INVESTIGATiONS N10 THE
PREPARATEON AND DECQMPOSIYION 0F
SAMARlUM TETRABOEEDE AND
SAMAREUM HEXABGRIDE

Thom GM “to Duqmo of DB. D.
IiiECHEGAN STRTE‘. UMVERSITY

Gordon L. Galloway
19 61

 

‘;
- ‘

5

THE

.frolli.vr.i.t.

'ABSTRACT

INVESTIGATIONS INTO THE PREPARATION AND DECOMPOSITION
OF SAMARIU M TETRABORIDE AND SAMARIU M HEXABORIDE

by Gordon L. Galloway

Attempts to prepare pure samarium tetraboride either by the reaction
between samarium sesquioxide and varying amounts of boron or by utilizing
the direct reaction between the elements have been shown to result only in
rmixtures of the tetra- and hexaborides. No evidence whatsoever was
obtained to suggest the stable existence of a lower boride such as samarium
diboride or a higher one such as samarium dodecaboride.

The extent of contamination of the products of any reaction used to
prepare a preponderance of one or the other of the borides was shown to
be a function of the starting composition of the reactants, the temperature
and duration of the heating, and, in some cases, the nature of the heating
cell used to confine the reactions.

Changing only the starting composition of mixtures of samarium
sesquioxide and boron examined at 16500 in a crucible of boron nitride
resulted in a smoothly continuous variation in the ratio of boronto oxygen
lost on heating. This result, coupled with accompanying weight data, has
been interpreted as evidence of the non-validity of any single equation
to describe the reaction simply.

A survey was made of some of the materials suitable for. use as
crucibles at these high temperatures and details of this study are pre-

sented in the thesis.

Gordon L. Galloway

The apparent stability of samarium tetraboride and samarium
hexaboride at various temperatures was shown to be dependent upon the
matrix material with which these borides were in contact.

The existence of the equilibrium:
3 SmB4(s) = Z SmB6(s) + Sm(1, g)

was ultimately verified as a result of observations on the diffusion and
evaporation phenomena occurring within the system.

When samarium tetraboride-hexaboride mixtures were heated in a
crucible of boron nitride, the top surface of a sintered compact was observed
to become depleted in samarium tetraboride and rich in samarium hexaboride;
complete conversion was brought about by continued heating. However, when
in contact with a matrix capable of absorbing boron into its interstices at
a sufficiently high rate, a sintered mixture of samarium hexaboride was
found to contain samarium tetraboride on its top surface after being heated
at temperatures higher than those at which this tetraboride is normally
unstable with respect to the hexaboride. These observations have been
explained in terms of a competition between equilibrium, rates of evaporation
of samarium from the heating cell and rates of diffusion of samarium to

evaporation sites .

INVESTIGATIONS INTO THE PREPARATION AND DECOMPOSITION
OF SAMARIUM TETRABORIDE AND SAMARIUM HEXABORIDE

By

Gordon L. Galloway

A THESIS

Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of

DOCT OR OF PHILOSOPHY

Department of Chemistry

1961

ACKNOWLEDGEMENTS

The author wishes to express his sincere thanks and appreciation
to Dr. Harry A. Eick for his guidance, encouragement, and cooperative
tolerance throughout the course of these investigations.

Appreciation is also extended to my collegue, James J. Mattis,
for his singular and devoted interest in a problem unrelated to his own
work and for his many perceptive and worthwhile suggestions.

‘ Thanks is also due to Mr. Gene Hood for all forms of his help,
to Mr. Frank Betts for his advice and assistance in the machine shop,
and t0»Mr. Paul Julien for the tedium of reading many x- ray films.

Finally a large token of gratitude is extended to the Atomic. Energy
Commission, to the Monsanto Chemical Company, and to the National
ScienceFoundation for their generous financial support.

**********>fi<**>§<***Jk

ii

I. INTRODUCTION ........................
:II. HISTORICAL ..................... . .....
A. Preface ..................... . . . .
B. The hexaborides ....................
C. The tetraborides . ...................
' D. The dodecaborides . . .................
E- Other borides ......................
F. . Methods of preparation ...................
1. Direct reaction of metal with boron .......
2. Reactions of rare earth sesquioxides with
boron carbide ...................
3.. Reactions of rare earth sesquioxides with boron
4. . Other methods of preparation ..........
G. Properties of the rare earth borides ..........
III. . EXPERIMENTAL . ............ . ......
A. Materials ........................
' B.. Sample preparation ................ . . . .
C. Analytical methods ...................
Samarium .....................
2.Boron ...... .. .....
D. . Reactions of samarium sesquioxide and boron in a
molybdenum crucible ..................
E. The samarium-boron system in almolybdenum
crucible . ........... . ............
F. The reaction between samarium sesquioxide and
boron inaboron nitride liner. . . . . . . . . .....
G. The reaction between samarium sesquioxide and
boron in a zirconium diboride crucible ...... . . .
H. . The heating of a sample of samarium hexaboride in a
low-density tungsten cell ....... . . . ......
I. Miscellaneous experiments ..........

TABLE. OF CONTENTS

' Page

iii

U.)

12
13
15
16

19
19
20
24
. Z4
25
27
33
37
39

41
45

TABLE OF CONTENTS - Continued

Page

IV. DISCUSSION .......................... 48
A. Crucible materials ................... 48

B.. Results and conclusions ................ 57

C. . Proposals for future work ............... 7O
LITERATURE CITED ....................... 73
APPENDICES . . . . ....................... 77
A. Supplementary experimental data ........... 78

B. Tabulation of physical constants ............ 95

iv

TABLE

II.

III.

IV.

VI.

VII.

VIII .

Pi

XI.

' XII.

XIII .

LIST OF TABLES

The samarium-boron system in a molybdenum
cruCible O 'O O O I O O O ..... O O O O O O O .....

The reaction between samarium sesquioxide and boron
in. a boron nitride liner .................

The reaction between samarium sesquioxide and boron
in a zirconium diboride crucible .......... . .

The heating of a sample of samarium hexaboride in a
low-density tungsten cell ......... . ......

. Melting point and vapor pressure data for rare earth

metals ..... . ....................

. Metallic and trivalent ionic radii of the rare earth

metals ............... . ........ . .

Lattice parameters of cubic rare earth hexaborides. .

. Lattice parameters of tetragonal rare earth tetra-

borides ................ . ......

. Lattice parameters of some cubic dodecaborides . .

Lattice parameters of tetragonal rare earth MBX
phase 0 C 000000000000 O O O O O O O O O I 0

Comparison of the interatomic spacin s of an experi-
mental sample with those of the 1550 polymorph of
samarium orthoborate . . r ............ . .

Comparison of the interatomic spacings of two experi-
mental samples with those of the delta polymorph of
molybdenum boride . . . . . . . . . ........

Raw data used for correcting observed temperatures
to true temperatures ..................

Page

36

4O

42

44

96

97

98

99

100

100

101

102

103

FIGURE

II.

III .

IV.

VI.

VII.

VIII .

XI.

LIST OF FIGURES

Radii of rare earth metal atoms and lattice parameters
of rare earth hexaborides versus atomic number . .

The structure of a cubic hexaboride (MB6) . . . . .

The structure of a hexagonal diboride (MBZ) . . . . . . .

The structure of a tetragonal tetraboride (MB4) - . .

The structure of a cubic dodecaboride (MBIZ) . ..

Heating cell assembly . . . . ............. .

Variation of pH in the vicinity of the end-point of
mannitol titrations ....... . . . . ........

Mole ratios of boron to oxygen and boron to samarium

versus composition of starting materials heated in

boron nitride . .. . ....................

Weight losses versus composition of starting materials
heated in boron nitride ..................

.Mole ratios of boron to oxygen and boron to samarium

versus composition of starting materials heated in
zirconium diboride .............. . . . . . .

Weight losses versus composition of starting materials
heated in zirconium diboride .............

vi

Page

10

22

28

60

60

62

62

I. . INTRODUCTION

Although very little is known about the actual chemistry of the
rare earth borides, they are nonetheless important from the stand-
point of two potentially useful considerations. As a first instance, since
both the rare earth elements and boron are individually effective neutron
moderators, some attention has already been devoted to the possibility
of combining these elements into control rods of refractory borides.
However, there must first be a collection of sufficient information
regarding the engineering involved in using the borides as well as an
assimilation of more specific details about their behavior at high
temperatures.

As a second instance, the trend toward higher and higher frequencies
has confronted the tube designer with the need for larger current densities
at the surfaces of cathode materials. . By nature, metals such as cesium,
barium, cerium, and thorium are not only the best electron emitters,
but also evaporate most readily. Therefore, operation of these metals
at sufficiently high temperatures to produce electron emission of high
current density causes so rapid an evaporation rate that the cathode be-
comes useless in high vacuum devices.

Because of the unique nature of the bonding in rare earth hexa-
borides, these compounds have not only high electrical conductivity, but
high thermal and chemical stability as well. If the borides can be
heated to temperatures sufficiently high to cause evaporation of the metal
atoms at the surface, and if these metal atoms can be replaced by diffusion
from underlying cells without the destruction of the boron framework,
then a mechanism is provided for constantly maintaining an active cathode

surface.. Such characteristics are ideal for cathode materials; thermionic

emissionproperties can be maintained without the loss of the very
materials which contribute to their high thermionicactivity. Lafferty (1)
has performed an extensive investigation on the boride cathodes of cal-
cium, strontium, barium, lanthanum, cerium, and thorium, with
particular attention devoted to lanthanum -hexaboride (LaBs). He has
convincingly demonstrated that when lanthanum hexaboride is heated in
contactwith a material capable of absorbing boron into its interstices,
the lanthanum atoms which evaporate from the surface are replaced by
A diffusion of the metal atoms from underlying cells. This process con-
tinues without destruction. of the boron framework until such time when
the boride becomes so depleted in lanthanum that only a boron residue
remains. -

Even though there is a need for more efficient control rod materials
as well as for cathodes for use in high vacuum devices, there is an
equally severe need for thermodynamic data on rare earth borides.
Once the formation and decomposition processes within the borides are
understood, it should be possible. to make precise measurements of the
free energies of formation for these compounds. Any introduction to
work of the sort contained in this thesis would be incomplete without the
assertion that, ultimately, collection of these fundamental data is con-
sidered to be more important than the possible uses to be‘made from

such measurements.

II. HIST ORICAL

A. Preface

 

It is frequently suggested that the rare earth metals represent a
class of elements whose chemistry is similar. While this is casually and
quite generally true, any serious scrutiny of the lanthanides readily
exposes the subtle, but definite, dissimilarities among them. One such
fundamental difference is found in the vapor pressure of these elements
in their free state, and for any given temperature, the vapor pressure
of samarium, europium, and ytterbium is appreciably higher than it is
for the other elements in this sub-group.

This study was undertaken initially to determine how the vapor
pressure of the metal component would affect the high temperature behavior
of the rare-earth boride. Specifically it was hoped that measurements
could be made on the vapor pressure of samarium tetraboride; but in
order to fully appreciate this most immediate and direct aim of the re-
search, it is necessary to consider the preliminary work on the rare earth
borides in general as well as a discussion of their properties and structures

in particular.

B. The hexabo rides

 

The first of the metal borides, categorically the hexaborides, were
prepared as early as 1897 by Moissan and Williams (2) who synthesized
calcium hexaboride, strontium hexaboride and barium hexaboride by
aluminum reduction of the corresponding borates in an electric arc.
Other workers (3-5) studied the calcium hexaboride preparation by melt-

ing together calcium carbide and boron sesquioxide, by the reaction of

calcium chloride or calcium fluoride with boron, or by the reaction of
calcium metal with boron sesquioxide. , However, with respect to yield
and purity of product, probably the most satisfactory of these early
preparative techniques was that used by Andrieux (6) who electrolyzed
molten mixtures of borates and fluorides of the corresponding metals as
well as mixtures of boron sesquioxide and the metal oxide or carbonate.
Using carbon electrodes, the borides were produced as fine crystals at
the cathode.

By the late 1920's, sufficient interest had developed in metal borides
to warrant studies into the preparation, structure, and properties of the
rare earth borides. In 1932, using the method of Andrieux, vonStackelberg
and Neumann (7) prepared the hexaborides of lanthanum, cerium, praseon
dymium, neodymium, and erbium, and showed that these compounds were
all isostructural with the simple cubic lattice of cesium chloride.

Allard (8) duplicated some of von Neumann's work as well as extending
its conclusions to include the hexaborides of yttrium, gadolinium, ytter-
bium, and thorium. The hexaboride structure was therefore firmly
established as one in which the metal atom was centrally positioned and
surrounded by a cage of boron atoms. The boron atoms are known to be
B6 octahedral units, with one such unit at each of the eight corners of the
cell and resulting in one molecule of MB6 per unit cell. The electrons
of the metal atom are not directly used in the bonding and are therefore
free to provide for the conductivity of the hexaborides. S. Flodmark (9)
has since completed a theoretical study of bonding in the MB6 compounds
and concludes that conductivity should be expected since several strongly
bonding orbitals are in the conduction band.

In view of the metallic nature of these compounds, it is therefore
not surprising that the trend in their lattice parameters (see Appendix B)
parallels the trend observed in the size of the rare earth metal atoms,

that is, unusually high. values are shown for euroPium and ytterbium.

Furthermore, the absence of any formal electron bonding between the
central metal atom and the octahedral boron units strongly suggests the
possibility of metal atom mobility in a situation where such mobility
might be the easiest path to decomposition of the compound.

Figure I provides a graphical representation of how both the lattice
parameters of the rare earth hexaborides and the radii of the rare earth
metal atoms vary with atomic numberr In Figure II, a diagram of the
MB6 lattice is presented. At any corner of the cube, only three of the

six boron atoms comprising the B6 octahedron are shown. .

C. The tetraborides

 

For about twenty years subsequent to these early structure determin-
ations, virtually no metal boride work was performed. . Then, in 1951,
. Zalkin and Templeton (10) prepared the tetraborides of cerium, thorium,
and uranium by heating the elements together in a vacuum at about 15000.
All three of these borides are isostructural and form a tetragonal lattice
which is a combination of the structure of uranium diboride and that of
calcium hexaboride. The boron atoms form a continuous three-dimensional
network among the metal atoms. There are sixteen boron atoms in the
unit cell, twelve of which are arranged in two identical octahedra and the
remaining four of which are used to connect these octahedra in a con-
tinuous two-dimensional chain. The actual continuity of the entire net-
work is sustained by the connection of atoms of the octahedra above and
below the aforementioned chain with other boron atoms in neighboring
cells. There are four molecules of MB,, in each unit cell.

Figures III and IV represent, respectively, diagrams of the lattice
of a metal diboride and a metal tetraboride. The diboride lattice is pre-
sented since the metal atom arrangement of MB, is intermediate between

that of MB6 and MB;.

 

0:: 2.10
3 2.00

:8

«,3 1.90

.‘3 1.80 ° “““

0 l
5 1.70

4.20

05’

2 4.15 0M _____
03: 4.10 '
0

«0|

 

 

 

57 58 59 60 61 62 63 64 65 66 67 68 69 70 71
Atomic Number

Figure I. Radii of rare earth metal atoms and lattice parameters of
rare earth hexaborides versus atomic number.

 

 

 

 

Figure II. The structure of a cubic hexaboride (MB6).

 

 

© Metal

0 Boron

 

 

 

 

 

 

Figure 111. The structure of a hexagonal diboride (MBZ).

 

 

 

 

 

 

 

 

 

 

 

 

   
 

I *0
O O
o
o
O O
O
8'o O 0
© 0
0 o
\L b O
c = O C = 0.2 C = 0 5
d = 1.0 .
1.0
0.8
0 'Metal “0 ”9009300 0-5
O 0.2
O Boron ®@ .94 0
(100)
Projection

Figure IV. The structure of a tetragonal tetraboride (MB4).

Since a tetraboride existed for cerium, the expectation that tetra-
borides of other rare earths could be prepared seemed reasonable.

In 1956, by reacting unspecified amounts of rare earth oxides. with boron
and carbon black in graphite crucibles, Post, Moskowitz and Glaser (l l),
succeeded in preparing the tetraborides of praseodymium, samarium,
gadolinium, and ytterbium.- Lattice parameters .were obtained for

these compounds Which were, as expected, completely ‘isostructuralwith
uranium tetraboride.

By 1959, lattice parameters had been obtained for all of the rare
earth tetraboridesvwith the exception of promethium, europium, and
thulium. Examination of the 3;? values of the lattice parameter for these
tetraborides (see Appendix B) shows a slow but regular decrease proceed-
ing fromlanthanum through. lutetium, quite devoid of the irregular in-
creases in size noted for europiurn hexaboride and ytterbium hexaboride.
It was this fact that led Post and his cot-workers (11) to make the observ-
ation that in the conducting hexaborides, the effective valence state of the
metal was very much like that of the uncombined metal, whereas the non-
conducting tetraborides, the metal atoms had muchmore of an."ionic"

natur e .

D .. The Dodecaborides

 

The reported existence of uranium dodecaboride (12) andzirconium
dodecaboride (13) provided ground for the speculation that the dodecaborides
of the rare earths might be synthesized. Post and his co-workers (11)
had been unsuccessful in their attempts to prepare any dodecaborides from
the reaction of the sesquioxides of lanthanum, cerium, praseodymium,
samarium, gadolinium, ytterbium, and/or thorium.with boron. . However,
in 1959,. Binder, LaPlaca and Post (14) did report such dodecaborides for
yttrium, .dysprosium, .holmium, erbium, thulium, and lutetium. These

compounds, were prepared by heating 99. 95% pure sesquioxides with

amorphous boron .in vitrified alumina crucibles at 1400~1500O in a
protective helium atmosphere.

. It is not surprising that this later attempt proved successful since
it involved the use of rare earth metals all of whose atomic radii are
smaller than those initially used. * Furthermore, the conspicuous absence
of ytterbium dodecaboride offers firm vindication to the argument that
dodecaboride formation is size-controlled.

. The cubic cell of the dodecaborides contains four molecules per
unit cell, is isostructural with the phases ZrBlz and UBIZ, and consists of
» a metal atom held in the center of a cage of 24 boron atoms. This cage,
a regular polyhedron with square and hexagonal faces, shares its atoms
with other similar cages to make a continuous three-dimensional network
of boron atoms. A diagram of the cell is shown in .Figure V.

The metallic atomic radius of the zirconium atom is reported by
Sanderson (15) as l. 597 X . Furthermore, Katz (16) reports that the
u-uranium atom may be considered as a nonspherical ellipsoid with major
and minor half-axes respectively 1. 65 and 1.4 X. , and the cubic
y-uranium atom as a Sphere of atomic radius 1.485 R. The metallic
atomic radius of dysprosium has been determined to be 1. 770 A: (17)
and no dodecaboride formation has been observed for any rare earth with
a. metallic radius larger than this value. It is of interest to note that the
atomic radii of zirconium and either form of uranium are both well below
the metallic radius of dysprosium, and that the size-controlled dodecau
boride postulation withstands this additional test.

Not much more is known about this difficultly prepared phase,
and it has been suggested that it can be prepared only by reaction of the

sequioxide with boron.

>:<
See Appendix B for values.

10

 

 

 

 

 

 

_-_-__.’J—-—-.n—-_--

—

— _-_--- _--_---—-—

a

 

 

 

 

 

   

_ --__- —-'

 

 

O s

The structure of a cubic dodecaboride (MBlz).

Figure V.

11

E. fiOther borides

 

In addition to the tetraborides of praseodymium, samarium, gado-a
linium, and ytterbium, Post, e: a}. (11) also have reported the prep-
aration of a contamination phase, MBx, where the value of x is suggested
to lie between 3 and 4, M. being lanthanum, praseodymium, gadolinium,
or ytterbium. A set of lattice parameters was obtained, although the
authors (11) suggest that the compounds are only boride-carbide phases
and that their stabilization was effected by carbon impurities introduced
from the graphite crucibles.

Nonetheless, the MBX compounds could be prepared reproducibly
and all are believed to be tetragonal even though no SOphisticated struc-
tural data are available. Gilles and co-workers (18) have presented
further evidence for these materials being. boride-carbide phases and
report the preparation of a Gde compound with atom ratios: Gd': B: C ::
0.10 i 0.05 00.55 i 0.05: 0. 37 i 0. 05. There appears to be no special
significance attached to this rare earth boride, and no further discussion

will be presented.

F. Methods of preparation

 

If one is interested in the high temperature properties of a particu-
lar rare earth boride, or even in the room-temperature physical proper:
ties of such a compound, it becomes important to work with the purest
material available. Since the initial goal of this study was an investiga-
tion of the vapor pressure of samarium tetraboride, the preparation of
a pure phase was of enhanced significance.

This study has provided us with information regarding the variables
involved in synthesizing a "pure" boride; it seems worthwhile to consider
briefly the methods that have been employed previously. No attempt will

be [made here to evaluate the effectiveness of these methods.

12

1. Direct reaction of metal with boron

 

The most direct means of synthesizing rare earth borides is the
reaction between the metal and elemental boron. Felten, . at 341. (19),
have reported the preparation of LaB4 "of a satisfactory purity" by
reaction of lanthanum metal with boron at about 15000. . No details are
given regarding the crucible material used to confine this reaction.

Post, (31: 11. (11), reported the preparation of the MB6 phase "when
the appropriate amounts of boron and metal were caused to react. " Using
x- ray powder techniques, they also reported the detection of this phase
when compositions were employed having a mole ratio of metal to boron
other than 1:6. . Although it is not completely clear from the literature,
it appears that these workers used graphite crucibles even when the
syntheses were performed using mixtures of metal and boron. Binder,
e_t 341. (14), have stated that it is impossible to prepare the dodecaboride
by this means.

a In this work, mixtures of samarium hexaboride and samarium
tetraboride were prepared from the direct reaction of the elements.
The important fact regarding this means of preparation, and one which
is too frequently ignored, is that the product is not independent of the
crucible material used to confine the reaction. This will be discussed

later in more extensive detail.

2. Reaction of rare earth sesquioxides with boron carbide

 

A second method of preparation, and one that has received wide
attention, is that described by the equations:
M203 + 3 B‘C = 2MB6 , + 3 CO (3.)

and employed by several Russian investigators (20-23) for the preparation

of the hexaborides of yttrium, europium, holmium, and lutetium, and the

l3

tetraborides of terbium, dysprosium, holmium, erbium, and lutetium.
These reactions were carried out in vacuo at 1400-16000 with varying
amounts of boron carbide. Samsanov (24) has reported the use of boron
and carbonin any desired mole ratio; however, he has cautioned that if
the starting oxide is of the formula, M203, as are most of the rare
earth oxides, then boron carbide of composition exactly B4C must be

used.

3. Reaction of rare earth sesquioxides yith boron

 

Another method ofsynthesiz-ing the borides is through the reaction
between the metal sesquioxide and boron. First mention of this method
awas made by Post, e]: a}. (11), who successfully used crucibles of
graphite or zirconium diboride to contain the pressed pellets of ses-
quioxide and boron. The formation of the volatile species, B203, .was
reported by Post, but no chemical proof was offered to substantiate the
existence of this species. Margrave, e_t a_._l.. (25), have reported mass
spectrometric evidence that between the limits of 1331 and 18080K. ,
the vaporazation of boron sesquioxide is satisfactorily described as a

monomeric vaporization shown simply as:

3203(1) ——> 1320, (g) (c)

However, even if boron sesquioxide is formed in the reactions between
boron and rare earth sesquioxides, one must allow for at least three
possible occurrences subsequent to the formation of this oxide:
1) Vaporization as shown
2) Reaction with a rare earth sesquioxide to produce a rare earth
borate
3) Reaction with excess boron to produce a lower boron oxide.

. Enough has been said regarding the first of these possibilities.

14

Levin, e1; a_.l. (26), have substantiated the possibility of the second
occurrence citedabove. They have Synthesized rare earth borates by the
reaction of boron sesquioxide with rare earth sesquioxides; their x- ray
diffraction data for samarium orthoborate are shown in Appendix B.

. Since samarium orthoborate was encountered in this work, its formation

as a result of the reaction:
ssz3(S) + B203(1) —-'>' 2 SmBO3(1, S) (d)

should not be excluded.

With respect to the third alternative, the existence of suboxides of
boron is well established. Ray and Sinha (27) have reported the formation
of the colorless oxide B405 from the decomposition of (NH4)ZB406, and
Kahlenberg (28) has presented considerable evidence for the existence of
the suboxide, B30, which was prepared by the complete oxidation of B70
with a 1% solution of potassium hydroxide. ,Kahlenberg has concluded
further that B70 and B50 are only mixtures of B30 and boron. However,
the unequivocal existence of B70 has been demonstrated recently by
Pasternak (29) who has determined the crystal structure of this compound.
This determination casts some doubt on the validity of Kahlenberg's con-
clusions.

Inghram, Porter, and Chupka (30) have examined the gaseous
species effusing from an alumina Knudsen cell containing a mixture of
solid boron and liquid boron sesquioxide. . The major peaks observed on
the mass spectrometer for the temperature range, 1300-15000K. , were
3202+.) BZO3+,. BO+, and B+. However, these workers (30) attribute the
latter two species only to dissociative ionization of BzOz(g) and BZO3(g).
Scheer (31-32) has used a torsion effusion cell to substantiate the hypothe-a
sis of Inghram, gt 11.; his measurements have shown that, in the 1300;-
15000 K. temperature range, the major gaseous suboxide formed in the

reaction between solid boron and liquid boron sesquioxide is gaseous B202.

15

Scheer did not detect any gaseous BO; however, his data requiredthe
postulation of an intermediate condensed polymer, (BO)X. Gaseous BzOz

can then be formed by two paths:
2/3 B(s) + 2/3 5203(1) ——> BzOz(g) (e)

Kanda, e_t a_L_1.. (33), have reported polymeric (BO)X prepared from a mix-
ture of solid boron and liquid boron sesquioxide, but found that the yield
of saublimed polymer depended on the atom ratio of boron to oxygen
initially used. On the basis of Inghram's mass spectrometric study (30)
and some x- ray diffraction data obtained in his own laboratory, Scheer
has felt compelled to emphasize the complications of the system and the
extent of possible interactions. He has written: "It is therefore suggested
that equilibration between the condensed phases [B(s), B203(1), and
possibly (BO)x] and gaseous B202 is not readily achieved in such effusion
measurements. "

' From this it would seem, that in any system where it is possible
for oxides of boron to couexist with boron itself, there is also the possi-
- bility of extensive interaction among the phases. This fact strongly
suggest that, without a mass spectrometer, it is virtually impossible to
study the mechanism by which a rare earth boride is formed from the

reaction of the rare earth sesquioxide with boron.

4. Other methods of preparation

 

The three principal and most widely used means of synthesizing
the rare earth borides have already been carefully outlined. .Aside from
the electrolysis of molten media, which was mentioned earlier, any other
' methods for successful preparation of the borides are merely combi-
nations or‘ modifications of the three principal methods. Assuming the

need for starting with the rare earth metal oxide, the modifications -.

l6

concern themselves with such questions as whether or not to use,
additionally, boron sesquioxide and carbon, boron and carbon, boron
carbide, or boron alone. [The reader is referred to Samsonov's extensive
review (24) on the subject of rare earth metal borides if a more compre-
hensive outline of any of the methods of preparation for any particular
boride is desired.] Each method used, however, is subject to possible
crucible interaction and the selection of any one method demands prudent
consideration of just how the vessel confining the reaction may affect the

nature, yield, and purity of the products.

G. Properties of the rare earth borides

 

Most of the structural properties of these compounds have already
been discussed in detail. However, since the keystone of this thesis was
initially an attempt to measure the vapor pressure of samarium tetra-
boride, the decomposition of the borides must be considered.

If a rare earth boride is heated to a sufficiently high temperature,
two paths of decomposition are possible. The first involves the loss of
metal atoms and the subsequent increase in boron content of the boride as

shown:
‘ A
MB6(s)———> xM(g) -+ M1_xB6(s) (g)
which, of course, in the limiting case is merely:
Mme—9+ M(g) + 6 13(3) (h)

The second possibility is a disproportionation of the hexaboride to give

a lower boride and free boron as shown:
A
MB.<s> -——> MB.(s) + 2 B<g> (i)

Naturally, the thermal stability of any rare earth tetraboride
is not independent of the stability of the corresponding hexaboride.

And so, for those elements for which the tetraboride represents the more

l7

stable phase, the tetraboride should be able to be heated to its melting

point without appreciable decomposition. The other course is:

3 M1345) ——->z 1.4136(5) + M(g) (3')

Consider the following hypothesis: that those rare earth metals with a
high vapor pressure would show hexaboride decomposition according to
equations (g, h), while those with low vapor pressure would follow the
path shown in equation (1). Since a hexaboride decomposition according
to equation (1) is incompatible with a tetraboride decomposition shown in
equation (j), any tetraboride which decomposes according to path (j)
should behave according to paths (g, h) on being heated to a higher
temperature, providing that the rate of escape of M(g) is sufficiently high
to bring about quantitatively measurable changes.

Since the hypothetical paths of decomposition are vapor pressure
dependent, it was decided to initiate this study assuming the truth of the
hypothesis that the high vapor pressure rare earth metals would show
boride decomposition according to equations (g, h) and (j). With the
acceptance of this hypothesis, the selection of samarium as the metal
component was made merely on the basis that samarium was the first
in line of the three rare earth metals which exhibited abnormally high vapor
pressure.

It seems worthwhile to note here that Samsonov (24) had written
prior to any of this work: "On the other hand, hexaboride phases are
always more stable than tetra- or diboride ones, as experiments on heat-

ing the lower phases i_r_1 vacuo show; in time they pass into hexaborides

 

with simultaneous loss of the metallic component. " Also, during the
course of these investigations, Leitnaker, et a1. (34), had written:
"In view of the regularity of the heats of formation of the rare earth

sesquioxides, one is tempted to estimate that the heat of formation of rare

18

earth hexaborides is -12 i 2 kcal. per gram atom of boron, with the
tetraborides perhaps 3 or 4 kcal. per gram atom of boron more stable. "

In order to investigate the vapor pressure of samarium tetraboride,
there were two fundamental requisites. The first was the preparation of
pure samarium tetraboride; the second was to find an inert crucible
material in which the boride could be decomposed.

This research, with all its experimental ramifications and, ultimately,
its conclusions, is an outgrowth of the failure to satisfy the two requisites

set forth above.

III . EX PERIMENTA L

A.. Materials

 

Samarium sesquioxide of 99. 9%: total purity was obtained from the
Michigan Chemical Company, St. Louis, Michigan- . Similar purity
samarium (> 99%) was obtained from the same source and also from
Research Chemicals, Burbank, California. . Crystalline boron (-100 mesh)
of 99. 2% purity and containing 0.4% iron and 0.4% carbon was purchased
from the U. S. Borax and Chemical Corporation, New York, New York.
.Amorphous boron of 99% total purity was provided by the Fairmount
Chemical Company of Newark, New Jersey. The molybdenum, as a
powder (-325 mesh) or as stock used for machining crucibles, was ob-
tained as 99. 9% pure molybdenum from the Climax Molybdenum Company,
New York, New York; boron nitride rods of 97. 0% purity and containing
2.4% boron sesquioxide, 0. 1% alkaline earth oxides, and 0. 2% alumina
and silica were purchased from The Carborundum Company, Latrobe,

. Pennsylvania. Machined crucibles and lids of zirconium diboride were
obtained from the Borolite Corporation, Pittsburgh, Pennsylvania.

. No purity Specificationsare available for this material. although the manu--
facturer does use an unspecified binder to bond the zirconium diboride.
An experimental sample of tungsten-copper alloy was provided by
Philips Electronics, Inc. ,. Mount Vernon, New York, for use in fashion-
ing a tungsten crucible. . No purity specifications were available; the
techniques of preparing this alloy for use are described in the next

section of this thesis.

l9

20

B. Sample‘preparation

 

For the reactions between samarium sesquioxide and boron, small
amounts of these components were mixed in the desired mole proportions
and carefully ground, to insure complete homogeneity, with alundum
mortars and pestles. The mixtures were then pressed into pellet-wafers
using a die of "carborized" steel with outside diameter of 2-1/2" and
pellet bore of 5/16'.‘. Average sample size was from 0. 5 to 0. 7 g. , and
pressing at 4000 - 8000 p. s.i. resulted in pellets ranging in height from
3/16" to 5/16", depending upon the boron content.

These pellets were loaded directly into a molybdenum crucible
fashioned from 7/8" molybdenum rod, 1 to 1-1/2" high, and bored with a
reaction well of 1/2" diameter by 19/32" deep. The crucible was also
provided with a lid of total thickness comparable to the wall thickness of
the crucible and fashioned with a lip to fit snugly into the crucible well.
Tantalum (0. 005") spacer rings were used between lid and crucible body
to prevent adherence of the lid to the crucible; these rings were especially
necessary in studies of the reaction between samarium sesquioxide and
boron.

In the series of reactions for which boron nitride was used as a
liner inside a molybdenum crucible, the liners were machined from rods
of boron nitridewhose outside diameter was 1/2". They were about 5/8"
in length and bored and milled with 3/8" tools to provide a bottom thick:-
ness of 1/8" to 1/16". Because of the fusion of a boron nitride lid to a
boron nitride crucible as a result of decomposition and re-formation of
the material at high temperatures, the heating cell was capped with a
lid of molybdenum rather than one of boron nitride.

These molybdenum crucibles, used either directly as reaction
vessels or as a means of heating boron nitride liners, were supported
with three tantalum legs (0. 060" in diameter), force-fitted at 120° inter-

vals around the base into holes 1/4" deep. Each lid was also provided

21

with a hole of approximately 0. 045" diameter to allow for escape of gases.
A black body radiation hole, 1/16" x 10/16", was drilled into the base of
the crucibles.

The low-density tungsten heating cell was fashioned from the
tungsten-copper alloy by machining a crucible of 1" length and was bored
with a well 5/16" in diameter and 3/8" deep. A lid was also machined
with an edge of 1/32" and a lip of l/l6", providing a total thickness of
3/32", and drilled with a gas escape hole of approximately 0. 040".
Tungsten legs of 0. 040" diameter were inserted into holes pre-drilled
in the base of the crucible; a black-body hole was also provided. After
machining this alloy into a crucible-lid combination, a tantalum spacer
ring (0. 005") was inserted between lid and crucible and the entire unit
was heated with a magnetic field at temperatures high enough to cause
expulsion of the copper from the alloy. Several heatings were necessary
to remove all of the copper and care had to be taken not to raise the
temperature too high too rapidly. Since the boiling point of copper is
23100 (35),. it was felt that heating the alloy for a long time, and finally
at the most extreme temperatures possible with the existing magnetic
field (N 22500), would» eventually remove all of the copper. The cell
and lid were heated several times at the maximum attainable temperature
until there was no' further visible deposition on the Vycor tube confining
this vacuum distillation of copper. Thus prepared, the cell was used to
study the vaporization behavior of samarium hexaboride.

Heatings were performed for varying amounts of time using a 20
kilowatt induction heating generator (Model 2000) supplied by the
Induction Heating Corporation of Brooklyn, New York. The existing glass
system used for containing the heatings is shown in Figure VI. Both the
samarium sesquioxide-boron compacts and samples of samarium hexa-

boride were heated in a vacuum of 10"4 to 10"6 mm. , while in the studies

Optical window, quartz

 

Vchr tube

W ate r outlet

 

Induction coil

u u n ufllflflu n".lr

Ianl

Reaction crucible

nun lllll 3' .N . y

 

‘——'“"“‘—‘Quartz table

 

S: 55/50 (Vycor inner)

103/60 (Pyrex)

. up...”
.. \

Water inlet

 

  

[UL-H...
- ‘uufiovd

Optical window, Pyrex

Optical shutter

“with“,
.. m 7‘, p
‘ '--:;_.

3-81. .2..---" _

I
I

1

3%
_=- I, ‘5‘::"

    

 

       

m
u
c
a
V
o
T

Heating cell assembly.

Figure VI.

23

involving samarium metal and boron, heatings were performed in an
atmosphere of helium equal to a vacuum of 100-125 mm.

The metal heating cells were outgassed for from 5 - 20 hours at
1800 - 21000 prior to heating and, according to directions from the
manufacturer, the boron nitride was outgassed for at least two hours at
16000 to remove most, if not all, of the boron sesquioxide impurity.

No evidence whatsoever was observed to suggest samarium nitride form-
ation in studies using boron nitride, and in no instance involving the
use of a boron nitride liner did the liner gain weight.

Temperatures were read with a Leeds and Northrup optical pyro-
meter (Serial No. 1524388) and unless otherwise noted are given in degrees
centigrade. X-ray photographs were obtained using Cu K0. radiation
“Eu: 1. 5418 A) and a Debye-Scherrer x-ray powder camera of 114. 59 mm.
diameter. Temperatures reported herein have been corrected for absorp—
tion due to the optical window and the prism when such corrections are
applicable.

The reaction products in all cases were always in the form of
sintered compacts regardless of how the samples were introduced into
the crucibles. Pellets remained intact; mixtures of samarium metal
and crystalline boron were introduced into the crucibles as a mound of
boron with the samarium confined within this mound in the form of tiny
chunks; these mixtures sintered on heating. The decomposition study in
the tungsten cell involved the introduction of a non-compacted material;
again sintering took place during heating.

The sample of samarium hexaboride used for the decomposition
study was prepared by heating a mixture. of samarium sesquioxide and
crystalline boron (in a respective mole ratio of 1:12) in a crucible of
zirconium diboride. The actual decomposition sample represented the
compilation of the material prepared from nineteen successive prepara-

tions. This material was washed several times with concentrated

24

hydrochloric acid followed by massive amounts of water. Dried and
photographed? the x- ray powder pattern showed no phase other'than
samarium hexaboride although the a lattice parameter of this cubic
phase was measured to be 4.141i 0. 002 A.., as opposed to the value of
4. 1333 i 0. 0005 .8. reported consistently by other investigators (36).
Chemical analyses were performed whenever there was a sufficient
amount of material for doing so and when it was felt that such analyses
justified the time involved in performing them. . Description of these

analyses follows .

C.- Analytical methods

 

1 . Samarium

 

The samarium analyses were only gross analyses for samarium
present in any form in the product. As a means of testing the analysis,
samples of from.0.05 to 0. 15 g. of freshly dried samarium sesquioxide
were dissolved in 4. 0 ml. of 6.0 M nitric acid and then diluted to 250
eml. with distilled water. . To these solutions, now approximately 0.1 l_\_rf
in hydrogen ion, five to ten ml. of a saturated oxalic acid solution were
added dropwise and with stirring. The precipitates which formed were
. allowed to digest overnight on a hot plate and then were filtered, washed,
and ignited in platinum crucibles. Results obtained using theisesquioxide
as described showed a consistent accuracy ofi'. 1% or better.

. It should be noted that samarium hexaboride is believed to be com-
pletely soluble in 6. 0 111 nitric acid. This is based on the observation
that on dissolution of hexaboride samples, the less the residue which
remains on treatment of a sample with nitric acid, the more nearly the
samarium content approaches the percent samarium in samarium hexa-
boride. . Some of the residues remaining from nitric acid treatment of

the samples were collected, photographed, and found to be boron.

 

:':In the entire text, "photograph" refers only to x-ray photographs.

25

' It was therefore assumed that samarium in any form (metallic, as
an oxide, or as a boride) would not escape dissolution on treatment With
6.0 M nitric acid. The tetraboride behaved similarly to the» hexaboride
onnitric .acid t'reatrnent; in both cases a large amount of nitrogen(IV)'j.
oxide gas was liberated indicating the strongly reducing nature of the
borides. .- In all cases, the samples were filtered subsequent to, dissolution
and prior to oxalic acid treatment. Analyses were run in duplicate

wherever possible and agreement of results was always within 1%.

2. Boron

Boron analyses were performed essentially as described ‘by
Blumenthal (37). . However, the modifications and elaborations which had
to be added are presented below.

The mere mixing of a sample of the boride with sodium carbonate
prior to the fusion is inadequate for the protection of the platinum crucible.
Even though sodium nitrate was added after the initial carbonate fusion,
total fusion was not effected rapidly enough to prevent the mixture from
attacking the crucible. . This was not a rapid attack, but one definite A
enough to cause tiny pits, and subsequently holes, in the platinum after
any repetitious use. Such attack could not be prevented even by fusing
the samples immediately with a sodium carbonate-sodium nitrate mixture.
. However, if the bottom of the platinum crucible was covered with sodium
carbonate which 'was then fused prior to the addition of any sample, attack
on the platinum became negligible. . Samples were introduced onto the
fused sodium carbonate; 0. 05 to 0. 15 g. (of sodium nitrate was added,
and the entire-mixture was coveredwith a fine sprinkling of sodium
carbonate sufficient to contain the vigor of the reaction that followed on
heating.

The crucibles were heated in a very low Meeker flame. .An initial

and vigorously exothermic reaction occurred, but as long as the heating

26

was gentle and the sodium carbonate present in sufficient amount, no
spattering losses were encountered. The fusions were continued at
moderate heat for about one-half to three-quarters of an hour to insure
complete decomposition of the sodium nitrate.

After fusion, the crucibles and contents were digestedwithw 50 ml.
of 6el\_ll hydrochloric acid. Twenty milliliters of 12.5 :31 sodium hydroxide
were added before slow addition of 8 g. of barium carbonate. . The order
of operations here is extremely critical and cannot be reversed.

The remaining operations were performed as described by Blumenthal
(37). However, while he has offered the choice of determining the end
points of the mannitol titrations using either a mixed indicator or a pH
meter, this work has shown that the potentiometric titration is the only
means of obtaining satisfactory results for samarium borides.

It was possible to use the prescribed mixed indicator successfully
for standardizations of sodium hydroxide with re-crystallized orthoboric
acid. Standardizations performed in this fashion with 0. 1 g. samples of
orthoboric acid showed an average deviation of from 0. 2 to 0.4%.
However, when samarium sesquioxide was added to the scheme as a con-
trol, and subsequently removed as a carbonate, the color changes in the
boric acid titration were so gradual and subtle that results ranged. from
2 to 10% high. (This must result from incomplete removal of the :‘I'
samarium(III) ion and resulting hydrolysis of that species.)

In view of this, the use of a potentiometric titration was chosen;
excess hydrochloric acid was titrated with sodium hydroxide to a pH of
6. 2 and the boric acid was titrated in the pH range of from 6. 2 to 9. 0.
The standardization of sodium hydroxide was also performed in this
fashion and the mixed indicator was completely abandoned. “Using a
Beckman Model G pH meter equipped with (large size) standard calomel

and glass electrodes, the average deviation in standardizations ranged

27

from.0. 2 to 0. 5%. . However, when a‘ titration using a mixed indicator
was compared to one using a pH meter, the difference in the value of

the grams of orthoboric acid equivalent to one milliliter of sodium
hydroxide was as much as 10%. Therefore, the sodium hydroxide was
always standardized over the same range of pH as the samples, just as if
it were a sample. . Figure VII contains four curves showing pH plotted
against milliliters of sodium hydroxide added in the region of the end
point. The slope of these curves shows the necessity of standardizing
the sodium hydroxide over precisely the same pH range as used for the
samples.

Using the entire scheme of the method, beginning with fusion of
samples of samarium sesquioxide and subsequent removal of the samarium
carbonate, it was possible to account for no less than 98. 9% of added
orthoboric acid with no result being higher than 101-. 7%. The same
general precision was obtained using samples whose x-ray photographs
showed no phase other than samarium hexaboride. It is therefore believed
that boron analyses performed in this fashion are accurate to-withini 2%,
and that greater accuracy can only be obtained using some more elegant

method at the expense of tedium.

 

.. D. Reactions of samarium sesquioxide and boron in a
molybdenum crucible* ‘

 

The first set of reactions in this category involved samples of _a
composition of l. 00 mole of samarium sesquioxide for every 10. 00'moles
of boron. . Using the same crucible for all experiments, investigations

were made in the temperature range 1360 - 18300. A summary of these

 

*In the experimental descriptions which follow, much of the rigorous
detail associated with early preliminary experiments has been deleted for
the sake of clarity. The data associated with these experiments are more
completely and thoroughly presented in Appendix A under the same general
headings as thoselisted in this section.

 

28

oo.wm

oo.>m

.mcoflmuufl Hofiqsme mo “50919.8 05 mo 31:0? oh» 5 mg no soflmwum> .HH> ouswwh

oo.om.

peep... 03088.3 838.... so 23552
8 .mm

.oo.¢m.

oo.mm

oo.Nm

OOeHm

oo.om

 

 

d

4

fi

—

1

q

A

f’\.

om.>

oo.m

om.w

oo.o

 

 

om.o

Hd 3° snun

29

experiments is presented below.

1.

Low-temperature heatings of these mixtures in the range of
1360 m 14050 for periods of aboutan hour resulted in the
formation of SmBO3 on top and bottom surfaces of the pellets

as well as on the molybdenum lid of the crucible.

.. Additional heating for one to two hours at temperatures of about

16000 resulted in the formation of SmB4 on the surface of the
pellets. The concurrent appearance of B~szO3 (high temperature

form) on the lip of the molybdenum crucible was also observed.

. Continued heatings at 18000 for one hour subsequent to the heat-

ings in (2) resulted in the continued existence of SmB4 on the
surface of the pellets, but caused the grow-in of some SmB‘ as
well. . A coating of what was principally B-szO3 appeared on

the inner walls of the crucible.

. Adding small amounts (~0.05 - 0.07 g.) of crystalline boron. to

the molybdenum crucible between some of the heatings resulted
in the formation of SmB6 on the surface of that boron when the
crucible was heated to temperatures in the range of 1590» - 17500.
Following one such use of boron to "clean" the crucible, a

pellet heated for an hour in the range 1675 - 18100 was: severely

diminished in size and showed Only SmB6.

. Another heating in the range 1675 - 18300 over a 0. 5 hour time

span, again subsequent to "cleaning" the crucible with boron,
caused the formation of alayer of 6 -MoB midway up the crucible,
blocking off the orifice in the lid. . The lid of the crucible» also
contained 6 -MoB. . X-ray photographs of both the top and the
bottom of the pellet showed the presence of SmB4 and SmB6.

Immediately subsequent to the phenomenon of the 6 -MoB grow-

in, the same crucible was used to heat a pellet for seven hours

at 14600. The top surface of the pellet contained SmB,‘ andrSmB6,

30

but six hours of additional heating at 1440°, followed by two
hours at 16500, yielded a pellet principally SmB4 but which did

contain small amounts of SmB6.

. Finally, the crucible had to be abandoned since an attempt to

clean it at 1900 - 19500 resulted in its destruction. . The des-
truction was a consequence of 6 «MoB formationwhich had
become a sufficiently intensive component of the crucible to

caus e interior melting .

The second set of experiments involving the molybdenumcrucible

and samarium sesquioxide-boron mixtures was performed with ratios

other than one mole of sesquioxide for every ten moles of boron. In what

follows ,

only a numerical ratio will be used to describe the starting

composition of the materials, using numbers such as 1:10, 1:8, 1:6, 1:2,

to describe the respective ratio of moles of sesquioxide to moles of boron

present in the initial compacts. . A new molybdenum crucible was used to

begin this series, and the following salient facts are set forth as a conse-

quence of these experiments:

9. A pellet of composition 1:6 was heated in an outgassed molybdenum

10.

crucible for ten hours at 14350. X-ray examination showed the
crushed pellet to contain SmB4 and S‘mB6; the lid of the crucible
was coated with SmBO3. The crucible was subsequently treated
with boron for one hour at 17500.

The operations above were followed by heating a pellet of 1:8
composition in the same crucible for 12. 5 hours at 14350,

this sample being 1. 25x heavier. The x- ray photograph of the
pellet showed the presence of SmB4 and SmB6. Again, the
crucible was treated with boron for an hour at 17000, (which
treatment caused a small crack in the crucible. There was

a concurrent deposition of B-szO3 on the crucible.

ll.

12.

13.-

14.

15.

16.

17.

18.

31

A final 1:10 sample was heated for 12.0 hours at l435°. . Only

8me was observed in the product.

Using the same crucible as in (3), a pellet of composition 1:6
was heated for two hours at 14350. . At the end of this heating,
the surface of the pellet was found to containSmBO3 and the
interior contained only SmBé. . No SmB4 could be detected by
x-ray powder analysis.

An intermediate 1:10 run for 1.5 hours at 14350 and l. 5 hours
at 15650 resulted in a homogeneously blue product of SmB6.
Invoking a more radical change of composition to 1:2 and heat-
ing the pellet for three hours at 14350, resulted in a grey mixture
whose x- ray photograph contained no lines of normally identifiable
phases.

Heating a 1:2 mixture at 15400 for three hours again resulted
only in the formation of yellow-white material. which collected in
the base of the crucible and which appeared to have once been a
molten mass. 1

However, whena 1:2 mixture was heated for 12. 0 hours at 15400
in the same crucible used for (6) and (7) and immediately subse-
quent to (7), the product x-ray showed the presence of SmB6

and B-szO3.

Heating a 1:1 mixture for three hours at 14350 resulted in a
product whose x- ray showed no lines other than those ascrib-
able to B-szO3.

Another 1:10 pellet was heated for l. 5 hours at 14350. Visual
examination of its surface at this point showed it to be the same
color normally associated with SmBO3. . Without any further
destruction of the pellet, it was returned to the molybdenum
crucible and heated for l. 25 hours at 1540°. . Only sms, was

found in the product after this secondary treatznent.

32

19. Using amixture of SmB6 and szO3 in a respective mole ratio

20

21

22.

of 11:2 and assuming the possibility of the reaction:

11 SmB6(S) + Z Sm203(s) =3 15 SmB4(S) + 3 BzOz(g) (k)

a pellet (N0. 15 g.) was pressed and heated in the molybdenum
crucible for one hour at 15400. . At the completion of this run,
the top surface of the pellet was found to contain SmB4 and
SmB6, while the interior was found to contain only SmB6.

No Sm203 was observed.

.. A new molybdenum crucible had to be fashioned for succeeding

runs since the previously mentioned crack became intolerable.
The new cruciblewas outgassed for several hours at 18500.

and initiated by heating a pellet of -325 mesh molybdenum metal
with szO3 (prepared in a respective mole ratio of 3:1) at 14600
for 3. 5 hours. . No phases other than B-szO3 and molybdenum

were observed on the x- ray photograph.

.— A final heating of a l: 10 pellet was performed in the new

molybdenum crucible for 14. 5 hours at 15100. Immersed in a
shell of glassy-looking material, some of awhichhad dispersed
on all parts of the lid, the pellet was a silver-grey color. The
material was exceedingly soft on crushing and the x- ray photo-
graph showed no lines of SmBé. The principal phase was clearly .
SmB4, but there seemed to be a few lines of another unidentifi-
ablephase.

Another 1:10 pellet was heated for 11. 5 hours at 15050 in the
same crucible used for (21). . In this case, the pellet was totally
nonhomogeneous on crushing and represented only a glassy kind
of nondescript residue. . No photograph was taken of the residue,
but material scrapedfrom the molybdenumlid was observed to

be B-SmZO3.

33

23. The crucible was heated with boron for 5. 5 hours at 14350 to
"clean" it. At the conclusion of this heating, the Vycor tube
was coated with a residue which looked like a metallic mirror
and which, after dissolution with 6 M nitric acid diluted to
pH = 1, formed a precipitate with a saturated oxalic acid

solution.

E. The samarium-boron system in a molybdenum crucible

 

The general nature of these reactions is quite simple. The com-
ponents were mixed on the assumption that samarium-boron* homogeneity
would result from the samarium becoming molten. Vacuum heating could
not be used due to the high volatility of the samarium metal and therefore
most of the reactions were performed in an atmosphere of helium.(575-
600 mm. ), bled into the line directly from the tank with no pre-purification.
Evacuation was used as a sort of purification technique for removing
excess samarium after runningthe reactions. .Since, at any temperature,
the vapor pressure of the samarium borides appears to be quite small
relative to that of samarium, it was possible to remove excess samarium
in this manner. Using samples of from 0. 5 - 1.0 g. , all of these experi—
ments led to essentially the same result; namely, success in the prepara-
tion of both borides of samarium. . In all instances, the product material
was light grey and in some instances the x- ray photographs were even
devoid of any lines of the samarium hexaboride phase. ' But in no instance
did analysis verify the preparation of pure samarium tetraboride. . In most
cases, the products had sintered to the shape of the crucible and all could
easily be ground using only alundum mortars and pestles.

No chemical analyses were performed on the products of the early

experiments involving samarium metal and boron. . Once again the

:4:
The ratlo of moles is now expressed as samarium to boron
respectively.

34

important aspects of these preliminary attempts to prepare pure
samarium tetraboride may be summarized as follows:

1.. A 1:4 mixture of samarium and crystalline boron was heated at
10450 i_1_1 wfor only 5.0 minutes. . The samarium effused so
rapidly that helium was added to provide aninert atmosphere.
The sample was reheated to 1365C) for 10.0 minutes and an
x-ray photograph of the product indicated SmB4 and ‘SmBs.
(Helium was used in all successive experiments unless otherwise
noted. )

2.. Heating another 1:4 mixture to 12500 for one hour again yielded
a product containing both borides of samarium.

3. Even when the composition of the samarium-boronmixture
was raised to 1: 10 and heated as low as 11000 for one hour, the
product was observed to contain SmB6 and SmB4.

4. Lowering the composition to 1:3. 8 to compensate for samarium
lost in vaporization still resulted in SmB4-SmB6 mixtures after
1.5 hours at 1275 - 1300°.

5. Again using the molybdenum crucible and heating 1:4 mixtures
of samarium with crystalline or amorphous boron at 11250 for
1. 0 .1 l. 5 hours resulted only in unidentifiable phases. . The same
experiment, when performed for 11. 5 hours, yielded a. product
containing SmB4, and a contamination phase, commonto the
heatings of short duration, but significantly devoid of SmB6.

6. Heating a 1:3. 5 composition for 2. 5 hours at 1100- 11250 and
then 1_n wfor 10. 0 minutes at 12750 resulted in a product
whose x- ray photograph showed SmB4 with. a trace of SmB6.

7.. Finally, heating another 1:3.5 mixture for four hours at 11500
in helium and then for one hour at 14550 in wresulted in a

mixture of SmB4-SmB6. When the sintered cake was replaced

35

into the molybdenum crucible and heated in helium for an

additional 8.0 hours at 12750, no significant changes were ob-

served. However, continued heating from 0. 75 - 2.0 hours

over the range 1390 - 1735o finally resulted in a phase whose

x- ray photograph showed only SmB4. The grow-in of the grey-

colored SmB4 as a consequence of the heatings began after a

heating at 15850 and was visually observed to commence at

the bottom of the sintered mass. I i

_ Additional experiments were performed to examine further the

effects of varying the temperature, the duration of heating and the initial
starting composition.

It was possible to prepare samarium tetraboride in a molybdenum
crucible using the 1:3. 5 starting composition even when the mixture was
heated to as high as 16400 in a helium atmosphere. It was further
observed that when the orifice in the molybdenumlid had become clogged,
the x- ray photograph showed no phase other than samarium‘tetraboride,
while an exact duplication of the experiment at 16150 (except for no orifice
clogging) resulted in a product whose phases were samarium tetraboride
and samarium hexaboride.

Thirteen final samarium-boron experiments were performed in
the same crucible. . The xsray photographs of3v_er_y product showed the
presence of samarium tetraboride. . Samples weighing from0-50 - 0. 52
g. were heated in helium at a pressure of approadmately 600 mm. and
chemical analyses of the products were performed in duplicate wherever
possible. All data are tabulated in Table I on the following page; special
notations have been made regarding the presence of samarium hexa-
boride in the products.

It should be noted that tiny bluish crystals were microscopically

observed only in the runs numbered E- 24,. ’E-41, and E-43. They are

.m #009302 E Amhonccsc 098w 2.3 H33: @0330me mucocfinomxo as. some.“ on. poswmuou one muongsc owoFH.

 

 

 

fl
¢.m~ m .N> a: .mmm- mom .oH+ amfim oosuBomme as 30d mvowvd m .NuH «$1M
a. .NN 5 .FN t. .wmml m: .NH+ amaw ooduhooma: «a god Nzom .o m .NHH 3.1M
0mm: avumafloum
1111 1111 mnvdmn 5.5+ ome vodka. Aaddomczfioé Hmfihwd magnum qum
wagon «£333de em. .pomt Am 6+ #255 noaflziomvmfi as no .N momfim .o m .NS 31m
Ema mow 2&2: 5.4+ oMEm ooduBoowmA as no; ofivomé .o.m: omtm
w .mm M .mh ma .th Awimaw omEm honmzoomf no sod Haemd o4”: wmum
6 0313 as so;
3 «1mm won. ométmfit mm.m+ omgm oodnHomv: no no; mmoomd 06: mmum
meow o.m\. w~.mm.. ~66: omgm oodnHommoH «a nod wmmamd oh: 21m
omwmfi as Ham .H
ofim mow can: oo.~m+ omEm nosfizomn: no Amd mmZmé on: Elm—
mfim wins Nwfiml 312+ cmam nocazoovma as H.8.~ woaomé oh: Slum
,. omo: Ho #04
imm ooh owned: wawé+ omcum yonflzomN: an #84 umpmmd m.m£ calm
ommmH um Amé
wfim m.m> mofim» ~>.Ho+ omEm sod—«20mg: no as; ooammd m.m£ mHIM
own m .2. Newman 2143+ «mam 3:0 mBOSmooomH an as on mvzmd «him: 31M
o .NN N .2. gins: om .¢N+ emHEm Eco moaogmoomma um Am .3 mNSmd m.m£ 21H
audpoum a: uodoounm Ewe: G3 7mg Ev menoanm >MMIN :oflmauomoflémvfimwmo? :oflwmomaoo pongz
Gouom am. 3 8m as mucouooom < ofifloduo 4 Go wudoEEoO mcflsom 035mm wan—noum a. cam

 

.OHQMUDHU Sundown—305 m a: Swami; conontgswumaom o£H .H 3an

37

apparently not samarium hexaboride inasmuch as isolation of one of these
led to a very complicated and unresolvable x- ray pattern.

Prior to initiation of the series of 21:2. 5 composition, several
attempts were made to heat phases rich in samarium tetraboride to a
temperature high enough to convert them to samarium hexaboride in a
-molybdenum cell. A portion of the residue from E-l6-was treated in
just such a manner. At pressures of about 1. 5 x 10"5 mm. and temperatures
4 as high as 17500, no conversion could be effected and SmB4-SmB6 mixtures
merely lost the SmB6 phase and coated the Vycor confining tube with a
heavy deposit of samarium containing material.

Subsequent to the series, E-41 - E-44, the molybdenum crucible
and its lid were heated to 19600. The lid fused to the body of the crucible
and the bottom boiled up within the cell. 1 The interior of the cell was
determined once again to be principally 6 -MoB, which is unaffected by
~ dilute hydrochloric acid or dilute sulfuric acid, but 'which dissolves

appreciably in concentrated nitric acid.

F. The reaction between samarium and boron in a boron
nitride liner

 

 

.Byusing a liner inert to boron, or at least one which showed far
less reaction with boron than molybdenum, it was hoped to determine the
mechanistic path by which SmB6 or SmB4-SmB6 mixtures were produced
fromheating pellet compacts of 5111203 and boron.

A general outline regarding the construction of these liners has
already been provided. . However a few more specific comments should
be made.

Since it was necessary to select a temperature easily attained and
one which would, at the same time, be high enough to prevent SmBO3
formation but low enough to preserve the life of the boron nitride, 16500

was chosen as the temperature for this study. There is no doubt that

38

oxide-boron compacts will undergo reaction atthis temperature, and it
was hoped that boron deposition on the surface of the boronnitride (from
heating) would lower its vapor pressure significantly enough, to prevent
rapid thermal destruction. Fortunately, it did. . In no instance didlthe
boron nitride liner gain weight. Using a liner weighing from ‘1. 7 to 2. 0
g. 10 - 20 mg. of weight were lost in every instance. It is' also note-
worthy that for all the runs performed, the interior of the liner remained
.white, even whenthe liner was heated devoid of contents. . Only when
the boron composition was very low relative to the amount of sesquioxide,
did there appear to be any darkening whatsoever on the interior walls
of the boron nitride, and in these cases, darkening was minor.

. Each heating was performed for one hour because it was belieVed
to represent an amount of time reasonable for the reaction to be complete.
In one instance, an especially large pellet (0. 92 g.) was heated for a
second hour at 16500. The weight loss at the end of one hour was equal
to 16. 0%. At the end of the second hour, the percent weight loss was 16.1.

Temperature measurements were made by sighting from above
the heating cell into the hole in the molybdenum lid.
All samples described below were heated for one hour at 1650 1 10°.

The pressure before heating was < 10"3 mm and after about 15 to 20
minutes, the pressure had again dropped to < 10"3 mm. The only variable
was the starting composition. . Samples were run in duplicate for the
compositions 1:12.5 and 1:7. 5, but all other samples were run in tripli-
cate or quadruplicate. X-ray powder photographs were obtained for
every product and all products were analyzed for boron and samarium.
Sample weights of the pellets, all of which retained their shapes during
the course of the reaction, ranged from 0. 58 to 0. 91 g.; sample weight
was a function of the amount of material needed for analysis. ‘

. For all instances except those in which the starliing ratio wasl:5,

the only phase shown in the x- ray photographs of the product was Sme.

39

In the case of the 1:5 mixtures, the products contained both Sme and
SmB4, with the former being the minor phase and the latter being the
phase which the pellet visually appeared to be. . In only one of the three
1:5 samples did the product appear totally grey; in the other two
samples, the interior of the pellets was grey and only the top surfaces
were blue.

Data regarding these experiments are summarized in Table II.
the columns represent, from left to right:

1. The starting composition of the pellet.

2. The average percent weight loss of the sample.

3. The average value of the ratio of moles of samariumlost
divided by the moles originally present.

4. The average value of x, where 35 represents the number of moles
of oxygen lost divided by the number of moles of boron lost as
a consequence of the heating.

5. The average mole ratio of boron to samarium in the product as
determined by analysis.

6. The average value of the percent samarium in the product.

7. The average value of the percent boron in the product.

G. The reaction between samarium sesguioxide and boron
in a zirconium diboride crucible

 

 

This section of the experimental results is included for the sake of
a more complete report of the investigation. . However, because the
zirconium diboride crucible and its lid were never subjected to very high
extremes of temperature, there is some question as to whether or not
the cell was properly outgassed before use. . Outgassing of the crucible
and lid was performed for two hours at 17500; the cell was then used for
nine experiments exactly comparable to the more extensive investigation
undertaken in boron nitride and reported in Part Fof this experimental

section.

40

 

 

chm $40 o.w nm.o m w.: com:

04m w.mc 0.5 mm.o MA oJN mKHS

m.om Ewe No de : odm Gama:

ion ~.>o 0.0 oo.o mm v.2... m.-§

mom mime m.m 2.6 pm m.mom odd:

oimm m.:. «use. Two 1v mg; mg:

Arm o.m~. mé No.0 :1 wow 04m:

m cs Em om pudoounm cw ”Om umoA Em mood 3:5 coaummomaoo
omduo>< owmno>< Em\m as on. no so .>< § .>< mcfludum

oflom 302 053.? .><

 

 

doc: oowufis Canon .m as economy was oofinowswmom Saunas...“ Cook/non coauomon 0AM. .3 23mm.

41

Once again each heating was performed for one hour at 16500 and
again the temperature was read by sighting from above the heating cell
into the hole in the lid. The emissivity of the lid was observed to be
only minutely different from that of the crucible body.

Sample weights ranged from 0. 60 to 0. 90 g. , heavier samples being
used for compositions lower in boron. The compositions studied are
shown in Table III. Three samples were run for the 1:20 composition,
while only two samples were used for the other three compositions
examined.

For all instances except those in which the starting ratio is 1:5,
SmB6 was observed to be the principal phase of the x- ray powder photo-
graph of every reaction product. In no instance was SmB4. observed, and
for one 1:5 composition, no SmB6 was observed either.

The zirconium diboride heating cell, weighing about 8. 75 g. , showed
weight losses of from three to ten milligrams during the heating of the
1:20 samples, but subsequent to these three experiments, it showed con—
sistent weight increases of from three to twenty-five milligrams.

In the range of low compositions, the heatings were observed to produce
perceptible blue color in the base of the crucible. Since the lower
composition mixtures were heated after the ones of higher composition,
the grow-in of this blue color may be partially a result of cumulative
effects.

The data are presented in Table 111, each column represents the
same things that were presented in Table 11, page 40, for which an ex-

planation has already been presented.

£1. The heating of a sample of samarium hexaboride in
a low-density tungsten cell.

—_._

 

This portion of the investigation involved only seven experiments.

The starting material was analyzed in triplicate for samarium and

42

 

 

©.mm 5.0m m.w ON.O o.® m.m~ 0.0NHH
O.wN «two HXV Pm.o ©.HH m.ON O.mHuH
NJN 9.00 mJu mv.o O.© mJ; O.OHHH
OJVH H.0w ©.N M.N Diva N..~N O.m1.nH
m as Em & podoOandw xOM umod Em mood 55> GoflfimoaaoO
ommho>< swayed/.4 SW\m. cw “mo o5 l>< so. 54 mcflumum
03.3w .><

039m 302

 

 

.opwsonfio Sassoon?“ a“ GOHO£ pom ooflxofldwmom Edwnogmm Cook/pon— Goflooou 0&8 .HHH oBdH

43

boron and found to contain: Sm - 64.0%, B - 30.8%; the remainder is
assumed to be oxygen. * Each of the runs cited. in Table IV was performed
in the order listed. The heating cell was heated for 1. 0 hour at 21700
subsequent to each experiment. . By the end of the experiments, both the
lid and the base of the reaction well containedfissures and the lower
portion of the crucible had swollen considerably. . Cracking was noticed
after the second experiment, and swelling in general was first noticed
after the fourth experiment. . In each instance, the Vycor tube confining
the reaction at pressures of 10"5 to 10"6 mm. was coated with a deposit
completely soluble in 6. 0 111 nitric acid, These acid solutions were
collected as quantitatively as possible and analyzed for samarium content.
Because all samples had sintered and were easily removed, it is believed
that weight data on the crucible and the sample arequite meaningful.
The weight of samarium in the Vycor deposit was added to the weight
increase in crucible lid and body to determine what percent of the sample
weight-change could be considered accountable. , Limits of precision in
the analyses were: Sm - 1'- 0. 5%; B - i 1. 5%. i

. Powder photographs were taken in every instance, but only in the
first two experiments was any SmB4 detectable in these x-ray photo-
graphs. . It is significant that a layer of SmB4 was visually detectable on
the top surface of the sintered compact :after the first and second heatings,
although. in the second experiment the layer was much thicker thanin the
first. The data for these experiments are tabulated on the next page in

Table IV.

 

3::
The excess oxygen is assumed to be in the form of amorphous

material of a formula BOX. While one might suspect the contamination
to be samarium orthoborate, such contamination would tend to raise
(rather thanlower) the percent samarium above 69. 9%, the percent
samarium in SmBé.

4
4.

 

. 05Hn :0

 

mode wououswm 3033 U 0053...
.6000 noo>> 0../000.32 01.00 0.0M 0.00 0.001 0000010 00.0H+ v0.0NH+ as #04 >13
0 oowmd
soon 003...?” 0003’s #0.: mam m.$ 0.2.. mommmé 000+ 2.0? as 00.0 0.3
O 000>H
0030.8 03900 34 0.00 0.00 0.00 0.001 00050.0 0H .:+ 0000+ um 30.0V 013
.303 noun—000a Ho 0000. D 00000
as messes 03080 are com o do was- 35.0.0 3.? we .51 no em; Tm
03500
30G n .003 ”MOHOU
>0u0 002600 03500
£00 303 US $30.30.“ 0 0000H
03 00000: 013 0.00 0.00 0.00 0.50.. m0~m0.0 0,030+ >0.00+ 00 00.0 .013
.Emflo.
0300.900 “:00 0:00.000 0 002.0
es broads 3020 use atom can ohm- 033.0 3.2+ 3.2+ as new mix
3302000
nook> 3,003 mmcfloga
M0000 GBOHA1G0UHom Doomed
00.3qu 0300an 0.00 0&0 0.2. v.31 00000.0 00.0+ N0.:.+ ”—0 #04 ~13
mxumgom 0090000 oaooussoooaq. m 80 03500 70 :3 708 as A08 :3 .0808 .02
soon a? so he so .5. as .>< 4 we as 308% 34 < .030 4 use case 03w

 

ill

 

. II‘I'IIIIOIR 1.

I'llltitt

.1.

to. u

. .0.

.300 Goummoa 33000133 .m fi.00€onox0£ 5300500 00 036.00 0 mo wcfioon 0:8 .>H 030R.

45

I. Misc ellaneous experiments

 

Some miscellaneous experiments which-are of major importance

to the conclusions of this study were performed and are summarized

below:

la.

1b.

2a.

2b.

About 0. 06 g. of samarium metal was heated for 0. 5 hour

at 15400 in boron nitride. Subsequent to this heating, the
interior of the boron nitride was as white as before the

heating was begun.

A second heating of O. 07 g. of samarium in boron nitride

was performed by raising the temperature slowly from 10200
to 17500 over a one hour period. . Only when the temperature
reached 17500 did the pressure within the system exceed 10"
mm. The crucible showed some visible thermal decomposition
and the internally coated with a thin layer of grey substance;
however, not enough sample remained for a powder photograph.
The samarium had volatilized without any apparent interaction
(or reaction) with the crucible.

A 0. l g. mixture of {-325-mesh) molybdenum and SmB6 was
prepared in a mole ratio of 1:1 and heated in boron nitride for
1. 5 hours at 14350. The x- ray photograph of the product
(A-711) contained lines of 6 -MoB and Sme. After the mixture
was returned to the boron nitride crucible and heated for an
additional hour at 15400, the photograph (A-713) showed-lines
ascribable to MoBz and Sme. No lines of SmB. were observed
subsequent to the first 25 second heatings.

Essentially the same experiment was performed using tung-
sten powder and SmB6 in a respective mole ratio of 3:1.

The duration of the heating was extended to 2. 5 hours at 16150.
The x-ray photograph. (A-7l9) revealed the presence of WBZ;

a very faint pattern of SmB6 remained, but no lines of SmB4

were observed.

46

3a. A crucible and lid, both of boron nitride, were made for this

3b.

experiment. The product from run E-l9, a grey-colored
mixture of principally SmB4 witha trace of SmB6 was placed
into the freshly outgassed crucible and heated at 14500 at

10" mm. pressure for 0. 5 hour. . After heating, the material
had sintered, but was visibly a two phase mixture. with blue
color confined to the top surface and to the sides of the wafer
while the bottom was totally grey. The x-ray powder pattern
(A-667) clearly revealed lines of SmB4 and SmBé. . Heating for
an additional 1. 5 hours at 14600 resulted in a totally blue product
whose diffraction powder pattern (A-668) represented only SmB6.
The residue from experiment 3a. was replaced into the same
crucible and. heated for one hour at 15400. The pressure within
the system remained essentially constant in the range of 10"
mm. , and no apparent deposition occurred on the walls of the
Vycor confining tube. X-ray examination (A-669) of the product
at this point showed only Sme. The slightly sintered material
was replaced into the boron nitride and heated for two more
hours at 17500. Since there was still no noticeable deposit on
the walls of the Vycor tube, the temperature was raised to
18500 for a final 15 minutes, after which the crucible and con-
tents were examined. The interior of the crucible was com-
pletely blue at this point, the color of SmBs, and the orifice of
the lid had become partially blocked by a tiny cluster of blue
crystals which were ground, photographed (A-6ZZO) as a powder

pattern, and found to contain no phase other than Sme.

4. A mixture of samarium metal and SmB6, weighing about 0. 6

g. , and prepared in a respective mole ratio of 1:2, was placed
into a freshly outgassed crucible of boron nitride fitted with

a molybdenum lid. The mixture was heated at 13800 for ten

47

‘minutes. . At this point, the product was examined bothvisually
and- withx-rays (A-829) and found to contain SmB, and-SmBé,
the former on the bottom of the sintered mass, the latter cons.
fined principally to the top surface. An additional heating for

tenminutes at 14850 produced no further observable change.

IV . DISCUSSION

The discussion section of this thesis is divided-into three major
sub-divisions. The first of these deals with the considerations to be
made in the selection of a crucible material for use with borides or
boron at high temperatures. Information has been presented regarding
the actual empirical problems encountered in this work. . However, con-
siderable attention has also been devoted to work performed by others
as it relates to the general problem of crucible selection.

The second sub-division of the discussion deals with. the immediate
results and conclusions of this thesis investigation.

In the third and final portion of the discussion are presented some
proposals and recommendations for continuing work on the borides of

samarium .

A . Crucible mate rials

 

There can be no doubt regarding the importance of selecting the
proper container for studying high temperature reactions. Since magnetic
induction offers a convenient means of heating i_n vacuo or in an. inert
atmosphere, the fundamental criterion governing crucible selection is
electrical and/ or thermal conductivity. Additionally, a reaction. vessel
must be able to be readily obtained (or the raw materials easily machined)
and must be inert to the species being studied at any chosen temperature.
Implicit’in these restrictions is the restriction of a high melting tempera-
ture or decomposition point. It is difficult tomeet all of these conditions,
and it becomes especially difficult once the species to be studied contains

free boron or boron available as a consequence of reaction.

48

49

'In this study, a variety of materials was considered and~ ultimately
employed, either directly as containers for the reactions, or indirectly
as some part of the constructed system necessary to achieve the tempera-
tures desired for reaction. Carbon, tantalum, tungsten, zirconium
diboride, titanium diboride, and boron nitride were examined; none of
these proved totally satisfactory for all investigations. A discussion of
individual utility and limitations, as well as a summary conclusion, is
presented herein.

Since molybdenum was the first crucible to be used in this research,
it will be the first to be discussed here. The choice of molybdenum was
made for several reasons. . It is a highly conducting, highmelting, easily
machinable material which was obtainable as a bar of high purity and
which seemed to offer some promise of being inert. The fact that
molybdenum would react with boron was well-known and has been investi-
gated (38-40). . However, it was hoped that although molybdenum was
known to react with boron this would not mean that it would react with
compressed pellets of samarium sesquioxide and boron, pellets which
were pressed in the belief that the samarium sesquioxide was the ingredient
in excess. . In this belief, the molybdenum was used as the crucible material
for containing the reactions between samarium sesquioxide and boron as
well as reactions between samarium and boron. When it was discovered
how actually non-inert molybdenum was, a change was made to boron
nitride.

Of all the materials investigated, only the boron nitride (41) fails
to meet the criterion of being an electrical conductor. .This fact is
particularly distressing since of the materials considered, boronnitride
is easiest to machine and comes experimentallytlosest to being ideal
for studies below 15500. . At higher temperatures, the decomposition of
the nitride reduces the crucible life; cracks and fissures are produced
from thermal decomposition, and there is a greater chance for chemical

reaction between crucible and contents.

50

Since boronnitride cannot be heated by direct means, a- radiation
shield of mate rial such as tantalum, carbon, or molybdenum must be
used. These shields heat the material effectively, but make heating very
cumbersome and clumsy. They are difficult to align, short-lived, and
rather dangerous as a modification to a glass system not originally de-
signed with shields in mind. An alternative is to use the poorer electrical
conductors as liners in materials which are good conductors. If the liner
is small relative to the containing body, temperatures of the containing
body can be considered to be temperatures of the liner as well. The boron
nitride was used in just this fashion--as a liner for a-molybdenum heating
cell.

Since boron nitride has an appreciable vapor pressure (42) at the
desired‘reaction temperatures, one has the problem of finding a-lid
material (different from boron nitride) which can be heated to the proper
temperature but which will not fuse to the crucible on decomposition and
subsequent re-formation of the boron nitride. . A lid of molybdenum was
machined for this purpose. Inusing it, one must consider the possible
temperature differential produced in heating the two different materials.
Empirically, it was determined that if the main body of the boron nitride
liner is confined completely within the molybdenum cell, the temperature
difference between the lid and the bottom of the crucible liner was never
greater than 500, and in most cases less than that. . It was necessary to
. prevent the molybdenumlid from coming into contact with the well Con-
fining the liner. . For this purpose, a 1/32" clearance was allowed to
prevent having to forcibly remove a lid from the main body of the crucible.

Additionally, the use of'a molybdenum lid necessitates a consider-

ation of the possible interactions induced within the system.
If: BN(s) + MBy(g) —-9 No reaction (1)

and if: Mo(s) + MBy(g) ——> ’Mon(s) + MBy_x(g) ‘ (m)

51

then one must not overlook the possibility of molybdenum interaction if

MB or any portion of an MB decomposition (or formation) involves a

Y Y
gaseous molecule containing boron. . Leitnaker (43) has asserted that
there is no evidence whatsoever of any gaseous metal boride molecules,
and thus far there has been no evidence to the contrary. . lnthis work,
both samarium orthoborate and samarium hexaboride were detected on
the lid, but no sufficient amount of data was collected to assert that
samarium hexaboride exists, or can exist, as a gaseous molecule.

Finally, boron nitride does have the very useful quality of forming
a protective layer of boron on its surface after initial use. This layer
is thin enough so that interactions between sample and liner are negligible,
but thick enough to radically inhibit later dissociation of the crucible,
dissociation which is synonymous with self-destruction.

Besides molybdenum and boron nitride, the only other material
used extensively in this study was a tungsten-copper alloy. As an alloy,
the metal could be easily machined whereas this would have been an
exceedingly laborious and time-consuming operation using tungsten alone.
While the tungsten appears to be more inert to boron than molybdenum,
it is still an essentially undesirable material. Brewer and Haraldsen
(44) have reported that within their uncertainty, no difference could be
distinguished between the thermodynamic stability of tungsten and
molybdenum borides.

The three materials discussed above were actually the only
materials to whichany major attention was devoted in this research.

The work comprising the largest portion of experimental data was per-
formed in molybdenum or boron nitride, but heating samarium hexaboride
in the tungsten cell was probably the most significant portion of the
research in terms of useful material extracted from performing a mini-
mum number of very careful experiments. NOnetheless, enough con-

sideration was given to the use of other materials such as carbon,

52

tantalum, and zirconium and titanium diborides that they are deserving
of inclusion in the discussion presented here. They are considered
below.

Graphite is easy to obtain, to machine, and to heat. The principal
reason for not using carbon as a crucible material was merely that its
reducing properties were well-known-with respect to boride preparation.

Inasmuch as the equation:

2 szO3(s) + 22 B(s) —-> 4 SmB4(s) + 3 B202(g) (n)

was believed to represent a possible truth, it was necessary to avoid
carbon in order to offer substantive proof that the reduction of the oxide
was done by boron. Furthermore, the porosity of graphite would have
provided an excellent route to rare earth carbide formation if any free
rare earth metal were liberated in the gaseous state.

Tantalum was avoided because of its tremendously rapid reaction
with boron. Leitnaker, at a_._l. (45), have reported that amorphous boron
and tantalum powder react rapidly at about 9000 with the evolution of a
great deal of heat. However, Johnson and Daane (46) have reported that
as long as pure lanthanum was used, the addition of mixtures of lanthanum
and LaB4 to crucibles of tantalum resulted in no attack on the crucible.
Brewer and Haraldsen (44) have estimated AH?” for tantalum borides
as < -26 kcal. per g. atom of boron, * and classify the tantalum borides
as being more stable than the borides of either tungsten or molybdenum,
whose AH?” values are estimated to range from -5 to -25 kcal. per g.
atom of boron. [There is the notable exception, WZB, for which AH?”
has been estimated to be within the range -28 to -20 kcal. per g. atom

of boron. ]

 

*
Brewer and Haraldsen (44) have asserted that because the entropy

changes within these systems are small and beca)use of the small changes
in léeat capacity with change in temperature, AHZ98 is assumed to equal
AFf at any temperature. This is an assumption, but presumably a valid
one.

53

In 1951, Brewer, e_t a}. (38) suggested that TaB offered considerable
promise as a very refractory material if sufficiently high sintering
temperatures were used to produce it in a compact and non-porous form.
This fact, coupled with the information (45) regarding the surprisingly
slow rates of reaction in the Ta-TaB system in the vicinity of 20000,
tempts one to suspect that TaB may be a good. crucible material for
examining the decomposition of rare earth borides even though tantalum
itself is unsatisfactory.

. Finally, the possibility of using the refractory materials, ZrBz
and/or TiBz, was considered as a means of escaping the dilemma of boron
diffusion into molybdenum and decomposition of boron nitride at high
temperatures. Not enough work has been donewith TiBz in this study to
say anything extensive about it. However, the absence of the phase,
TiBlz, and the stability estimate (44) of AH?” for TiBz (about -36 kcal.
per g. atom of boron), suggests that TiBz might serve as a crucible
material for studying decomposition of rare earth borides. Brewer,

e_t a_._l. (38), have reported that when titanium was heated with more than
66 atomic percent boron, the metal was converted to pure hexagonal
titanium diboride and any excess boron was lost. The titanium borides
are estimated (44) to be slightly less stable than the corresponding
borides of zirconium by about 3 kcal. per g. atom of boron.

The phase diagram for the zirconium-boron system has been re-
ported by Post and Glaser (47) and shows that starting with material
having a composition, ZrBz, at temperatures in the range of 1525 -.
23200, there exists the possibility of forming ZrBlz if boron can somehow
be introduced into the ZrBz lattice. The value of AH§98 for ZrBlz has
been estimated (44) to be about -10 kcal. per g. atom of boron, while the
value estimated for zirconium diboride is approximately -30 kcal. per

g. atom of boron.

54

There is some possibility for using crucibles of zirconium or
titanium diboride, but only if these borides can be obtained in sufficiently
pure form to assure that reactions which occur are not reactions between
reactants and crucible impurities. There can be no doubt regarding their
undesirability with respect to machining since only electro-machining
can be used. . In a magnetic field, these materials do not couple nearly
as well as the metals, andwhile it is easy enough to attain temperatures
of 18500, in decomposition studies,’ the more important temperatures are
higher ones. Using these materials, it is difficult to attain temperatures
from 1850 - 22000. Crucibles of zirconium diboride and titanium diboride
have been used in our laboratory to confine liners of boron nitride and
their utility for this purpose is excellent. Little more can be said except
that these borides do warrant additional examination for use as containers
which may be inert to boron, at least under special conditions.

The selection of crucible materials is therefore seen to be controlled
by both kinetic and thermodynamic considerations. Boron is such a
reactive element at high temperatures that it seems practically impossible
to find a material completely satisfactory for confining syntheses of rare
earth borides or for studying their decompositions. While a material
may be essentially inert to a boride over a wide range of low temperatures,
any boron produced in the thermal decomposition of that boride produces
the instantaneous need for the crucible to be inert to boron as well.

It appears that diffusion processes are exceedingly important in
high temperature boron chemistry, and in view of the small size of the
boron atom relative to the size of the metal atoms within a matrix, this
is not surprising. .

The information obtained by workers at the Vallecitos Atomic
Laboratory in Pleasanton, California, may also be ultimately useful in

the selection of crucible materials. These workers have examined the

55

stability of metal borides relative to the metals-which might be used to
contain these borides as control rods. In one such study, Antony and
Cummings (48) conclude that of the metallicumatrix-boride systems
studiedt(matrices: titanium, zirconium, iron, nickel, copper, silver,
aluminum, and stainless steel) only copper, silver and aluminum were
determined to be compatible with boride compounds at elevated
temperatures. The obvious difficulty with using materials of this sort
is their low melting points. Unlessa system can be engineered to
produce high temperatures with rapid cooling, as in arcumelting, use of.
these metals becomes quite difficult- These workers also have reported
that titanium and zirconium were found to be the least reactive of the
refractory matrices; the diborides of hafnium, zirconium, and titanium
were found to be the least reactive boron compounds regardless of the
containing matrix.

Some additional conclusions of the work of Antony and-Cummings
(48) are cited below:

a) The thermodynamic stability of metal borides appears to be

partially dependent upon the boron density of the compound.

b) If boron density is defined as:

 

density of the compound grams of boron per
molecular weight of the mole of compound
compound

c) Then (b) implies high temperature reaction between a boride
and a matrix if the matrix boride is of lower boron density
than that of the boride to be dispersed. '

d) Diffusion of boron into the matrix is rate controlling, and
reaction is more complete in matrices of a lower’melting
temperature.

The statements as presented are acceptable in those situations in which
there is no interaction between any species other than boride and matrix.

If the reaction between a Species suchas samarium hexaboride and

56

zirconium (results in. the release of samarium concurrently and if that
samarium can possibly react with any of the boride species remaining or
formed then the conclusions are invalid ones and cannot be employed.
Samarium tetraboride has been found to be less stable than samarium
hexaboride; but the boron densities for these compounds may be calcu—
lated to be 1.36 and 1.53 g./cc. respectively.

. Aside from this sort of a qualification, the statements above are
reasonably useful in evaluating crucible materials.

. In summary conclusion, it appears that the use of any metal crucible
for studying the chemistry of rare-earth borides is unacceptable unless
plans are made to utilize the crucible's reactivity with boron as part of
the synthesis or decomposition investigation.

Boron nitride seems. suitable for inertness at lower temperatures
as long as nitride formation is negligible, or at least minimal. The
possibility of using refractory borides offers much hOpe if these materials
can be conveniently sintered, machined, and heated to sufficiently high
temperatures. Once the engineering difficulties are overcome, the rare
earth borides themselves may be the materials in which. to examine the
decomposition and formation of other rare earth borides.. In addition to
zirconium and titanium diborides, tantalum boride in the form of TaB
may be an excellent material for inclusion in this category-

Hoyt, it a_l- (49), have found that rare earth tetraborides and hexa-
borides react severely with carbon when hot-pressed in graphite dies,
the products of the reaction being boron carbide and rare earth carbides.
However, Lafferty (1) has noted that tantalum carbide is a surface im-
pervious to boron. Tantalum may be carbonized by packing a crucible
with sugared characoal and heating in a vacuum for one minute at 23000.
Lafferty has reported that this produces a carbonized surface about
0. 002" thick which is not only extremely stable at 15000, but which also

prevents boron from diffusing into the tantalum.

57

No discussion of this sort is ever complete; but its inclusion,
however lengthy, is necessary to understand the- scope of the problems
besetting the individual who wishes to examine borides at high tempera-
tures. ._ The selection of a suitable crucible material is controlled by the
chemistry of the materials under study. However, it is equally rigorously
controlled by the power of the magnetic field used for heating, by the
availability of facilities for machining and glass-blowing, and even by the
sheer ingenuity of the experimenter. . In an area of chemistry whichis
just beginning to expand, there remains much. to be learned before the
boride chemist can select a particular crucible to study a reaction involving
boron and. still be reasonably certain that it is the best crucible, best in
the sense that it is capable of being heated to the desired temperatures,
and best in being inert to reactants and reaction products at those

temperatures.

B. . Results and Conclusions

 

Every chemical reaction is dependent upon the thermodynamics and
kinetics involved in going from a mixture of reactants to their existence
as products. The very firm conclusion of this study is that the processes
of the formation and decomposition of the borides of samarium are
potentially capable of being more rigorously controlled by kinetic effects
than by thermodynamic ones. It is further a conclusion of this work that
the reaction between mixtures of samarium sesquioxide and boron is not
described simply but that it is a very complicated process dependent not
only upon the composition of the starting materials, but also upon the
nature of the heating cell, the time allowed for reaching a temperature,
the duration of maintaining that temperature, and the extent to which,
and the rate at which, samarium escapes from the vicinity of the reaction.

A more .specific analysis follows.

58

In the first instance, » it is believed that the failure of a molybdenum
crucible to sustain long life when used for heating sesquioxide-boron
compacts is principally the result of the formation of phases of the nature

BO When used in the studies involving just samarium and boron, the

x°
molybdenum crucible does not qualitatively appear to be attacked nearly
as rapidly as when used for compacts of boron and samarium sesquioxide.
The conclusions regarding rapid attack by a phase containing boron and
oxygen is not unreasonable in light of the consideration that only a
minimal amount of the boron content of the pellet is in direct contact with
the crucible. Furthermore, the flow properties of oxides of boron are
much more extensive than those for boron alone, and a considerable amount
of evidence exists for concluding that the reaction of oxide-boron compacts
involves a stage where molten phases containing boron and oxygen exist
and/or volatilize. The importance of these BOx phases should not be
minimized. The mode by which they attack the crucible and the duration
of that attack seems to influence the boride products resulting from heat-
ing various mixtures of samarium sesquioxide and boron in a molybdenum
crucible.

The presence of samarium orthoborate detected on the surface of
the pellets as well as on the molybdenum lid of the heating cells suggests
that the orthoborate may be a volatile species. This provides a mechan-
ism to account for the deposition of high temperature samarium sesquioxide
on the sides and lids of the molybdenum crucible. Hot molybdenum may

catalyze the decomposition of samarium orthoborate as follows:

SmBO3(g) = Sm203(s) + B203(g) (o)

The samarium sesquioxide stays behind and the boron sesquioxide vola-
tilizes if it can escape faster than it can undergo reaction with the lid
01‘ body of the heating cell. The Vycor tube confining the reaction always

carries a white amorphous deposit early in the reactions which turns

59

brown as the reaction proceeds and which ultimately gives a positive
methyl borate flame test. The Vycor appeared to collect more white
material quite early in the reaction for those instances where the initial
boron content of the sesquioxide-boron mixtures was high, rather than
low. As long as any brown color was associated with the Vycor deposit,
dissolution of it in 6 l\_d_ nitric acid was always accompanied by the evolu-
tion of c0pious amounts of nitrogen(IV) oxide. (This result could be
caused by samarium or suboxides of boron present in the deposit.)

At one point in the study, a sample of samarium sesquioxide was
inadvertently heated in a crucible of boron nitride which had not been
outgassed. The observation that most of the samarium sesquioxide was
lost in the heating and the occurrence of tiny beads of material on the
sides of the crucible is understandable if one considers that the principal
impurity of the boron nitride was boron sesquioxide and that samarium
orthoborate may be a volatile species.

When pellets of samarium sesquioxide and boron were heated at
16500 in boron nitride (Experimental Section F), there were no indi-
cations of the complicating effects of borate phases. * In fact, if one plots
the value of 35 for BOX as a function of the starting composition of the
pellet, the result is essentially a straight line with slope nowhere equal
to zero. This result, shown in Figure VIII, suggests that the oxide of
boronliberated from the reaction mixtures of samarium sesquioxide-
boron pellets heated in boron nitride is 121’. the same regardless of the
starting composition and that since no plateau area is observed, one must
conclude that no one simple equation can possibly describe the reaction.
The greater the percentage of boron available for reaction with samarium
sesquioxide, the more radical is the reduction in weight loss . 1 Table ’II

(and Figure IX) shows that the percent of overall weight-loss very closely

 

* .
Samarium orthoborate is never observed if the composition of the

starting materials is greater than ten moles of boron for every one mole
of samarium sesquioxide and the temperatures used are in excess of 15400

60

 

 

 

 

 

01.0 0
48.0
.75-
e7.0
x z
0 9.
(D
“2 .50- -6.0.,
8 2’.
>4 .0.
18' “5.0
, w
3 25v e
"a -4.08
:>
I I l L
5 10 15 20

Mole ratio: Boron to samarium sesquioxide respectively.

Figure VIII. . Mole ratios of boron to oxygen and boron to samarium
versus composition of starting materials heated in boron nitride.

Q 100 - 100 o

 

”‘75

25"

 

Percent total weight loss

 

 

I I l 1
5 10 15 20

Mole ratio: Boron to samarium sesquioxide respectively.

 

mntreures ut 9901 iqfitsm quested

Figure IX. Weight losses versus composition of starting materials
heated in boron nitride.

61

parallels the percent loss in samarium in the range of. low "boron
composition. One may therefore conclude that if the percentage of boron
in the starting materials is increased, the samarium losses are sharply
diminished concomitant with a marked increase in the amount of weight
loss ascribable to oxides of boron. Since this is true and since for the
material lost oxygen ratios are not reflective of simple species such as
BzOz(g) or B203(g), it appears that changing the composition of starting
materials results in a change in the ratio of two or more competing
reactions rather than a change in merely the actual oxide of boron lost.
Yet in view of the complex chemistry of this system, one must not exclude
the possibility of species such as B30 or B70 being volatile at the. tempera-3
ture of the investigations.

. In the above discussion, no consideration was given to oxygen lost
in the form of SmO or 02. There should be no difficulty in accepting
the absence of any appreciable amounts of molecular oxygen, but even if
all the samarium lost were accompanied by an equimolar amount of
oxygen, the boron to oxygen ratio would still show the same trend even
though the slope of the curve would be lessened.

The same general results were observed for the work done using
zirconium diboride except that here the conclusions seem less certain.
The ratio of boron to oxygen lost was more closely constant, as seen
in Figure-X, but the value of x in BOX still increased with a decrease in
the boron content of the starting composition. . Aside from the series using
a. 1:5. 0 starting mixture, the value off was less than for comparable
compositions examined in boron nitride.

The most striking fact about the studies performed in zirconium
diboride was the low percentages of samarium lost on heating (Table III,
and Figure‘XI). . Coupled with the observed weight increases in the
crucible and the grow-in of the blue color in the base of that crucible, the

low percentages of samarium lost may reflect the formation of samarium

62

 

 

 

 

 

 

02-60 9.0 o
p
2.20— I --8.0
/
/
/
1.80— / -7.o§
/ (‘D
l/ H
1.40_ ,0 -16.0 2’.
>4 ’ 5'
o l// ..
m1.oo— / -5.0 w
‘5 / E
w
M 3
1H 60— 44.0
0
<1) .
:3
76 .20» ‘O\0 e 3 o
:>
J l l I 2.0
5 1o 15 20

Mole ratio: Boron to samarium sesquioxide, respectively.

Figure XV Mole ratios of boron to oxygen and boron to samarium
versus composition of starting materials heated in zirconium diboride.

‘O

 

25 25

 

 

 

 

umpeures ssoI iqfitem 1U9319d

U)

8

._. 15 - — 15
4.)

.1:

.2.”

Q)

3

76‘

‘5‘

*4 5 - i 5
4.:

c:

Q)

g 1 1 1 1

Q)

C1. 5 10 15 20

Mole ratio: Boron to samarium sesquioxide, respectively.

Figure XI. Weight losses versus composition of starting materials
heated in zirconium diboride.

O

63

hexaboride if samarium could be released from the pellet of boron and
samarium sesquioxide, and if any excess boron were present in the

zirconium diboride. The reaction:

Sm(g) + 3 ZrBz(s) = SmB6(s) + 3 Zr(g:1 (p):

is not likely either thermodynamically or inlight of the observations made
on boron density (48). However, since no analysis of ZrB-fig was performed
or provided, the first possibility is not to be excluded.

Data presented in Table I for the "E" series also shows the marked
effect of the boron concentration of starting materials upon the weight
losses within a system, even over a variety of temperatures. The weight
loss of samples having compositions of 1. 0 mole of samarium for every
5. 0 moles of boron is about one—eighth of the weight loss of samples
one-half as rich in boron.

Before leaving the discussion of the studies performed in zirconium
diboride and in boron nitride, mention should be made regarding the
implications of the variation in the ratio of moles of samarium to boron
observed in the products. Two assertions are worthwhile.

No phase higher than samarium hexaboride was observed in the
x-ray powder photographs, and no boron was ever detected in any of the
photographs obtained in these two series of reactions. The first asser-
tion is neither surprising nor exceptionally significant since samarium
dodecaboride was never knowingly obtained at any time. . However, the
second assertion is important since it demonstrates the difficulty in
detecting free boron with x- ray analysis. Even when powder photographs
were made bf boron-samarium sesquioxide starting. mixtures of 11:10
composition, no boron could be detected, although boron was observed on
the powderphotograph of a sample whose starting composition was 1:20.

Returning to the reactions occurring inimolybdenum, it may be

asserted that the longer the extent of heating of an oxide-boron compact

64

in molybdenum, the greater the chance of obtaining a product rich in
samarium tetraboride. Inthe early work, a mixture of the two borides,
prepared in seven hours, was observed to become depleted in samarium
- hexaboride asa consequence of eight hours of additional heating. Even a
l: 10 starting composition resulted in the preparation of samarium tetra-
borideewith no detectable samarium hexaboride when a 14. 5 hour heating
was employed at temperatures as low as 15100.

The general observation that long duration heatings in molybdenum
favor production of the tetraboride is borne out clearly in experiments
E-l3 and E-l4. X-ray photographs of the products showed no phase other
than samarium tetraboride when heatings were 10. 5 and 20. 0 hours
respectively.

This particular generalization cannot be extended directly to the
study performed in the tungsten cell since none of these heatings was
longer than 4. 0 hours. 1 However, it is noteworthy that the amount of
samarium tetraboride produced in the 4. 0 hour heating at 17500 (H-Z)
was ostensibly greater than that amount produced in the 1. 5 hour heating
at the same temperature, and that in none of the instances to follow (H-3
through H—7) was any tetraboride detected either visibly or in the
diffraction powder patterns of the products.

At this point, an analysis of diffusion phenomena seems necessary
and pertinent. The data of Tables I and IV show that in every instance
but one (E-l9) where weight data were recorded for the molybdenum or
tungsten crucibles, the crucibles gainedweight. It is primarily significant
that for the one instance where an overall weight lo_siwas recorded, the
heating represented the longest time a reaction of the "E" series was
held at a temperature as high as 16950. . If boron is the element reSponsible
for these weight gains, then it seems plausible to suggest that in the in-
stance where a weight loss was recorded for the crucible, the conditions
of time, temperature, and previous treatment were all favorable for this

result. The -molybdenum borides are capable of decomposing in this

65

temperature range, and arproper set of conditions could easily result in
loss in crucible weight as boron diffuses from the crucible into the
contents.

For the studies performed-in the tungsten crucible, the greatest
mass gain in crucible and lid (under 21700) was observed for the 4. 0
hour run, ,H-Z, an experiment performedvat a full 2000 less than two other
comparable (4. 0 hour) experiments. Since the product of experiment H-Z
contained the highest percent samarium and the lowest percent boron of
any of the tungsten heatings, the high weight gain in crucible and lid for
this experiment seems to be more a consequence of being the second
experiment performed in this heating cell than as a consequence of being
at 17500. Since the crucible is still close to being "new", boron can
diffuse into the cell at a very rapid rate. The crucible weight gain observed
in experiment H-l is about the same as that observed in H-2 even though
there is a 2. 5 hour difference in the duration of heating. ~Yet under the
same conditions of time and temperature as those employed for experi-
ment H-Z, the total weight gain observed in» H-5 was only about seven-
ninths of the amount observedin H-Z; furthermore, there was a marked
difference between the two experiments in. the percent samarium in the
product.

- It should be noted that the datain Table IV do assert that an increase
in weight gain is favored by either higher temperatures or increased dura-
tion of heating and that increase in both of these variables has an associ-
ative effect.

The H-6 (experiment clearly represents the low extreme of the study.
The total weight gain in crucible and lid is lowest; the percent loss in
sample weight is the lowest; the percent samarium in the product is lowest,
and the percent boron is the highest. It is of interest to note that for the

two instances where samarium tetraboride was observed in the products

66

after heating (H-1 and H-Z), the percent samarium in the products was
the highest for any of the seven heatings.

Regardless of how many numbers are associated with an experiment,
they are of essentially no value unless they can be interpreted meaningfully
and with internal consistency. In attaching a finalizing significance to
these. experiments, the two most pressing questions are: Why does
samarium tetraboride decompose in boron nitride at temperatures far
lower than those at which it exists stably in molybdenum? Why is

samarium tetraboride observed as a constituent of the upper surfaces of

 

several pellets of boron and samarium sesquioxide heated in molybdenum,
of some samples of samarium hexaboride heated in tungsten, and, indeed,
of some of the sesquioxide-boron pellets heated in boron nitride, while

samarium hexaboride is clearly the top surface component after a sample

 

of samarium tetraboride has been heated in boron nitride? There is also
a need to justify why samarium hexaboride is the only boride of samarium
present in addition to borides of molybdenum or tungsten after heating a
mixture of samarium hexaboride and molybdenum or tungsten powder in a
crucible of boron nitride. [Similar phenomena are independently reported
by Antony and Cummings (48).]

In answer to these questions and in view of the experimental results,
samarium tetraboride must be thermodynamically less stable than

samarium hexaboride; but the following equilibrium exists:
3 SmB4(s) == Sm(g) + 2 SmB6(s) (q)

It is suggested that the mutual existence of samarium tetra- and hexa-
boride is possible not only as a consequence of equilibrium but also
because of the inter-relating effects of diffusion and surface evaporation.
It has been noted that the reaction between samarium metal and samarium
hexaboride does, indeed, result in the formation of some samarium

. o
tetraboride at temperatures as low as 1435 .

67

Lafferty (1) has shown that heating lanthanum hexaboride in contact
with a base material capable of absorbing boron by diffusion into its
interstices can effectively permit the evaporation of lanthanum atoms.

»The mechanism is simply that the base material will absorb boron from
the lattice of the hexaboride and that as the boron framework of the hexao
boride collapses, the metal atoms are released.

The conclusion of this work is that exactly the same phenomenon
is demonstrated by samarium hexaboride. . If heating samariumhexaboride
in tungsten results in the production of the less stable samarium tetra-
boride on the surface of the sintered product, and if this process ceases
to be observable as the crucible material is used again, even though the
crucible weight gains continue, then it seems reasonable to suggest that
the gains in crucible weight are due to boron diffusion and that the observa-
tion of samarium tetraboride on the surface is only a consequence of the
diffusion of samarium atoms to the surface at a rate faster than they can

actually evaporate. In boron nitride, samarium tetraboride only partially

 

decomposed to the hexaboride after 0. 5 hour and in this case, the top
surface was blue (SmB6) rather than grey (SmB4). Clearly then, evapora-
tion of samarium from samarium tetraboride heated in boronnitride is

» primarily a surface phenomenon whereas in the tungsten heatings,
samarium released from the hexaboride lattice—collapse must diffuse to

an area of evaporation faster than it can actually evaporate- When. the
temperature is very high, e. g. H-7, large weight gains in the crucible

may occur without observing any tetraboride on the surface of the "product"
since now the evaporation rate equals the rate of diffusion.

Some discussion should be devoted to the ease with which samarium
tetraboride can remain the principal phase of mixtures prepared from
heating samarium and boron (even at a variety of temperatures and compo-
sitions) in armolybdenum crucible. Since the tetraboride is less stable

than the hexaboride, then subjecting samarium tetraboride to high

68

temperatures should result in conversion to samarium hexaboride unless
some mechanism exists for preventing tetraboride decomposition. The

only instances for which, samarium tetraboride appears to remain stable
when heated in molybdenum are those in which the material was prepared
from a direct reaction of the elements under-a helium pressure of about

600 mm. or those in which a very long duration of heating was employed.
Since an equilibrium exists between the two borides, the effect of the

helium inhibiting the vaporization of samarium may be responsible for the
ease with which samarium tetraboride is prepared in molybdenum from a
direct reaction of the elements. Heating these same mixtures of tetraboride

in vacuo results in their conversion to hexaboride unless boron diffusion

 

into molybdenum becomes so great that the tetraboride appears to reduce
its weight without changing phase.

The absence on the powder patterns of lines of a lower boride of
samarium after samarium hexaboride was heated with molybdenum and/or
tungsten, and the continued presence of lines of SmB6 in these same powder
patterns that show MoB or WBZ, demands that the rate of evaporation of
samarium from the areas of its production equals the rate of diffusion of
the gaseous atoms through the solid compacts.

It must be recognized and appreciated that the ultimate result of
any of these experiments says nothing about the time dependence of the
rates of diffusion and/or evaporation, but merely reflects the gross
amount of change that occurs in a given amount of time at a given tempera-
ture.

The hypothesis present in Part I of this thesis must now be recalled.
It was suggested that rare earth metals possessing a high vapor pressure

would show a tetraboride decomposition according to the path:
A
3MB4(s) ——> .M(g) + 2 MB6(s)

and a hexaboride decomposition at a higher temperature according to the path:

69

MB6(s) -—A—->.M(g) + 6 B(s)

It is easy to see, now, the complications involved in proving or dis-
provingthis hypothesis. There must exist a container inert to the borides
and to boron at temperatures sufficiently high to permit observation of the

overall sum of the reactions:
3MB4(s) ——> 3 M(g) + 12 B(s)

Yet at the temperatures needed for this, we are faced with another possible
reaction:

13(6) --> B(g)

whose rate will determine not only the remaining product to be examined
after materials are subjected to high temperatures but will also determine
the extent of possible recombination between M(g) and B(g) in a cooler area
such as a lid orifice. . It may be that this is the process which“ led to the
formation of crystals of samarium hexaboride in the orifice of the boron
nitride crucible (as described in Part III). One is faced with choosing
between a gaseous metal boride molecule, a recombination of gaseous
boron and samarium, or the reaction between gaseous samarium and solid
boron produced in the decomposition of the boron nitride. . Considering the
extent of the crystal growth in the orifice, the perfection of the crystals
themselves, and the heretofore absence of gaseous metal boride molecules,
the recombination process seems the most likely one.

Nonetheless, no absolute assertion can be made without a mass
spectrometer and the need for this kind of examination should be obvious.

The decomposition of samarium tetraboride into samarium hexa-
boride when the former is heated in a cell or boron nitride and the form-
ation of the hexaboride when samarium and boron are caused to react
directly in boron nitride offers substantial proof to the notion that samarium

tetraboride is the less stable of the two borides and that its decomposition

70

product is the hexaboride. Unfortunately, the existence of the tetraboride-
hexaboride equilibrium, the lack of a mass spectrometer, and the absence
of a container totally inert to boron at the high temperatures necessary to
decompose samarium hexaboride, whose melting point has been estimated
to be about 28000K (49), make proof of the second decomposition path a
rather severe problems.

Daane and Johnson (46) have reported the partial conversion of
lanthanum tetraboride to lanthanum hexaboride when the tetraboridewas

heated for fifteen minutes in vacuo in a tantalum crucible. -They observed a

 

surface layer of lanthanum hexaboride, just as we observed a surface layer
of samarium. hexaboride. Complete conversion to lanthanum hexaboride
could be affected when a pressed mixture of boron and lanthanum tetraboride
was heated. i_r_1 wfor fifteen minutes at 15000.

It is both interesting and satisfying to note the consistency between
the work of Lafferty (1) in 1951 and the work of Daane and Johnson (47) in
1961. . Without evidence to: the contrary, the samarium-boron system appears
to behave quite similarly. Yet no definite proof has been obtained to justify
the assertion that its particular behavior is an immediate consequence of the
high vapor pressure of samarium metal relative to the vapor pressure of

other rare earth metals.

C.- Proposals for future work

1.. Not enough attention has been devoted to the crucible materials .
for use in high temperature studies of the rare earth borides. There is
a severe need for quantitative measurements on boron and rare earth
interaction with a variety of crucible materials. Until these fundamental
investigations are made, crucible materials will continue to be a variable

with effects of undetermined magnitude.

71

2. With particular reference to the samarium-boron system, the
gases"”evolving fromthe reactions between various mixtures of samarium
sesquioxide and boron should be examined with a mass spectrometer if
any unequivocal conclusions are to be made regarding the mechanism by
which samarium tetra- or hexaboride is formed.

3.. A thorough examination should be made of mixtures of samarium
hexaboride and boron. . One could compare the lattice parameters of pure
and impure samarium hexaboride and determine the relationship between
the parameters and the percent boron in the materials. Since boron seems
so insensitive to powder diffraction detection, it would be of interest to
know what percent of boron could be added to a sample of pure samarium
hexaboride before detection of the boron could be made. The compilation
of these data might result in a more complete characterization of the
nature of samples whose x-ray powder photographs show only samarium
hexaboride. There is still uncertainty regarding the form of sample
impurity and its effect on sample behavior. Such fine points could be
carefully and neatly examined and should prove quite useful towards pre-
venting future misinterpretations.

4. If a truly inert crucible can be obtained, then studies should
be performed aimed at measuring the equilibrium constant of the samarium—
boron system. These studies might be approached using arKnudsen cell
to measure the vapor pressure of samarium escaping from a sample of
samarium tetraboride. . Of course, the use of this method is precluded by
a need to know the exact molecular weight of the vaporizing species.

. There is a possibility that the equilibrium constant might be more
easily measured by quenching a series of samples heated in a closed
system. Since the crystal structure of the two borides is different,
perhaps an x-ray diffractometer could be used to quantitatively analyze
the composition of the quenched phases.

The ultimate goal in equilibrium and vapor pressure studies would

naturally be to obtain values for the free energy of formation of the borides.

72

5.. Finally, the use of an appropriately balanced—magnetic field and
the consequent levitation of a sample of material. offers exciting possibilities
for a rigorous examination of the) decomposition of rare earth borides with-
out the problem of crucible interaction. . However the experimental details
of this sort of technique are not simple and a» levitation approach to the

problem is probably. still only‘a. distant possibility.

LIT ERATU RE CIT ED

1.. J. M. Lafferty, J.-Appl. Phys., _2_2_, 299 (1951).

2. .H. Moissan and P. Williams, Compt. rend. ,. L25, 629 (1897).

3- M. Geelmuyden, Com-pt. rend., L30, 1026 (1900).

4. E. Jungst and R. Meines, Chem.. Zentr., I, 195 (1905).

5. A. Stock and W. Holle, Ber. deut. chem...Ges., if: 2095 (1908).
6. L. Andrieux, Compt. rend., _l_§_2_, 126 (1926).

7. S- M. von Stackelberg and F. Neumann, Z. physik. Chem., 19B,
314 (1932).

8. G. Allard, Bull. soc. chim- France, _5_1_,. 1213 (1932).
9. S. Flodmark, Arkiv Fysik, E, 513 (1959).
10. A. Zalkin and’D. H. Templeton, Acta Cryst., 6, 269 (1953).

11. B. Post, D.. Moskowitz, and F. W. Glaser, J.. Am- Chem. Soc.,';l_8,
1800 (1956). h

12.. F- Bertaut and P. Blum, Compt. rend., 229, 666 (1949).
13.. B. Post and F.' W. Glaser, J- Metals, 4, 631 (1952).
14. I. Binder, 5.. La Placa, and B. Post, Proceedings of the Conference on

.Boron, Plenum Press, Inc. ,. New York, New York (1960).. Also,
private communication.

 

15. R.. T. Sanderson, Chemical Periodicity, Reinhold Publishing Corporation,
New York, New York (1960).

 

16. J. J. Katz and E- Rabinowitch,. The Chemistry of Uranium, Part I,
McGraw—Hill Book Company, Inc. ,. New York, New York (1951).

 

17. F. H. Ellinger and W. H. Zacharaisen, J. Am. Chem. Soc., 15,
. 5650 (1953).

73

74

18. P. W. Gilles, Private Communication.

19. E. J. Felten, I. Binder, and B. Post, J. Am. Chem. Soc., _8_0_, 3479
(1958).

20. G. V. Samasonov, V. P. Dzeganovskii, and I. A. Semashko,
Kristallografiya 4, 119, (1959); C. A. 55, 7125g (1961); also Doklady
*Akad. Nauk 5.5.5.11. ,, Ural- Filial 1193—506 (1958); C'..A.~.5_3_,.— 3962a
(1959).

21. G-A. Kudintseva, M. D.. Polyakova, G. V. Samsonov, and B- M.
Tsarev, Fiz- Metal'l. Metalloved., _6_,, 272 (1958); C. A. 22, 987l_i_
(1959).

. 22. V. S. Neshpor and G. V. Samsonov, Zhur. Fiz. Khim. _3_2_, 1328
(1958); C. A._5_2_, 4370h (1958); C. A. 23, 9533 (1959).

23. Yu. B. Paderno, T- I. Serebryakova, and G. V. Samsonov, Doklady
Akad. Nauk- S.S.S.R., 125, 317 (1959); C..A. _5_5_, 1115911 (1961).

24. G. V. Samsonov, Uspekhi Khim., _2_8_, 189 (1959). (Translated by
F. Hilton.)

25. J. L- Margrave,. J. R. Soulen, and D. Sthapitanonder, J. Phys- Chem.,
_52, 132 (1955).

26. E. M. Levin, R. S. Roth, and J. B. Martin, Am. Mineralogist,
(1961, to be published).

27. R. Ray and P. Sinha, J- Chem. Soc., 742 (1941).
28. H- H. Kahlenberg, Trans. Am. Electrochem. Soc., 41, 23 (1925).
29. R... A. Pasternak, Acta Cryst. , .1_2, 612 (1959).

30. M. G. Inghram, R.. F. Porter, and W. A. Chupka, J. Chem. Phys.,
35, 498 (1956).

31. M..D. Scheer, J. Phys. Chem., fl, 1184 (1957).
32. M. D. Scheer, J. Phys. Chem., .6_Z_, 490 (1958).

33.. F. A. Kanda, A. J. King, V. A. Russell, and W- Katz, J. Am- Chem.
Soc., _7__8_, 1509 (1956).

34..

35.

36.

37.

3.8 .

39.

40.

41.

42.

43.

44.

45.

46..

47..

48..

49.

75

J. M...Letinaker, N. H. Krikorian, and M. G. Bowman, Los Alamos
Scientific Laboratory, LosAIamos, New Mexico. Paper presented at
a Symposium on High. Temperature Poisons, Albuquerque, New—'Mexico.
February 1961. ‘

 

N. V- Sidgwick, The Chemical Elements and Their‘ Compounds,
Clarendon Press, Oxford, England (1950). T

H. A. Eick and P. W. Gilles, J. Am. Chem. Soc., 81, 5030 (1959).

H. Blumenthal, Anal. Chem., _2_3_, 992 (1951).

.L. Brewer, D. L. Sawyer, D. H. Templeton, and C. H. Dauben,

J. Am. Ceram. Soc., 34, 173 (1951).

ClimaxMolybdenum Company Bulletin. Cdb-8, March 1956. Available
from the Climax Molybdenum~ Company, 500 Fifth Avenue, New York
36,. New York. ‘

A. Chré‘tien and J. Helgorsky, Compt. rend., 252, 742 (1961).

A. A. Giardini, Boron Nitride, a review; Bureau of Mines Information
Circular 7664 (1953).

 

R. Lorenz and J. Woolcock, Z. anorg. Chem., 176, 289 (1928).

J. M- Leitnaker, Los Alamos Scientific Laboratory, Los Alamos,

. New Mexico. Technical Report No. LA-2402 (1960).

. L. Brewer and H- Haraldsen, J.. Electrochem. Soc. , 102,. 399 (1955).

J. M. Leitnaker,. M. G. Bowman, and P. W. Gilles, J. Electrochem.
Soc., 108, 568 (1961).

R. W. Johnson and A. H. Daane, J. Phys. Chem., _6_5_, 909 (1961).

. Also, private communication.

F. W. Glaser and B. Post, AIME Trans. ,. 197, 1117(1953).

K- C. Antony and W. V. Cummings, Vallecitos Atomic Laboratory,
Pleasanton, California. AEC Research and Development Report
GEAP-3530.

E. W. Hoyt, J. Chorné, and W. V. Cummings, Vallecitos Atomic
Laboratory, Pleasanton,. California. AEC Research and Development
Report GEAP-3548.

51.

52..

53.

54.

55.

76

. F. H- Spedding and A. H. Daane, ~The Rare Earths, John Wiley and

 

Sons, Inc. , New York, New York (1961).

J- O’M. Bockris, J. L. White, and J. D- Mackenzie, Physico-;Chemical
Measurements at High Temperatures, Butterworth‘ Scientific Publications,

 

 

London (1 959) .

D. H. Templeton and C. H. Dauben,. J- Am. Chem. Soc., 16, 5237‘
(1954).

N. N. Tvorogov, Zhur. Neorg. Khim., 4, 1961 (1959); C. A. _5_‘_l_, 117883:
(1960).

I. Binder, J. Am. Ceram. Soc., :13, 287 (1960).

S. La Placa, I. Binder, and B. Post, J. Inorg. 81 Nuclear Chem.,lfi,
113 (1961).

56. X-Ray Powder Data File, American Society for Testing Materials,

 

Philadelphia, Pennsylvania (1961).

APPENDICES

77

APPENDIX A
SUPPLEMENTARY EXPERIMENTAL DATA

The series of experiments that are subsequently discussed repre-
sent attempts to prepare pure samarium tetraboride. Because of the
importance of maintaining the identification of the various phases observed
in these experiments, the original numbers assigned to the x- ray powder
photographs have been retained. An A-number associated with a phase
implies that the x- ray photograph has been taken for that phase even
though direct mention of the photograph may have been avoided. In. all
context, the word photograph is to be understood to be associatedvonly
with x- ray powder diffraction photographs; no other kind of photography
was employed.

Although compositions are casually referred to as 1:10, 1:8, 31:6,
etc. , in all cases they represent mixtures of samarium sesquioxide and
boron carefully prepared in respective mole ratios of 1. 00:10. 00,

1. 00:8. 00, 1. 00:6. 00, etc. , and weighed to the nearest one-tenth of a

milligram.

D- Reactions of samariugsesquioxide and boron in a
molybdenum crucible

 

 

1. A 1:10 pellet-mixture (0. 29711 g.) of samarium sesquioxide and
crystalline boron was placed-into a freshly outgassed molybdenumcrucible.
(This crucible had been used only once prior to this run and only for an
experiment in which the temperature of reaction was too low to result in
any identifiable phas'esis) The crucible was equipped with a strip of
tantalumfoil (0. 005") between its lid and the main body of the crucible.

The temperature was slowly increased over an interval of 15 minutes to

78

79

14000 and maintained at that temperature for an additional 45 minutes.
Subsequent to this heating, the pellet appeared light grey on its top
surface which was scraped and photographed (A-342). The lid orifice of
the crucible had clogged with a slight amount of yellow white material
(A-343). Both of these photographs show the same phase, samarium
orthoborate, but the photographs also show the absence of SmB4’, .SmB6,
and 6 -MoB. . X-raydata for SmBO3 and 6 -MoB is supplied in Appendix B.
It is interesting to note that the phase of samarium orthoborate observed
was the "15500 form" as recorded by Levin e_t a_l. (26).

Since the mixture was only light grey, it was reheated in the same
crucible at 15400 for one hour. The pellet had darkened considerably
and a photograph of the top surface (A-349) showed the principal phase
to be SmB4. (The photograph leaves some question as to whether or not
there was any SmB6 present, and if there was it must have been in. very
small amounts; the cubic SmB6 phase is much more sensitive to x-ray
detection. than the tetragonal SmB4 phase.) The pellet was returned to the
crucible and was heated for a third time for slightly under one hour at
15800. The lid had become stuck to the crucible by a ring of light yellow
material (A-345); the pellet itself was removed and the surface photo-
graphed (A-346). . Photograph A-345 appears to contain some lines of
B-SmZO3 (high temperature form). . However, photograph A- 346 shows
almost completely SmB4; only the slightest trace of SmB6 can be detected.

Still a fourth heating was performed by returning the pellet to the
heating cell and heating for 1. 5 hours at 17800. The pellet was again
scraped for x- ray analysis (A-348) and enough of the dark deposit from
the walls of the Vycor containing-tube was removed for a photograph
(A-347). The pellet had diminished in size ostensibly. Examination of
A-348 showed a faint trace of SmB6 contaminating the principal phase,
SmB4; however, A-347 could not be classified from the phases shown on

the photograph.

80

2. The molybdenum crucible used above was heated with 0. 07 g.
of boron at 14250 for one hour- The crucible and contents experienced a
weight loss of 0. 00086 g. , and the boron had become coated witha dark
blue layer; a second heating for another hour at 15900 produced some
additional blue onthe surface of the boron and a flaky transparent deposit
on the lid of the crucible. The lid deposit was photographed (A-a351) but
the very few lines on that photograph are unidentifiable.

Still a third heating was performed for one hour at 18800. .At the
end of this third heating, the inside of the crucible had become quite
bright; the dull film,which had coated the innards of the reaction well prior
to the heating, was gone. The residual boron had turned blue, and the
photograph (A-352) of some of the material from the surface of this
residue showed only lines of SmB6. There was additionally a bit of
white-yellow deposit at the top of the crucible (A—353); this material was

mostly B-szO3.

3. Using the crucible pre-treated as outlined in (2),, a 1:10 mixture
of (0. 21337 g.) of szO3 and crystalline boron was pressed into a pellet
and heated (without opening the system betweenheatings) as follows:
10.0'minutes at 14800; 10.0 minutes at 15650; 15.0 minutes at 16750;
20.0 minutes at 17350; 20.0 minutes at 17650; and 10.0‘minutes at 18100.
The deposit on the Vycor tube at the end of this series was very thin.
What remained of the pellet resided at the bottom of the crucible in a
snowflake-like disk and was photographed as A-367. The only phase
shown in this photograph is SmB6. There were three additional photo-
graphs associated with this exPeriment: a crystalline deposit on the inside
of the Vycor tube (A-366) showing no identifiable lines, and A-368 and
A-369 which represent respectively the deposits on the inside surface and
lip of the molybdenum lid. . Both A-368 and A-369 are devoid of the

B-szO3 pattern, but otherwise cannot be identified.

81

4- Once again the crucible was heated with 0. 06 g. of boron.
Heatings-were performed for 30 minutes at 14350 and 30 minutes at 17800.
The surface of the boron powder had become coated with a blue color,
presumably SmBé, but the interior of this partially sintered mound of
boron was distinctly black and merely boron.

5. Another 1:10 pellet (0. 18466 g.) of szO3 and boron was prepared
for heating in the same molybdenum crucible used above, but this time
at lower temperatures.

The first heating was performed for 20 minutes at 13600 and the
second for 30 minutes at 14050. The Vycor tube had become coated with
a barely discernible white deposit which was not removed prior to the
second heating. When the system was opened to the atmOSphere, the
inside of the crucible lid was observed to be coated with a glassy material
which looked as if it may have been molten at one time. The lid, still
warm, was a bit sticky, with an amount of yellow material clogging the
orifice. The pellet was grey on top, the color of SmB4, but tam—white on
the bottom. . Material was scraped from both the top (A-371) and the
bottom (A-372) of the pellet. Both A-371 and A-372 show the same phase,
the 15500 form of SmBO3, although A-371 appears to contain a few more
lines of this phase than A-372. Both of these photographs represent the
same phase that appeared in A-342, described in (1).

The crucible and pellet were again returned to the line and heated
for one hour at 15700; the white deposit on the Vycor tube had become
thicker and had turned to a golden brown. The pellet appeared black on
both the t0p and bottom surfaces. The only difference between the sur-
faces was that the bottom was flaky and contained blue specks, while
the top appeared more homogeneous. The top of the pellet was scraped
and photographed (A-37 3) showing the presence of both SmB6 and SmB4.

.Once more the same pellet was returned to the vacuum line and

. . 0 . . .
heated in 15 m1nutes to 1790 . Examination of the material subsequent

82

to this brief heating revealed the presence of grey color in the pellet
and a deposit (A-376), principally SmBO3, on the lid of the crucible.
The inside of the crucible was coated with material photographed as
A-377.-. The pellet proper was photographed as A~374 which shows the
presence of both borides, but in which the SmB6 lines appear weaker
than they do in photograph A~373. The photograph (Au377) of the material
, coating the innards of the crucible shows only lines accountable as
B~szO3.

-6- Another 1:10 pellet (0,. 02955 g.) was prepared and initially heated
for 45 minutes at 15500 in the same molybdenum crucible used above.
On examination, the pellet was observed to be stuck to the bottom of the
crucible with material which appeared to have oozed out of the reaction
mixture. The lid was coated with a flaky, glassymlooking material, and
a white sandy residue was on the bottom of the crucible. The following
photographs were prepared: A-388, from the material which comprised
the lid deposit; A-389, from the material with which the pellet was stuck
to the crucible; and A~390, comprising the sandy residue. . Photograph
A-388 shows exactly the same phase as that previously recorded for Ae376.
Photograph A—389 and A-390 have not been characterized but do not
appear to contain any lines of B-szO3.

The pellet was returned to the line and heated from 17150 to 18300
in about 30 minutes. At the end of this heating, the molybdenum lid
orifice had clogged with the yellow material previously encountered.
Much deposit had lined the inside of the Vycor; this was scraped, ground,
. and photographed as A-394. .Most striking of all developments was the
inside of the crucible. While the lid was again coated with a deposit, a
layer of crystalline material had grown across the entire interior of the
crucible proper, completely closing off the path of any exit material.
. Nonetheless, there was still enough of the pellet at the bottom of the
crucible from which to prepare a powder photograph. Four photographs

were taken of the components of the reaction: A-397, representing the

83

top of the pellet; A—400, using scrapings from the bottom of the pellet;
A-401 and A-402, from what appeared to be two different phases deposited
on the lid of the crucible. The layer of deposit that had formed midway
up the crucible was later identified as 6 -MoB (A~410).

Characterization of A-394 was never accomplished; the photographs
A-397 and A-400 both show the presence of SmBé, SmB4, and a few lines
from another unidentifiable phase. The materials photographed as A-401
and A-402 both contained 6 -MoB, but A-402 shows the additional presence
of another contaminating phase. Subsequent to this run, the crucible
was heated for a total time of about four hours, two of which were at 14350-
17050, one at 17450, and one at 19000 - 19500. The crucible was bright
and shiny after this treatment, but a portion of the lid had cracked and
was photographed as A-411. This photograph shows only lines accountable
as 6 -MoB.

7. Still another 1:10 pellet (0. 14985 g.) was heated in the same
crucible for seven hours at 14650. The top surface of the resulting pellet
was photographed at A~457 which shows a mixture of SmB4 and SmB6.

The material was returned to the crucible for another heating at 14400

for six hours; the temperature was then raised to 16400. for an additional
two hours after which the heating was stopped. This time, the photograph
(A-456) shows only lines of SmB4 with a minute trace of SmBé.

8. The crucible was returned to the vacuum line and heated (without
boron) to a temperature of approximately 18500. . At about 16400, a
deposit began to collect on the Vycor confining-tube. After about 0. 5
hour at this temperature, the cleaning operation was stopped and the
Vycor was scraped for enough material to photograph as A-486, material
whose photograph shows the presence of some C-szO3 (cubic low-
tempe rature (modification) .

9. Subsequent to this cleaning, another 1:10 pellet (0. 28048 g.) was
prepared and heated for 2. 5 hours at 14350. . No x-ray photograph was

84

made of this material for it appeared to be SmBG. The crucible was
immediately re—heated for 20 minutes at temperatures of 1900-19600.
This time however, heating caused the crucible to begin melting.
Apparently the boron concentration of the crucible had become sufficiently
high to cause the molybdenum boride to begin flowing. Some material on
the lid was scraped and photographed as A-491. This photograph shows
lines of what is principally MozB. The crucible had deteriorated to the

extent that it could no longer be used and therefore was abandoned.

D1. . Use of other compositions of samarium sesquioxide and boron

 

10. . A new crucible was machined and outgassed for two hours at
18500 to initiate these studies. Control x-ray photographs: A-495, A-497,
A-498, A-499 and A-500 were prepared for comparison purposes and
refer to, respectively, molybdenum metal, compositions of samarium
sesquioxide and boron in mole ratios of 1. 00:8. 00, 1. 00:6. 00, and
1. 00: 10. 00, and C-szO3.

11. A 1:6 pellet (0. 19870 g.) was prepared and heated at 14400 for
10 hours. This pellet, removed with reasonable ease, appeared dark
with no visible traces of blue. The inside of the crucible was coated with
a very light grey material, not uniformly, although the lid bore a uniform
coating of what appeared to be this same material. . It is suspected that
this grey material is the same that precedes any evolution of black Vycor
deposits; the pellet was crushed and photographed as A-501. The lid
was scraped and photographed as A-502. . Photograph A-501 shows a
mixture of SmB4 and Sme while A-502 contains lines of the same phase
of SmBO3 as shown A-342, A-371, A-376, A-388, and A-508.

. Subsequent to this first heating, crystalline boron was added to the
crucible which was then heated to 17500 in an attempt to clean it. ~ At the
end of this operation, the crucible looked bright; amild deposit on the lid
was discarded; the blue SmB6 had formed in small measure on the surface

of the boron inside.

85

12.. Subsequent to the "cleansing" operation described above, a
1:8 sample (0. 24424 g.) was prepared and heated for 12.6 hours at 1435°.
Since the sample was 1. 26x larger than in-(l l), heating was done for
l. 26x as long as in (11). After the heating, the Vycor tube carried a
mild cloudy grey film with no black deposit at all. . The pellet from
this experiment (A-503) seemed to be more full of blue specks this time
than the last time. The photograph shows a mixture of SmB4 and SmB6.
The crucible was treated with boron again for one hour at 17000
and, after cooling, was observed to be cracked. . It not only hadncracked
but also had undergone some sort of reaction with the material which it
had contained. Some of the exterior of the crucible was scraped and
photographed as A—504. The photograph of these crucible scrapings shows
lines ascribable to molybdenum metal and what appears to be B-szO3.
13. As the end of a natural sequence to the 1:6 and 1:8 experiments,
a 1:10 mixture (0.24915 g.) was heated at 1435° for 12 hours. At the
end of this run, no visible deposit appeared on the Vycor although; washing
the tube with. 0. 5 M_ hydrochloric acid did. produce bubbles. The pellet
was brittle, containing irregular blue blotches on the surface; on crushing,
the blue color was observed to be present throughout most of the interior.
The photograph (A-506) shows only lines of SmB6.
14. Since the heating time of the experiments described in (11),
(12) and (13) is long, in this experiment, a 1:6 pellet (0. 23863 g.) was
heated for only two hours at 14350... At the end of this heating, the pellet
was coated with a grey-white material on its top surface which was
scraped for a photograph (A-508), while the remainder of the pellet was
crushed completely and representatively photographed as A-509. . After
the top layer of white material had been scraped from the surface, the
pellet was observed to be a dark color underneath. The principal phase
observable on A-508 is SmBO3. . In A-509, the lines are principally ones

of SmB6; there are a few lines of the A-508 phase, but none of SmB4.

86

15. An intermediate run was performed using a l: 10 mixture and
heating-the pellet at 1435° for 1. 5 hours and at 15650 for an additional
1. 5 hours. At the end of this heating, visual examination of the product
revealed only a homogeneously blue material of a color normally associ»
ated with SmB6. . No powder photograph was obtained.

16.. Since decreasing the time of heating produced no marked changes
in the product other than the undesirable absence of pure SmB4, a more
radical change in composition was invoked.

A 1:2 pellet (0. 51223 g.) was heated in the molybdenum crucible
for three hours at 1435°. . At the end of the heating, the color of the pellet
was not uniform on the surface, but on crushing in a diamond mortar, the
powder was alight grey throughout. The photograph of a portion of this
material shows no identifiable lines other than those believed to be due
to B—szO3.

Immediately subsequent to this experiment, another 1:2 mixture
(0. 23726 g.) was heated for three hours at 15400. This time there was no
homogeneity whatsoever in the reaction product, which in fact was fixed
as a gooey mass in the base of the crucible; no x-ray photographs were
obtained.

17.. In view of the results of (16), it was decided to extend the heat-
ing time of a 1:2 mixture (0.97571 g.) to 12 hours at 13300. . The product
was crushed and photographed as A-512 which shows the presence of SmB6 I I
and B-szO3. . Using a 60x microscope, it was possible to observe tiny
blue flecks of material against a background of grey-white material.

. Another pellet (0. 22867 g. ), inadevertently prepared in aratio of
1:1 rather than 1:2, was heated for three hours at 1435°. The resultant
x-ray powder pattern (A-513) shows only lines of B-szO3.

18. A 1:10 pellet (0. 20651 g.) was prepared and heated in the
molybdenum crucible for l. 5 hours at 14350; both the top and the bottom

of the pellet were a grey-white color at the end of the heating, the same

87

color normally associated with. SmBO3. The pellet was returned to the
vacuumline directly without crushing and re-heated for an additional

1. 25 hours at 15400. At the end of this second heating, the pellet was
entirely blue-black in appearance save for one light grey spot on the

edge of the top surface. Crushing, followed by x-ray examination. (A-514),
showed no phase other than Sme.

19.. Inasmuch as all previous attempts to prepare pure SmB4 had
proven futile, it was decided to make a change in experimental approach.
The product of the experiment described in (13), whose photograph shows
only SmB6, was mixed with szO3 in a respective mole ratio of 11. 00:2. 00.

The assumptiOn was that:

115mB6(s) + ZszO3(s) + 15 SmB,(s) + 3 3202(3)

might conceivably represent a reaction Which would occur.

A pellet (0. 15713 g.) was pressed from this mixture and placed in
the molybdenum crucible for heating at 1540C) for one hour. . At the com-
pletion of this experiment, the sample had been reduced in weight to
y 0. 13323 g. and both the top and bottom surfaces were coated with a blue-
black layer photographed as A-515. The blue interior was photographed
as A-516. Examination of both of these photographs showed that in the
case of A-515, SmB, is a contamination phase of the principal SmB6
phase. . In A-515, no SmB4 is discernible and the phase appears to be all
SmB6.

20. Subsequent to the experiment described in (19), the crack first
noted in (12) became intolerable and a new crucible had to be fashioned.
This crucible was again fashioned of molybdenum, outgassed _for several
hours at about 18500, and initiated as described in (21).

21.. A pellet was prepared using 0.13137 g. of (-325 mesh)
molybdenum and 0. 15915 g- of Sm203 (a respective mole ratio of 1:3).
The pellet was heated at 14600 for 3. 5 hours and later photographed as

A-521. The photograph shows only lines of molybdenum and B-szO3.

88

22. By now it was not believed possible to prepare pure SmB4 from
any mixture of szO3 and boron in a molybdenum crucible- However it
was decided to. attempt this once'more in afresh crucible.

. A 1:10 pellet (0. 38543 g.) was prepared and heated for four hours
at 1510 - 15400. The blue-black product was weighed, replaced in the
line, and heated for an additional 10. 5 hours at 1510°. The appearance
of the product subsequent to this treatment was not at all what was
expected. . Immersed in a shell of glassy-looking yellow material, some
of which seemed to have been dispersed on all parts of the lid, the pellet
was a silver-grey color. The coreumaterial (A-523) was soft oncrush-
ing, hinting of SmB4; the x-ray photograph showed lines of this phase and
_of another unidentifiable phase, not SmB6. Total weight loss in crucible
and pellet was 0. 01081 g.

Immediately subsequent to the run described above, the experiment
was repeated in the same crucible for 11.5 hours at 15050. The total
weight loss for this heating was 0. 03410 g. , roughly three times the
weight loss encountered before; however, the pellet was silvergrey only
on its underside. . A large amount of yellow material had collected on
the base of the crucible and the pellet appearedtotally non-homogeneous.
The yellow material which had dispersed on the lid was photographed as
A-524. The only identifiable lines in this photograph are some which
may be attributed to B-szO3.

23. . The'crucible was heated for 3. 25 hours at 18000; this resulted
in the deposition of a large amount of black metallic material on the Vycor
confining-tube. . Some boron was added to the now-bright crucible and
heating was performed at 14400 for 5. 5 hours. A small amount of SmB6
had formed on the surface of the boron; the crucible and boron were
heated to 18000 for an additional hour with no results other than the pro-
duction of a thin layer of deposit on the Vycor. The mixtures of SmZO3

and boron were abandoned in favor of a system of boron and samarium.

89

 

B.. The samarium-aboron system in a molybdenum'crucible

Again it should be noted that although mixtures are referred to as
1:3. 8, 1:4, 1:5,. etc. ,. all mixtures have beenicarefully prepared weighing
each component to the nearest. one-tenth of a milligram.

1. The first of the experiments using samarium and crystalline
boron was performed in the same molybdenum crucible used for the last
. experiment _of the sesquioxides-boron work... Initial heating of the first
1:4 mixture of samarium and boron (0.48300 g.) was done in a vacuum
of 10" mm. for only five'minutes at 10450. The rapid vaporization of
samarium was observed immediately and the final heating of this first
mixture was done at temperatures no higher than 13650 for only ten
minutes. The product of the reaction was a sinteredicak-e but quite soft
to grinding. The x— ray photograph (A-530) of this product shows SmB4
as the principal phase, with SmB6 contamination.

2. Using another 1:4 mixture (0. 52498 g.), but in this case using
amorphous boron, a second heating was performed over an interim of one
hour at a helium pressure of about 600 mm. . (Unless otherwise noted,
this is the pressure of the helium in all successive experiments described.)
The temperature was raised quite slowly in an effort to ascertain at what
temperature the Vycor deposit would beginto show a condensate... Only
after 0. 5 hour at 11500 did any appreciable deposit appear on the walls
of the confining reaction tube. The maximum recorded temperature
between the first half hour and the end of the one hour heating was 12500.

Examination of the products suggests that the material inside the
crucible never underwent any definite melting. However, the same grey
substance which-had caked into the shape of the crucible proper in a
previous run came out of the crucible again as a single piece. . It was very
soft to grind and contained some specks of a .clearly discernible blue

material although the blue material was present in very small amount.

90

The photograph. of the sample (A-533) shows principally SmB4 with some
SmB6 contamination. There appears to be another phase present in
A-533, but there are not sufficient lines on the film for any genuine
characterization.

4. The mole ratio of samarium metal to boron was increased to 1:10
and a reaction mixture (0. 71494 g.) of this ratio was heated in helium at
1075 - 11250. . The experiment was carried out for: just under one hour.
The product of the reaction was a non-homogeneous blob of material which
nonetheless was observed to contain three distinct phases: a powder sur-
rounding a central core and a silver material which stood out from the rest
of the agglomerate.

The powder was photographed (A-540) and observed to containSmB4
and SmB6. However, the silver material (A-536) and the core (A-541)
represent phases which could not be characterized, although neither appear
to contain any SmB4.

. 5.. In an attempt to compensate for samariumlosses, a 1:3.8 mixture
(0. 50369 g.) was prepared. , The mixture was heated in a helium atmosphere
in the molybdenum crucible for 1.5 hours at 12750. The x- ray photograph
of the product (A-542) shows only SmB4 with a trace of Sme.

. After an. intermediate treatment with boron (45 minutes at 18000),
another 1:3. 8 mixture (0.49466 g.) was added to the crucible and heated in
an atmosphere of helium for 1.5 hours at 1300 - 13150. Once again, the
product (A- 543) proved to be SmB4 with a trace of SmB6.

6. Lowering the temperature of 11250 and returning to a 1: 4 compo-
sition a mixture of samarium and amorphous boron.(0. 56244 g.) was heated
in an atmosphere of helium for 1. 2 hours. . At 11250., the product of the
reaction was principally grey, but contained clusters of blue, flaky crystals
throughout. An attempt was made to isolate one of these crystals and

photograph-it as A-546. This photograph shows the same phase(s) as A-548

91

which is the photograph of another "isolated" crystal of the blue material.
In neither case was it possible to make any phase assignments. The gross
material proper was photographed as A-552 which does not show any SmB4
or Sme.

7. Returning to crystalline boron, a repeat (0. 56331 g.) was made
of the experiment described in (6). . Again the product appeared grey to‘the
naked‘eye, but microsc0pic examination once more revealed the presence
of the blue crystals (A-548). The material proper was photographed as
A-549, but the only thing that can be said of it is that it looks the same as
A-552 and is equally unidentifiable.

8.. Another 1:4 mixture (0. 51309 g.) was heated in the molybdenum
crucible for 11.5 hours at 11250 in a helium atmosphere. The product of
this reaction appeared lumpy and non-homogeneous at the end of the reaction
and the x- ray photograph (A—554) shows the presence of SmB4 with con-
tamination by the phase common to A-549 and A—552.

9. This experiment is the initial one on the idea of attempting to
prepare pure SmB4 by heating a 1:3. 5 mixture (0. 53818 g.) in a helium
atmosphere and then. "cleaning up" the sample of excess samarium by heating
for a time in a vacuum. The run was performed for 2. 5 hours at 1100 - 11250,
followed by ten minutes at 127 50 (111 M. . The composition of the product
was principally a black grey material with a few of the blue crystals again
present. However, the photograph. of the product (A-557) shows only a
faint amount of SmB6 within the principal phase of SmB4.

. A portion of the above sample showed violent reactionwith nitric acid
resulting in an undissolved residue (A-558) which turned out to be boron.

10. This experiment represents a crucial one with respect to the work
done using samarium and boron for it is the experiment in which, for the
first time, a sample of SmB4 was prepared for which the photograph showed

no Sme.

92

A 1:3. 5 mixture (0.42576 g.) of samarium and crystalline boron was
prepared and placed into the molybdenum crucible for heating in an .atmos-
phere of helium... The mixture was maintained at 11500 for four hours.

It was then evacuated to a pressure of 10‘“5 mm. and heated for an additional
hour at 14500. The bottom surface of the reaction mixture was devoid of
blue after this treatment, but the interior was filled with blue crystals.

The photograph of the material proper (A-559) shows a mixture omeB4

and SmB6 with no predominance of either phase. The material was placed
back into the crucible and heated once more in helium for eight hours at
12750. . After this heating, a portion was photographed as A-560. , It is dif-
ficult to decide whether SmB4 or SmB6 is the principal phase of this photo—
graph, but it seems safe to assert that the heating had caused some change.

The sample was returned to the furnace line and heated once more in
helium for two hours at 13850; the xuray photograph of the resultant product
(A-567) shows the SmB6 phase as the more predominant one. The material
seemed homogeneous and was a dark blue-black color although a few iso-
lated lighter blue crystals could be observed microscopically. The sample
was ground easily and was reheated for another 1. 5 hours at about 14800.
The orifice was allowed to remain partially clogged with what was believed
to be szO3. . (Actually, the origice had been clogged with blue needles of
a material which turned white on standing. .. Such needles were observed
earlier in some of this work, but their photograph (A-555) is believed to
represent SmO.)

A fourth heating was undertaken for two additional hours at 14800.

A portion ofthe resultant sample was photographed as A-568 which shows
both boride phases and seems no different from A-560.

Again the mixture was returned to heat for l. 25 hours at 15850.
Examination of the product (A-569) revealed that much of the blue color
had diminished and that a good dealpof grey color had suddenly developed
in the; mixture although the grey was clearly confined to the lower regions

of the sintered mass. The photograph shows a mixture of SmB4 and SmB6.

93

Two final heatings were performed. . The firstof these was done
at 17000 for 1.5 hours and led to a mixture (A-570) totally devoid of blue
color although the x- ray photographcshows SmB4 as the major and SmB6
as the minor phase. The final heating was performed for 0. 75 .hour at
17500 and resulted. in a completely grey product (A-571) showing only
lines of SmB4 in the x- ray photograph. . However, the presence of tiny blue
crystallets could be observed microscopically in spite of the x-ray
detection of only SmB4.

No additional details are provided in the appendix for the remainder
of the experiments (Sections E, F, G, H, and I) described in Part III of
this thesis. However, in keeping with the policy of presenting A-numbers
for x- ray photographs associated with particular experiments, a compilation
of such numbers is presented below under a letters-heading corresponding

to the headings in Part'III.

, E. The samarium-boron system in a molybdenum crucible

 

E-13 A-636 E-23 >A-642
E-l4 A-Z:; E-24 ‘A-644
E-15 A-e640 E-26 A-647
" E-16 A-641 E-41 'A-657
E-17 ‘A-643 E-42 A-658
E-18 A-645 E-43 ' A-659
E- 19 A-646 E-44 A-660

94

F. The reaction between samarium sesquioxide and boron

 

in a boron nitride liner

 

 

 

Compositions Photographs
1:5.0 rA-751, 752, 753, 759
1:7.5 A~773,. 774
1110.0 A-734, 737, 739
1:12.5 A-769, 770
1215.0 A-730, 732, 733
1:17.5 A-762, 764, 765
1:20.0 Au742, 743, 745

G. The reaction between samarium sesguioxide and boron
in .zirconium diboride

 

 

 

 

Compositions Photographs
1:5.0 A-780, 788
1:10.0 A-783, 784
1:15.0 A-779, 781
1:20.0 A-775, 776, 778

H. The heating of a sample of samarium hexaboride in a
low-density tungsten cell 7

 

 

 

 

Experiment . VPhotograph
H-l A-soz'
H-2 A-803
H-3' A-804
H-4 A-805
H-S A-807
H-6 A-808

H-7 A9809

APPENDIX B

TABULATION OF PHYSICAL CONSTANTS

95

96

 

 

 

 

Table V. . Melting point and vapor pressure data for rare earth 'metals
(50, 51)
Melting point Temperatures in 9K for vapor pressure; in
M Element at 1 atm. mm- Iig W
0C 1 mm 100 mm 760 mm.

La 920 2420 3180 . 3742

Ce 795 2430 - 3219 ' 3741

Pr 935 2120 2820 3400

Nd 1024 2050 2800 3300

Pm (1027) '- an an (3000)

Sm 1072". 1225 1580 2173

Eu 826 1100 1455 1712

Cd 1312 (1960) ‘ (2580) 3273

Tb 1356 2290 2940 3073

Dy 1407 1680 2140 2873

Ho 1461 (1655) (2215) 2873

Er 1497 (1872) (2480) 3173

Tm 1545 1368 1749 2000

Yb 824 (1095) (1500) 1700

Lu 1652 (2380) ' (3010) 3600

 

Table VI. Metallic and trivalent ionic radii of the rare earth metals

97

 

 

 

(17, 52)

Element Metallic radius (R) Trivalent ionic radius (3.)
La 1.871 1.061
Ce 1.818 1.034
Pr 1.824 1.013
Nd 1.818 0.995
Pm ----- 0.979
Sm 1 804 0.964
Eu 2.084 0.950
Gd 1. 795 0. 938
Tb 1.773 0.923
Dy 1 770 0.908
H0 1.761 0.894
Er 1.748 0.881
Tm 1.743 0.869
Yb 1.933 0.858
Lu 1 738 0.848

 

98

> Table VII. . Lattice parameters of cubic rare earth hexaborides

 

 

 

Element 39 in Bingstriims Respective references
Y 4.128, 4.126, 4.08 53, 21, 54
La 4. 153 54 and 53
Ce 4.141 53 and 54
Pr 4.130,. 4.132 54, 53
Nd 4.1260, 4.128 36 and 53, 54
pm ................
Sm 4.1333, 4.129 36 and 53, 54
Eu 4.178, 4.169 19 and 53, 24
Cd 4.1078, 4.121, 4.14 36 and 54, 53, 24
Tb 4.1020, 4.092, 4.13 36, 53, 24
Dy 4.0976, 4.407, 4.13 36, 54, 24
Ho 4.096, 4.12 36 and 53, 24
Er 4.101, 4.110 53, 54 and 24
Tm 4.10 53
Yb 4.1468, 4.140 36, 53, 54 and 24
Lu 4.11 24

 

99

Table VIII. . Lattice parameters of tetragonal rare earth tetraborides

 

 

 

Element a_o in Singstr'dms $9 in Xngstrams References
Y 7.09 4.01 54
La 7. 30 4.17 54
Ce 7. 205 4.090 10
Pr 7.20 4.11 54
Nd 7.219 4.1020 36
Pm ---------- --
Sm 7.174 4.0696 36
Eu ---------- --
Gd 7.144 ' 4.0479 36, 53, 54
Tb 7.118 4.0286 36, 53
Dy 7.101 4.0174 36, 54
Ho 7.086 4.0079 36, 53
Er 7.071 3.9972 36, 54
Tm ---------- --
Yb 7.01 4.00 11

Lu 6.98 3.93 24

 

100

Table IX. . Lattice parameters of some cubic dodecaborides (55)

 

 

Element 30 in gngstrbms
Zr 7.408
Y 7. 500
Dy 7. 501
Ho 7. 492
Er .7. 484
Tm 7. 476
Lu ' . 7.464
U 7. 473

 

Table X. . Lattice parameters of tetragonal rare earth MBx phase (14)

 

 

Element 39 in xngstr'dms _c_o in Xngstriims
La 3. 82 3. 96
Pr 3. 81 3. 81 (7)
Cd 3. 79 3. 63
Er 3. 77 3. 51

9 5

 

101

.xmom umowconum one Ouo>fldflou xmom nod“... mo awn—«mamas: one .1- H

*

 

 

32. SH 2 28H - -- o cow-N
32 2H NH 38H 32 ommd 2. Hmmd
32. EH 3 SF; - -- om omm..~
.2; EH Hm EL; - -- oH mix-N
H5 Hm; om How; 2 mfm em. awed
3> $4 om 2w; -- -- ON Nata
.3 SH HN 2.x; 32 Nos 3 Sad
.2;- 8; Hm 8a; - -- 2.. $9M
3 oo.~ 3 :3; w 8.... 2: $56
.2. 8.~ om 823 - -- 2 2....
35. mod NH 03d -- -- mm 86
- -- m HSJ 2 $5 2. SM
2 2 .N 3. m2 .~ , Hz .314 am. 3.4.
s; m~.~ S «3.... I. -- HH 8....
3> £1~ H: 03.... - -- 2 8.5
men-4 nomfim mam-4 nomHHum

H 6 His H 2% His H 8 His *H 8535

 

 

 

.33 ouduonofiuo gamma-mm Ho aduonbfiom

oommg 0:» mo omoafi 5H3 mag-mm fiducoofiuomxo no mo mmcwommm 3530935 03 mo nogudmfiou .Cn man-oh.

102

Table XII. Comparison of the interatomic spacings of two experimental
samples with those of the delta polymorph of molybdenum
boride, ~15 -MoB (56)

 

 

dhkflx) dhkfix) dhkl(x)
6 —MoB A-410 A..-411
1.20 1.20 1.20
1.21 ---- _---
1.25 1.25 1.25
1.28 1.28 1.28
1.34 1.34 1.35
1.35 ---- ----
1.38 1.38 1.38
1.41 1.41, 1.45 1.41, 1.45
1.56 1.55 1.55
1.73 1.73 1.73
1.91 1.91 1.90
2.12 2.12 2.11
2.28 2.28 . 2.28
2.72 2.71 2.70
3.06 3.04 3.03
4.2 4.23 ----

 

103

Table XIII. .Raw data used for correcting observed temperatures to true

 

 

 

 

temperatures.
’ Observed

Observed temperature True
temperature expressed as Corrected to temperature
°Centigrade l/OK x 10+4 l/K—wo x 10+4 O-Centigrade

1000 7.855 7.737 1019

1100 7.283 7.165 1123

1200 6.789 6.671 1226

1300 6. 357 6. 239 1330

1400 5. 978 5. 860 1434

1500 5. 640 5. 522 1538

1600 5. 339 5. 221 1642

1700 5. 068 4.. 950 1747

1800 4. 824 4. 706 1852

1900 4. 602 4.484 1957

2000 4. 399 4. 281 2063

2100 4.214 4.096 2168
Tungsten lamp temperature as a direct sighting: 1810°K = Kwo
Tungsten lamp temperature with prism and window: 1772°K = Kw

Since the theory of pyrometry demands that l/KW - l/K be constant
over all range of temperatures, the corrections are made using the con-
stant obtained from the value of Kw and Kwo' ., Specifically:

(5.643 x104 - 5.525 x10“) = 1.18 x10'5

694' ’ 1- . ,
‘ L -.
* ‘ ‘-. w
'33}.—

he

9" i 1 Y QT‘RY ”BRA“!

. _ ‘3...”
‘8" ““3:

 

g-

' \

\