inclusions have been introduced by samarium depletion. Bending 2 17 13
of domain walls around these inclusions or termination of domains at
internal inclusions has never been observed in this laboratory. In all
cases the domain walls seem to be oblivious to the presence of the cobalt-
rich inclusions, A typical micrograph of domains in SmCo grains with 5
Sm, Co precipitates is shown in Figure 20. A small domain apparently
growing out of the grain boundary which is a region of Sm Co can be
seen in this micrograph.
Small oxide particles, ^1 micrometer, such as those shovn in the
quenched specimen of Figure 17 also do not appear to have a preferred
locus coincidental with domain walls and, therefore, are evidently not
pinning aom '..i wuiii. However, we do at' ;ia"L much evlllfenoc on t i , yet.
An inclusion domain-nucleation model can explain most if not all
of the confirmed phenomenological relations between samarium-cobalt
magnet processing and the resulting intrinsic coercivity. By confirmed
phcnomcnological relationships we mean behavior that has been observed
in more than one laboratory or is generally accepted. An example is the
observed drastic decrease in coercivity if excess samarium is not present
during sintering. According to this model, both Sm 0 and Sm Co
inclusions, which may be submicroscopic, can heterogeneously nucleate
new domains. Magnetic coercivity of a grain and of a sintered aggregate
will decrease as the population of small inclusions increases.
Rather than present a lengthy discussion of these phenomena and
their possible association with inclusions, we have chosen to list effects
with postulated causes in Table 1.
45
![Page 54: Best Available - Defense Technical Information · PDF fileApproved for public release; distribution unlimited. Reproducpcl by NATIONAL TECHNICAL INFORMATION SERVICE U S Daparlmtnl](https://reader031.vdocuments.mx/reader031/viewer/2022030413/5a9f16157f8b9a7f178c57e4/html5/thumbnails/54.jpg)
•-^^mrmm^m^rm^^m <*m^mmmi^^^™9^^mfm^^mmmri^^^mt " "\*\"^^m^m^^m^*^m*:^qfi^w*i
SA R731 99
FIGURE 20 MAGNETIC DOMAINS PASSING THROUGH Sm^Co^ PRECIPITATES WITHIN SmCo5 GRAINS
46
mam ÜMmehMfa—^^^-.^.■■■.-.-..:■.,- ^ tjixiim irm
![Page 55: Best Available - Defense Technical Information · PDF fileApproved for public release; distribution unlimited. Reproducpcl by NATIONAL TECHNICAL INFORMATION SERVICE U S Daparlmtnl](https://reader031.vdocuments.mx/reader031/viewer/2022030413/5a9f16157f8b9a7f178c57e4/html5/thumbnails/55.jpg)
■W" ■uNefiviivfU'iii -<u^ifi|.iuuipi.im^iffr?<?iwi • -iiJm''mmmmriimvi-i,ipmmimi^mmmm^'»r** ^^^mmmmm» mmiimfm* 11 11 ""•'" wwm^^^m^^vmmmrmw^mmmm
>, w
• in
H
tn
■p 0 0 p
in •H y
■a 3
X! t) W CD S 3 O XJ in E £ ca in 3 JH > u P -H H P t^ 3 y
lO 0 hfl ■H P rH ca 3 u 0 <H bfl 0 y
0 5 E 0) in
a ■H -P 8 0
0) ca ca
IH •H O 3
•H IH O
•H u in y
E t» JH o o c tN N U t> tn fn 0 y rU y rH in w r-H Ü tu tu e "w B 3 rH •rl 3 P 3 § 3 8 p • 3 73
a ■a 1 U 'H W 0 M +J 0 in y 3 "rl ü y c y ca y
§ G a u ca o Ä in a rH !H c rH y id 0 > N cu m -H tu Ü tu U (N u y E a O bfl y y bfl ■p U
CO w
g E &
h <H hfl O Ü o B •H H 3 ca > in ■o & n 0 <H W in <H 3 H a a w x o y xi • bfl y X p 0 tn
w c « T3 3 0) 3 o s ? -H U 3 P 3 H B c ca •H XI o T3 fH ■H E c in > 0 XI y "O p o P y •H -H a. P c < 0) 0 CD ■H ca d rt in P 3 y h O hfl 3 ? 0 ^ T3 ca
^j ^ JH H u M ca i-t ca E 3 >> 3 O y E 0 y U c. 0) bn hfi a m in o 3 .3 ca a CO X rH A o 0 y
i JJ o C 0 P y y •a o U rH p 3 1/5 3 a y li •3 0) <D u • 0 •H •H T3 tn -p SH rH y O U ■p ■o ca XI
w 0) > U X! E m p in p jt -H Si ca 0 9 &• C 'H c 3 c y u
> M 0) ■o i-H G P co ß •H a X 3 O 3 •H E 3 y in CO tyi 0 CU 0 tu •a in o in 3 3 S ca tn u c ca CO 3 ■p in .-( E 3 -H rH (U •H y 0 0 «, y o bfl in E g >. 3
2 ca n X C •H in Ä K X in w y -H •H m <H in 3 ■H ca ^3 tj £ a) c u 3 « 0) c ca o 73 P +J -H O ca bfl ■H ca in
+J •H <H in IH rH X! y •H ca ca ja y 3 •a c JH E •H
PEJ (D in (U o •H +J >, u tn x y •OH« B 'H 3 y y tn W X! HH H tn -p cow > 3 CD X) O 3 O rH •d ?
•H .H
•H bfl in ü p
s 0) 0)
0N§ rH 'H
O > * ca o p
3 O •H 'H
y y 3
X h ?? in y
•H ■p
■H O
i , H 00 in > E •H in tu in in in JO 3 y tn 0 rH s | >1
in •H CD W in in "O in H 0 >> 3 +J y rH >, 3 bfl c tn H Q <£ C !-i 3 ■H -H •H < rH 4-> rH 3 y ca H O • WH C •H tn 0 rH ^ 0) a in rH ■o X •H Ü in -H ca y 3 •H bfl !-i c ■H ^ 0 tp ca < f—1 01 Ü 0 0) P rH C y ca <H in SH P 3 •H rH y
D. T3 , bO c 0) h ■ d * •rl 'H > B 1H Ä -P ca 'iH « t/1 hi 3 0 •H
iO 0) C in C ■H ■a s 3 w 3 ■o ,Q O O 3 3 "5 b p 3 O ■ h ■p X!
8 B
^ CD P > o
•H 0 x u
■P 3 ca 'H
P y p 5 s 1 -a
y in in y •H 3
X ■8 0 p §
•rH
■a •H ■P g c
tn a ca
e c ca a ■0 O Ü !H ca X a O 0 U <H y y y O 0 3 01 3 c c in w 0 p o ■H •H tu h in p in B • 0 > -o ? rH K E ■H •a P 0 0 bfl
•H •H tn X 0) 6 a h£ •H ■rl ho hfl •rl -H y ca M ca U •H ca h
•rl •H 0 in +J a 0 y X! E 3 a 3 tn 3 3 y +J X E y -H 3 tL X P in rH
in CD • H b£ CO 3 IU 0 tu •H y o -H •H y y o y ca u l/i rH 0 y 3 3 rH
v ■a U C 0) <H in P hfl rH O hfl tn rH M u y +J <H y c rH c a "O rH 3 H 5«
EH D h-i
u ■r-t 2 •H c u U a y & 1 8 U 3 rH y in ^ -p & rt ca Ü B •H Ü P X y M •H 3 in u ca 9 C ■P O y ä ^ 3 ca X E A ,3 y M 3 y u 0 Q. < VH M B) O H M c. QUO « in w P in M E 0 En H J )
O •H S M >
•a H <a o ■0
8 5)
XI bfl
•o y
in y
73 3 p
ü c p c
•H 3 ft -H
E -P o ca
rH Ü IH
•H u
0) c
s K
•l-l •H •H p fi •H 0 bfl b£
•^1 in 0) -H as in
y h y
rH P in ■ny
ca E 1 c
•H •H tS ■P tn
rH u •rl
■p y ca y
y
ca a as
M 3 p
73 y
■^ ^ •^ P 0) y P hfl ca y in ca u
i M
£3
■a •H n G 1
O 0 § u cd
3 •H in g SH
y y p
a c
in 8)
X) H in
3
rH
y
3
y
bfl -H ca
in
a •H in
2 a y
3 •H in
| U) ■H ca
in C3 & § H
3 rH Ml
y y 1 rH 3
E •H
■p
0 g +J +J 0) E IU Si ■P 3 •H y •H 3
E H u a E 1/3 > rH ca •H +J -rl
4) o u a E in ■P •H y 0 ^ W <H d '/I •a y hfl ti 3 «ü bfl SH f >,
3 <H o •H in 3 3 a 3 3 3 0) X P a P
W +J 0) CD -H E •H o a he M 3 •r
u u y rH >, tn •P P o > "O 0 X hfl hfl ■p ■H y x y •H •H o •rl
n g o (U 3 3 y Ä rH 0 in 3 ? N O j] o •H -H hfl as ■P hfl C3
I E •H o to hfl rH C •H 3 "H ? u "O y
H 3 •J)
rH ■H
CD CD 0) in
•H in y
»•HO in
U u
•H 3 g **•* 0
y
1 2 ■p
tu O
4-> a
■H 3 M
IU rH O
y 4-» 3
in ca
•H y u 3 y
y •a
9 •H rH
ca a §
rH H
0 in hu o P •p •H •H u y hfl
3 0 rH
•H BJ o c 2 B P in y 3 iu tn C y in E 00 0 u 3 ■H IH y •H •H
h in
E y 0) 3 ca y •H 3 U V in ia "O p a p a > y +J a 3 hfl y y J3
Ul 2 8) in P
S cd tu U tu in |
tn ca ca SH y
•H 3 •H
rH u
p 3 tn
in 0) w 0) P 0) rH in ■H 2 8.5 ü ß -
hfl 1 •H 0 y
s ■p c ■H
1 ho
a ^ 8 ^ fi t y
3 H
•H
w •H in < B 3 5 ■H
in J3 P M
y y c Q P -H
> o a
ca a 0
x>
47
mmm—•—— ■ ■Illl 'Ill "---
![Page 56: Best Available - Defense Technical Information · PDF fileApproved for public release; distribution unlimited. Reproducpcl by NATIONAL TECHNICAL INFORMATION SERVICE U S Daparlmtnl](https://reader031.vdocuments.mx/reader031/viewer/2022030413/5a9f16157f8b9a7f178c57e4/html5/thumbnails/56.jpg)
< ii.iiwiwwiui.LJii^mvji^wfijp^ii!JIIU|W^^^
VII SINTERING
A. Introduction
Benz and Martin studied the mechanism of sintering of rare-earth
14 cobalt alloys, and postulated that the sintering kinetics are controlled
by the diffusion of samarium atoms in grain boundaries via a samarium
atom-cobalt vacancy cluster exchange mechanism. This conclusion was
based on an observed one-third time dependence of volume shrinkage
calculated f .'om end point densities of samples that initially contained
a liquid phase, and on a comparison of the total volume shrinkage or
density with the alloy composition. Benz and Martin further noted that
a correlation existed between the conditions that produced rapid shrinkage
and the conditions that yielded high intrinsic coercivity, and since
this correlation has important processing implications, it was decided
to follow the solid phase st^ge of sintering of SmCo_ under conditions o
in which kinetic models could be applied thus alloving a study of the
effect of changes in stoichiometry on the kinetics of sintering.
The sintering kinetics of solids have been studied extensively 15 16 17
beginning with Kuczynski and then followed by Kingery and Berg, Coble,
and Johnson and Cutler. 18
Refinements in the assumptions relating the
mathematical models to the physical model have been made primarily by
19 Johnson, From these studies the general equation for solid state
sintering is given as
Preceding page blank
AL
L
Kvno
kTr
m
(4)
49
^.^tttmimm
![Page 57: Best Available - Defense Technical Information · PDF fileApproved for public release; distribution unlimited. Reproducpcl by NATIONAL TECHNICAL INFORMATION SERVICE U S Daparlmtnl](https://reader031.vdocuments.mx/reader031/viewer/2022030413/5a9f16157f8b9a7f178c57e4/html5/thumbnails/57.jpg)
■UPlilli I iuiwuii ^iJHJi||MHfH!JHI|tiUiii!iy||Llpujj|JI|!JliPili|Jii ^pffl mmmmwmimmiwvm. m u • ■ ü 11 i i
where
AL/L = the fractional shrinkage o
V = the surface free energy
Q = the volume transported per atom of the rate-controlling species
D = the self-diffusion coefficient of the rate-controlling species
k = Boltzmann's constant
T = absolute temperature
r = the particle or powder radius
K, n, m = constants that depend on the sintering mechanism.
Table 2 lists the values of the constants versus the sintering
mechanism.
Table 2
SINTERING CONSTANTS FOR EQUATION (4)
Sintering Mechanism K n m Reference
Volume Diffusion 5.34 3 0.49 5
20/V2 3 0.40 2
2 3 0.50 3
Grain Boundary Diffusion 2.14b 4 0,33 5
3b 4 0.33 3
The symbol b in Table 2 is the grain boundary width. The constant m is
usually determined by plotting the logarithm of fractional shrinkage
versus the logarithm of time, and a sintering mechanism is deduced based
on the value of m. This method of determining the sintering mechanism
50
mmmmtm ite — i^M ~—'- ■
![Page 58: Best Available - Defense Technical Information · PDF fileApproved for public release; distribution unlimited. Reproducpcl by NATIONAL TECHNICAL INFORMATION SERVICE U S Daparlmtnl](https://reader031.vdocuments.mx/reader031/viewer/2022030413/5a9f16157f8b9a7f178c57e4/html5/thumbnails/58.jpg)
Is not precise unloss time und IcnKth corrections are applied, because
m varies with small errors in time and shrinkuKe. The errors result
from the inability to heat the sample instantaneously to the sintcrinn
20 temperature. Lr.y and Carter have outlined a relatively ea.sy method
of correctin« the time and 'onuth that involves calculating L from
thermal expansion data and extrapolating a length versus time plot to
Intersect the calculated L value, thus determining the effective initial 0
t ime,
21 Johnson later pointed out that the method of Lay and Carter is
strictly valid only if an array of identically sized spheres are used in
the study, and if it is known that oil of the mass transport is due to
u single mechanism.
The sintering studies described in this report are incomplete in
that a complete range of itoichiometries has not been investigated and
the validity of assuming u single transport mechanism has not been tested.
D. Experimental Procedure
Samurium-cobalt alloy powders were made for this study by arc
melting the elements, Slight changes in the alloy compositions were
obtained by adjusting the samarium/cobalt ratio used in the arc melting.
The arc-melted alloys were annealed in an oxygon-gettered argon atmosphere o
at 1110 C for four hours, and were then individually ground in an alumina
ball mill for three hours using sodium-gettcred hcxane as the milling
fluid. The final alloy compositions were determined from EOTA titration 22
data, ' and the compositions were adjusted for loss of samarium due to
the formation of Sm 0, . The oxygen concentrations in the alloys were
determined by neutron activation analysis and by vacuum fusion analysis
(courtesy of General Electric Company). The oxygon content in solid
solution was assumed to be zero.
51
^■^M
![Page 59: Best Available - Defense Technical Information · PDF fileApproved for public release; distribution unlimited. Reproducpcl by NATIONAL TECHNICAL INFORMATION SERVICE U S Daparlmtnl](https://reader031.vdocuments.mx/reader031/viewer/2022030413/5a9f16157f8b9a7f178c57e4/html5/thumbnails/59.jpg)
Disk-shaped specimens 1.0 inch in diameter by U.13 inch thick,
were fabricated by die pressing at 30,000 psi in the absence of a mag-
netic field. The specimens were placed on a molybdenum support in a
furnace employing a graphite resistance heater surrounded by oxygen-
gettered argon. The change in sample diameter was recorded by time-lapse
photography as functions of time and temperuture, and the data wore
read fron the films by means of a Teleroadex film analyzer.
Sample length corrections were made following the method of Lay
20 and Carter using the thermal expansion data presented in Figure 21,
which were obtained by using a commercial dilatometer modified so that
a static argon atmosphere could be used to surround the sample to prevent
oxidation.
Heating the sintering snmples from room tempi'raturc to the sintering
temperature requiroü approximately 90 seconds. Shrinkage below approxi-
o o mately 1100 C is almost negligible and the time required to reach 1165 C
from 1100 C was approximately 5 seconds. Calculated L values (initial o
sample diameter) corresponded exactly with the experimentally measured
o o L values at 1100 C. At 1120 C a small correction in L was required
0 o and these corrections increased in magnitude as the sintering temperature
o was raised to 1165 C, indicating that shrinkage was occurring as the
o sample was heated from 1100 to 1165 C. The corrections ranged from 0.0%
at 1100OC to less than 2.0% at 11650C.
C. Results und Discussion
Shrinkage data were obtained for four different alloy comparisons,
and these data are shown In Figures 22, 23, and 24. The value of m
(sec Eq. 4) as determined by a least-squares analysis ranged between
0.24 and 0.37. The correlation coefficient except for one isotherm was
above 0.93. Thus the shrinkage data approximate a slope of m = 0.33 and
indicate that the shrinkage is controlled by grain boundary diffusion.
52
i^MMiM —
![Page 60: Best Available - Defense Technical Information · PDF fileApproved for public release; distribution unlimited. Reproducpcl by NATIONAL TECHNICAL INFORMATION SERVICE U S Daparlmtnl](https://reader031.vdocuments.mx/reader031/viewer/2022030413/5a9f16157f8b9a7f178c57e4/html5/thumbnails/60.jpg)
iMWHM
1 4 1 1 1 1 1 1 1 1 >
1 2
§
'—
, ,0 — y
EX
PA
NS
ION
o
as — v^
<
-
< 06
■ Ul z
yr *- 04
< a Z y* ' 3 0.2
0 sS 1 1 1 1 1 J__ 1 1 1 100 200 300 400 5UU 600
TEMPERATUBE - 700 800 900 1000 1100
SA 8/31 103
FIGURE 21 LINEAR THERMAL EXPANSION OF lJOLYCRYSTALLINE RANDOMLY ORIENTED SmCoc
53
■MB^Maa IIHIHBIII ■ h^< ^■M - -*- ^■■■~ -'^ JM
![Page 61: Best Available - Defense Technical Information · PDF fileApproved for public release; distribution unlimited. Reproducpcl by NATIONAL TECHNICAL INFORMATION SERVICE U S Daparlmtnl](https://reader031.vdocuments.mx/reader031/viewer/2022030413/5a9f16157f8b9a7f178c57e4/html5/thumbnails/61.jpg)
^^
8 8
I I z
\
cc O u. in 5 CC
8 HI o < z cc z (/)
< 00 o o T s £ < s < CO
CM CM
CC
o d
0-inv
8 d
54
■ ■ ■ 1111 m liiiiiiinin r
![Page 62: Best Available - Defense Technical Information · PDF fileApproved for public release; distribution unlimited. Reproducpcl by NATIONAL TECHNICAL INFORMATION SERVICE U S Daparlmtnl](https://reader031.vdocuments.mx/reader031/viewer/2022030413/5a9f16157f8b9a7f178c57e4/html5/thumbnails/62.jpg)
^^^w^^^
o
<
I O
UJ
I i i
S
-inv
55
MaM MütfiüitaiaMaii^^kM^MM^MMiMiM •aMMMMMM
![Page 63: Best Available - Defense Technical Information · PDF fileApproved for public release; distribution unlimited. Reproducpcl by NATIONAL TECHNICAL INFORMATION SERVICE U S Daparlmtnl](https://reader031.vdocuments.mx/reader031/viewer/2022030413/5a9f16157f8b9a7f178c57e4/html5/thumbnails/63.jpg)
■ IL-I 1 ' •^m^^
8 S
E 1/5
n CO
n Q 2 <
CO <■>
a z z <
o u to > o
cr O LL
(/) s tr LU
g
X 10
3
UJ
nnv
56
MMM ii »iiii—MiMMinriii ,
![Page 64: Best Available - Defense Technical Information · PDF fileApproved for public release; distribution unlimited. Reproducpcl by NATIONAL TECHNICAL INFORMATION SERVICE U S Daparlmtnl](https://reader031.vdocuments.mx/reader031/viewer/2022030413/5a9f16157f8b9a7f178c57e4/html5/thumbnails/64.jpg)
^^»-v^Mm^ ■■ - •^m
The data presented in FiRUres 22, 23, and 24 were force fitted to lines o
drawn with slopes equal to 0.33, and the data near 1100 C exhibit con-
siderable scatter due to the small amounts of shrinkaRO involved. The
rate of shrinkage can be calculated from the intercept values for each
isotherm, and the rate of shrinkage is shown plotted versus reciprocal
temperaturos in Figure 25, All four alloy compositions exhibit identical
rates ol shrinkage within experimental error and this result is in conflict 14
with the results of Benz and Martin.
SmCo alloy compositions are very sensitive to the oxygen impurity 5
concentration. Therefore thr oxygen content of the alloys used in this
study were determined by both neutron activation analysis and by vacuum
lusion analysis. These methods of analysis for oxygen in SmCo give
conflicting results with the vacuum fusion method always exhibiting the
higher value. The results of the oxygen analysis are listed in Table 3.
Because of the discrepancies shown in Table 3, very brief descriptions
of the two oxygen analysis techniques are given below.
The neutron activation analysis was conducted by Gulf Radiation
Technology Company. Each sample was irradiated for 10 seconds in a 14-MeV 8 2
neutron flux of approximately 10 neutrons/cm -sec. An oxygen standard 16
was irradiated with each sample and the induced activity of N was
counted on a single channel pulse-height analyzer using a pair of
5 x 5-inch Nal(Tl) scintillation crystals. The N is formed by oxygen 16 16
interacting with neutrons according to the reaction 0 (n,p)N . The
oxygen concentration was determined by comparing the intensity of the 16
6.13-MeV gamma-ray photo peak of N from the sample with the same peak
from the oxygen standard.
The vacuum fusion analysis was conducted by General Electric Company
and an excellent description of this technique has been given by Horten
57
MM t^mm mmmmmm*immm .M^MMH .
![Page 65: Best Available - Defense Technical Information · PDF fileApproved for public release; distribution unlimited. Reproducpcl by NATIONAL TECHNICAL INFORMATION SERVICE U S Daparlmtnl](https://reader031.vdocuments.mx/reader031/viewer/2022030413/5a9f16157f8b9a7f178c57e4/html5/thumbnails/65.jpg)
•^wmmmmmm mmmmm **rm
0.10
0.09
0.08
0.07
0.06
0.05 -
0.04 -
0.03 -
i r
^"y 0.02 -
0.010
0.007
• -27.7 w/o Sm
■ 31.4 w/o Sm
A 33.0 w/o Sm
A 33.3 w/o Sm
1/T x 10
FIGURE 25 ARRHENIUS PLOT OF THE SHRINKAGE RATE VERSUS RECIPROCAL TEMPERATURE
58
mmmm ^^^aMMMMi
![Page 66: Best Available - Defense Technical Information · PDF fileApproved for public release; distribution unlimited. Reproducpcl by NATIONAL TECHNICAL INFORMATION SERVICE U S Daparlmtnl](https://reader031.vdocuments.mx/reader031/viewer/2022030413/5a9f16157f8b9a7f178c57e4/html5/thumbnails/66.jpg)
■ II II I III! PI
■
/^ t' b » ■ >^ 0
>> i m 0 5 rH n pH
pH • • t
(—4 pH 0) s 1 < E M M
3 c 3 •^ U
a c > 0 •H
a u *i
G c ■ 0 0 •H e 4->
8 a >
•H m s tf CO c H t*
w Ü • • • • < n n rH tv
•o n n co N • c •p p M K
■H 4-> 3 3
CO U V >> pH 2 O ■ a a o
w 1 ~ <
m c 0 0 o •H E tn
M m 5 *
m u. U • • •
01 O 1 g pH pH o 1
PH ^/ 3 Si w 3 a M c B H
1 0
■H
cs
(4 >
3 < e
0 •rl
% B il H D a
e > g 8 m «0 «f CO Q U (0 O) * CO
c < 00 in m 0)
V • « • • bC e o o o o >> 0 R (H
o -P 3 i c 0 •rl
>. P A U
C9 t* t*
m -o ^ PH p-l
a> u ■H in m m o f 8 «1 -P "H CS U -H 8 8 U M .; o Q B e E E E E a +J w m CO CO W CO
0) >> a d «V —*>»
H 1
X
>> f- 00 OJ o 0 • CO «5 <c t*
pH 0 1 1 1 1 PH K s s ■ S <
59
— —^
![Page 67: Best Available - Defense Technical Information · PDF fileApproved for public release; distribution unlimited. Reproducpcl by NATIONAL TECHNICAL INFORMATION SERVICE U S Daparlmtnl](https://reader031.vdocuments.mx/reader031/viewer/2022030413/5a9f16157f8b9a7f178c57e4/html5/thumbnails/67.jpg)
" " «■...- —^-. .-..-...- . -.-™L. - . J I.,. ,..„,.. L I RMIIM
23 and Cnrson. The technique cssentJally consists of melting the SmCto
D
alloy in a platinum-tin bath in a vacuum in the presence of an excess of
carbon. Carbon monoxide is produced and the amount produced is taken as
a measure of the oxygen content in the alloy. The alloy sample is placed
in a platinum foil-tin container and introduced into a vacuum side arm
above an outgassed graphite crucible containing 95% platinum plus 5% tin
o -5 held at 900 C in a vacuum of approximately 10 t^rr. The sample is
dropped onto the solid ilatinum bath and the tempen'ture is raised to
o o 1950 C requiring approximately 12 minutes to reach 17.>0 C and approxi-
o mately 15 to 18 minutes to reach 1950 C. The quantity of gas evolved is
determined manometrically using McLeod gages.
Oxidation of powder samples or porous samples may occur during the
time the sample resides on top of the platinum bath and this might account
at least in part for the higher oxygen values obtained when analyzing
by a vacuum fusion technique.
Table 3 also lists the phases detectable by x-ray diffraction. If
the oxygen contents determined by vacuum fusion analysis are correct, we
would expect that alloy M-68 would contain a greater amount of Sm Co
than the M-69 alloy. Since this was not confirmed by x-ray analysis,
all of the alloys were examined metallographically. M-67 and M-68 con-
tained very small amounts of second phase and M-69 contained much more
second phase than did M-68, but less than M-70. Therefore, the x-ray data
and metallographic examinations correlate more closely with the neutron
activation data. It should be realized that the sintered specimens sub-
mitted for oxygen analysis were very porous because only the initial
stage of the solid phase sintering process was studied, and the possibility
of oxidation occurring during the vacuum fusion analysis does exist.
Additional work is necessary to determine the best method of oxygen
analysis in SmCo alloys; however, in this report the neutron activation 5
analysis data have been used to calculate the samarium contents of the
60
.—^^^^^^^ . .__. i—m ii—im i : üäi
![Page 68: Best Available - Defense Technical Information · PDF fileApproved for public release; distribution unlimited. Reproducpcl by NATIONAL TECHNICAL INFORMATION SERVICE U S Daparlmtnl](https://reader031.vdocuments.mx/reader031/viewer/2022030413/5a9f16157f8b9a7f178c57e4/html5/thumbnails/68.jpg)
■ •-■""I wmmnmimmmmwwi'iim>inmpmm «i.ii"infim**m~***mmmmmm**™**^v*»*u***v™ ^mm^m^mmm
alloys (see Figure 25). The samarium content has therefore been varied
almost entirely across the SmCo phase field and into the Sm2Co
17 Phase
field with the exception of the postulated hyperstoichiometric SmCo
24 region. Since we do not observe any difference in the ratio of
shrinkage as the samarium/cobalt ratio is varied, we are forced to con-
clude that deviations in stoichiometry in this system do not alter the
grain boundary defect structure sufficiently to cause a change in the
grain boundary transport of the rate-determining species.
Additional work will be completed during the next report period
that will allow a quantitative evaluation of the solid phase shrinkage
rate, including experiments to extend the samarium/cobalt ratio into
the Sm Co phase field. This work is necessary before the final con- 2 7
elusions from the solid phase sintering work are made.
61
mmmmt^mm «a^mfm tmmmm
![Page 69: Best Available - Defense Technical Information · PDF fileApproved for public release; distribution unlimited. Reproducpcl by NATIONAL TECHNICAL INFORMATION SERVICE U S Daparlmtnl](https://reader031.vdocuments.mx/reader031/viewer/2022030413/5a9f16157f8b9a7f178c57e4/html5/thumbnails/69.jpg)
|!IWI<|I|II'I,'III<I'I"<11111 immmmmmmmmmmmm*** r**mm mmum ^mmmmmmmmn
REFERENCES
P. Kofstad, Nonstolchiometry, Dlifusion, and Electrical Conduc-
tivity in Binary Metal Oxides, Wiley-Interscience, p. 28G,
New York, 1972
2. T. L. Felmlee and L. Eyring, J. Inorg. Chem., 1, 660-66 (1968)
3. P. J. Jorgensen and R, W. Bartlett, "Materials Processing of
Rare Earth Cobalt Permanent Magnets," Tech, Report AFML-TR-72-225
(September 1972)
5.
C. E. Wicks, and F, E. Block, Thermodynamic Properties of 65
Elements - Their Oxides, Halides, Carbides, and Nitrides,"
U.S. Bureau of Mines, Bull. 605, Govt. Printing Office (1963)
P. J. Jorgensen and R. W. Bartlett, "Materials Processing of
Rare Earth Cobalt Permanent Magnets," Tech. Report AFML-TR-72-47
(April 1972)
6. G. D. Stone, Ph.D Thesis, Arizona State University, 1968
7. C. D. Wirkus, M. F. Berard, and D. R. Wilder, J. Am. Ceram. Soc.,
50, 113 (1967)
8. M. F. Berard, C. D. Wirkus, and D. R. Wilder, J. Am. Ceram. Soc,
51, 643 (1968)
9. V, B. Tare, and H, Schmalzried, Z. Phys. Chem. N.F., 43, 30 (1964)
10. M. G. Benz and D. L. Martin, J. Appl. Phys., 42, 2786, (1971)
11. J. D. Livingston, "Present Understanding of Coercivity in Cobalt-
Rare-Earths, to be published in AIP Conference Proceedings (1972)
12. K. H. J. Buschow, P. A. Naastcpad, and F. F. Westendorp, IEEE
Trans. Mag. 6, 301 (1970)
13. H. Zijlstra, J. Appl. Phys. 41, 4881 (1970)
Preceding page blank 63
-■ - ■- ■
![Page 70: Best Available - Defense Technical Information · PDF fileApproved for public release; distribution unlimited. Reproducpcl by NATIONAL TECHNICAL INFORMATION SERVICE U S Daparlmtnl](https://reader031.vdocuments.mx/reader031/viewer/2022030413/5a9f16157f8b9a7f178c57e4/html5/thumbnails/70.jpg)
piWiMUIII 1.11)1 .1111 w^****miimmm^m m^^mmmm^nmmm^mm !■ '
REFERENCES (Concluded)
14. M. G. Benz and D, L. Martin, J. Appl, Phys. 43, 3165 (1972)
15. 0. C. Kuczynski, Trans. AIME 185, 169 (1949)
16. W. D. Klngery and M. Berg, J. Appl. Phys. 26, 1205 (1955)
17. R. L. Coble, J. Am. Ceram. Soc. 41^, 55 (1958)
18. D. L. Johnson and I. B. Cutler, J. Am. Ceram. Soc. 46, 541 (1963)
19. D. L. Johnson, J, Appl. Phys. 40, 192 (1969)
20. K. W. Lay and R. E, Carter, J. Am. Ceram. Soc. 52, 189 (1969)
21. D. L. Johnson, J. Am. Ceram. Soc. 52, 562 (1969)
22. P. J. Jorgensen and R, W. Bartlett, "Materials Processing of Rare Earth-Cobalt Permanent Magnets," Tech. Report AFML-TR-71-188, p. 59 (August 1971)
23. W. S. Horton and C. C. Carson, "Treatise on Analytical Chemistry," Vol. 10, pp 6057-6091, edited by I. M. Kolthoff, P. J. Elving, and E. B. Sandell, Wiley-Interscience (1972)
24. K. H. J. Buschow and A. S. Van Der Goot, J. Less-Common Metals 14, 323 (1968)
64
MAMflMMr UMi iiiiMiiiiiiin«-