PHYSICAL CHEMISTRY OF PYROMETALLURGICAL PROCESSES. PART II. THE INTERACTION OF LIQUIDS WITH GASES AND SOLID PHASES

• Document Type: CREST
• Collection: General CIA Records
• Document Number (FOIA) /ESDN (CREST): CIA-RDP81-01043R003400070002-1
• Release Decision: RIPPUB
• Original Classification: U
• Document Page Count: 164
• Document Creation Date: December 23, 2016
• Document Release Date: May 1, 2014
• Sequence Number: 2
• Publication Date: April 14, 1959
• Content Type: REPORT

Declassified in Part - Sanitized Copy Approved for Release @ 50-Yr 2014/05/01 : CIA-RDP81-01043R003400070002-1
STAT
Declassified in Part - Sanitized Copy Approved for Release @ 50-Yr 2014/05/01 : CIA-RDP81-01043R003400070002-1
Declassified in Part - Sanitized Copy Approved for Release ? 50-Yr 2014/05/01
CIA-RDP81-01043R003400070002-1
AEC-tr-3439
CHEMISTRY ?GENERAL
PHYSICAL CHEMISTRY OF PYROMETALLURGICAL
PROCESSES. PART II. THE INTERACTION OF
LIQUIDS WITH GASES AND SOLID PHASES
(Fizicheskaya Khimiya Pirometallurgicheskikh
Protsessov. Chast Vtoraya. Vzaimodeistvie
Zhudkostei s Gazami)
B,y
0. A. Esin
P. V. Gerd
TRANSLATED FROM A PUBLICATION OF THE STATE SCIENTIFIC?
TECHNICAL PUBLISHERS OF,LITERATURE ON FERROUS AND NON-
FERROUS METALLURGY, SVERDLOVSK?MOSCOW, 1954
Book 2
UNITED STATES ATOMIC ENERGY COMMISSION
Technical Information Service Extension, Oak Ridge, Tennessee
.et
Declassified in in Part - Sanitized Copy Approved for Release ? 50-Yr 2014/05/01 ?
DIA-RDP81-01043R003400070002-1
Declassified in Part - Sanitized Copy Approved for Release ? 50-Yr 2014/05/01 : CIA-RDP81-01043R003400070002-1
This is Book 2 of two books, pgs. 360-676.
A translation of: Fizicheskaya khimiya pirometallurgicheskikh protsessov:
chastvtoraya. Vzaimodeistvie zhudkostei s gazami. Gosudarstvennoe Nauchno-
Tekhnickeskoe' Izdaterstvo Literatury po Chernoi, i Tsvetnoi Metallurgii.
Sverdlovsk, Moskva, 1954.
Translated by the Language Service Bureau, Washington, D. C., under Con-
tract AT(40-1)-2274:
In the interests of expeditious dissemination this publication has been repro-
duced directly from copy prepared by the translating firm.
AEC Technical Information Service Extension
Oak Ridge. Tennessee
ktS
:t.
I.
Declassified in Part - Sanitized Copy Approved for Release ? 50-Yr 2014/05/01 ? CIA-RDP81-01041Rnmannn7nnno
Declassified in Part - Sanitized Copy Approved for Release ? 50-Yr 2014/05/01 : CIA-RDP81-01043R003400070002-1
.?
?
?
0. A. IN and P. V. GEL,D
PHYSICAL CliEMISTRY OF FYROMETALLURGICAL PROCESSES:
PART II
THE INTERACTION OF LIQUIDS WITH GASES AND SOLID PHASES
(FIZICHESKAYA KHIMIYA PIRommALLURGICHESKIKEI PROTSESSOV:
MAST VTORAYA
VZA1MODEISTVIE ZHUDKOSTEI S GAZAMI)
STATE SCIENTIFIC-TECHNICAL PUBLISHERS OF
LITERATURE ON FERROUS AND NON-minus METALLURGY
SVERDLOVSK MOSCOW
? 1954
1
Declassified in Part - Sanitized Copy Approved for Release @ 50-Yr 2014/05/01 : CIA-RDP81-01043R003400070002-1
Declassified in Part - Sanitized Copy Approved for Release ? 50-Yr 2014/05/01 : CIA-RDP81-01043R003400070002-1
Chapter VI.
INTERACTION OF GASES WITH METAL AND SLAG.
Gases dissolved in metals are known to exert a substantial influence
on the properties of he latter. It is only natural, then, that a great
number of investigations was dedicated to the study of this influence.
Zhukov, Chizhevsky, and Sieverts, who accomplished classical studies on
the solubility of nitrogen and hydrogen in steel, are the pioneers in this
field. Later, the works by Ka.rnaukhov and Morozov, Turbin, Samarin,
Yavoysky, Chuyko, Nossyreva and others provided a broad analysis of the
gas saturation of commercial and pure metals and alloys, contributing there-
by to the elucidation of a number of very important production problems.
The present chapter reviews only the simplest systems: gas-metal,
gas - slag, and metal - slag - gas, both from the standpoint of their
equilibrium conditions and in relation (whenever this was possible) to the
dissolution rate of gases. As regards the more complex non-equilibrium
system gas - slag - metal, a detailed description of this issue will be
found in the monographs by Morozovrlj and Yavoysky L72.7.
I. THERMODYNAMICS OF GAS INTERACTION WITH METAL
We shall first consider the simplest case - the dissolution of pure
gas in metal
Gm(gas) ;7-2Z? 9i(solut1on)
Equilibrium conditions here are determined from the equality of the
increments of the molar F and the partial-molar p free energies of gas
= a
g g?
(v111)
The first increment can be expressed through volatility fi of gas in
the following manner:
dlg = R511nfil
(VI, 2)
and the second may be represented as a total differential with respect to
pressure P, temperature Ty and concentration Ng of gas
--(D.4 dP .) )P
dT
g P T,N T P,N p,Tdllg? (v1, 3)
As is known, the first partial derivative of equation (VI, 3) is equAl
to the partial-molar volume Vg, the second - to the entropy Sg with a minus
sign, while the third is related to the activity of dissolved gas by the
ratio:
?
?
?
?
?
?
?
?
?
?
R
(OF T9)
k?
olV91P.T ag
The equation (VI, 1) will then assume the aspect of
RTd In fs cIN
us ?
(lay 5)
We shall consider the equilibrium at any given, though fixed, tempera-
ture, i.e., at dT = 0. Since the solubility of gases in metals ordinarily is
small, the behavior of these systems will deviate relatively little from
the behavior of the infinitely dilute solutions, in consequence of which one
can assume that
Whereupon
or
Vi
lnf
RT 1
f
Ns RT
The resulting equation (VI, 7) relates the volatility and concentration
of gas to temperature. In conformity with the rule of phases -- according
to which the system under survey is bivariant (two components and two
phases) -- this equation reflects the dependence of gas solubility Ni upon
temperature and pressure.
In a number of cases the first item in equation (VI, 7) is small in
absolute value and can he neglected, while volatility can be replaced by
pressure. Then
or
f . P
In C ,
1\1,-
P9 l(N9.
-361-
(1a, 9)
Declassified in Part - Sanitized Copy Approved for Release ? 50-Yr 2014/05/01 : CIA-RDP81-01043R003400070002-1
Declassified in Part - Sanitized Copy Approved for Release ? 50-Yr 2014/05/01 : CIA-RDP81-01043R003400070002-1
In other words, the solubility of gas under these conditions is pro-
portional to its pressure (Henry's law).
If the solution of gas is accompanied by a chemical reaction, for
example
Gm(gas) 4? n n(solution)
then, with arguments similar to those preceding, it is easy to show that,
instead of equation (VI, 7), the following expression will be valid
f4, 1,p
1,1 9 9 +c.
.v,3 Rt.
(vi, 10)
In particular, if the solution of diatomic gases ( m = 2) is accom-
panied by their dissociation On = 1), i.e., if
then
G2 (gas);==t2G(solution)
in _f`;_ ._C_791' 4- C.
N9 RT
For small pressures in this particular case, in passing simultaneously
to partial pressures, we obtain instead of equation (VI, 9) the following
relationship
N9- K tric ?
It indicates that the solubility of gas is now proportional not to the
first power, but to a fractional power of gas pressure.
The equation (VI, 12), which was experimentally confirmed by a number
of authors, is an approximate one, as this was first emphasized by
Krichevsky L-3_7. It is wlid for relatively small pressures.
A few investigators nj, L75_7, however, found a direct proportion-
ality in their experiments between the solubility of hydrogen in iron and
vriTIT2 at pressures up to 200 atm. and temperatures up to 1550?C, i.e.,
instead of expression (VI, 11), formula (VII 12) proved to be valid.
Krichevsky attributes this either to the small influence of the item
containing VH, or to the fact that the error which is connected with its
-362-
?
?
?
?
?
4
?
?
rejection, is offset by another one - that of the replacement of volatility
by pressure. In the similar way he interprets also Perminov's data 2:67,
according to which the rate of hydrogen diffusion through iron at tsipera-
tures of 320-55000 and pressures up to 350 atm is proportional to VW/
2
In general practice such simplifications at high pressures are, of
course, inadmissible. It is imperative that the difference between fi and
pi.p the presence of the addend containing .1121 and the variability of V2,
be all duly taken into account. The latter is usually mall and it is be-
ing disregarded. Actually, as shown by Krichevsky and Khazanova t27, who
investigated the iron - nitrogen system under high pressures, the two first
factors are of primary importance. On the other hand, it is well to
underline once more that in the region of moderate pressures the relation-
ships (VI, 9), (VI, 12) and those sim4lar to them prove to be fully
applicable.
As to the behavior of gases whose molecules are composed of different
atoms, this has not yet been sufficiently investigated. There are grounds
to believe that here too dissociation takes place during dissolution.
Thus, for instance, the concentration of SO2 in copper (see fig. 130)
for a variety of temperatures Droved to be proportional tq/976- L$7.
2
This justifies the belief that the dissolution of 802 is accompanied by
the disintegration of the molecule into three atoms - S and 20.
It is reasonable to expect that upon the solution of CO in iron there
also occurs a dissociation into carbon and oxygen. The strong catalytic
activity of iron with respect to reaction 2COgas = C CO2 (gas), in
particular, speaks in favor of this fact.
3011
200
A
/00
2 j 4
/0
Fig. 130. Effect of pressure and temperature on the solubility
Sgn of sulfurous gas in copper.
Legend: A) Solubility of sulfurous gas s.?02-
-v2 loog.
1/q
B) gn mm Hg.
-'2 -363-
Declassified in Part - Sanitized Copy Approved for Release @ 50-Yr 2014/05/01 : CIA-RDP81-01043R003400070002-1
Declassified in Part - Sanitized Copy Approved for Release ? 50-Yr 2014/05/01 : CIA-RDP81-01043R003400070002-1
Further discussions will deal only with data pertaining to hydrogen,
nitrogen, and oxygen.
HYDROGEN AND IRON
According to equation (VII 12) the dependence of the solubility SH
of hydrogen in metal upon its partial pressure in the gaseous phase con-
forms to equation
Sti K
(VI, 13)
TO what extent this relationship is operative can be seen, for example,
from KarnaUkhovis and MOrozovis work zg. They experimented with liquid
iron at a temperature of 1685 ? 15?C, while changingpu within
"2
5.72-31.4 mm. Hg. It appeared thereby that the exponent of p" u is fairly
2
constant, varying within 0.46 and 0.50. Considering the fact that measure-
ments were made at a high temperature, the constancy of the exponent
should be assumed as fully satisfactory.
Analogous results were recently obtained also by other investigators
D.OT In these tests total pressure was maintained as constant, while
pH was being changed through introduction of helium into the gaseous
2
phase.
The parabolic relationship (VI, 13) of hydrogen solubility in iron
versus pH2 persisted, as shown by measurements, over broad temperature and
pressure ranges. This is illustrated in fig. 131, the curves of which
pertain to temperatures varying from 500 to 1550?C and to pressures
attaining 130 atm Lg.
Let it be noted that the square root law sometimes L117' is being in-
terpreted as a consequence of atomic hydrogen distribution. Indeed, a
formal combining of equation (VI, 13) with the expression for the constant
DH2 of the dissociation of H2 in the gaseous phase
leads to the formula of distribution
(vi, 14).
I% K
PH lac. Nil 15)
where 5E7is the equilibrium contents of hydrogen in metal, %.
-364-
?
?
?
?
V
?
?
'7
NO
150
A
BO
4'8ea 120 atm.
Fig. 131. Effect of pressure and temperature on the solubility
of hydrogen in iron.
cm
Legend: A) Solubility of Hydrogen Sul
100g
However, for moderately elevated temperatures, at which hydrogen is
practically undissociated into atoms, pH is very small and can be hardly
attributed the physical significance of pressure. The same can be said
about oxygen and nitrogen, the diatomic molecules of which are even more
stable.
As to
KarnaUkhav
formula
the temperature dependence of hydrogen, for liquid iron
and Morozov Z97 suggested the use of the following empirical
Ig S11.0,5Ig
1745
+ 0,888.
7'
(vi, 16).
Here SH - hydrogen contents in iron, cm3 of H2 per 100 g. of Fe;
pH2 - partial pressure of hydrogen in the gaseous phase, mm. Hg.
A simple transformation of equation (VI, 16) gives a temperature de-
pendence of the hydrogen solubility constant K in equation (VI, 13)
S.
," Ig K
PH,
1745 -+0,888.
Dyakonov and Samarin L127, on the basis of other author' data, have
selected similar relationships for the different phase states of iron
7'
-365-
(VI, 18)
Declassified in Part - Sanitized Copy Approved for Release ? 50-Yr 2014/05/01 : CIA-RDP81-01043R003400070002-1
Declassified in Part - Sanitized Copy Approved for Release ? 50-Yr 2014/05/01 : CIA-RDP81-01043R003400070002-1
1220
T ?2'335'
1340
7,
1530 ? 1,71 .
A graphic representation of equations (VI, 18) - (VI, 21) is provided
in fig. 132. Phase transitions of iron are accompanied by jumps of lg Ky
i.e., in the final count, by jumps in solubility of hydrogen.
/
lp; 45 41.1c 47 48 49 f le3
-2,5 _Nic4:t
-2,8
-449
-42
-44
-4d
Tre
ire
2900 /OAT 428 /2501/88 1088 980
A
Fig. 132. Isobar of hydrogen solubility in iron.
Legend: A) Temperature, ?K
Let it be noted that the data of a number of investigators for solid
iron are close enough. For liquid iron, on the contrary, they diverge
noticeably. Thus, for example, according to an investigation/./.1371 at the
temperature of 1550?C, SH = 29.2 cm3/100 g., while Karnaukhov and Morozov
report SH = 23.6 cm3/100 g. at 1560?C and SH = 27.2 em3/100 g. at 1685?C.
This probably is the reason for the divergence of formulas (VI, 17) and
(vi, 21).
The same can be said also with regard to the thermal effect AH of
the solution of hydrogen, which according to the data reported by various
authors amounts to, Kcal/g mole H2:
?
?
?
Karnaukhov and Morozov /97
16
000 t 1500
Dakonov and Samarin Z).2
14
560
Yavoysky L1247
17
800
Smitells LL57
15
600
Lapp Acir
14
500
Liang and others L1.37
14
550
-366-
?
?
?
?
?
For solid iron .11H is somewhat smaller and, according to Armbruster's
L177 tabulation, varies within 11300 - 14400 Kcal/g mole H2.
In conclusion, may it be said that the solubility SD of deuterium in
iron is inferior to that of hydrogen (SH). In the temperature region of
500-1450?C this difference comprises 0.1-0.9 cm3/100 g. A87. Yet, when
Su
dissolved in palladium SD > SH, while -,4= varies from 0.67 at a tempera-
ture of 300?C up to 0.91 at 1000?C b.97.
HYDROGEN AND IRON ALLOYS
Fe - Cr - H System.
Until recently there was no unsnimity in the evaluation of the effect
of chromium upon the solubility of hydrogen in iron. According to some
authors it increases with chromium contents, according to others it drops.
Experiments show that in the solid state iron absorbs hydrogen in
greater quantities than chromium L47. The same data also indicate that
with increasing temperature the solubility in chromium grows faster and
becomes greater than in iron.
This points to the possibility of a similar relationship of solubility
(SH, or> SH, Fe) for liquid metals.
Morozov and Gluskin L217. have examined SR in liquid alloys with low
chromium contents (up to 12% Cr). They find that at 1685?C temperature
SH drops with the increase of Cr contents.
On the other hand, the only experiment carried out by them with a
rich alloy containing 50% Cr revealed a considerable increase in solubility
of hydrogen even if compared with pure iron. This fact again evidences
that the relationship SH Cr ?SH,
Fe is possible for liquid metals.
The isotherm plotted by them obeyed the equation (VI, 13). The ex-
ponent varied within the narrow limits of 0.47- 0.52. The solubility
thermal effect 41H for an alloy with 5% Cr proved to be equal to 17500 cal.,
i.e., differing little from AH for pure iron (16000 ? 1500 cal.).
Kurochkin, Yavoysky, and Geld L227 effected measurements of SH both
for poor and rich alloys (up to 90% Cr) at temperatures ranging from 1560
to 1700?C. The results obtained by them are illustrated in fig. 133. In
all examined cases solubility grew with temperature. Aa regards the
effect of composition, it proved to be quite complex.
Thus, for example, at a temperature of 1600?0 an increase of chromium
percentage from 10-12% causes a drop in SH. Yet, a further increase of
chromium contents leads to a rapid growth of SH. In other words, the
-367-
Declassified in Part - Sanitized Copy Approved for Release ? 50-Yr 2014/05/01 : CIA-RDP81-01043R003400070002-1
Declassified in Part - Sanitized Copy Approved for Release ? 50-Yr 2014/05/01 : CIA-RDP81-01043R003400070002-1
surveyed isotherm shows a minimum. The foregoing is true for the curves
corresponding to temperatures of 1600, 1550, and 500?C. At higher temper-
atures (1650 and 17000C) the minimum is absent and addition of chromium
leads only to further growth of SH.
With the lowering of temperature, a broadening of the interval of con-
centration is to be observed, within the limits of which SR drops with
the growth of chromium contents. This is in perfect agreement with the
fact of lower hydrogen solubility in solid chromium than in iron.
EXtrapolation of the experimental data permits one to establish SH
for liquid chromium in the conditions of supercooling. At a temperature
of 19100?C and a pressure of 760 i Hg it turned out to be approximately
110 =3/100 g, whereas at 16500- it amounted to 80-85 coN100 g.
EXperiments conducted at three different pressures (277, 495, and 760 mm
Hg) confirmed the validity of equation (VI, 13). The value of the exponent
varied from 0.47 to 0.60 altogether regardless of the percentage of chromium
in the alloy. Moreover, in ten cases out of sixteen the exponent varied
within 0.49 to 0.52, ime.jocoinciding with the values previously determined
by Morozov and Gluskin.
Finally, it should be noted that in case of rapid cooling a consider-
able proportion of hydrogen evolves from the alloy. As it may be seen
from fig. 133, the concentration of the residual hydrogen, while being pro-
portional to its solubility in liquid metal, is still approximately 3 ?
3.5 times below it. The considerable evolution of hydrogen during cooling
is responsible for the blistered condition of low carbon ferro-chromium
-368-
?
?
?
?
?
?
?
?
?
?
/Mk
90
10
70
67/
A
50
30
20
0
m v.ke3
q,
.? ?
?
?N.
?
09
5r44
4-
;oa at: 760,fd 05
15511?C,.175r/f H9 1.11 ?
568?C, 03t)
f Zaell. H5
/0
20
.10
40 50
60 70 10 51'
Fig. 133. Hydrogen solubility in ferro-chromium alloys at different
temperatures and pressures.
Curve K for 1685?C, 760 mm Hg and curve for 156o?c, 76o mm ug ?
according to Morozov's and Oluskints data. Curve EF - hydrogen
concentration in cooled alloy.
cm3
Legend; A) Solubility of hydrogen SH,
166g
B) Chromium contents, %
Fe - Ni -H and Fe - Go -H Systems
In solid state nickel dissolves hydrogen in a greater proportion than
copper, iron, or cobalt (fig. 134). At the melting point there occurs a
sharp rise in solubility in each of these metals 4157. The validity of the
square root law, i.e., of equation (VI, 13) was experimentally confirmed
for nickel over the temperature interval 300-600QC. The expanded relation-
Ship with the effect of temperature included Lig assumes the following
aspect
645
Ig H = 095 Ig p )4At f4 +1,732?.
wherein ln is the number of hydrogen micromoles per 100 g of metal.
For liquid Fe-Ni alloys, according to MorozovYs and Gludkints data
2:227, as well as those of Kurochkin and co-workers 2:227, hydrogen solubility
-369-
Declassified in Part - Sanitized Copy Approved for Release
50-Yr 2014/05/01 CIA-RDP81-01043R003400070002-1
Declassified in Part - Sanitized Copy Approved for Release ? 50-Yr 2014/05/01 : CIA-RDP81-01043R003400070002-1
also increases with temperature. As to the effect produced by the composi-
tion, the first two authors record a drop in SH with the increase of
nickel percentage (Fig. 135), whereas the second group produced plots with
a maximum which corresponds to approximately 4% Ni. All these investiga-
tors proved the applicability of expression (VI, 13) to each of the examined
alloys. For 1iauid nickel, as well as for iron, this was also confirmed
in other works j1.o7.
30
20
A
/0
488
1200
Fig. 134. Effect of temperature upon the solubility of hydrogen
in Fe, Ni, Co, and Ca at pn.2 = 1 atm.
Legend: A) Solubility of hydrogen Su,
100 g
B) Temperature, ?C.
In all these investigations the values of the exponent were found to
be sufficiently Close to 0.5.
The solubility of hydrogen in liquid nickel at 1590?C temperature
amounts to 34.2 cm3/100 gi2;ZwhiCh is noticeably below Sieverts' data
(41.7 cm3/100 g at 16000C). The latter, though, seem questionable since
they sharply set nickel aside from the Closely related analogues.
The heats of H2 dissolution in Fe-Ni allays are estimated by Morozov
and .uakin at 15200 cal. at 0.85% Ni, 17000 cal. at 5.65% Ni, and
15500 cal at 10.36% Ni, which corresponds fairly Closely to the values for
pure liquid iron.
As it may be seen from fig. 134, SH for solid cobalt is lower than
for nickel and iron, but it grows faster with temperature. The effect of
small additions of cobalt (up to 4.5%) upon hydrogen solubility in liquid
alloys is demonstrated in fig. 136. Like nickel, cobalt in small con-
centrations also augments SH.
-370-
41'
0
A
20
/0
ic
/68.5? 74- Pvif.
/J50?1; 7.00 mAt 113
?7,1.9117471:3-tm HT ? ? ? ? ?7,0
7/VIV , Hs
/070?C, 7SIJAhv #1*,
13.90`r Ztia
s_ _ ---
? * --
/0 /2 14 to 10
21g. 135. Hydrogen solUbilkby in Fe and Ni alloys ?227.
Curves AB and CD - according to L217 data. Curve IF -.hydrogen
contents in cooled alloy.
Legend; A) Solubility of hydrogen 8H,
bog
B) Nickel contents, %
It is possible that a further rise in the percentage of cobalt will
lead to a decrease of S. The solubility of hydrogen in pure cobalt speaks
in favor of such a tendency of the curve. The former proved to be equal
L227 to 37.9 00/100 g at 760 mm. Hg pressure, and a teeperatare of 1590?C,
i.e., somewhat greater (rather than lover) then for liquid nickel (34.2
000/100 g) at the same p. and temperature.
The data on hydrogen solubility in liquid Fe-Co alloys at 480 and
735 mm. Hg pressures lead to oscillating% exponent values. They change
irregularly within 0.4 and 0.48. The authors /27 believe that this is
ceased by errors in measurements and assomithat the square root law is
applicable Also for these alloys.
-371-
Declassified in Part - Sanitized Copy Approved for Release ? 50-Yr 2014/05/01 : CIA-RDP81-01043R003400070002-1
Declassified in Part - Sanitized Cop Approved for Release ? 50-Yr 2014/05/01 : CIA-RDP81-01043R003400070002-1
Ii
1 I
A
Fig. 136.
solubility
Curve A
Effect of cobalt on the
of hydrogen in Fe-Co
alloys.
- concentration of
hydrogen in cooled alloy.
Legend: A) Solubility of hydrogen
SH' cm3
100 g
B) Cobalt contents, %
A
Fig. 137. Hydrogen solubility in
Fe-Si alloys: 1 - according to
Yavoysky's data; 2 - according to
Morozov's data.
Legend: A) Solubility of hydrogen
cm,
ql' 100 g
B) Silicon contents, %
Fe - Si -H System
The isotherm of hydrogen solubility in solid iron-silicon alloys is
known tcuossess a maximum corresponding to approximately 2% Si for gamma
iron ZA/./. Yavoyaky was first to study the effect of Si on SE in liquid
Fe-Si alloys L47,/g57. Contrary to the universal belief prevailing at
that time, that silicon increases hydrogen solubility in iron, he estab-
lished the opposite, i.e., that the growth of Si contents actually leads
to a decline of SH. The author attributes this to the formation of
suicides which decrease the concentration of free iron.
Later on, Yavoysky's ideas and data were confirmed by other studies
Lgq as well as by Morozov LIZ His data and the results produced by
Yavaysky show an identical character of influence of the Si contents upon
SH (fig. 131).
Fig. 13e represents a family of isotherms of hydrogen solubility in
liquid Fe-Si alloys for a broader c? ? ? sition range (from 0 to 65.7% Si)
at temperatures of 1400 to 165000 Lg. . All of them have a clearly defined
minimum corresponding to the chemical compound FeSi (50 atomic percent).
This fact confirms the opinion that the formation of silicides impedes
the solution of hydrogen.
It is possible to draw a conclusion on the basis of the plots in
fig. 138 that hydrogen solubility in pure liquid silicon is considerable
and exceeds somewhat that for iron.
-372-
?
?
?
?
?
Jo
20
A
I
I/ 7d 4d 51/ PP
Fig. 138. Influence of the temperature and the composition of
Fe-Si alloys on the solubility of hydrogen in them.
cm3
Legend: A) Solubility of hydrogen, sH, 155-g
B) Silicon contents, %
According to Morozovis data the thermal effect of dissolution grows
with the contents of silicon. Thus, for instance,
A /10.14% Si = 15 000 cal.,A H147 = 19 800 cal.
,
H A /11.3,usi = 25 060
si
Pursuant to the data of other authors A371 ,N7, which are not as
accurate but sufficiently so to permit discussion of the quantitative
aspect of the problem, the thermal effect increases initially and then
begins to drop with the growth of silicon contents
% Si
0
1,78
11,0
21,7
45,7
51,5
63,7
AH
14550
19500
17100
17500
14250
7700
5310
Noteworthy is the fact that a sharp drop of A H is observable upon
the appearance in the Alloy of 'Tree silicon not combined into Pea
silicide. This also shows that the bond of hydrogen atoms with silicon
is more stable than with iron.
Apparently, between free silicon and the atoms of dissolved hydrogen
there arise directed bonds responsible for the formation of complexes
akin to those of silanes. The possibility of their formation in liquid
iron was discussed, for one, by Yavoyaky
2)-A2=
-373-
Declassified in Part - Sanitized Cop Approved for Release ? 50-Yr 2014/05/01 ? CIA-RDP81 01043R003400070007-1
Declassified in Part - Sanitized Copy Approved for Release ? 50-Yr 2014/05/01 : CIA-RDP81-01043R003400070002-1
40
a ?1-8-
zoo 100 500 700 900 llaa
Fig. 139. Effect of temperature upon the solubility of hydrogen
in manganese (during heat treatment)
ca3
Iegend: A) Solubility of hydrogen SE, 155-g
B) Temperature, 00
Fe - Via -H System.
EMperiments Show that sharply defined jumps in hydrogen solubility
(fig. 139) correspond to the phase changes ?are manganese, just as this
is the case with iron. These jumps occur i2 at temperatures 727
( ), 1138 (1--> 6), and 124500 ""(
Apart from this, an increase of hydrogen solubility with temperature
decline was registered below 5000C L1217, ,4207-. The authenticity of this
observation, though, calls for additional confirmation. Finally, for the
binary Mn-H system the applicability of the square root law as per eve:-
tion (VI, 13) was established.
According to lavoyskyt147, in liquid iron-manganese alloys a rise
in the percentage of the latter component leads to increased hydrogen
solubility. The validity of equation (VI, 13) was also confirmed for these
alloys. On the basis of the data obtained at temperatures 1548 and 1700?0,
Iamoyaky was able to determine that the heat of hydrogen solution in the
alloys containing 2% Mn, comprises 19700 cal., i.e., somewhat more than
in the ease of pure iron (17800 cal). For pure manganese it attains
40000 cal/Mole according to Morozov's calculations.
?
Fe -C-H System.
The complexity of the experimental procedure for a long time impeded the
study of the influence of carbon upon the solubility of hydrogen in liquid
?
I) This circumstance is analogous to the change of the sign of the
temperature coefficient of electrical conductivity as observed during heat
treatment of manganese.
-374-
Declassified in Part - Sanitized Copy Approved for Release
Iron. The first data pertaining to this matter were obtained by Karnaukhov
and Morozov Z97. They conducted investigations into alloys containing up
to 1.5% 0 over an interval of pressures from 9 to 25 mm, at temperatures
ranging from 1530 to 1585?C. The validity was thereby established of the
square root law, with the exponent varying within 0.45-0.56, as well as a
decline in hydrogen solubility with growing carbon contents.
Similar results were produced later 1.1227. It was also revealed that
the addition up to 7% Cr has practically no effect on SH in Fe-C-Cr alloys.
Carbon, thus, produces an influence upon hydrogen solubility in iron
similar to that of silicon, which apparently should be attributed to the
similarity of the status of these elements' atoms in the melt. It is true,
though, that silicon reduces SE somewhat more intensely than carbon in
consequence of a greater bond energy with iron.
Other Alloying Elements.
The elements V, Ti, Nb, and Ta are interesting because they are apt
to produce comparatively stable complexes containing one or even two atoms
of hydrogen. According to the data recorded in Uthansky's monograph zo,
these complexes may be described by formulas VH0.72, TiH1.75, BbH0.4.7, and
TaH0.78.
In spite of the fact that the number of hydrogen atoms varies in
them (from 0.47 to 1.75), a number of experiments have revealed that for
pure solid vanadium, titanium, niobium, and tantalum, the law of the square
root is operative. This circumstance remains unclear and requires further
study.
Inasmudh as the formation of hydrides is accompanied by liberation
of heat, their stability decreases with growing temperature. It is,
probably, on account of this that the considerable solubility of hydrogen
In these elements (tens of thousands 'times greater than in iron) begins
to abate with the growth of temperature.
According to KarnaUkhov and morozovb7, the introduction of Ti, Kb,
and Ta into liquid iron at temperatures of 1560 and 1685?0 leads to a
strong rise of SH. The quantitr of heat absorbed during the solution of
a mole of oxygen in the alloy decreases sharply thereby. This fact is, no
doubt, connected with the exothermic nature of the formation of complexes.
The magnitude of the exponent n of pH in equation (VI, 13) serves also
2
as an indication of their emergence in the allay. Even with small addi-
tions of the element, it exceeds 0.5 and grows along with the percentage
of the addition. For instance, n = 0.55 at 0.45% Ti and n = 0.72 at 3.41%
Ti.
The cited authors, on the basis of the inequalities 0.5 4: n < 1.0
found by then, deduce that diatomic hydrides, partially dissociated as
per reaction Me1 Me + 2n, must be contained in the allay.
-375-
50-Yr 2014/05/01: CIA-RDP81-01043R003400070002-1
Declassified in Part - Sanitized Copy Approved for Release ? 50-Yr 2014/05/01 : CIA-RDP81-01043R003400070002-1
They consider them as groups within the confines of which directed
bonds emerge for a definite period of time.
However, for pure tantalum and niobium, no stable hydrides with two
hydrogen atoms are known to exist even in the solid state L307. It,
therefore, remains unclear whether it is possible to assume their presence
in alloys with iron at high temperatures.
In the conclusion of this paragraph we shall remark that a number of
authors 2:317, 4327, /337have made an attempt, through application of the
methods of physical statistics, to write formulas for the solubility of
gases in metal. The model of an intrusion solution was pat at the basis
of the deduction. The particles of gas are arranged along the interstices
of the metal lattice; each of them oscillates here until it accumulates
sufficient energy to pass into the neighboring interstitial space.
In constituting the distribution function for the system consisting
of hydrogen molecules and atoms in the gaseous phase and hydrogen ions in
the metal, after a number of approximations, the following expression for
solubility is obtAined L.327i
1gS11= 2,774+ 0,51gp- 0,25 Ig T - lgp
A H
9,1487'
(VI, 22)
Here p - pressure, mm Hg;
e? density of metal;
411- heat of solution, cal/g mole H2.
This expression is compared with the experimental data for SR in iron,
nickel, cobalt, copper, and so on.
In the cases dealing with the solubility of hydrogen in hydride-
forming metals it is necessary to take into account the localization of
the protons, which, in the final count, leads to the equation
lg
s ) -0,242 0,5Igp-1,751gT+
H
in which So relates to the room temperature.
A If
9,148 7'
(vi, 23)
Regrettably both these formulas give only approximate results, with
calculated values quite often exceeding 10 times those obtained by experi-
ment.
The works dedicated to the quantum-mechanical analysis of the problem
concerning the status of the atoms of a metalloid dissolved in metal are
scarce.
Reference should first of all be made to the qualitative interpreta-
tion of the effect of hydrogen upon the paramagnetic susceptibility of
palladium L33.7. It is well known that the latter drops linearly and attains
the zero value at 63 atomic percent of hydrogen. In explaining this fact
-376-
?
?
?
?
?
?
?
?
?
it is assumed that the hydrogen atoms, when getting into palladium, become
ionized and give away their electrons into the d-band. The filling of the
latter leads to the compensation of the spin of the unpaired electrons,
i.e., to the decrease of paramagnetism. In inking use of a model of free
electrons, it is possible to demonstrate L47 the validity of the assump-
tion pertaining to the ionization of hydrogen during its dissolution in any
metal. However, the semi-classical character of the model, i.e., disregard
of the interaction between the electrons in metal, reduces the value of
this argument.
Several works Z357, L367; Z377 are confined to the survey of the
electron motion in the periodic field of the lattice distorted by the atom
of the admixture. In the presence of a perturbing potential there appear
local levels. If they happen to get into the zone of the solventts con-
duction, then the electron of the admixture atom (e.g., hydrogen) is being
collected. And conversely, if the local level remains between the zones,
then the so-called bound state arises and the electrons are retained by the
atoms of the impurity. Moreover, under appropriate conditions the metal
electrons may pass onto these levels forming a negative ion instead of a
neutral admixture atom (for example of oxygen, or sulfur).
However, this reasoning is based on the so-called zonal approximation
whereby the interaction of electrons is also disregarded.
In one of the studies Z387 a detniled analysis is made of the ques-
tion dealing with the dissolution of metalloid atoms in metal within the
framework of the zonal theory. The work incorporates a wealth of experi-
mental material available both for the postulation of the initial premises
and for the evaluation of the results produced.
The energy of the bottom of the conduction band of the solvent metal
is composed of three quantities: the energy of sablimation (Es), the
first potential of ionization (Ei) and the Fermi energy of the metal (EF).
E0 - E+ E1 EF
All of them are negative, since the energy of an electron at rest
outside the metal is being taken as the basis of reckoning.
If one compares the absolute values of E6 and the ionization energy
(tE) of the atom of the addition, then at 1E01> 1E1 the admixture electron
will proceed into the conduction band of the solvent. On the other hand,
at 1E014 1E1 the electron will remain in the admixture.
A calculation made for hydrogen dissolved in copper does not yield
an unambiguous answer. Consequently, the energy of the solution of the
admixture. atoms is being computed. Two Alternatives are hereby being
considered: 1) the atom is ionized, and 2) the electron remains in a
bound state in the atom of the admixture.
The energy of solution is being calculated as a sum of four addends:
-377-
Declassified in Part - Sanitized Copy Approved for Release @ 50-Yr 2014/05/01 : CIA-RDP81-01043R003400070002-1
Declassified in Part - Sanitized Copy Approved for Release ? 50-Yr 2014/05/01 : CIA-RDP81-01043R003400070002-1
=E-1--E'+E"-I-E2,
where E - variation of the energy of Coulomb interaction between the
proton and the bound electron as a result of dissolution in
the metal lattice;
El - variation of the energy of Coulomb interaction for conduction
electrons (since they are now located in the additional field
of the proton and the bound electron);
the exchange energy between the bound electron and the Fermi
electrons of parallel spins;
correction for the interaction with the ions of the solvent
(copper).
En -
E2 _
The calculations are being made by quantum,mechanical method. For
the ease of non-ionized atom the computation produced Es =:- 2.0 electron-
volts, and for the ionized EL = -2.85 electron-volts.
In other words the second configuration turned out to be preferable
from the point of view of energy, since
Es' 10%) the following rela-
tionship is fairly well conformed to L547:
1g; =a + b lg (Cr],
Samarin and Korolev L537 were the first to register this fact.
On the other hand, for poorer alloys ( 4. 10% Cr) the relationship
(VI, 31) is not being observed as it may be gathered from fig. 143. Further-
more, here, as over the entire examined range of compositions, solubility
fails to obey the rule of mixing, in consequence of which
(vi, 31)
SN. Fe. Cr SN. Fe [Fe]+ SN. Cr [Cr] .
(VI, 32)
The investigation of the effect of pressure upon nitrogen solubility
in liquid Fe-Cr alloys proved the validity of the exponential equation
SN =Kn, , (VI, 33)
in which n grows together with the contents of chromium.
Moreover, according to Morozov and Gluskin 1.217, it amonts to 0.9 -
0.95 already at 10% Cr, whilst conforming to other data L54/ these values .
are being reached only at 40-50% Cr.
Fig. 143.
CM'
/00e
ZS
Vg17.
/75
'
-0,125
I
475 40 /,25 45D 475 4fiCr1.%
Influence of chromium concentration upon the salability
of nitrogen in Fe-Cr alloys.
Curve AB was plotted on the basis of the rule of additivity.
The postulate pertaining to the emergence in the allay of a nitride
containing two nitrogen atoms, i.e., Cr112 L217 encounters with difficulties
referred to in the course of the discussion of the Fe-Ta-H and Fe-Nb-H
systems. The fact of the matter is that the possibility of the existence
of the CrN2 compound, as emphasized by Proxvirnin /557, has never been
-383-
Declassified in Part - Sanitized Copy Approved for Release ? 50-Yr 2014/05/01 : CIA-RDP81-01043R003400070002-1
Declassified in Part - Sanitized Copy Approved for Release ? 50-Yr 2014/05/01 : CIA-RDP81-01043R003400070002-1
i
experimentally proved. In other words, no phase of this composition has
ever been produced.
On the other hand, Mozgovoy and Samarin 1567 on the basis of X-ray
diffraction studies of solid alloys admit the possibility of the formation
of Cr2N nitride. This question still remains unclear and calls for further
elaboration. All investigators acknowledge the drop of nitrogen solubility
with the rise of temperature and associate this with a partial decomposi-
tion of nitrides.
Fe - Si -Nand Cr - Si -N Systems.
Silicon produces with nitrogen a number of sufficiently stable nitrides
which were thoroughly investigated and synthetized by Chizhevsky-bZ He,
in particular, established that these nitrides lose their stability in the
presence of liquid iron, i.e., they easily decomposed by it1).
The data produced by a number of authors L517, L567 on the solubility
of nitrogen in liquid Fe - Si alloys are recorded in fig. 144. The
solubility isotherms show a maximum when the concentration of Si is
As to the initial ascent of the curves, its causes are not sufficiently
clear (see, however Z517). The drooping branch, which composes the prin-
cipal part of the isotherm, on the contrary, is caused by the formation of
stable bonds between Fe and Si, i.e., by the decrease of the number of
"free" iron atoms, capable of retaining nitrogen in the melt. Actually,
the curves intersect the X-RXiS (SN = 0) near the composition correspond-
ing to the FeSi compound.
Fig. 144. Solubility of nitrogen in Fe-Si alloys according to
the data of:
1 - Karnaukhov and Morozov; 2 - Chipman and Vaughan.
cm3
A) Solubility of nitrogen, --
100 g
Legend:
B) Silicon contents, %
It should be noted that the drop of nitrogen contents in electric
1) For the thermal characteristics of Si2N4 see L517.
-384-
?
?
?
?
?
?
?
steel with the growth of silicon concentration long ago began to attract
the attention of investigators 2597; LW%
The studies of Karnaukhov and Morozav,[517disclosed that the solu-
bility of N2 in Fe-Si alloys changes with pressure in accordance with the
square root law. Over the interval of compositions ranging from 1.07 to
12.9% Si the exponent n in equation (VI, 33) varied within the limits of
0.49 - 0.52. This lead the authors to the conclusion that unitrogen forms
with iron and silicon chemical compounds of the FexN and Sy types -which
are dissociated into atoms in a greater or lesser measure".
The study of a related Cr-Si-N system was the basis of the work of
Morozov and Samarin L56/. They found that, just as in the case of hydrogen
solubility in Fe-Pr alloys, equation (VI, 31) proves to be valid in this
instance, i.e., there is a linear relationship between lg SN and lg Lpx7.
It must, however, be noted that, as in the system Fe-Si-N, the rise
in silicon concentration leads here to a decrease in nitrogen content.
The solubility of nitrogen, therefore, also is determined by "free"
chromium unbound into stable suicides, for example CrSi and Cr381 2g7.
In view of the above it is hardly possible to expect a strict con-
formity to expression (VI, 31), since the relation between lgSN and ler:7
loses in this case its linear character. Finally,according to Mozgovey
and Samarin, the temperature dependence
IgSN -
10980
+B
4,575 T
(VI, 14)
comprises the same thermal effect (AH =10980 cal) independently from
the composition of the allay. On the other hand, the quantity B is a
function of the silicon percentage. This circumstance is similar to the
one established for the Fe-Cr-N system 2:627.
In both eases it is apparently caused by the fact that SH is deter-
mined primarily by the presence of chromium. However, in Fe-Cr alloys the
entire amount of chromium is, it seems, participating in the absorption
of nitrogen. In the Cr-Si-N system, nitrogen, on the contrary, engages in
intensive interaction only such Cr atoms whose bonds are not satisfied
by silicon.
Somehow or other, the thermal effect of solution, calculated per one
gram-atom of nitrogen, will be determined mninly by the bonds of Cr with
Ny i.e., depending little on the presence of the second component in the
Alloy. Moreover, the magnitude of AH must be Close to its value for the
Cr - N system. Yet, according to MOzgoveyts and Samarinis determination,
it appears to be approximately 1.5 times qmPiller and amounts to only '7600
cal.
-385-
Declassified in Part - Sanitized Copy Approved for Release @ 50-Yr 2014/05/01 : CIA-RDP81-01043R003400070002-1
Declassified in Part - Sanitized Copy Approved for Release ? 50-Yr 2014/05/01 : CIA-RDP81-01043R003400070002-1
2, KINETICS OF GAS SOLUTION IN METAL
The process of gas dissolution in metal may be broken down into three
successively developing stages:
1) movement of the substance from the nucleus of the gas stream to
the interface;
2) adsorption of the molecules, which is usually accompanied by
their dissociation, and the event of dissolution proper which is
Closely connected with it;
3) movement of the dissolved particles into the mass of metal.
Let Us first consider experimentally the most thoroughly studied case,
when the main inhibitive factor in the kinetics of gas absorption is con-
stituted by the thin surface layer of metal, within the limits of which
there develops a two-way diffusion of the gas and metal atoms. In other
words, let un suppose that the third stage is the limiting one.
Lotus further assume that the gase phase is uni-component, and that
the process becomes stationary when its rate u is equal to
Here g -
N; and N -
S- z=!=
ky-
dgk'S(N;?N).
dt
quantity of dissolved substance;
its salability at pressure p and concentration in metal;
interface area;
coefficient of mass transfer, accounting for convection and
diffusion.
Further, if V is the volume of metal, then
dg = VdN,
dN = ?k'S (N. ?N) = k (N; ? N) .
dt V P
(VI, 36)
For the given pressure and temperature, N? also remains constant.
Hence
?In (N; ? N)= kt + const.
(VII 37)
If, at the same time, at the initial moment (t = 0) the metal was
pure (N =0), then
N0
In P kt
-386-
?
?
?
?
?
?
4
?
As to the deduction of a relationship for the other boundary condi-
tions and for the mixture of gases, it can be found, for example, in the
monograph by Ramm Z637.
We shall now revert to the constant k. It depends both on the ratio
8e6 the surface of metal to its volume, and on 30, i.e., on the properties
of the dissolving substance and the boundary film. These comprise, for
instance, the coefficient of diffusion, viscosity, the surface character-
istics of the film, the size of the vessel, the rate and character of mix-
ing. Of great importance for the latter during melting in an induction
furnace is the current in the primary circuit, the magnetic properties of
metal, and the mutual location of both nwindings".
The influence of surface tension a presents particular interest.
Ekperiments on the absorption of a number of gases (H2, N2, CO, CO2,
Ar, etc.) by water and other liquids, conducted at room temperature, show
N7that Cr produces an effect upon the event of dissolution proper only
in the case of weakly interacting particles (e.g., Ar and 1120). Far great-
er is the significance of cr for the hydrodynamic properties of the film,
and consequently, also for the coefficient k.
As demonstrated by Kapitzalg67 the wave character of the flow of
liquid, even with small Reynolds! numbers (Re = 20 25), is more stable
than the laminar; with the wave-length proving to be proportional to
cr?45. According to Frumkin and Levich L667, &Or-the capillary active
substances, which reduce the surface tension, produce a damping effect on
the movement of the surface and thereby considerably decrease the rate of
interphase interaction (dissolution, evaporation, and so on).
This was borne out, for instance, by the tests conducted by Sklyarenko,
Baranaev, and Mezhnyeva L687, L6097, who were able to establish that the
rate of vaporization depends on the presence of the adsorption film only
in the ease if the evaporating substance reduces surface tension. Further,
Ternovskaya and Byelopollalcy,06,7 demonstrated that small additions of
surface-active substances (up to 0.01%) produce a considerable effect on
the SO2 dissolution rate in water. Moreover the subsequent increase of
their concentration ordinarily affects the kinetics hardly at all.
Kryukova L7a7; while investigating the behavior of the
venous mercury electrode (the hydrodynamics of which is in many respects
similar to the movement of metal in induction furnace), revealed a slow-
ing down of the surface movement under the influence of capillary active
substances. For the additions of strong adsorbates the effect may be ob-
served even at very mall concentrations (10-8m).
Ekperimental material dealing with the kinetics of gas absorption
by molten metals is very scarce. Most interesting in this respect are the
data pertaining to the rates of hydrogen and nitrogen solution by liquid
iron obtained by Karnaukhov and MorozovLsg, /97; 1517. They heated the
metal in an induction furnace, while judging the intensity of the process
according to the rate of pressure drop in a closed vessel.
-387-
Declassified in Part - Sanitized Copy Approved for Release ? 50-Yr 2014/05/01 : CIA-RDP81-01043R003400070002-1
Declassified in Part - Sanitized Copy Approved for Release ? 50-Yr 2014/05/01 : CIA-RDP81-01043R003400070002-1
In view of this equation (VI, 36) was modified in the following manner.
First of all
N = b (P be,37-
(vI, 39)
where Poose and P are the pressures at the beginning of the test and at
the instant of time t;
b is the factor of proportionality.
Further, according to the square root law
Then
Np= 63/77.
7.7
? b ? r k [8 P ? b (Pbe3 P)).
dt
The latter expression may be developed to assume the aspect
Here
B m ? al
d[? Ig (m2? at) + ?2 a 1g m a I
df (P)
dt dt 2,303
Pbe ?P
?
.17-1c15,
m=1/15 1- ;
a
Pb".-1- (-2-B)2 ;
(v1, 40)
(vi, 42)
P is the final pressure in the conditions of equilibrium.
pend
The validity of the diffusion equation (VI, 36) was proved by the
authors through the fact that the linear relationship (VI, 42) of f (P)
versus time was experimentally corroborated. As to the direct cheek of
the applicability of expression (VI, 37), it was effected for nitrogen
by other authors L57, who determined the existence of a direct proportion-
ality between In(11; - El) and t at certain temperatures (1540, 1600, 1680,
and 17600.
The basic results obtained by Karnaukhov and MOrozov canbe reduced
to the following:
The coefficient of mass transfer k related to unit volume does not
depend on pressure. For instance, for hydrogen at the temperature of
?388-
?
?
1685?C and the pressure varying from 9.72 to 412 mm Hg it fluctuated ir-
regularly-from 3.0 to 3.4.
The value of k changes with the intensity of current I in the inductor
and is in direct proportion to the square of the current intensity, i.e.,
k k1/2.
This they attribute to increasing mixing of metal the intensity of
which also grows with I.
It should be noted that a number of authors 2:507, 2:727; bq did not
register any noticeable effect of the temperature upon the coefficient of
mass transfer. In this they perceived a confirmation of the diffusion
character of the process, overlooking the fact that the rise of tempera-
ture in induction furnaces is achieved by means of increasing current
intensity I. This in turn leads to intensified mixing which, considering
the diffusion character of the process, unavoidably results in its acceler-
ation.
Karnaukhov and MOrozovhave found that the growth of the volume of
metal V, with a constant interface area S, decreases k approximately so
as it follows from the ratio k = kV4 in the expression (VI, 36).
The recorded facts evidence without a Shade of doubt that the movement
of the substance from the surface of the metal into its bulk constitutes
the limiting stage of the process.
By experiments of the same authors it was established that the rate
(more precisely-10 of hydrogen dissolution is 8-a0 times greater than that
of nitrogen. This fact can apparently be explained by the relationship of
the coefficients of hydrogen and nitrogen diffusion in iron, i.e., by the
fact that DH > DN.
Indeed, under the inflUence of the electromagnetic field there occurs
a movement of metal in the crucible of the induction furnace. Fresh por-
tions of metal move along the axis of the crucible (fig. 145) towards the
interface renovating thereby the surface layer. On getting into contact
with gas, the metal becomes saturated with it as it moves towards the
crucible wails whereupon it mixes up with the general mass.
With the temperature constant and the mixing intensive, the principal
kinetic barrier is constituted of the thin film of metal in which only
diffusion is taking place. It is precisely this circumstance that permits
one to think that the difference in the dissolution rates of H2 and N2 is
caused by the difference of the diffusion coefficients.
According to Karnaukhov and Morosov, the introduction into iron of
alloying additions (Al, Ti, Nb, Cr, Ni) hardly affects the kinetics of
solution. Only silicon presents an exception, its additions in proportions
of around 1% approximately double the values of kH and kle According to
the data of other authors 2,507, /587 the effect of silicon is considerably
-389-
Declassified in Part - Sanitized Copy Approved for Release @ 50-Yr 2014/05/01 : CIA-RDP81-01043R003400070002-1
Declassified in Part - Sanitized Copy Approved for Release w 50-Yr 2014/05/01 : CIA-RDP81-01043R003400070002-1
stronger - kN grows 20 times upon admixture of 0.7% Si. An appreciable
disparity exists also with regard of aluminum - according to the same
experiments BC7after addition of 0.13% Al the value of kN was reported
to have increased 12 times.
As was stated earlier, the effect of admixtures on the kinetics of
solution may be caused by a variety of reasons. They can affect viscosity,
surface tension, and specific resistance of metal, as well as the mode of
the existence in it of the dissolved gas.
Mere silicon additions are concerned the effectiveness of the in-
troduction of small doses (^-1 1% Si) is noteworthy. In this connection
it is well to mention that the solubility of nitrogen (but not of hydrogen)
also increases at such concentrations of Si. Possibly, the increase of SN
constitutes one of the reasons for the growth of kN upon addition of Si.
This, however, cannot explain the increase of kH.
Fig. 145. Diagram of the movement of metal in the crucible of an
inductive furnace.
Somewhat surprising is the constancy of k (kH = 2.9 - 3.0; kN = 0.79)
during further increase of silicon contents ( > 1%). As a matter of fact,
silicon is surface-active 2:747 with respect to iron, besides, it augments
the resistivity of metal. Both these circumstances should retard dissolv,-
ing: the first - because of the decreasing tuqulence of the surface layer
movement, the second - because of the drop in 14, i.e., of the intensity
of mixingl). Therefore, the independence of k from the subsequent increase
of silicon up to 10%, or nickel up to 20%, or chromium up to 50%, requires
elucidation.
It should be noted in conclusion that the kinetics of gas dissolution
in metal, apparently, is not determined by diffusion processes in All cases.
Let it be remembered that the rate of nitrogen solution in iron and its
alloys (e.g., Fe-Cr) is many times greater when NH3, and not N2, is present
1) It should be borne in mind that at a constant temperature (for
example 1685?C) the magnitude of IT a.:714,11 is constant, in consequence of
which the growth of 11 leads to a drop of I.
-3906-
?
?
?
?
?
in gas.
If one considers that this is caused by the difference in the energies
required to disrupt the bonds in NH3 and in 112, than the event of chemical
adsorption of gas should be recognized as constituting the retarding stage.
Probably, in connection with analogous circumstances the rate of
steel saturation with gases in arc furnaces is greater than in reverbera-
tory units. Indeed, the arc causes a partial dissociation of 242, H2, 1120.
The presence of atoms substantially facilitates and accelerates the process
of solution.
Chuyko Z592; .4757, ShatalintV, and Ageyev,[777'were among the
first to register this fact. Moreover, for the calculation of nitrogen
and hydrogen solubility Chuyko suggested the use of the square root law
which also takes due account of the partial dissociation of gas
SN = KNV P
I "N ? SH = K1117 P " ?
vis 43)
In this connection the works D.57 and L7 dealing with the satura,
tion of metal with hydrogen are also interesting. It turns out that the
rate of this process is appreciably greater than in the ease of normal
dissolution of hydrogen. The reason here is, apparently, the same as re-
ferred to above, namely the presence of the atoms of H along with the
molecules of H2.
3. GASES IN SLAG
The interaction of gases with slags is still far too insufficiently
studied. A more or less systematized material is available only for
hydrogen, water vapors and partially-for nitrogen. This material will be
described below.
SOLUBILITY OF GASES IN SILICATES.
Glasses.
A comparatively voluminous experimental material pertaining to the
degree and the avenues of saturation of silicates with gases was accumulated
as a result of studies of optical and other special glasses. It permits
making the assertion that steam and other gases possess a certain degree
of solubility in glasses.
Furthermore, with increasing temperature and pressure the solubility
of H20 grows rapidly, and in a number of cases when the contents of 1120
rises up to 20% and more, said brittle glasses change into thick pastes,
the viscosity of which drops with further growth of 1120 contents. Thus
the change from solid glasses to common soluble (aqueous) glass may, it
seems, be achieved continuously, which fact points to the close affinity
-391-
Declassified in Part - Sanitized Copy Approved for Release w 50-Yr 2014/05/01 : CIA-RDP81-01043R003400070002-1
Declassified in Part- Sanitized Copy Approved for Release ? 50-Yr 2014/05/01 : CIA-RDP81-01043R003400070002-1
of the vitreous and liquid states.
In the course of a number of old experiments Z797 confined to the
study of gas contents in sodium-lime glasses by means of vacuum extraction,
the following was revealed.
The volume of the dissolved gases attains four times the volume of
glass. The basic components of gas are: 212, 11201 002, SO2, 02, and N2.
A preliminarily outgassed glass dissolves 11201 but is, allegedly,
incapable of absorbing 002, 02 and N2.
On the basis of this assumption, a conclusion was made to the effect
that, with the exception of water vapors, all other gases are "physically"
insoluble in glasses and evolve from them only as a result of irreversible
decomposition of Chemical compounds. However, subsequent investigations
have exposed the fallacy of this assumption.
KrassikovL007found, in particular, that the blowing of moist CO2
through livid glass increases the degree of gas saturation of the latter
pp to the level characteristic for industrial samples. For instance, in
the glass (7)% Si02, 11% Ca?, il% NA20, 4% K20) melted in a 140 1-capacity
pot, the concentration of gas amounted to around 16 cm3/100 g; the cola,
position of gas being 75% 1120, 7% 00, 6% 142, 5% 002, 5% 02, and 0.3% N2.
In laboratory melting the contents of gas before blowing with moist 002
did not exceed 3.6 cm3/100 g, whereas after blowing it increased up to
12 cm3/100 g.
Appreciable absorption of H20, 002, 02, and air by liquid glasses was
also Observed by Slavyankl3y2:8a2r. He succeeded in revealing that degassed
glass after contact with the enumerated gases bubbles again during sub-
sequent degassing.
Thus, this and a number of other experiments proved that not only
1120, but also 002, S02, 02, and N2, are being absorbed by glasses. It is
true that their solubility, particularly that of N2, is far lower than
that of water vapors. This is also evidenced by the data of Krassikov,
and the results of Varginis and Skobelerts measurements /$17. The latter
have found that a barite glass (45% HaO, 13% B203, 33% SiO2 3.5d ZnO
' /1" '
3% A1203, 1.6% As203), at a temperature of 1350?C, contained about 80 cm3
of gas per 100 g of glass, with the gas consisting of 76-85% 1120 and
9-15% 002.
Similar results were obtained 2$27 for many sodium-lime glasses.
The principal component of the gases (from 42 to 83 cm3/100 g) dissolved
in them were water vapors. The share of this component in the products
of extraction amounts to 40-97%.
Somewhat different results were obtained by Kondrashava,637 who
-392-
?
?
?
?
?
?
?
?
?
?
?
investigated glasses contairdng B203(2%) or PHO (10.72%). The total gas
saturation of the samples comprised, as usual, 50 - 100 cm3/100 g, but
the contents of 112 (36-67.5%) and N2 (7-44%) wgs exceptionally high, while
that of H0 (075-4.3%) was small. The results which she produced require
additional verification.
Particularly high values for solubility of H20 in liquid glasses
were revealed by Eavoyaky. According to his data the contents of 1120 in
window glass, at 1250?C temperature and pH.20 = 26 mm Hg, amounts to 0.15%,
and in glass designed for chemical vessels -.0.19% (or 236 cm3/100 g) at
130000 and Py =35 mm Hg. On the other hand, KarnaUkhov and Morozov in
a discussion with Yavoyaky indicate that with pli.20 =100 mm Hg no more
than 10-20 cm3/100 g, or 0.016 1120 can dissolve in glass.
We believe this question should also be thrashed out experimentally.
An extraordinary high solubility of water vapors is characteristic for
molten K28iO3 which at normal pressure can contain Z027up to 14 mol. %
1120 (i.e., 1.87% or 2320 =3/100 g) reducing its melting point approximate-
ly by 300. Solubility of water vapors is also high in K25i205(7mol.% 1120).
It is well to mention here that in the Si02 - K28103 - 1120 system a
number of hydrates (K28103.0.5H20; K2SiO3-1120; Koi205.H20; y1e9.H20)
was revealed, of which the last one melts congruently.
Finally, for Na2S1205 as well as for K2SiO3 the water vapor pressure
(up to 140 kg/cm?) was found /847to cause a considerable drop of the
melting point (from 850 to 755?C).
Magmas
Natural silicate fusions - magmas - not infrequently contain noticeable
quantities of dissolved gases, mainly, water vapors.
EXperiments show that granite glass or volcanic glass ("J 75% Si02;
14% A1203; 9% K20), for instance, absorb &57ilio to 10% 1120, at a tempera-
ture of 90000 and the pressure of 5000 megabare).
Moreover, the conditions of glass crystallization change with in-
creasing 1120 contents, namely a drop of temperature and a rise of pressure
are to be registered.
1) 1 megabar - 1.0198 kg/cra - 0.98697 atm.
-393-
npelassified in Part - Sanitized COPY Approved for Release ? 50-Yr 2014/05/01 ? CIA-RDP81-01043R003400070002-1
Declassified in Part - Sanitized Copy Approved for Release ? 50-Yr 2014/05/01 : CIA-RDP81-01043R003400070002-1
5
A 4
2
Fig. 146. Effect of pressure and temperature on the solubility
of hydrogen in albite.
Legend: A) Solubility of hydrogen,
B) Pressure, 10-3 kg/cm2
An investigation of the equilibria L857 in the systems NeA1Si308
(albite)-H20 and KA1S1308(orthoclase)-H20 conducted under pressures of
3500 kgiam2 and at temperatures up to 12000C also points to a considerable
solubility of H20. TO illustrate this a number of isotherms of 0 for
molten albite') are shown in fig. 146.
In processing these data, Ywayskyt7arrived at the conclusion that
the solubility of water vapors obeys the law of the square root
SH,n = K 1.1,0117i-1,0
(VI,4 4 )
Volarovich and Korchemkin L887 studied the effect of gaseous NH3,
H20, I72, air, and other gases on the crystallization of molten rocks
(basalt, diabase, diorite). They disclosed thereby that after passage of
1120, 1101, HF, and particularly lawm, the viscosity of the melt decreases
considerably, a fact which is not observable for N2 and air. After pass-
ing water vapors through crystallized basalt they found up to 0.5% 1120 in
it.
At a later date, Volarovich L897 examined the question of 1120 solubil-
ity in basalt glass at pressures up to 1000 atm and temperatures attaining
1300?C. It appeared that at 13000C, for example, and a pressure from 500
to 600 atm the 1120 contents in basalt amounted to about 1%.
Slags.
The solubility of gases in slags and the permeability of slags remain,
1) These studies were repeatedly discussed in the periodical litera-
ture (see, e.g.. Z867, LEVF). _394-
?
?
?
?
?
?
?
?
? ? ?
?
?
up to this date, insufficiently studied. Through the efforts, predominant-
ly. of the Soviet scientists (Karnaukhav, Morozov., Yavoyaky and others), it
was established that the widespread opinion of slag as an insulator of
metal from gases is wrong. Laboratory and industrial observations show
that slags in one or another measure dissolve gases, that they are perme-
able by gas and can sometimes serve as media for metal degassing.
On the basis of the works of a number of investigators (Chuyko L757,
ShatalinL7q, Maksimenko L907, and others) the conclusion can be reached
that nitrogen is soluble almost in all alags. Its concentration is the
greatest in reduction process slags (from electric and blast furnaces) and
the least in those of oxidizing processes.
According to Chuyko L757 the concentration of nitrogen attains 0.0633%
in white carbide slags containing 0.71% C. According to other data L917
an even higher content of nitrogen (up to 0.23%) is possible in electric
furnace slags, whereas in the basic and acid open-hearth slags it is con-
siderably lower and comprises 1 - 3 .10-4 and 1 - 3 -10-3 % respectively.
In the judgment of Yavoysky L927 the percentage of nitrogen in acid
slags is readily determinable. In chili ed state they contain 0.0094 -
0.0121% N. The solubility of nitrogen obeys the law of the square root
SN= KNII-11:,, ?
(VI, 45)
Moreover, KN increases with growing temperature faster than in
the case of nitrogen solution in iron. Increase of basicity and the re-
placement of MnO by FeO also contributes to intensification of solubility.
Contrary to the aforesaid, Karnadkhav and Morozov believe that
nitrogen does not dissolve in oxidizing slags, since its concentration
never exceeds the limits of experimental accuracy. In view of this, it is
highly desirable for this matter to be further explored.
The total percentage of hydrogen for acid reduction Slags amounts to
12-16 ml per 100 g, and to 7-13 m1400 g for basic slags, according to
certain data L937. Substantially differing data are recorded iv-Yaw/y*7
(/927and others) - the concentration of hydrogen in basic Slags is higher
than in acid, and reaches in certain cases 50 m1/100 g.
On the basis of his laboratory work Iavoysky L27 came to the follow-
ing conclusions. The salability of hydrogen in reduction Slags drop with
growing temperature. The increase of partial water vapor pressure augments
the value of SE according to equation
SH KEIT/ P14,0 .
(vi, 46)
The growth of acid slag basicity involves an increase of solubility.
Especially effective in this respect is the addition of CaO, while MnO,
MgO, and FeO produce a weaker effect.
nprlassified in Part - Sanitized Copy Approved for Release ? 50-Yr 2014/05/01 : CIA-RDP81-01043R003400070002-1
e$,
Declassified in Part - Sanitized Copy Approved for Release ? 50-Yr 2014/05/01 : CIA-RDP81-01043R003400070002-1
It is interesting to note that the substitution of MnO for FeO causes
a rise in hydrogen concentration both in acid and in basic Slags, in spite
of the decreasing water vapor pressure. It is true that for basic slags a
maximum is to be observed. The subsequent lowering of SH is provoked here
by a considerable drop of pH20.
Karnankhav and Mbrozov, in their polemics with Tavoysky, totally re-
ject the possibility that any Vinci of relationship could be established
for hydrogen solubility in reduction Slags versus temperature, composition,
and pressure. They assume that the existing methods for the determination
of hydrogen contents are too imperfect. Moreover, they state that hydrogen
was never found to be present in industrial Slags in proportions greater
than 0.5-1 cm3/100 g Z947.
In his rebuttal, Yamoyskypoints out the inadequacy of the measures
used against adsorption and the condensation of water vapors in the com-
ponent parts of the installations utilized for the determination of hydrogen
ealubilitor in slags. The underrated results produced borMbrozov and
Gluskin L927 are to be attributed, in his opinion, to a possible underesti-
mation of these circumstances.
The problem pertaining to the solubility of hydrogen in slags requires
further experimental study.
DIFFUSION OF GASES THROUGH SOLID SILICATES
Influence of Pressure
The investigation of gas diffusion through glasses, carried out by a
number authors L957; revealed that its rate, with a few exceptions, grows
linearly with increased pressure drop. From this point of view, the studies
by Kondratyev Z967 who accomplished a survey of hydrogen diffusion through
quartz are quite interesting.
His apparatus contained two co-axial quartz cylinders. Hydrogen was
introduced into the inside cylinder under pressure P. It diffused through
the 'all into the space between the two cylinders from which gas was
evacuated beforehand. The speed of the process was estimated according to
the growth of pressure p in this volume.
A few of the plotted curves are recorded in fig. 147. Their numbers
correspond in sequence to the order in which the tests were carried out. In
the initial period of the first test (curve 1) the inflow of hydrogen was
somewhat retarded, but later its pressure changed linearly with time. In
the second experiment which followed the first, and was conducted without
degassing of the vessel, the linear relationship was registered right from
the beginning. The same is generally valid for the third and the fourth
tests, carried out at lower pressure. However, the initial section of
curve 3 is anomalously characterized by increasing rates of gas ingress.
-396-
?
ci
4
?
?
A
20 40 60
Fig. 147. Variation of the rate of hydrogen diffusion through quartz
depending on pressure at 690?C.
Legend: A) Pressure pp nun Hg.
B) Time, min.
dp
The anomalous values of at the beginning of both experiments are
at
governed by the critical solubility of hydrogen in quartz glass. Airing
the first test a saturation of quartz takes place which Slows down the dif-
fusion process. In the third experiment, on the contrary, it proves to be
supersaturated in relation to the decreased pressure. Establishment of a
steady distribution of hydrogen along the cross-section of the wall is ac-
companied by its partial outgassing.
On the basis of the magnitude of the anamolous sections, Kondratyev
was able to evaluate the solubility of hydrogen SH in quartz. Thus, for
example, at the temperature of 690?C it proved to be equal to 6.1041 g/cm3
for 788 mm Hg and 2.48 '10-8 dud for 321 mm Hg. In comparing these quan-
tities with the concentrations cH2 of hydrogen in the gaseous phase (2.63 ?
? 10-5 giv0 and 1.07 *10-5 gicm3 respectively), he demonstrated a good
conformance with Henry's law.
Actually the ratio of concentrations in both vases is almost identical:
(CH
SH)788 wki
438:
01-1)321 if
432.
Hence, not only the permeability, but also the solubility, of hydrogen
in quartz change linearly with pressure. Both these circumstances furnish
-397.-
Dad. - Aniti7ecl Coov Aooroved for Release
? 50-Yr 2014/05/01 CIA-RDP81-01043R003400070002-1
Declassified in Part - Sanitized Co .y Approved for Release ? 50-Yr 2014/05/01 CIA-RDP81-01043R003400070002-1
-
grounds for the belief that hydrogen is in molecular state in quartz.
Indeed, the linear relationship of hydrogen pressure in the external
cylinder can be easily deduced should one proceed from the fact of the
existence of a proportionality between the solubility of hydrogen and the
pressure. Then
(v1, 47)
Here C and c are hydrogen concentrations in quartz along the edges
of the wail;
k and lc' are constants.
Moreover, the first item characterizes the diffusion flow through the
wall of the internel tube, and the second - through the outer cylinder.
Let us designate, as above, the given pressure of' hydrogen in the
internal cylinder by P, and the variable pressure in the outer tube - by
p. In relation to them the edges of the internal wall are saturated with
hydrogen, i.e., C and c are solubilities which are proportional to the
pressures.
Whence
cap. (VI, 48)
=ka(P .?"p) ?k'ap.
dr
Disregarding pp as compared to P, we have
dP kP; p.kPt = ,
df
(VI, 49)
(VI, 50)
i.e., p is proportional to t when P is fixed.
On the other hand, should the passage of hydrogen into quartz be ac-
companied by dissociation of its molecules into atoms, then its solubility
would have to comply with the law of the square root, whereupon
C a 17-15 ; c = a
(vi, 51)
In substituting these expressions in equation (V/, 47), we obtain
(VI, 52)
== a d 167 kali P? I') P
'It
and upon a series of simplifications we find a quadratic, and not a linear,
?398-.
?
?
?
ii
nIfd in Part Sanitized C
?
?
relationship of p to t
d,
=KVP; p=k2P12.
dl
(VI, 53)
It should be noted that a linear dependence of the diffusion rate
upon pressure is to be observed not only for hydrogen, but also for a
number of other gases (Be, Ne, N2, etc.). Moreover, it turns out that
hydrogen diffuses less rapidly than helium but faster than nitrogen.
This is in good agreement with the ratio of their effective sizes (He -
2.18 54 H2- 2.74 34 11'2 - 3.75 S).
Tb the extent that the structure of quartz is far from being Close-
peeked, the diffusion through it of molecules of some substance or other
is not particularly surprising.
Influerce of Temperature
Fig* 148 shows that equation (VI, 50) is valid at different tempera-
tures. Moreover, the value of k varies with temperature (fig. 149) accord-
ing to expression
92w,
= koe RT
(VI, 54)
In other words, the activation energy for molecular diffusion of
hydrogen through quartz amounts to 9200 cal/mole, according to Kondratym
46
42
A
08
0,4
70 30
Fig, 148. Effect of temperature
upon the rate of hydrogen dif-
fusion through quartz.
.I2ggglig.l. A) Pressure p, mm Hg.
B) Time, min.
Fig. 149. Curve of the rate of hydro-.
gendiffusion through quartz vs.
temperature.
Aoiroved for Release ? 50-Yr 2014/05/01 ? CIA-RDP81-01043R003400070002-1
Declassified in Part - Sanitized Copy Approved for Release ? 50-Yr 2014/05/01 : CIA-RDP81-01043R003400070002-1
Close values were produced also by other authors (from 9300 to 12000
cal/mole). Furthermore, the activation energies for other investigated
gases are, as a rule, in direct relation to their effective sizes. Thus,
for instance, for he4um it is smaller (5700 cal), while for nitrogen it is
considerably higher t29900 cal) than for hydrogen L957.
Making use of the obtained values of activation energy, Kondratyev pro-
duced a theoretical estimate of the coefficient of hydrogen diffusion D
through gas. On the basis of expression
and replacing
we have
B3
D = - ?
:2
IR
e
9200
D r_- a2v ? C RT
(VI, 55)
(VI, 56)
Here tt- - time of the "sedentary life" of the molecule H2;
6- distance apart between the interstices;
frequency of oscillations.
Assuming for quartz & =3.3 A, and Yr.: 1012 sec-11 he finds that
D SY 10-5 am sec-1 for the temperature of 690?C. By order of magnitude
this value coincides with the observed value (1.7 ? 10-5 cm2 sec-1).
This coincidence constitutes an additional substantiation of the molecular
state of hydrogen in quartz.
It is important to note that the investigations of other authors L957
show that in the region of low temperatures ( 4( 400?C) a deviation is to
be registered from the simple logarithmic rule of the variation of k with
T (fig. 150). In particular, at the temperature of 193?C the activation
energy of hydrogen diffusion drops to 4300 cal. Analagous deviations were
also revealed for He and other gases.
This fact indicates that the character of diffusion processes under-
goes substantial changes with decline of temperature. According to Berrer
L957 the diffusion of gases through quartz can occur in two ways. One
course is via the crystal lattice from interstice to interstice. High
activation energy is characteristic for this type of diffusion. It is
caused at elevated temperatures by gases with small molecules (B2, He).
-400-
?
?
?
Fig. 150. Effect of temperature on the rate of hydrogen diffusion
through quartz.
In the second case diffusion develops following the defects and mole-
cular crevices along the surface of the grains. The transfer of larger
molecules (Ar, N2, 02) takes place thereby at moderate temperatures.
The cited anomaly may be, probably, explained by the fact that at
temperatures higher than 4000C diffusion of hydrogen through quartz develops
mainly the first way (inside the grains), and at temperatures below Anon
the second type of diffusion (along the grain boundaries) prevails.
Effect of Glass Composition
Since the molecules of gases shift in the process of diffusion from
interstice to interstice, the peeking of the structure should diminish(the
permeability of the solid substance. As is known, the change from quartz
to silicates leads to the destruction in the lattice of the largest (dif-
fusion) channels, i.e., makes it more compact. This involves a considerable
decrease of the rate of diffusion (or, more precisely, of the constant k).
For example, for Be at 300?C temperature, with other conditions being
equal Z957, the diffusion rate depending on the material of the glass com-
prises
Quartz
3.15
Pyrex
038"l0-9
sodium glass
Lead
o 009a ? lo-2
0 0037 ? 10-7
Jena
0 0036 ? 10-9
Thuringian n
0 00084 ? 10-9
In smmaingup, it appears that the diffusion of gases (particularly
of He, H2, Ne) through quartz develops comparatively easily-without a
change of their molecular state. The absence of any influence of the
electric current which passes through glass on the diffusion rate of He
and H2 L9T7 also speaks in favor of thie fact.
-401-
Declassified in Part - Sanitized Copy Approved for Release ? 50-Yr 2014/05/01 : CIA-RDP81-01043R003400070002-1
Declassified in Part - Sanitized Copy Approved for Release ? 50-Yr 2014/05/01 : CIA-RDP81-01043R003400070002-1
The possibility of such process diminishes during the change from pure
quartz (or Bob) to silicates with a more compact structure. Consequently,
the molecular permeability of industrial Slags seems to be quite low.
THE MODE OF EXISTENCE OF GASES DISSOLVED IN LIQUID SLAGS
The question as to the form of the particles present in liquid slags
still remains insufficiently clarified. The high polarity of molten Slag
or silicate and the relatively close packing of its particles suggest that
it is in the form of ions, and not of molecules, that the dissolved gases
exist in liquid slags.
Even in solid minerals the dissolution of volatile components usually
leads to the formation of OH-, 003-, Cl- anions and compound silicon-oxide
complexes L907. Below, we shall consider only the mode of existence of
hydrogen and nitrogen.
Hydrogen.
Three alternative assumptions were made with respect to the ionic
form of hydrogen. Some people believed that it is present in liquid slags
in the form of proton H' L937, others thought that it has the form of
hydroxyl L17, L237, 1927, 1997, the third assumed it to be in both forms -
114-and OH-L7.
A stable existence of a proton in a free form in Slag is highly im-
probable. Its small size and the elementary unit charge produce a con-
siderable potential and lead to very intensive interaction with the sur-
rounding anions L997.
Since 02- constitutes the main anion-forming component of slag, the
proton solidly combines with it into an ion of hydroxyl. Moreover, the
linkage occurs either with a free anion of oxygen
H+ + 02- = OH- .
or with a bound anion to form a silicon-oxide (or alumoslicon-oxide) com-
plex.
SO + H+ SiA.0(j,t712)- + OH- .
Thus, the anion of hydroxyl seem to be the most likely form of the
existence of hydrogen in liquid Siege).
This point of view is consonant with the opinions voiced by
1) The possibility is not ruled out of a e addition of the pro-
ton to the complex anions Six0;7 t p which, apparently
takes place predominantly in highly acid Slags.
-402-
?
?
?
?
?
?
authors concerning the structure and properties of hydrogen-bearing
silicate's. Crystal chemistry teaches that in the lattice of talcum
(Mg3514010 .2 OH), serpentine (Mg3Si29.5 ? 4 OH), hornblende (Ligg,F275
2S18022 ? 2 OH), montmorillonite (g, C1,7 A1281506 16 OH), kaolinito
(A125i205 ? 4 OH), muscovite (LE, 147 4101/4020 4 OH) and other alumino-
silicates (with the exception of zeolites containing molecularly bound
water) hydrogen is present in the form of OH L907.
Analogous considerations were repeatedly expressed in connection with
the study of steam adsorption on glasses. Thus, KysselyavL1007 reports
that there exist groups of OH in the adsorption layer which break up as a
result of desorption according to the following pattern:
I 1 I t
? Si? O?Si? OH 4- OH ? Si ?0 ? Si -
1
1 1 I
zt ? Si ?0?Si ?0?Si ?i i ?Si ? + H20.
1 i t
In one of his works, Voilanovffogwrites that the presence of OH
groups in porous glass, and the decrease of their number upon heat treat-
ment, was convincingly demonstrated by N. G. Yaroslavtzav on the basis of
his investigation of infrared absorption spectra.
As concerns the liquid Slags directly, one of the experimental
evidences which speaks in favor of the presence of OH anions is the direct
proportion, revealed by lavoysky, between SH and according to
equation (VI, 46) as compared to equation (VI, 44). He states that the
process of steam solution is accompanied by a breakdown of its molecules
H2O(gas)4? ?t-slag)=
or
2 ?Hi-slag)
H20(gas + 2s1o4(an) = 2 OHCslagy? 81206-7(siag)
and so on. Similar considerations were expressed also by other authors
2987with reference to magmas.
By applying the ideal law of mass actiQn as a first approximation to,
say, the first of the reactions referred to above, we obtain
N011 = K 1/NO2 ? y
PH)'
58)
whereupon it immediately follows that, with constant composition of Slag
(N-
= const), the solubility of water varies with the pressure of its
0
vapors according to the square root law.
-403-
Declassified in Part - Sanitized Copy Approved for Release ? 50-Yr 2014/05/01 : CIA-RDP81-01043R003400070002-1
Declassified in Part - Sanitized Copy Approved for Release ? 50-Yr 2014/05/01 : CIA-RDP81-01043R003400070002-1
Furthermore, Yavoyskyl;9rdemonstrated that the experimental plots for
N include not onlyi---", bat alscici73 Thus, for a four-component slag
OH- H20
(CaO, FeO, MnO, Si02) he obtained an equation
Non = (7.22+ 1,26157)10 1-
-0,81 ? 10 5(I- 1370?)1 Trici .
(VI, 59)
It further appears that during the extraction of gases from slags it
is watery and not hydrogen, that evolves (as is also the case during the
desorption of H20 from glass). Moreover, hydrogen, as compared to water, is
considerably less soluble in molten silicates and Slags which do not con-
tain easily reducible oxides (FeO, MnO). This also conflicts with the
theory concerning the presence of 114'protons or even free atoms of hydrogen.
It is necessary to remark that only through admission of the existence
of hydroxyl anions it becomes possible to explain the observed dependence
of hydrogen solubility upon the composition of Slag. Indeed, the greater
is the basicity of the latter, the higher is the activity or concentration
of oxygen ions which leads, according to equation (VI, 58), to a growth of
hydrogen solubility.
Furthermore, the process of water vapor solution in slag consists in
its interaction with the anion of oxygen. The activity of the latter does
not depend only on the concentration, but also on the interaction energy with
the surrounding cations. The greater is their charge and the smaller their
size, the more stable become the bonds of the oxygen ions, and the more dif-
ficult becomes the process of slag hydration.
Consequently, the replacement of cations Fe2+ by mh2+, and even more
so by Ca2*, intensifies the activity of 02- anions, which shifts the equi-
librium of reaction H20 02- = 2 OH- towards an increase of hydroxyl con-
centration. Particularly strong, from this standpain-t2W, must be the
increase of H20 solubility in slag upon the introduction into it of large
K', cations, i.e., alkali metal oxides.
FinAlly, the capillary activity of hydrogen, reported by Yavoysky,
could also serve as an argument in favor of the OH- anion. It can be ex-
plained by the fact that the weaker OH- anions (as compared to 02- and
Si041 are being driven off towards the slag-gas interface. Had hydrogen
4
been in the form of proton, it would have not been able to reduce surface
tension. In this case it would not have been hydrogen, but the lees power-
ful cations, that would have been expelled into the surface layer.
However, revue:lc:7's conclusion concerning the capillar activity of
hydrogen was made exclusively on the basis of the spreading out of a drop
of slag after introduction of water vapors from the outside. Be failed to
take into account the change of surface tension at the solid body-gas
-404-
?
?
?
interface.
Meanwhile Slavyansky L817 showed that the spreading of silicates over
platinum is related not so much to the presence of 002, H20, H2 and N2, as
to 02 which forms oxide films with Pt and Ir. These results contradict
Yavoyably's conclusions and it is deemed that in view of this fact this
problem should be made subject to further examination.
Nitrogen.
No sufficiently reliable data are available at the present time for the
judgment of the move of nitrogen existence in liquid Slags. Moreover, as
noted by Trubin Z607, even the studies concerning the solubility of N2 in
melts are still in the embryonic state. Yet, it still seems worthwhile to
record the considerations expressed by individual authors in order to stim-
ulate further investigations into this matter.
In Chuykols opinion L597, of the two combinations - CaCN, and SilN2 -
possible for nitrogen in Slag, preference should be given to The first,
because the second is subject to energetic decomposition by iron, which
eliminates the chance of its existence in the bath of an arc furnace.
Maksimenko L907, on the contrary, categorically denies the existence of
CaCN2 in molten slag on the basis of the commonly known thermal instability
of cyanamide which decomposes at high temperature according to the following
pattern
CaCN2+ CO -3- Ca0 + 2 C N, .
Referring to other investigators ff0.37, he considers it to be more
probable') for calcium azide (Ca3N2) to form in slag. It is a thermally
stable compound and is commonly-present as an impurity in commercial calcium
carbide (CaC2).
It should be noted, however, that CaC2 and CaCN2 form a low-melting
eutectic by virtue of which fact a complete thermal decomposition of CaCK2
up to CaC2 is not feasible by normal methods LT057. Moreover, even a step-
by-step process does not assure D.0.67 the derivation of a product cont.on-
ing more than 99% CaC2. CaCN2 is always contained in carbide as an un-
avoidable impurity.
It is true that in the conditions prevailing in electric furnaces the
CN2- ion partially changes into other anions under the influence of excess
2
calcium, namely into CN"; N3- or even (in the presence of hydrogen) into
NH2-. Yet, these reactions develop incompletely, in view of which fact
1) A similar point of view is shared by Fdneral ff0g.
-405-
Declassified in Part - Sanitized Copy Approved for Release ? 50-Yr 2014/05/01 : CIA-RDP81-01043R003400070002-1
Declassified in Part - Sanitized Copy Approved for Release ? 50-Yr 2014/05/01 : CIA-RDP81-01043R003400070002-1
there always are admixtures of calcium cyanamide and calcium azide in
commercial carbides. The solubility of calcium azide in molten slags has
not yet been investigated.
Thus, it appears to be Rossible to exRect to find the following
nitrogen-bearing anions - CN, CN-; and N-1- - in highly basic, and par-
ticularly in carbide slags. Moreover, taking into account Yavoysky's ob-
servation (which still calls for additional verification) that the solu-
bility of nitrogen in Slags obeys the law of the square root, it would
seem that preference should be given to the groups with one atom of NI, i.e.,
to the anions CN- and N3-.
The mode of nitrogen's existence in acid and semi-acid slags is
absolutely unclear. Reference is usually made to the solubility of nitrides,
carbonitrides and cyanamides in them but without sufficient substantiation
to this effect.
Ekperiments show O.27 that the contents of nitrogen in blast furnace
slags grows regularly with increasing silicon and manganese percentage in
cast iron. This, of course, does not imply that nitrogen dissolves in
slag in the form of silicon nitrides (1) because the latter are very easily
decomposed by iron. It is most probable that "more deoxidized" slags are
produced in these conditions, in which the stability of CN, CN-, and N3-
anions increases. The cited anions are known to be easily oxidized; it is
probably because of this that the nitrogen contents in the oxidizing slags
is =SU /91.7.
4. DISTRIBUTION OF HYDROGEN BETWEEN SLAG AND METAL.
The problem of hydrogen distribution between- the liquid phases of a
heat began to attract the attention of metallurgistis, especially after it
became clear that slag is capable of dissolving measurable quantities of
gases and is, therefore, permeable to them.
The first and unsuccessful attempt to establish the quantitative
regularities for hydrogen, on the basis of electrochemical concepts, MUS
made /_937 on the erroneous premises that hydrogen is present in slag in the
form of proton. In recording the reactions between metal and slag
2+
2 H4lag). Fe(met) = Fe(aag) 2 H(met),
2 Htslag)+ mil(met) = itaf+slag)+ H(met)
the authors obtained an expression for the equilibrium constants
KM. Fe
11-1117(Fe2+)1/(Mn2+)
I 'H. Mn
0-I+) (Hf)
-406-
?
In elaborating the data of the industrial heats, they determined also
the temperature dependences for K, for example,
KH.mn=3,97 ?0,0164(t ?1550). (VI, 61)
The most essential errors of this work are the inferences concerning
the proton form of hydrogen in slags and the existence of equilibrium in
the conditions of industrial heats.
A. more deliberate approach to the analysis of this problem was made
by ravoysky W. First of all he applied the so-called method of gradual
faiminntiOn og for the evaluation of the equilibrium conditions from
the investigation data of the commercial heats. The essence of this method
consists in the simultaneous determination not only of the metal and slag
compositions but also in establishing the direction of the process. In
this particular case it became clear which one of the phases is supersat-
urated with hydrogen in relation to the other. Knowledge of the manner of
the change of the supersaturation sign (direction of the arrow in the
diagram) permits estimation of the conditions of equilibrium.
Furthermore, Yavoy*y proceeded from a more reliable assumption that
hydrogen exists in slag in the form of hydroxyl and admitted the presence
in basic slag of the following anions:
4?
FeO.,, 29 Cr0 A102, OH , PO4 SiO4?
9
In view of this the interation between the metal
according to the following reactions:
20H- + Fe2+ = Fe
(slag) (slag) (net)
20H- + Mn2+ Mn
(slag) (slag) (met)
and slag will proceed
.20 (met) + 2H (met)
t20 +2H+ 2H(met)
the equilibrium constants of which may be recorded as
(OH?) 12 (Fe241
KFeH
[0] Ifil J [Fe]
f (OH?) 12 (Mn2+)
Kmn. H=1. 1 [01 [H) f fMnJ
(vi, 62)
Here the parenthesized values represent the ion fractions of Slag,
calculated on the basis of the theory of perfect ionic solutions LI0S7.
In computing the value of X on the basis of the a priori non-equi-
librium concentrations and using the methods of gradual elimination, he
-407-
Declassified in Part - Sanitized Copy Approved for Release ? 50-Yr 2014/05/01 : CIA-RDP81-01043R003400070002-1
Declassified in Part - Sanitized Copy Approved for Release ? 50-Yr 2014/05/01 : CIA-RDP81-01043R003400070002-1
determined graphically (fig. 151) the equilibrium constant and its depend-
ence on the temperature. Time for the reaction with the participation of
Fe24. it followe that
and for Mh2+
25 800
ig Ki:e. = 13,4 ,
42 200
1g Km?. H = ?19,2.
(vi, 63)
(us 64)
Yavoeicy remarks that the low accuracy of the determination of
(OW), LIV, and LOT concentrations gives an orientation character to the
recorded relationships, and assumes that the possible deviation from the
real values may be 10 to 20-fold.
2
10
f
/
'I 1
J_
.?-7
-21
1 _r_
.
? 7
I
I
___
tl
171
___
i
1 r
V
r-
li
I ,
y
f 1
X 4?
, ,
A
L
1--.
oiq
_1;7 2 4-4 l5 ?7 le 4-9 6,0 4/71 /0
Fig. 151. Dependence of lg r
-Fe upon temperature:
1 - 185-350 ton capacity open-hearth furnaces; 2 - treatment of slags
in the ladle; 3 ? small open-hearth and electric furnaces; 4 - heats
run in induction furnace.
As it may be seen from equations (VI, 63) and (VI, 64) rising temper-
ature involves a drop of the equilibrium constant values, which is
equivalent (see equation (VI, 62)) to the increase of hydrogen contents in
metal. This conclusion is in conflict with the deduction which follows
from the formulas (VI, 6(j) and (VI, 61). Indeed, according to relationships
(VI, 60) and (VI, 61) the growth of temperature reduces K and LOTWhich,
apparently, contradicts the facts.
It may appear at first glance that a singular result follows from the
expressions for the equilibrium constants, proposed by Yavoysky, namely,
that the greater is the oxidizing Rower of the slag (for example, the
greater is the concentration of Fe4+) the higher is the contents of hydrogen
in metal. Yet this is not so.
-408-
?
4,
?
Account must be taken of the fact that the Fe2+ ions "accompany" the
passage into metal not only of the anions OH-, bat also of 02-, i.e.,
F 2+ 4. 02- = Fe * 0
e(alag) (slag) (met) (met).
The exclusion from the equilibrium constant of this passage
1011F0 (VII 65)
IC.Fe 0 (02?)(Fe24)
. =
of the quantity LO7 by means of the expression for KFelii produces
== (OH-2 jFel (VI, 66)
(0212(Fe2+)(H12
The relationship between (02-) and (OH-) is established by the process
1120(gas) Otaag) 20HTa1ag),
for which
or
(oH12
(02?)PH,0
=1 1 and equal to 2 or 3.
Of course, the recorded evaluation methods of the magnitude of z2 bear
a very approximative character. There is no doubt, however, that carbon
dissolved in iron remains in the form of cations possessing more than one
charger') .
TRANSFER OF HYDROGEN IONS
For the identification of the nature of the intrusion phases, which
form during the solution of H, C, N in iron, of great importance is the
study of the properties of the hydrogen-saturated transition elements
(for example, Pd).
As shown by theoretical calculations, ordinary hydrogen, the atoms of
which are co-valently bound into H2 molecules, changes into 4metallic"
hydrogen under very high pressures. In other words, the valence electrons
of its atoms become common not for two, but for a large assemblage of
nuclei. Both forms can co-exist at pressures of the order of 250,000 atm2).
Such pressuresoo far, have not been realized. However, the metallic
properties of H may come into display in its solutions with metals, as a
result of a peculiar internal pressure caused by the intrusion of hydrogen
into the lattice.
Indeed, if one es into account the compressibility factor of
palladium (0.4 ? 10 cm2/kg) and the expansion of its volume upon satura-
tion with hydrogen (")11%), then the "internal pressure" produced could
be estimated at 270,000 atm 1.537.
In these conditions, hydrogen should behave like metal, as, in fact,
has been confirmed by experiments.
As a matter of fact, experiments show that in this case direct current
provokes quite an intensive migration of hydrogen towards the cathode Z547
This was ascertained by a variety of methods, including measurements of
resistivity, of the electromotive force, and so on.
1) In view of this, the average charge of all carbon cations must not
necessarily be integral.
2) Incidentally, according to other data 1.527, this pressure is even
greater and comprises at least 400,000 atm.
Declassified in in Part - Sanitized Copy Approved for Release ? 50-Yr 2014/05/01 : CIA-RDP81-01043R003400070002-1
Declassified in Part - Sanitized Copy Approved for Release ? 50-Yr 2014/05/01 : CIA-RDP81-01043R003400070002-1
The mobility of hydrogen ions in Pd grows rapidly with temperature
and greatly (by hundreds of times) exceeds that for carbon in iron. Its
high values in the electric field are in full accord with the very small
sizes of the hydrogen cations, i.e., proton?).
In this connection, it is well to note also the exceptionally high
coefficient of hydrogen diffusion D (cm2/sec) in metals. Its magnitude
cannot be explained, unless one assumes that hydrogen is present in the
ionic, and not in the atomic state.
According to Chu,yko, the coefficient of hydrogen diffusion D in Gamma-
iron can be calculated at various temperatures with the aid of equation
1,9363 .
(VI', 42)
Table 21 provides a comparative listing of the coefficients of diffu-
sion in Gamma-iron for hydrogen, carbon, nitrogen, oxygen, silicon, and
sulfur L541
Igale 2J Diffusion CoeffIciants of certain elements dissolved in Iron.
Diffusion Coefficient, 105- cm2/24 hours, at
Element temperature, ?C
1000
1100
0
Si
105000
1730-2330
1170
6.5
149 000
3900-10000
3400
100
8,5
The small activation energy of the process of hydrogen diffusion in
iron is noteworthy. Thus, for diffusion in Alpha-iron it totals only
3750 cal/g-atom, while in (amma-iron it is 11500 cal&-atom, while, for
example, for carbon in Gamma:Aron it equals 32000 - 33000 cal/g-atom. /577
This circumstance points to the negligible energy obstacles encountered
by hydrogen during its migration from one interstice into another, i.e.,
emphasizing thus the very small sizes of its particles (nprotonsn).
1) It should be noted here that a positively charged hydrogen par-
ticle present in the metal could only conditionally be termed as proton.
Its properties are not identical to those of a 'free " proton, since it
remains in interaction with the whole crystal. Yet, it differs just as
substantially from the atom of hydrogen, because of the surrender of the
valence electron into the general cauldron.
-448-
?
?
4
?
?
4
?
Finally, Baymakov and Yevirinntkov Z587, on the basis of X-ray diffrac-
tion and electrochemical analyzes of nickel containing hydrogen, came to
the conclusion that the latter may be present in the metal in two forms.
Thgy believe that in the process of electrolytic deposition of nickel,
one part of the protons is adsorbed by the surface of nickel crystals,
neutralized by electrons, and forms atoms which dissolve partially in metal
or combine into molecdles. The other part of the protons enters into the
crystalline lattice together with the ions of nickel and forms a solid in-
trusion solution.
According to their data, the formation of a staid solution leads to
the distortion of the crystalline lattice kf,nickel and to the growth of
its parameter from a = 3.490 to a = 3.500 X*). Upon annealing in vacuo
the metal acquires normal parameters. Subsequent saturation with hydrogenx
(lasting cathodic polarization) leads to the intrusion of atomic hydrogen'
which does not affect the lattice parameters.
5. SUSCEPTIBILITY AND SPLITTING OF 111.6 X-RAY SPECTRA LINES
OF METALS IN WHICH HYDROGEN WAS DISSOLVED.
SUSCEPTIBILITY
Palladium is known to be paramagnetic. This indicates that on the
average there exists a certain effective number of unpaired electron spins
in the d-shell in all atoms. A simfflar condition prevails also in the
atoms of the other transition elements. The addition to them of other
metals with filled d-shells (Ag, Au, Cu) causes the migration of the
valence electrons, which previously belonged to them, on to the 4d level.
2 0,4 46 08 h'hge
800
100
400
200
0
/8,7?C
\20?C
20 40 60 80 am %/fa
Fig. 161. Influence of the concentrations of hydrogen and gold
upon the paramagnetism of their alloys with palladium.
3.&?aexttenof deformation resulting from hydrogen intrusion is
unusually high and requires checking.
2) which, according to the authors, is more rapidly adsorbed on the
developed surface of the nickel crystals.
-449-
Declassified in Part - Sanitized Copy Approved for Release ? 50-Yr 2014/05/01 : CIA-RDP81-01043R003400070002-1
Declassified in Part - Sanitized Copy Approved for Release ? 50-Yr 2014/05/01 : CIA-RDP81-01043R003400070002-1
As a result of this, the number of unpaired electrons diminishes and
paramagnetism is consequently reduced. Such is the situation which arises,
for instance, upon the addition of Cu, Zn, Al and other elements to nickel.
The number of unpaired electron spins in a gram-atom of palladium
is equal to 0.55 N (where N is Avogadro's number). In other words, not
every Pd atom has an unpaired electron in the 4 d band, since a proportion
of electrons passes from the 5S level to 4 d. In view of the above, pal-
ladium's paramagnetism vanishes for instance, in alloys with gold when the
percentage of the latter amounts to 55 atom %. As it may be seen from
fig. 161, hydrogen, dissolved in Pd, produces such an effect.
With increasing hydrogen contents, paramagnetism of palladium drops
almost linearly and disappears when the ratio of H to Pd equals 0.5 - 0.65.
The liquidation of paramagnetism of palladium with the help of Ag is
achieved upon the introduction of 73% of silver. It is interesting to
note that with increasing Ag percentage the solubility of hydrogen in Pd-Ag
alloys decreases rapidly and reaches the zero point at about 65-70% Ag.
The recorded data also underline the metallic character of hydrogen
dissolved in palladium, namely, the ionization oflIand the transition of
its valence electrons into the vacant 4 d-levels of Pd. This, in particular*
explains the high solubility- of H in Pd its decrease upon addition of Ag,
which reduces the number of unpaired spins in the &band.
SPLITTING OF THE LINES OF THE X-RAY SPECTRUM.
Krassnikov L597 studied the condition of hydrogen in metals by means
of a method elaborated by him for the measurement of the distance Aft.
between the lines of the ir.(1,,,(xdoublet in the x-ray spectrum. Accord-
ing to his data, the stated distance for a number of metals (Cr, Mn, Fe,
Co, Ni, Cu, Zn) decreases more or less smoothly with the increase of the
atomic number (fig. 162).
Along with this, a hyperfiae structure (splitting of the doublet in
two) appears in the spectra ZW. It farther transpires that
diminishes during dissolution in the hydrogen specimen and grows after
the replacement of pure metal by its oxide (see the dotted curves in fig.
162). For each metal the variation ofA,E1 comprises about 3.5%, while the
width of the lines varies from 25 - 30%.
,
In this connection Krassnikov L61/ observes the inadequacy of the
accepted concepts, according to which the magnitude of the doublet differ-
ence is determined only by the atomic number and the structure of the
electron shell. He believes that the variation of ..a .E under the influ-
ence of hydrogen points out to the intrusion of protons into the atomic
electron shell?). Only this, he feels, can explain the effect of
-----a) One should rather assume that the ionized hydrogen atoms are
located in the intersticial spaces (as well as the particles of C, N, etc.)
and change the energy level of the electrons in the lattice of metal.
-450-
?
a
4
dissolved Huron the x-ray spectra which are governed by the behavior of
the electrons Closest to the nucleus.
Cr No re to /ye Cu 1/7
Fig. 162. Distances apart between the lines of the doublet
][042)0( for a :lumber of elements:
Me - pure metals; H -metals saturated with hydrogen; 0 - metal oxides.
According to the data (see fig. 162), the drop of.e..e , caused by
hydrogen, is equivalent to the increase of the atomic number Z or the
atomic weight A. of the element. And conversely, the growth of .A.2. after
introduction of oxide is equivalent to the decrease of A and Z.
Inasmuch as the observed drop of A.2 is small and comprises only
10-20% of the effect produced by the increment of the atomic weight per
unity, Krassikov assumes that there is a small and short-term growth of A
at the expense of the statical and continuous redistribution of protons
within the interior regions of the various atoms. This distribution of the
protons in metal he terms as "proton gas" L62/. The /htter constitutes
"the protons of the given substance, since it is difficult to call them
protons of ordinary hydrogen, or protons which present the fixed part of
all atoms". Both the atomic bonds and the condition of the electron gas
in metal are determined by the condition of the proton gas.
In this manner, Krassnikov comes to the conclusion not only to the
effect that hydrogen, dissolved in a number of metals, remains in the form
of protons, but that there exists a fairly deep and Close bonding between
them and the remaining atoms.
Later on Galaktionova L637 succeeded in accomplishing a series of
x-ray diffraction studies of iron samples out gassed at high temperatures
lavacuo and of specimens especially saturated with hydrogen. She was un-
able to reveal any difference in their structure, which fact repudiates the
theory concerning the existence of hydrides in metal.
-451-
Declassified in Part - Sanitized Copy Approved for Release @ 50-Yr 2014/05/01 : CIA-RDP81-01043R003400070002-1
Declassified in Part - Sanitized Copy Approved for Release ? 50-Yr 2014/05/01 : CIA-RDP81-01043R003400070002-1
Moreover, within the limits of precision of her measurements (4. 0.0015 2)
the parameter of the Alpha-iron lattice remained unchanged. On the basia
of this fact, she rejects the possibility of an intrusion of hydrogen atoms
into the interstitial spaces, and believes that the entire quantity of
hydrogen is present in the form of protons.
It should be mentioned, however, that according to the data of a num-
ber of other authors,[647the intrusion of hydrogen into the interstitial
spaces of iron causes a deformation of the lattice to the extent of
0.0005 A (with the contents of hydrogen amounting to 30 cm)/100 g).
6. INTRUSION PHASES OF THE METALS OF TRANSITION GROUPS.
The above data refer to the ionized state of hydrogen dissolved in
transition elements. It forms with them the phases of_Antrusion which were
examined in detail by numerous investigators 2:657, LW. In them the
metalloid particles M are arranged in-the interstices of the metal (Me).
Such distribution of U without substantial disruptions of the arrangement
of the Me atoms becomes possible when rM 4:0.59.
rMe
The intrusion phases can be considered either as subtraction phases
(solid solution of the III type) with respect to the metalloid, or as solid
solutions of la intrusion into metal The first interpretation is
based on the fact that these phases produce homogeneous systems over a com-
paratively wida interval of compositions 207. However, as demonstrated
by Makarov L68/, such contradistinction is hardly justifiable. He believes
that they should be "incorporated into a single type of solid solutions
with a variable number of atoms in the unit cell".
Proceeding from the fact that H solutions in Pd are the most tyrpical
representatives of these systems, Umancy&gputs forward the following
surmise. In metals of transition groups, hydrogen, carbon, and nitrogen,
which constitute with them intrusion phases, are to be found in the metallic
state, i.e., "they surrender a par4) of their valence electrons to the
d-band of the crystal".
This inference can be substantiated by a series of facts. First of
all it is important to indicate the sharp reduction of paramagnetism during
the formation of the intrusion phases, which can be traced, for example,
from the following data:
Phase Zr Zr,C Zr,N Ti Ii,C TIN W W,C
Susceptibility +91 -13 +30 +57 +4 +24 +52 +5
It is interesting to note that the more readily ionizable carbon2)
1) More precisely, "the effective part".
2) The ionization potentials for C and N in relative units (1 is adopted
for H) are equal to Ic = 0.83, and IN = 1.07, respectively. For conversion
into Kcal. per gram-atom they must be multiplied by 312.
-452-
?
I.
4
?
4
reduces paramagnetism stronger than nitrogen. Thus, C and N, like H which
was discussed earlier, are present in these phases in a partially ionized
state.
The same is also evidenced by the high values of specific conductivity
of such solid solutions, as well as their metallic luster. Noteworthy is
the fact that in individual cases the formation of intrusion phases is ac-
companied even by a drop of the specific resistance with the preservation
of the same sign of its temperature coefficient. TO illustrate the above,
the following data are hereby recorded:
Phase
Specific resistance
e.10-4 ohm.cm
V V,C: V.N1 Zr Zr.0 Zr.NI Ti TIC TIN Ta Ta,C Ta.N
0,07 1,56 0,86 0,45 0,63 0.14 0,90 I 93 0,22 (i,15 1.0 1.35
Other circumstances also point to the ionized state of carbon and
nitrogen. As is well known, the first ionization potentials of carbon,
nitrogen, and hydrogen do not differ very much (Ic : IH : IN= 0.83 : 1.00 :
: 1.07). For this reason the collectivization of their valence electrons
proceeds in Close energy conditions.
Moreover, the deformation of the metal lattice during the intrusion of
the metalloid leads to a reduction of the energy of the valence electrons,
which fact creates favorable prerequisite for the filling of the d-band.
It is precisely this that explains the relatively great thermal effect
of the formation of intrusion phases:
Metals Ti Zr V Mo Ta
Thermal (
effect (
cal/mol (
for nitrides
80000
82500
60000
-17000
58000
for carbides
50000
5800
49500
- 4200
38000
The fact that the heats of nitride formation are higher than those for
carbides is to be attributed to the correlation of the sizes of the metal-
loid atoms. indeed rN > r which causes a greater deformation of the
lattice during nitrogen intrusion.
Finally the considerable diffusion coefficients of carbon and nitrogen
in these phases also indicate that these elements are in ionized state.
Actually the covalent radii of C, N, and 0 are quite close; they are
equal to 0.77, 0.74, and 0.74 24 respectively. Had these particles been
existing in a monotypic form, one might have expected the diffusion coef-
ficients to be approaching each other. However, they are far greater for
C and N, than for O.
This fact emphasizes the singular status of C and N atoms: their high
mobility both in the absence, and -- in view of the foregoing -- also in
the presence, of an external electric field. In other words, they should
be considered as mum-size cations of a metal alloy.
?453?
Declassified in Part - Sanitized Copy Approved for Release @ 50-Yr 2014/05/01 : CIA-RDP81-01043R003400070002-1
Declassified in Part - Sanitized Copy Approved for Release ? 50-Yr 2014/05/01 CIA-RDP81-01043R003400070002-1
7, GENERAL CONCLUSIONS ON THE MODES OF METALLOID EXISTENCE.
The question dealing with the mode of the existence of metalloids dis-
solved in metals still remains in the very embryonic stage of solution.
Within the scope of the technology of ferrous metals, the most important
impurities of iron are known to be constituted by C, 0, S, P, H, N, Si, Cr, Mn,
Ni, and Co.
The preceding survey has shown that carbon, hydrogen and, probably,
nitrogen are present in the solid and liquid iron - as well an Cr, Mn, Ni,
and Co - primarily in the form of positively charged particles.
Oxygen, sulfur, phosphorus, and silicon, apparently exist in iron in
a different state. Moreover, judging by the data pertaining to chemical
equilibria, oxygen and sulfur are to be found in monatomic form. The small
value of their diffusion coefficients imply that the sizes of these parti-
cles are comparatively-large. It is natural then that they do not surrender
their electrons to the d-band and do not form intrusion phases with metal.
On the basis of the fact that the electron cloud in the lattices of
iron oxides and sulphides is shifted towards the metalloid, it is possible
to assume that the mode of their existence approaches that, of the anions
(02-, S2-). In particular, Kulikov, in comparing the heats of solution ?
approximately estimated by him -- of the various forms of oxygen in liquid
iron, comes to the conclusion that the most stable state in which oxygen
is to be found in iron is the 02- anion.
The very large sizee of 02- and S2-, as compared with the cations of
carbon and nitrogen, hinder their movement and make it practically impos-
sible for them to shift in the metal under the influence of the external
electric field.
Like silicon, phosphorus occupies an intermediate position, since it
is neither a cation, nor an anion, and forms particles with the atoms of Fe,
in which a considerable role is played by homeo-polar bond?). In respect
to silicon, this is confirmed by the shape of the e.m.f. and surface tension
curves, with regard to phosphorus - only by a break in the e.m.f. plots
A further theoretical and experimental development of the problem will
permit verification of these inferences and to reveal a number of additional
characteristics. One of them already can be dwelt upon at this time. The
matter concerns carton.
The fact that carbon exists in metal in the form of a cation affords,
1) As is well knownt the energy of loosening during the diffusion of
phosphorus through iron (E= 4k100 cal/mole 1579), is considerably smaller
than during self-diffusion of iron (E = 69000 calliaeLW). This implies
that the diffusion event occurs, in all probability, mainly between the
phosphorus atoms which break the bonds with their neighbors and build a
new environment composed of the particles of iron.
-454,
acciflrfr1 in Prf C
A
?
3
????
4
in particular, the explanation of the reason why it produces an intrusion
phase with Gamma-iron, although this does not correspond to the relation-
ship between their atomic radii (0.605, i.e., greater than 0.59).
In the solid phase of cementite the carbon ion is surrounded by six
iron atoms. It is located in the octahedral pore of the Gamma-lattice.
Yet, the coordination number of iron remains the same 2:127 as it was in
pure metal.
Fe3C is known to constitute a thermodynamically unstable phase. This
ordered solid solution tends to decompose and separate graphite. Conse-
quently, it is reasonable to expect, that in the liquid state -- due to a
lover energy of loosening') -- in alloys with high contents of C, there
should also appear, along with evenly distributed carbon atoms, groups of
them similar to the elements of the graphite lattice.
Such incomplete molecular miscibility in eutectic melts was confirmed
by x-ray diffraction studies described by Dpnilov finable ?0-workers in
their works dealing with non4wraus metals D.87, L74/. Amin carried
out a number of experiments which illustrate tbe posgbility of polymeri-
zation of carbon atoms in liquid cast iron 2'35/, ZW.
A stream of east iron diffused at the surface of ice water into
minute droplets with a diameter up to 0.005 cm. The rate of cooling was
estimated thereby at several thousand degrees per second. The increase of
the degree of fragmentation, and consequently also -by acceleration of the
rate of freezing, resulted in a reduction of the contents of the precipitated
graphite.
Yet, even in the minutest droplets graphite was steadily preserved in
small quantities. Besides, it vas also revealed that the graphite contents
decreases with increased superheating of oast-iron. These facts provide a
justification for Bmin to speak of the possibility of a heterophase fluctu-
ation in liquid Fe-C alloys. In the limiting ease it attains the sizes of
the minutest particles of graphite.
It is Clearly evident that such deviations from complete atomic mis-
cibility are caused by a diversity of the interparticle forces.
Apparently, the interaction energy of the carbon atoms among them-
selves somewhat exceeds that between C and Fe. This causes the emergence
of cybotactic groups, rich in carbon, and also facilitates the formation
of graphitic heterophase fluctuations. It is possible that the latter in
the initial form constitute the basic planes of the graphite lattice.
1) Samaria and Schwartzmann t727on the basis of the data for east
iron viscosity determined the energy of loosening to be equal to 17600 oil.
If one considers that the same magnitude also refers to diffusion, then
it will appear to be far smeller than the energy of loosening for the solid
state.
-455-
d for Release ? 50-Yr 2014/05/01 ? CIA-RDP81-01043R003400070002-1
Declassified in Part - Sanitized Cop Approved for Release ? 50-Yr 2014/05/01 CIA-RDP81-01043R003400070002-1
BIBLIOGRAPHY TO CHAPTER VII
1. A. N. Morozov. Vodorod i azot v stali (Hydrogen and nitrogen
in steel), Metalurgizdat, 1950.
2. M. N. Karnaukbov and A. N. Morozov. "Izvesyiya Akademiyi nauk
SSSR", otdeleniye tekhnicheskikh nauk, (Etilletin of the USSR Academy of
Sciences, Department of Technical Sciences), 12, 1845, 1948.
3. H. Lepp. "Metals Industrial", 45, 315, 341, 1935; "Journal of
Iron and Steel Institute", May 1940; ref. see "Stal'" (Steel), 10, 62,
1940.
4.A. Kyakonov and A. Samarin. "Izvesyiya Akademiyi nauk SSSR",
otdeleniye tekhnicheskikh nauk, (Bulletin of the USSR Academy of Sciences,
Department of Technical Sciences), 9, 813, 1945; 1, 121, 1946.
5. J. Chipman et al. "Journal of the American Chemical Society",
55, 3133, 1933; "Transactions of the American Society for Metals", 29,
313, 1936.
6. C. Sherman, H. Elvander and J. Chipman. "Journal of Metals",
188, 334, 1950.
7. A. N. Volt sky. Osnovy teoriyi metallurgicheakikh plavok
(Principles of the Theory of Metellurgical Heats), Metallurgizdat, 1943.
8. A. N. Volt sky and Maas Plobodakoy. "Izvetniye metaly" (Non,
Ferrous Metals), 1, 102, 1936.
9. 0. A. Yessin and I. T. Sryvalin "Zhurnal fizicheakoykhimiyi"
(Journal of Physical Chemistry): 25, 1503, 1951; 26, 371, 1952.
10. Ia. I. Gerassimov. Ibid, 13, 1436, 1939; "Uspekhi KhimiYi"
(Achievements of Chemistry) 14, 282, 1945.
il. C. Smithells. Gasy-i metaly (Gases and Metals), Metallurgizdat,
1940.
12. C. F. Floe and J. Chipman. "Transactions of the American
Institute of Mining and Metallurgical Engineers", 143, 287, 1941.
13. S. Marshall and J. Chipman. "Transactions of the American
Society for Metals", 30, 595, 1942.
14. B. V. Stark. "Izvesyiya Akademiyi nauk SSSR", otdeleniye tekh,
nicheakikh nauk, (Bulletin of the USSR Academy of Sciences, Department of
Technical Sciences) 5, 655, 1948.
15. M. I. Temkin and L. A. Schwartzmann. "Zhurnal fizicheakoy
kbmiyi" (Journal of Physical Chemistry), 23, 755, 1949.
16. 0. A. Yessin and L. K. Gavrilov. "Izvesyiya Akademiyi nauk SSSR",
otdeleniye tekhicheakikh nauk, (Bulletin of the USSR Academy of Sciences,
Department of Technical Sciencies), 7, 1040, 1950.
?456?
?
?
4.
4
?
4
?
17. A. M. Samaria and L. A. Schwartyrnnn. "Izvesyiya Akademiyi
nauk SSSR", otdeleniye tekhnicheskikh nauk, (Bulletin of the USSR Academy
of Sciences, Department of Technical Sciences), 8, 1231, 1949.
18. V. I. Danilov and I. V. Badchanko. "Zhurnal eksperimental'noy
teoreticheskay fiziki" (Journal of bcperimental and Theoretical Physics),
7, 1158, 1937.
19. Ye. I. Akhumov. "Zhurnal prikladnoy khimiyi" (Journal of
Applied Chemistry), 21, 226, 1948.
20. Ya. T. Gerassimov. "Acta Physicochemica USSR", 18, 275, 1943.
21. F. Koerber and H.
5, 569, 1932.
22. Watley and Lord.
5-8, 1932.
Oelsen. "Archly fuer da s Eisenhuettenwesen",
"Journal of the American Chemical Society",
23. R. P. Smith. "Journal of the American Chemical Society", 68,
1363, 1946.
24. L. S. Darken. "Transactions of the American Institute of Mining
and Metallurgical Engineers", 140, 204, 1940; "Metals Technology", 2,
1940.
25. O. A. Yessin and L. K. Gavrilov. "Izvesyiya Akademiyi nal* SSSR",
otdeleniye tekhnicheakikh nauk, (Bulletin of the USSR Academy of Sciences,
Department of Technical Sciences), 8, 1234, 1951.
26. P. V. Geld, N. V. Zaimakikh, N. N. Serebrennikov and Yu. P.
Nikitin. "Zhurnal prikladnoy khiriyid (Journal of Applied Chemistry),
25, 687, 1952.
27. P. V. Geld, S. I. Pooel and 7U. P. Nikitin. "Zhurnal prikladnoy
khiriyi" (Journal of Applied Chemistry), 25, 592, 1952.
28. N.11. Ageyev. Priroda khimicheskay svazi v metallicheakikh
splavakh, (Nature of Chemical Bonds in Metal Alloys), ed. USSR Academy of
Sciences, 1947.
29. F. Koerber and W. Oelsen. "Mitteilungen K.-Wilhelm Institutes
fuer Eisenforschung", 18, 219, 1946.
30. J. Chipman and N. Grant. "Transactions of the American Society
for Metals", 31, 36, 1943.
31. V P. Yelyutin and B. B. Levin. "Stall", (Steel), 9-.10, 554,
1946.
32. H. Liang, M. B. Bever and C. F. Floe. Metals Technology",
T. P. 1975, February 1946.
33. I. R. Krichevsky. nhurnal fizicheskoy khmiyi" (Journal of
Physical Cehmistry), 9, 867, 1937.
?457?
n
ifd in Part Sanitized CoIDV Approved for Release ? 50-Yr 2014/05/01 ? CIA-RDP81-01043R003400070002-1
Declassified in Part - Sanitized Copy Approved for Release ? 50-Yr 2014/05/01 : CIA-RDP81-01043R003400070002-1
34. S. Debsian. Nauchnye annoy vakuwmmoy tekhniki (Scientific
Principles of Vacuum Technology), GILL (State Publishing House for Foreign
Literature), 1950.
35. R. Berrer. Biffaziya v tverdikh telakh (Diffusion nf Solid
Bodies) GILL, 1948.
36. S. S. Nossyryova. "Stalin (Steel), 6, 542, 1948.
n P. Bardenheuer and A. Ploum. nAitteilangen.K-Milhelm Institutes
fuer basenforschung", 160 129, 1934.
38. M. Bodenstein. "Zeitschrift fuer Elektrochemie", 28, 5114 1922.
39. M. Borelius and P. Lindblom. "Annalen der Physie, 82, 201,
1927.
40. K. P. Romadin. Elektrolitiehmekiyperenos vmetallicheakikh
zhidkikh i tverdfth rastvorakh, Trudy V.-Vozdushaoy Akademiyi imeni N.E.
Zhukovskogo (Electrolytic Transfer in Liquid and Solid Metallic Solutions,
Transactions of the N. E. Vaukovsky Air-Force Academy), ed. 167, 1947.
41. K. E. Schwartz. Elektrolyticheskaya provodimosti v zhidkikh i
tverikh metallakh, (Electrolytic Conduction in Liquid and Solid Metals),
1940.
42. V. I. Prosvirnin. "Vestnik metallopromyshlennosti" (Bulletin
of the Metal Industry) 12, 102, 1937.
43. W. Zeitz and T. Deur. "Zeitschrift fuer Blektrochemieu, 44,
256, 1938.
44. I. V. Grevenshchika7 and G. Ya. Tarassov. "Vtoraya konferentziya
po fiziko-khimicheskim voprosam", Nauchnoye khimiko-tekhnicheskoye
izdateltstvo ("Second Conference on Physicochemical Problems, Scientific
Chemical and Technical Publishing House), 159 - 172, Leningrad, 1928.
45. J. G. Dorfman and I. K. Kikoin. Fizika metallov (Physics of
Metals), GTTI, Leningrad, 1934.
46. T. A. Lebedev. uMetalIurgu (Metallurgist), 5, 5, 1934.
hir. V. M. Gutermann. "Izvestiya sektora fiziko-khAmicheskogo
analiza" (Bulletin of the Sector for Physicochemical Analysis), 19, 452,
1949.
48. W. Zeitz and 0. Kubaschewaky. "Zeitschrift fuer Elektrochemie",
41, 551, 1935.
49. T. A. Lebedev and B. M. Guterman. "Doklady Akademiyi nauk WSW,
(Reports of the USSR Academy of Sciences), 60, 1201, 1948.
50. W. Clemm. "Magnitokhimiya (Chemistry of Magnetism), GoOkhimizdat,
1939. -458-
?
?
?
?
?
4
4
51. N. V. Ageyev. Prirodakhimicheskoy svyazi v metallicheakikh
splavakh, (Nature of Chemical Bonds in Metal Alloys), ed. USSR Academy of
Sciences, 1947.
52. E. Wigner and I. B. Huntington. "Journal of Physical Chemistry",
3, 764, 1935.
53. A. R. Ubbelode. Uspekhi Khimiyi" (Achievements of Chemistry),
7, 1969, 1938.
54. A. Kohen and W. Specht. "Zeitschruft fuer Physik", 62, 1, 1930.
55. P. L. Chang and W. D. Bennett. "Journal of the Iron and Steel
Institute", 170, 205, 1952.
56. Ya. S. Umansky. Karbidy tverdikh splavov (Carbides of Solid
Alloys), Metallurgizdats 1947.
57. Ya. S. Umanaky, B. N. Finkelstein and M. Ye. Blanter. Fizicheskiye
osnovymetallovedeniya (Physical principles of Metallography), Metal-
lurgizdat, 1949.
58. Yu. V. Baymakov and L. M. Yevlannikov. "Zharnal fizicheakoy
khimiyi" (Journal of Physical Chemistry), 25, 483, 1951.
59. A. I. Krassnikov. "Zhnrnal eksperimentalinoy i teoreticheskoy
fiziki" (Journal of Experimental and Theoretical Physics), 9, 11941 1204,
1209, 1345, 1939.
60. A. I. Krassnikov. "Doklady Akademiyi nauk SSSR" (Reports of the
USSR Academy of Sciences, 49, 346, 1945.
61. A. I. Krassnikov. "Zharnal eksperimentalinoy i teoreticheakoy
fiziki" (Journal of Erperimental and Theoretical Physics), 14, 285, 1944.
62. A. I. Krassnikov. "Izvegyiya Akademiyi nauk SSSR", otdeleniye
tekhnicheskikh nauk, (Bulletin of the USSR Academy of Sciences, Department
of Technical Sciencies), 1, 133, 1946.
63. N. A. Galaktionova. "Izvesyiya Akademiyi nauk SSSR", otdeleniye
tekhnicheskikh nauk, (Bulletin of the USSR Academy of Sciences, Department
of Technical Sciences), 11, 1666, 1949.
64. Weber and Pfarr. "Mitteilungen K-Wilhelm Institutes fuer
Eisenforschung", 15, 1971 1933.
65. H. Haegg. "Zeitschrift fuer Physicalische Chemie", B6, 221, 1929;
7, 339, 1930; 8, 445, 1930; 9, 43, 1931; "Metallwirtschaft"; 10, 387, 1931
etc.
66. Ya. S. Umanaky. Trudy Moakovtkogo Instituta stall imeni I. V.
Stalina (Proceedings of the I. V. Stalin Moscow institute of Steel), 20th
ed., 1940.
65C see, e.g., H. A. Meyerson. "Redkiye elementy", (Rare Elements),
Al 6, 1935. _459...
Declassified in Part - Sanitized Copy Approved for Release ? 50-Yr 2014/05/01 : CIA-RDP81-01043R003400070002-1
Declassified in Part - Sanitized Copy Approved for Release ? 50-Yr 2014/05/01 : CIA-RDP81-01043R003400070002-1
68. le. S. Makarov. Stroyeniye tverdikh faz s premennim chislom
atomavv elementarnoy yacheyke (Structure of the Solid Phases with a
Variable Number of Atoms in the Unit Cell), ed. USSR Academy of Sciences,
26, 1947.
69. Ya. S. thansky. "Izvestiya sektora fiziko-khimicheakogo analiza"
(Bulletin of the Sector for Phyaiocochemical Analysis), 16, 127, 1943.
73. P. L. Gruzin4 lu V. Kornev and G. V. Kurdyumov. nokledy
Akademiyi nauk SSSR" (Reports of the USSR Academy of Sciences), 80, 49,
1951.
71. O. A. lessin, L. K. Gavrilov and N. A. Vatolin. "Doklady
Akademiyi nail SSSR" (Reports of the USSR Academy of Sciences), 85, 117,
1952.
72. A. M. Samarin and L. A. Schwatrzmann. NIzvedyiya Akademiyi
net* SSSRN, otdeleniye tekhnichedeikh flank, (pdlletin of the USSR Academy
of Sciences, Department of Technical Sciences), 6, 891, 1949.
73. V. I. Danilov. "Izvesttya Akademiyi neUk SSSRN, seriya fizichedkaya
(Bdiletin of the USSR Academy of Science, Series of Physice), 5, 30, 1941.
74. A. I. Danilova, V. I. Danilov, Ye. Z. Spektor. NDokiady
Akademiyi nail SSSRN (Reports of the USSR Academy of Sciences), 82, 561,
1952.
75. K. P. Bunin. Trudy Utallskogo IndustriyalTnogo institute,
imeni S. NC. Kirova (Proceedings of the Ural S. M. Kirov Industrial
stitute), 19, 80, 1944.
76. X. P. Bunin. Vaelezo-dglerodistiye splavy, (Iron-Carbon Alloys),
Mashglz, 1949.
3
4
?
Chapter VIII
IhitRACTION OF LIQUID METAL AND SLAG.
The interaction processes of liquid metal with slag are of great
technological significance. Innumerable investigations both of general and
more specific character have been dedicated to the study of this phenomenon.
It appears to be altogether impossible to embrace all these studies in the
present chapter. Hence, it became necessary to confine this review only
to a few most important problems. Considered as such are the electrochem-
ical theory of interaction and the reactions of desulfurization, dephos-
phorization, and decarbonization.
I. ELECTROCHEMICAL THEORY OF INTERACTION OF LIQUID METAL
(or MATTE) WITH SLAG.
It was revealed in the preceding chapters that the approach to liquid
slags as to ionic solutions is more realistic than that of the molecular
hypothesis, in that the former permits one to explain a greater number of
experimental facts.
If this is sop then it is correct for the interaction of slags with
liquid metals and mattes to be considered as electrochemical in character.
Vanyukov2:17 was the first to formulate this idea with respect to matte
and slag in 1912.
GALVANIC CFTTS AT HIGH TEMPERATURES.
Among the serious proofs supporting the idea of the electrochemical
character of interaction between metal and slag is the experimental evidence
pertaining to the existence of a jump in the electrochemical potential at
the interface of the two liquid phases mentioned 27. It consists in the
construction of galvanic cells in which molten slag is the electrolyte.
Moreover, liquid cast-iron, steel, ferrous alloys, or mattes, must be used
to serve as cell electrodes in order to corroborate the electrochemical
theory of the ferrous and non-ferrous pyrometallurgical processes.
From this point of view the work 257, in which CaO, A12031 and 8i02
slag was used as electrolytes, with graphite and carborundum functioning
as electrodes, is inconclusive since it leaves unanswered the question
concerning the existence of the potential jump at the interface of molten
ferrous and non-ferrous metals with slag. Apart from this, the very con-
siderable distortions caused by the thermal electromotive force constituted
the basic deficiency of the measurements.
Cells with Metallurgical Alloys
of Various Concentration.
For the substanttdion of the electrochemical theory in application to
cast-iron and steel, the iron-carbon system was investigated as most
important for the metallurgy of ferrous metals. A galvanic cell Lilcom,
posed of two liquid iron-carbon alloys (electrodes) was subjected to
-461-
Declassified in Part - Sanitized Copy Approved for Release ? 50-Yr 2014/05/01 : CIA-RDP81-01043R003400070002-1
Declassified in Part - Sanitized Copy Approved for Release ? 50-Yr 2014/05/01 : CIA-RDP81-01043R003400070002-1
analysis with a varying concentration of C (from 0.2 to 4.7% C), and of a
molten synthetic slag (electrolyte), containing 49% A1203, 43% CaO, and
about 8% aluminum and calcium carbides. Schematically it could be repre-
sented as follows:
Fe,CICaO,A120,,CaC2,A1,C3IFe,C
The arrangement of the cell is illustrated in fig. 163.
2
(v111, 1)
Fig. 163. Schematic diagram of the measuring cell:
1 - fused magnesium container; 2 - graphite tube; 3 - graphite rad;
4 and 5 - graphite conductors; 6 - tungsten-molybdenum thermocouple;
7 - fireclay lid; 8 - graphite stopper; 9 - magnesite packing;
10 - carbon alloys, 4% C; 11 - alloys with different carbon contents;
13 - slag; 14 - compartment for the thermocouple.
Since the concentration of iron ions in this type of slag is very low
in comparison with the concentration of carbon ions, the cell functioned
as a result of the difference in the activities of carbon in the metal
alloys. The process which determined here the potential at both electrodes
was probably as follows:
C22 lar-_-_.2Cime.m+ 2f9.
(VIII, 2)
According to expectations, the electromotive force of the cell,
measured at 1600?C, turned out to be the greater, the greater was the dif-
ference between the concentrations of carbon in the metal alloys (fig. 164).
On the basis of its determined values, evaluations were made of AF of the
free energy at 1600?C, during the solution of graphite in the Alloys of
a
?
4
?
?
?
iron with carbon, and the activity of the latter at different concentra-
tions. They proved to be approaching the figures found by different methods,
namely by means of a study of chemical equilibria 2,57.
This fact confirms the correctness of the exposed theory as to the
reasons for the display of differences in the potentials of the investigated
cell, particularly so, if one considers that, according to the data obtained
by methods, the Fe-C melt deviates substantially from
an ideal solution.
NO less important then the Fe-C system for the metallurgy of ferrous
metals are the alloys of iron with silicon. For this case a concentration
cell was built of Fe-Si-C alloys and Slag, containing 50% CaO, 10% MgO, and
40% Si02, as follows
Fe, SI, CiCaO, MgO, Sia, Fe, Si, C.
(VIII, 3)
The silicon contents in the alloys varied from 0.65 to 43%, while the
concentration of carbon in all instances approached saturation. This fact,
as well as the absence of iron ions in the slag and the high contents in
it of silicon-oxide anions, justify the belief that the reason for the
emergence here of the electromotive force is entirely due to the difference
in the activities of silicon in the alloys.
Considering the slag composition and the probable absence from it of
812 07 6- anions Jr, it appears to be possible to describe in the following
manner the process which develops at both electrodes and determines the
potential
3S101- 4 2
4 csIr__ 7 ka10_ 3 + 4H.
(vin, 4)
The experiments were conducted at a temperature arepnd 1470?C in order
to avoid reduction reactions which become noticeable LO/ in these conditions
only upwards of 1500?C.
The curve plotted from these data and reflecting the dependence of
the e.m.f. upon the concentration of silicon in the alloy (fig. 156) is
composed of two branches separated by a sharp break at 50 atom. % Si. It
is similar to the carve of potentiometric titration and points to the
existence in the metal alloy of a sufficient/7 stable FeSi compound. This
fact, as well as the character of the silicon activity dependence (deter-
mined from the e.m.f.) on Si concentration are in accord with the results
obtained by other methods (heats of solution, chemical equilibria, surface
phenomena) 297, 157; 2J.077.
Declassified in Part - Sanitized Copy Approved for Release ? 50-Yr 2014/05/01 : CIA-RDP81-01043R003400070002-1
-463-
Declassified in Part - Sanitized Copy Approved for Release ? 50-Yr 2014/05/01 : CIA-RDP81-01043R003400070002-1
Fig. 164. Dependence of the e.m.f. of a high-temperature galvanic
cell upon the concentration of carbon.
A similar galvanic cal was producedDJ7for the Fe-C--:P alloys which
can be schematically represented as follows
Fe, C, P J CaO, MgO, Si02, Ca3P2, Mg3P3 I Fe, C, P.
(vizi, 5)
One of its electrodes had a constant composition (in partiaUlar-,,
24% P), while in the second the concentrations of phosphorus varied from
1.5 to 23%. All alloys were carbon saturated. Slag served as the
electrolyte, and contained 45% CaO, 14% MgO, 40% 5i02 and around 0.5%
phosphides. Measurements were effected at 1470?C. The following was
apparently the process which determined the potential at both electrodes
P(met) -I- 3 43 P(5-1)
(v1111 6)
The cell's e.m.f. varied only with the activity ratio of phosphorus
in the metal alloys.
The resulting curve (e.m.f. - phosphorus contents), like that for the
Fe-Si-0 alloys, shows a break (fig. 165) indicating that there exists a
stable Peg compound in metal. This agrees with the fusibility diagram
for the Fe-P sxstem, with the values of the heats of formation of iron
phosphides 2.).2/1 and the leek of electrolytic effect in these alloys 22.3/.
5 /0 /5 a (7.P1
Fig. 165. Dependence of the e.m.f. upon the phosphorus contents in
metal at 1470?C.
-464
?
The cited facts confirm not only the existence of a jump of the electro-
chemical potential at the metal-slag interface for the most important metal-
lurgical alloys of ferrous metals, but they also show that there are suf-
ficiently stable groups of atoms, corresponding to PeSi and Peg compounds,
in cast-iron, steel, and ferrous Alloys.
For non-ferrous pyrometallurgy the question of the existence of over-
voltage at the matte-Slag interface presents great interest. kyhole
series of concentration cells was prepared for this purpose 124/. Instead
of multicomponent mattes, binary alloys of Pb-PbS, Cn-Cu28, Ni382--CU2S, and
FeS-Cu28 were used as electrodes in them. Moreover, one of the electrodes
was a sulfide with a constant composition (Cu25 or PbS) while a melt with
variable man= concentration constituted the other.
Instead of industrial slag containing highly aggressive iron silicate,
glass composed of 72% SiO2, 17% Na20, and 9% CaO was used to serve as
electrolyte. Moreover, a quantity of Na2S was added to the glass in a
proportion of"-, 10% to the latter's weight. Thus, the e.m.fs were measured
in galvanic coals
Pb, S I Na20, CaO, Si02, Na2S I PbS
Cu, SI Na20, CaO, Si02, Na2S Cu2S
Cu, Ni, S I Na20, CaO, Si02, Na2S CurS
Cu, Fe, S I Nall:), CaO, SiO2 Na2S CuIS
(vizi, 7)
Experiments were carried out at temperatures ranging from 1180 -
1300?C.
The element, which in all eases determined the potential, Was sulfur
2+ 2-
S one) S(51) ?
(VIII, 8)
The results obtained by means of e.m.f. measurements coincide with
the data produced by investigators who used methods other than electro-
chemical ones. In particular, in those systems where separation in two
liquids (Cu4u25; Pb-PbS) was registered, the e.m.f. isotherms display a
horizontal seciion.
Aa it may be seen from fig. 166 pertaining to the Ou-Ou2S system, the
length of such a segment corresponds approximately to the extent of the
separation region as reflected in the fusibility diagram. On the contrary-,
in eases where there is no separation (FeS-Cu2S; Ni3SECu2S), the e.m.f.
changes smoothly with the composition (see fig. 167).
-465-
Declassified in Part - Sanitized Copy Approved for Release ? 50-Yr 2014/05/01 : CIA-RDP81-01043R003400070002-1
Declassified in Part - Sanitized Copy Approved for Release ? 50-Yr 2014/05/01 : CIA-RDP81-01043R003400070002-1
45 Cu:, 5
% S
Fig. 166. Comparison of the
fusibility diagram for the
Cu-Cu28 system with the e.m.f.
dependence upon the melt compos-
ition.
Fig. 167. Dependence of the e.m.f.
upon the composition at 1180?C (1) and
the fusibility diagram (2) of the
Cu2S-Fe8 system.
The activities of sulfur calculated. from the e.m.f.s and according
to the data for chemical equELLIWALsib.5/ do not, as a rule, differ very
much. For Fb-PbS and Cu-0u25 melts the activity coefficients deviate
hardly at all from unity (diluted solutions).
On the other hand, in On-Ni-S and Cu-&e-S systems, where the atoms of
the different metals differ by their binding energies with sulfUr, the
activity coefficient for the latter grows sharply (11 or 20 times) with
rising concentration of the weakly interacting element (Fe or Ni, respect-
ively). Filially the activity of sulfur in 0U26-YeS and Cu2S-Ni3S2alloys
fails to obey the rule of additivity.
The cited regularities confirm Venytkovis assumption concerning the
electrochemical character of interaction between liquid mattes and Slags
and also shed additional light on the nature of sulfide melts. They in-
dicate that the latter, while being semi-conductors, still approach metal
alloys by their properties. The ion-atoms of these melts are not statis-
tically evenly distributed in space. Same preferential groups of them may
be found here, which cause the deviation from the laws of perfect solutions.
Cells with Slags of Different
Composition.
The presence of a potential jump at the metal-slag interface is con-
firmed not only by the cells with alloys of varying concentration, but
also by galvanic cells with &Aga of different composition.
Furthermore, if the above-cited investigations of the ooncentration
cells of the amalgam type permitted establishing of the existence in liquid
-466-
?
?
?
?
?
?
?
?
4
metals of stable FeSi and Fe2P compounds, then the study of the cells with
a varying composition of slags enables one to reveal in the latter the
presence of complex silicon-oxide anions. In order to determine whether all
calcium silicates found in the solid state also remain in liquid slag, the
following galvanic cell was constructed Z.77
or
Fe, Si I CaO, Si02, A1203, MgO ! C
MgO I CaO, Si02, A1203, MgO I Fe, Si,
(VIII, 9)
the e.m.f. of which was determined only by the ratio of the silicon
activities in both slags.
In order to avoid the emergence of diffusion potentials, the cell was
made to constitute a concentration chain wlthout transfer. It was composed
of two chemical chains with slags of different composition, but with the
same electrodes.
In each of these cells the energy developed as a result of the silicon
formation reaction. Hence, the alloy containing silicon (Fe-Si) constituted
one of its poles, while the other pole was formed by the oxygen electrode.
In a number of tests graphite was used to Serve as the latter, because
graphite is capable of retaining a certain proportion of combined oxygen
even at high temperatures. In cases when slag is apt to be easily oxidized
through contact with oxygen, or reduced by carbon, the oxygen electrode
must be inert and resemble the go-called glass electrode which is being
used in aqueous solutioRs. This condition in molten slags is apparently
satisfied by solid magnealum oxide. At high temperatures it msnifests an
appreciable conduction and quite a definite activity of the oxygen which it
contains.
Az follows from fig. 168, measurements with graphite and magnesia
electrodes produce similar results. Apparently, the errors caused by
some degree or other of their irreversibility as oxygen electrodes are
negligible.
Furthermore, the foregoing permits consideration of the transition
cell /7 as being composed of oxygen (graphite) and silicon (carborundum)
electrodes. Indeed, upon appropriate conversion, these data do not deviate
too much from the section of the curve which refers to acid Slags.
-467-
Declassified in Part - Sanitized Copy Approved for Release ? 50-Yr 2014/05/01 : CIA-RDP81-01043R003400070002-1
Declassified in Part - Sanitized Copy Approved for Release ? 50-Yr 2014/05/01 : CIA-RDP81-01043R003400070002-1
/0
o-/
? -2
x
Jo io 4/0112
Fig. 168. Dependence of the e.m.f, on the contents of silicon:
1 - data for the magnesia electrode; 2 - same for graphite electrode;
3 - smile according to the data recorded in N.
There are two inflections in the isotherm of e.m.f. (see fig. 168).
Their presence indicates the existence in the melt (at 14/0?C) of stable
silicon-oxide anions of different degree of complexity. The first break
corresponds to the composition where practically the entire oxygen, with
the exception of the proportion uped for the formation of A10-2-, is com-
bined in the form of SiOtion. The second break corresponds to metasili-
cate MeSiO3' i.e., here silica is to be found mainly in the form of
(S102) anions.
3 n
In view of the a3ove, the first s
the melt containing 0, A10-' and Si
2
pertains to liquid slag with A10-2-, Si
ction of the curve characterizes
- anions, whereas the second section
and (8102-) anions.
3n
As to the third section, the SiO4- and A10-- ions are absent from the
4 2
melt; instead of than, along with the existing (Sil)n, there appear here
compound complexes of the 8i03- type, the composition of which is, ap-
parently, variable.
Thus, the cited data not only confirm the electrochemical theory of
interaction between metal and slag, but Also point to the existence in
the melt of anions of the Ca2S104 and CaSiO3 compounds. On the other hand,
compounds characteristic for their low stability in liquid state (which
melt incongruently), for instance Ca351207, do not manifest themselves in
any appreciable way in the isotherms of electromotive forces.
?
?
?
14
?
POLARIZATION AND ELECTROCAPILLARY PHENOMENA.
The electrochemical theory is further substantiated by the polarization
of the cell, i.e., the change of the difference of the electrode potentials
depending on the strength of the flowing current. This phenomena was dis-
clved (fig. 169) in a voltaic cell (VIII, 3) composed of Fe-Si-C alloys
L6/.
Fig. 169. Polarization of a galvanic cell at different concentrations
(in atomic fractions) of silicon in its alloys with iron:
1 - anode Nsi = 0.6, cathode Nsi = 0.025; 2 -.anode Nsi m 0.10,
cathode Nsi = 0.6; 3 - anode Nsi = 0.36, cathode Nsi = 0.6; 4 -
anode Nsi = 0.48, cathode Nsi m 0.6.
The investigation of polarization by commutation method showed that
it was caused by slow diffusion of ions in metal. This is also supported
by the fact that the angular coefficients of the polarization curves (re-
duced to common geometric characteristics) remain practically constant,
in spite of a strong change of the composition of the metal alloys (from
2.5 to 48 atom.% Si), and depend on the composition of slag.
The existence of electrode potential and polarization lend confidence
to the belief that the superposition of an external electric field may
under favorable conditions cause a change of interphase tension at the
metal-slag interface. Actually, such phenomena were revealeiLlqin the
ferrous metal - slag system at temperatures of 1350-1500?C. For this
purpose a method was used consisting in the determination of the size (per
x-ray picture) of a liquid metal bead located under a. layer of slag and
subjected to cathodic and anodic polarization.
-469-
Declassified in Part - Sanitized Copy Approved for Release ? 50-Yr 2014/05/01 : CIA-RDP81-01043R003400070002-1
Declassified in Part - Sanitized Copy Approved for Release ? 50-Yr 2014/05/01 : CIA-RDP81-01043R003400070002-1
Fig. 1?). Dependence of interphase tension cr at the Fop? - slag
interface from the jump (pa' potential, and the curve
of the strength of the polarizing current (I).
Fleetrocapillary phenomena were studied for iron allays both with
carbon and phosphorus in contact with Slags composed of CaO, Na20, A1203
and 402. A. few of the electrocapillary curves obtained are displayed in
fig. 170.
It was revealed thereby, both in the presence and in the absence of
the external electric field, that the surface of the mentioned metal alloys
is negatively charged, that the sodium ions at these interfaces possess a
greater capillary activity then the ions of calcium, and that carbon ad-
sorption fails, whilst that of phosphorus remains practically unchanged
with growing negative value of the metal potential.
The foregoing provides still another confirmation of the electro-
chemical theory of interaction of ferrous metals with slags. Moreover,
this offers an explanation, for example, of the influence exerted by the
sign of the pole, carried to the electrode, upon the size of the metal
beads separating from it during welding with direct current.
QUALITITATIVE PATTERN OF INTERACTION.
Lotus now examine how the universally known phenomena observed during
metal and Slag interaction can be explained in terms of the electrochemical
theory. We shall first dwell on the question why is it that the oxidizing
power of slag cannot be determined in a simple form by all the oxygen ions
existing in it.
Oxidizing Power of Slag.
. Let us take a basic Slag containing, say, Ca01 MgO, FeO, and a small
proportion of 8102, i.e., consisting of Ca24, mg2., Fe2*, 02-, and Pit
ions. Suppose it is brought into contact with molten liquid iron. Each
type of the ions of Slag will tend to pass partially into the metal phase.
-470-
?
S.
4
?
.e
)
?
This tendency will be determined by the difference in the binding energy
of the particles with both phases and by temperature.
Since the binding energy of the exygen ion with metal is sufficiently
high, a fUllymeasureable number of 04- ions may penetrate into liquid
iron.
02-
(slag) = 0 (met.) 2 (1
(nil, 10)
However, this will disturb the electro-neutrality of both phases.
Owing to the shortage of negative ions, the slag will be charged positively
and the metal negatively because of the presence of oxygen anions. Since
excess charges in conductors are being expelled onto the surface, there will
appear at the metal-slag interface a double electric layer and in connec-
tion with this a jump in the potential will manifest itself.
The existence of a double layer will impede the further migration of
oxygen anions into liquid Iron. If the interface surface is not too ex-
tensive (i.e., considerably geniler, for example, than in emulsions), then,
as is evidenced in electrochemistry, a very smnil amount og excess charges
will be sufficient to inhibit completely the transfer of 04- anions. Con-
sequently, their tangible macroscopic transfer will become possible only
provided it is not accompanied by a further increase both of the density
of the double layer envelope charge, and of overvoltage.
This condition materializes, in particular, when an equivalently
charged number of some or other cations migrate from the slag into metal
alongside with oxygen anions:
Me24. = Me - 2 e (VIII, 11)
(slag) (met.)
This transfer is Also facilitated by the arrangement of the double
layer charges: the excess of positive ions in slag contributes to the
expulsion from it of cations attracted by negative charges which abound in
the metal.
According to the elelctromotive force series, the equilibrium con-
centration in liquid metal must be the greatest for cations of those
metals which in the given system are relatively the highest. In the case
under consideration these will be the cations of iron which actually
accompany the migration of 02- anions
Fe24' -F 2 8
(Slag) (met.)
Declassified in Part - Sanitized Copy Approved for Release ? 50-Yr 2014/05/01 : CIA-RDP81-01043R003400070002-1
-471-
(VIII, 12)
Declassified in Part - Sanitized Copy Approved for Release ? 50-Yr 2014/05/01 : CIA-RDP81-01043R003400070002-1
Regardless of whether calciln?) and magnesium are highly or weakly
soluble in molten iron, their transfer into slag in appreciable quantities
is impossible here. The position of Ca and Mg (highly base) in the
electromotive series is such that they, while remaining in the metal,
must expel the iron ions from slag and pass practically completely bank
into Slag:
2+
Ca(mat)--1- Ca).
(lull, 13)
Thus, in the examined example only iron cations can accompany the
transfer of oxygen anions into metal
(VIII, 14)
Fe9Fe + 0
st tc() Omit) (met)*
One should not presume, however, that in consequence of this the
double layer and overvoltage will disappear. This will not occur because
the tendencies for migration into metal of 02- and Fe2* ions (determinable
by their binding energies with the phases) are different. If this tendency
is stronger with 02- than in Fe2+, then a double layer will be formed with
a negative charge in metal (and positive in slag), which will inhibit the
transfer of 02- and facilitate the migration of Fe24".
On the contrary, if the tendency towards migration in 02- ions is
weaker than in Fe24., then the arrangement of the charge signs in the double
layer will be inverse (an excess of cations in metal, and of anions in
slag) which will also evaTize the transfer rates of both types of ions.
It becomes clear from the aforesaid thnt the oxidizing power of slag
cannot be determined only by the concentration of all oxygen anions which
exist in it. It also depends on the concentration in slag of the Fe2*
anions. If the latter are absent, then oxygen practically does not pass
from slag into metal at all. Its transfer increases in proportion with
the growth of the iron cation contents in slag.
The oxidizing power of slag depends therefore not only on its content
of FeO, as follows from the molecular theory, but is simultaneously deter-
mined by the concentration of all oxygen ions and the concentration of
iron ions. In other words, in the simplest of cases (without any allow-
ance for the activity coefficient) the oxidizing power of slag must be
measured not by the mole fraction of ferrous oxide Npeo, but by the product
of the ionic fractions of oxygen and iron, i.e., Npe2+ ? 1102-.
1) As shown by tests Z177 with application of tracers (radioactive
isotope Ga45) the solubility of calcium in liquid iron saturated with
carbon and covered by a layer of slag composed of CaO,A1203 and Si02 is
inferior to the detection sensitivity of this method (6.10-5P-
...47a-
?
?
Effect of Silica on Oxidizing
Power of Slag.
Let us now consider the reason why the oxidizing action of slag for
one and the same concentration of ferrous oxide declines upon the addition
of silica.
Had there been no 5102 and other acid and amphoteric oxides contained
in Slag, i.e., if it consisted, for example, only of FeO, CaO, and MgO,
then the concentration of oxygen ions in it would have been constant regard-
less of the changes in the Rontents of the individual components. The
degree of the transfer of 04- anions into metal would in this case be
determined, as a first approximation, only by the concentration of Fe24
ions in slag. More precisely (and this actually reflects the difference
between the real slags and perfect solutions), the tendency towards migra-
tion would not be solely dependent on the ionic fraction of Fe2+, but also
on the concentratIon of other cations, inasmuch as they retain, with vary-
ing energy, the 04- anions in Slag.
The addition of silica leads to a considerable increase of the energy
binding the oxygen ions with slag. In consequence thereof, even with a
constant contents of FeO, the introduction of 5102 diminishes the tendency
of 02- anions to migrate into metal. This is to be attributed to the
formation of sufficiently stable silicon-oxide anions resulting in a de-
crease of the o2 ion concentration. For instance, in basic Slags SiO4-
4
anions emerge upon the addition of 5102
SA L+. 202- (VIII, 15)
and in in the general case - Si_O ions.
y
x Si024- (y ?2 x) 02- SixOyz- .
(VIII, 16)
Thus, if in the molecular theory the effect of 8102 is attributed to
a reduction in the degree of iron orthosilicate dissociation into molecules
of free oxides
Fe, SiO4 =t 2 Fe0 Si02
(VIII, 17)
then, according to the ionic theory, it comes as a result of a drop of the
oxygen anion concentration following the formation of silicon-oxide com-
plexes.
Similar too is the action of other acid oxides (P205, TiO2 and so on)
as well as of Fe203'Al203 and their analogues, if they are present in the
form of anions binding thereby the ions of oxygen.
?473-
Declassified in Part - Sanitized Copy Approved for Release ? 50-Yr 2014/05/01 : CIA-RDP81-01043R003400070002-1
Declassified in Part - Sanitized Copy Approved for Release ? 50-Yr 2014/05/01 : CIA-RDP81-01043R003400070002-1
Desulfurizing Power of Slag.
We shall now consider the question of why the desulfurizing power of
slag is primarily a function of the Ca0 contents in it.
If dissolved oxygen is absent in iron, then the transfer of S2- anions
from Slag into metal, as in the ease of 02- ions, must be accompanied by
a migration of Fe2f cations,x_ The introduction of CaO into slag weakens
the migratory tendency of er" ions, in the first place, because the con-
centration of both S2- and 42+ decreases, and in the second place, in view
of the increasing penetration of 02- anions into metal. The latter also
demand the transfer of an equivalent number of Fe2+ cations, causing thus
a reduction of the share of Fe21' ions accompanying the S2- anions during
their passage into metal.
On the other hand, should the MCO contents be increased in Slag, in-
stead of Ca?, then the concentrations of 82-- and Fe2f ions would decrease
as in the preceding case. However, since metallic manganese is closer to
iron in the electromotive force series than calcium or magnesium, its ,
cations, to a certain degree, might accompany the transfer of S2- and 0."-
anions into metal. This would weaken the effect of the second factor
(pertaining to the reduction of the share of cations accompanying the anions
of sulfur) and somewhat increase the chances of the S-4- anions to migrate
into metal. In this respect the addition not of MnO, but of FeO, would be
even more effective. Then the number of cations capable of accompanying
the anions would not be changed at all.
If there is a definite concentration of dissolved oxygen in metal,
then the charge compensation required for the macroscopic transfer of the
sulfur anions may be achieved not only as a result of the migration of
iron cations
Fe24 S' ? Fe + S
(51 (5I 1.4? Irma/ (met) ?
(nil, 18)
A similar effect will be produced also by the transfer of oxygen
anions in the reverse direction
S7 +0 s +02?
(sti (.44 (..) (0).
(VIII, 19)
The above equilibrium should be shifted more to the left, the weaker
the bond between the oxygen anion and Slag.
-474?
Since the 02- ion is being retained firmly enough?) by the Fe2+ cations,
leas gelidly by MC2f cations, and comparatively weakly by Ca2f ions (owing
to the change of the cation radius), it would seem reasonable to expect
that, at the same concentration of MOO, the equilibrium shift (VIII, 19) to
the left should be the greatest for CaO, intermediate for MnO, and the least
for Fe0.
However, the bond stability of the 82-anions Also grows with the
change from Ca2+, to Mia;+ and Fe2f. It is true though that the increase of
interaction energy is somewhat less here than in the ease of 02- ions since
their radius is smaller than that of SZ-. Yet, in spite of this, the exam?
ined equilibrium shifts but a little /a7 upon replacement of CaDA tgrItnO,
ad 740 in view of a simultaneous strengthening of the bonds of (Y4- and
S4- ions with Slag.
Oxidation of Impurities.
Let us tarn now to the oxidation processes.
The elimination of carbon, silicon, and phosphorus will depend not
only an the concentration of oxygen anions in elag, but also on the contents
in it of iron cations. Indeed, in all these cases, on the one hand, as it
may, for instance, be seen from the following possible transfer schemes2)
Ct)sl + 02- ? 4? CO + 2 A ,
(me (74
Si +4 02- 4-=.- SiO4- + 4 0,
one.t) (0) 4 (St )
02? --*" P03? t 9
Ppner) ?+ 4 01)4? 4 (51 )+5
the metal receives an excess negative charge. Hence, for a tangible macro-
scopic development of these processes, compensating transfers of Fe4*
cations from Oleg into metal are required.
On the other hand, the completeness of the shift of the mentioned
equilibria will also be determined, apart from the concentration of oxygen
anions, by their binding energy with the slag. The effect of the 8102 add-
ition can be ir4erpreted in two ways - either as a decrease of the con-
centration of 04-ions, or as a strengthening of their bonds with the melt.
1) It should be borne in mind that the stability of ion bonds, in
the first approximation, is determinable by their charges and by their
radii according to the law of Coulomb. On the other hand, the interaction
energy of the oxygen atoms will be greater with Ca atoms than with those of
Fe. Such change in the arrangement of metals is due to the difference in
the magnitude of ionization potentials, which can be easily demonstrated
by means of thermodynamic cycle.
2) The schemes are given without any clear-cut factual indication as
to the existence of FeSi and Pe2p compounds.
Declassified in Part - Sanitized Copy Approved for Release ? 50-Yr 2014/05/01 : CIA-RDP81-01043R003400070002-1
Declassified in Part - Sanitized Copy Approved for Release ? 50-Yr 2014/05/01 : CIA-RDP81-01043R003400070002-1
In either case the introduction of S102 will contribute to the shift-
ing of eauilibria to the left. For silicon the shift in this direction
becomes conspicuous only withconsiderable 3102 contents. In this case the
reaction should be more correctly rezritten in the following form
Oy('-(4r)ii + 4 SixOyz-. (s Si( mat ?
(VIII, 23)
This direction if interaction, i.e., migration of Si from Slag to
metal, is more commonly known as the process of silicon restitution.
The dephosphorization reactionl as exhibited by tests, is more sensi-
tive to the change of silica contents.
The substitation of one oxide by
energy- of the oxygen anion with slag.
stead of FeO, on the one hand, should
rium to the right, for example (VIII,
another changes only the binding
Hence, the introduction of CaO in-
contribute to the shifting of equilib-
22).
On the other hand, the reduction of the concentration of Fe2+ cations,
which effect the charge compensation, will retard the oxidation reaction.
Consequently, there should exist an optimum correlation of FeO and
CaO concentrations which is most advantageous for the eliminativa of the
impurity-from metal.
This situation was recently experimentally confirmed for phosphorus
?197.
In limiting ourselves to these few examples, let us remark that the
distribution of elements (iron, copper, niCkell cobalt, oxygen, and
sulfur) between liquid mattes and Slags, as well as in a number of other
eases, can be examined in a similar my.
EQUILIBRIUM EQUATION.
Assuming that equilibrium establishes itself as a result of a cor-
related process of cation and anion distribution, let US formulate the
general relationships on the basis of electrochemical thermodynamics.
To put it differently, in defining the conditions of equilibrium let
us proceed from the experimentally established fact that the ions of each
type, in distributing themselves between metal and slag, participate in
the formation of a double electric layer and in the origination of over-
voltage.
Supposingifor example, that there is a basic slag consisting of CO,
MgO, Feu, MnO, Si02, P205, and OaS, which is in a state of equilibrium
with the metal containing Fe, Mn, Si, P, 0, and S.
The general equations covering the distribution reactions of indiv-
idual ions, or as more commonly referred to in electrochemistry - the
-476-
individual electrode processes - may be recorded schematically (without
explicitly reflecting the presence of FeSil Fe2P and other compounds in
metal) in the following manner:
Fe(2710 + 20 Fe(0,440,
Mni2st + 2 e M n(met) P
? 2 0 0(04?t) ,
(knet))
S2?) ? 29 S
S1044- :140 Si(mej + 4 0(2s-i) ,
P?43- + 5 0 Z- P(met) + 4 o(25?t ) ?
(yin, 24)
(VIII, 25)
(VIII, 26)
(VIII, 27)
(VIII, 28)
(VIII, 29)
These equations reflect only the initial and the final states, i.e.,
that which is indispensable for the determination of the equilibrium con-
ditions, and which is usually insufficient for the identification of the
mechanism and the kinetics.
The equilibrium difference of the potentials for each of the enumerated
electrode processes will comprise respectively:
(VIII, 24')
(VIII, 251)
PT a? 2+
EFe = E? ?In
Fe 2F a
Fe
RTamn
cAin ?=_? shin +
2 F aAin
, RT ao
= 1- ?In
0 0
=
2F a
RT as
+ ?1n
2 F a _
, RT
? in asio4?
4
4F a ? si a02_
a
. R7' , P03?
4
c p = e?
5F aP ? a04 2?
(VIII, 260
(VIII, 270
(VIII, 280
(VIII, 290
Since a single electric potential is being established for the
equilibrium at the metal-slag interface, then
eFe = E = e0 s
S
= i
E, =e =
Mil
state of
(VIII, 30)
Hence, in equating the right sides of the recorded equations, we find
the equilibrium conditions for various metallurgical interactions.
Thus, for instance, from equations (VIII, 24') and (VIII, 261) we ob-
tain the coefficient of oxygen distribution
-477-
Declassified in Part - Sanitized Copy Approved for Release ? 50-Yr 2014/05/01 : CIA-RDP81-01043R003400070002-1
Lo---=
aFe2+ ? a02?
a ? a
Fe 0
Declassified in Part - Sanitized Copy Approved for Release ? 50-Yr 2014/05/01 : CIA-RDP81-01043R003400070002-1
eo? eepe
? exp 2 F1,
R7'
(v111, 31)
and from expression (VIII, 241) and (VIII, 270 the same is derived for
sulfur
Ls
a 2
Fe+ S exp Fe 2 F1 .
=
RT
aFe. as
(VIII, 32)
A combining of the relationships (VIII, 240 and (VIII, 251) gives
the equilibrium constant for the reaction of manganese
K
aFe
? a Mn2
= exp En
Fe Mn
RT
CMn ? aFe2+
? 2 Fl , (VIII, 33)
while the same for relationships (VIII, 241) and (VIII, 29') produces the
equilibrium constant for the reaction of dephosphorization
5 2 3_ e --e?
a ? aon
Fe ? 4 IFeP ? 10F.
Kp ? exp
n2 ? a8 2? ? a5 2+
1) 0 Fe
(VIII, 34)
Without dwelling on other possible expressions, we shall only note the
result of the comparison of equations (VIII, 26?) and (VIII, 27')
K
0. s
a
0.a.2_
Eo
exp eS
a02._.as
R7'
? 2F 1 .
(VIII, 35)
If the concentration of a substance dissolved in iron is not large,
then (according to the theory of infinitely dilute solutions) the activity
of this admixture can be approximatively replaced by percentage in weight,
whereas the activity of the solvent aFe may be assumed to be equal to
unity. In this case the equation will include by way of unknown quantities
only the activity of the ions in Slag. In order to pass from them to the
concentrations, either experiment.ally established ratios, or reference to
the quantitative theory of Slags-Via required.
In those eases where the metal phase cannot be considered as infinitely
dilute solution (cast-iron, special steels, ferrous alloys, mattes) the
substitution of activity by weight percentage becomes inadmissible. Along
1) From the words of semi-empirical character; see, for example, 5717,
zpg. -478-
't
?
?
?
?
?
?
with experimental studies of the degree of the metal phase deviation from
the ideal solution, there arises a need for the development of a quantita-
tive theory for such solutions.
PLFCTROMOTIVE FORCE SERIES.
Aa may be seen from the equations recorded above, the equilibrium
conditions, i.e., the distribution coefficients and equilibrium constants,
may be found experimentally, and not only by routine chemical methods, but
also by means of e.m.f. measurements.
A fUrthAr devlopment awl more. precise elaboration of the measurement
procedure LAir, A/2 04 AV, will permit the production of an electro-
chemical series of electromotive forces for liquid Slags, if the normal
potential for oxygen e or iron E? for example, be taken for a standard.
0 Fe,
On the other hand, the electromotive series for molten slags, i.e., a
selection of magnitudes of normal potentials of elements (Mn, 0, Si, 5, PI
Al, etc.) with respect to that for iron (or oNygen) could be calculated
according to the above-mentioned equations from the distribution coefficients
and equilibrium constants determined by regular methods other than electro-
chemical.
Thus, for instance, if for reactions
1
Met) + Me'0(s1) .:-' Mei naetri- Me"0(51.)
Ale,ati ? Me'20,51 ):?-'-' 2 Me;mtv + Me"0(si )
(VIII, 36)
one should make use of the experimental data L207 for the equilibrium
constants' calculated on the basis of the ideal law of mass action
xomptiej. 1( (meo)tme,i2
_
wommer I omonmel '
then, for example, from the equation
e? =.K,
hie' --- A4e. I00001T1g
(VIII, 37)
(VIII, 3)
which is analogous to expression (VIII, 38) it would become possible to
evaluate the difference of the standard potentials for molten oxidee, i.e.,
the quantity (Erten - ELI) required for the compilation of an electro-
motive force series.
A comparison is made in table 22 of the values - Elle?)
calculated in this manner with the corresponding values for aqueous solu-
tions and molten chlorides Z217. In the latter case the data for Eio
Fe
were lacking, in view of which, for the purposes of comparison, the ratio
was taken for oxides of equilibrium uconstants" for interaction reactions
-479-
Declassified in Part - Sanitized Copy Approved for Release @ 50-Yr 2014/05/01 : CIA-RDP81-01043R003400070002-1
Declassified in Part - Sanitized Copy Approved for Release ? 50-Yr 2014/05/01 : CIA-RDP81-01043R003400070002-1
of various metal oxides with iron.
It can be seen from this tabulation that the sequence order and the
sign of the difference of potentials remained the same in each of the three
series, with the exception of the Ni, Co pair. The potentials of the latter
for aqueous solutions differ very little. Consequently a small difference
in the temperature coefficients could easily change their position in the
series with a rise in temperature. The absolute values of the difference
of the standard potentials for the oxide and chloride melts approach each
other and change regularly with increasing temperature.
Table 22. Comparison of differences of the standard potentials
for aqueous solutions and molten oxides and chloridesc_,
(E E 0 ) v at the temperature ?C of
men Met
Met
Men
Oxides
1600
Fe24-
Co- '
Fe2+
Fe2+
Fe2+
CuNi
2+
-0.10
+0,96
+0,80
1200
+0,51
-0,08
-0,25
-0,33
-0,56
+1,07
+0,84
Chlorides
Aqueous so-
lutions,
25
+0,06
+0,66
-0,007
-0,18
-0,19
-0,96
Similar calculations may be effected also for other elements, provided
the corresponding values of Li and Ki are known.
KINETICS OF METAL AND SLAG INTERACTION
Introductory Remarks.
The advantages of the ionic theory become conspicuously evident when
one examines such phenomena as cannot be explained by the molecular
hypothesis. Listed herein should be, for instance, good electrical con-
ductance of liquid Slags, their electrolysis, the existence of over-voltage
at the metal-Slag interval, electrocapillary phenomena, and the special
form of the equilibrium laws (replacement of the activity of a compound by
the product of the activities of its ions). No doubt, similar phenomena
should also be expected to appear in the domain of the kinetics of metal
and slag interaction.
Under definite conditions the intensity of interaction may, as is well
known, be limited by the rate of any of its stages. At the present time
the difference of the ionic theory from tho molecular in metiers related to
kinetics can be demonstrated with the least difficulty in those cases when
the penetration of the metal-slag boundary constitutes a process determining
-480-
?
?
?
?
V
the rata 2'227.
Indeed, according to the molecular hypothesis there occurs an exchange
of uncharged particles between these phases. Moreover, apart from the
adsorption field, it is not necessary to take into account any other fields
at the metal-slag interface. Furthermore, if the substances which take
part in the transfer form dilute solutions, then the rates of the processes
are directly proportional to their concentrations in the corresponding
phases.
On the contrary, according to the ionic theory it is charged particles
which migrate across the phase boundary. Thus, for instance, as a result
of carbon, silicon, and phosphorus combustion, oxygen ions pass from slag
into the metal; sulfur ions move in the opposite direction during desulfur-
ization; in the process of deoxidation and alloying it is the ions of the
added elements which migrate, and so on. The fact to be considered in all
these cases is that there appear excess charges in metal and slag, while a
jump of the electric potential occurs at the phase interface.
The latter would have rapidly stopped these migrations, had.no other
processes been developing along with them which liquidate the further ac-
cumulation of charges. Among these processes the following should be men-
tioned in particular - migration of iron ions from slag to metal and back,
reduction of the trivalent iron ions to divalent, movement of the oxygen
ions from metal to slag, and so on.
As a result of this the macroscopic process which actually takes place
constitutes a combination of two such stages, with the following possible
alternatives:
1) simultaneous migration of cations and anions in the same directions,
e.g.,
2)
Fe2s+1 ) ? 2 Fe
(tviet)
02- - 2 H -> 0Oriet)
(si
Fe(2s+1 + 0(2-6-1) Fe(met) +?(,b
of anions in different directions
S
2 H -> S2-
ontt
02- ->
(SI) 0 * + 2 H
rnav
SJ- S2-i 1- 0 f?
(M12.0 (Si Si I I (Thtut 9
3) of cations in opposite directions
MnmAtd ? 2 AMn)
Fe2$+ 2 H Feanet)
Mn(m4 Fe(2s+i) mn(2siirri Fe
(third ,
(VIII, 39)
40)
(VIII, 41)
Declassified in Part - Sanitized Copy Approved for Release ? 50-Yr 2014/05/01 : CIA-RDP81-01043R003400070002-1
Declassified in Part - Sanitized Copy Approved for Release ? 50-Yr 2014/05/01 : CIA-RDP81-01043R003400070002-1
?
A) combining of the trivalent iron reduction events with the migra-
tion of anions
2 Fe07(51 2 9 2 Fef2;11") + 4 0(2s-/
02- ?2 C16-fleW
(sl )
2F'e02 (so 2 Fe(2;1")+30(2s-1)+0(,,04,
5) same in conjunction with the transfer of cations
2Fe07 + 29 -* 2 Fe(2:0 ? 4 0i6f)
Mnwet) ? 2 Mn
2 FeOT Mn(het) -9- 2 Fe(2's4f)--1-MQ-1-4 0(25-1
(viii, 42)
(VIII, 43)
The presence in one process of the two stages mentioned, and the exis-
tence at the phase interface of overvoltage affecting substantially the
development of these stages, impose a singular imprint on the aspect of the
kinetic equations distinguishing it somewhat from the one pertaining to the
transfer of uncharged particles.
We shAll deal in detail only with one of the unemerated cases, since
the analysis is similar in every respect for all other alternatives. Let
us consider the combination (VIII, 40), i.e., the transfer of sulfur and
oxygen in different directions. Preference was given to this alternative
in view of the experimental material available for subsequent comparison
with the regularities determined.
Deduction of Kinetic
Equations
Let us assume that sulfur ions migrate from metal to slag, while the
subsequent accumulation of charges in both phases and the increase of the
potential jump at the interface is eliminated by the inverse movement of the
oxygen ions.
Since the process is in development, the equilibrium potentials of
sulfur Epls and oxygen ;1,0, corresponding to their activities ai in both
phases, are still to be attained. They can be represented by equations
. PT
E
2F
a o
PT a
0
, eo -- In
C.
P.
2F a
in which Ei? and Ei? = standard potentials of sulfur and oxygen;
0
R = gas constant, joule/degree;
T =absolute temperature;
F = 96500 Coulombs.
-482-
( mix 44)
(VIII, 45)
?
s
(
?
0
?
?
At any given moment throughout the process there arises some single
potential Eat the phase interface. It differs-from the equivalent
potentials
E= s
E ==E
S p. 0 I U
by quantities 72 - and
(VIII, 46)
The latter were assumed to have different signs for the following
reason. For a directed transfer of sulfur from metal to Slag, it is neces-
sary to diminish the potential as compared to the equilibrium ones. In
other words, the density of the positive charge in the metallic envelope
of the double layer must be reduced. On the contrary, for the migration
of oxygen ions from Slag to metal an increase is required of the potential
over the equilibrium potential, i.e., intensification of the density of the
positive charge in metal. It follows directly from the above that the
signs of -rz and should be different.
LO
The constant accumulation of charges discontinues with the establish-
ment of a steady macroscopic process. The ions, which move more rapidly,
accelerate the migration of those which move slowly, while the latter slow
down the movement of the former. Therefore, the transfer rates of sulfur
and oxygen ions are equal at any given moment
?+ 4?
V = s V 0 .
(VIII, 47)
The mathematical expression for the rates of such transfers was dev-
eloped in electrochemistry relatively long ago L237. Furthermore, it was
applied to the analysis of similar processes, namely to the solution of
sodium from amalgam
and in general to
media
By utilizing
review
Na(Hg) Na 0
Na_)
1-1+
(p-m
Na(Hg)--1--H+ Na_ HI,
(p?i') (p fl)
2
the corrosion of metals and allays by aqueous aggressive
these expressions we obtAin for the processes under
[Slexp
(gp. s ?1s) 2Fai
v =vs= k s. s
PT
1
I (i_ s? T,$) 2F a2. s
? k j. s (SI? ) CX1) i P
-483-
RT
(VIII, 48)
Declassified in Part - Sanitized Copy Approved for Release ? 50-Yr 2014/05/01 : CIA-RDP81-01043R003400070002-1
Declassified in Part - Sanitized Copy Approved for Release ? 50-Yr 2014/05/01 : CIA-RDP81-01043R003400070002-1
and
+10)2 a2.0 I
V Vu = k2, l`?1 / "t" P? ?
?k1 01?Jexp (tp. 0+ 71)) 2al. F
1 ?
. RT
(VIII, 49)
Here grand Z.07= atomic fractions of sulfur and oxygen in metal;
(S2-) and (02-) = their ionic fractions in Slag;
0( = constant coefficients, depending on the correlation
of the interaction energy of a given particle with
both phases (more precisely upon the shape of the
potential curves at the point of their intersection),
whereupon according to the deduction
1. s az, s = 1
and a + a = 1;
1, 0 2. 0
(VIII, 50)
k and k2,s are the constants of the direct and inverse reactions
1,S
rates,
SOneD) + 2 0 = sr2S-1 r
They include the energy of activation E of the corresponding event
k" = s exp
RT ?
(VIII, 51)
%
The same applies to the rate constants k120 and k2,0 of the reaction
00,1e14 4- 2 Et =, .
Since the transfer is realized by the particles located at the phase
boundary, it is determined, in the first place, by their concentrations in
the adsorption layer.
Because the contents of sulfur and oxygen in metal, and of their ions
in acid slags, is usually small, one may consider that the adsorption layer,
is far from being saturated, as far as they are concerned. In other words,
the concentrations of these substances in it are directly proportional to
the volume concentrations. The factors of proportionality are included in
the quantities ki20 and k.
To find the explicit form of dependence of vupon the composition and
-484.
?
?
?
?
?
temperature, it is necessary to exclude (Ep2s, Ep20,-r2.s and Tzo from
equations (VIII, 48) and (VIII, 49). We shall first substitute in formula
(VIII, 48) the value of epos from expression (VIII, 44). At the same
time we shall replace the activities of S and 82- by the products of con-
centration by the activity coefficient, for example,
as = [S] .
(VIII, 52)
Than, considering that 71= 0 when v = 0, we find, after a number of
modifications, that
Here
--?
f 2F ris
k [Si a" (S2-
= ? Vs - S ; I ? S exp RT
r 2 F yis
1
? exp
PT
?
al
1
s j (VIII, 53)
7' f 2F 2 s
k k Ar s== k 42.s ; A- s exp '
s s S I kr
7
Proceeding in a similar manner with E p20
we have
where
(VIII, 54)
and the equation (VIII, 49)
V =.v0 ==k0 [0122.0(02-)a1.?{eXp )
RT 2,O}
2F f 2
RT 1.
?
exp o)1 a (
= k1.0 B1. 0 = k2. 0 B-a 2,0 ;
I 2? 2F E?
B=01
eXp
10
RT
(VIII, 55)
(VIII, 56)
In passing over to the replacement of rz s and 770, we determine their
interrelation from equations (VIII, 44), (VIII, 45), and (VIII, 46)
17 RTin 11 2_
2F [0] (S2?) 2F To
and substitute forlh in formula (VIII, 55) its value from expression
-485-
(VIII, 57)
Declassified in Part - Sanitized Copy Approved for Release ? 50-Yr 2014/05/01 : CIA-RDP81-01043R003400070002-1
(VIII, 5/). Then
Declassified in Part - Sanitized Copy Approved for Release ? 50-Yr 2014/05/01 : CIA-RDP81-01043R003400070002-1
RA B [S] o 2F 1
? 0 ' s2_) (0 ) CXp
v k s RT a2, }?
(VIII, 58)
0
{2 Theis
( BA (SI2s?]) 11. ofOlexp
RT cli.?1] ?
The system of two equations (VIII, 53) and (VIII, 58) permits one in
principle to exclude n s and to express the rate of the process v through
concentrations and temperature. However, inasmuch as all c4i are usually)
proper fractions, it appears to be impossible to do so in a general way.
The problem is to be solved for every particular case separately when the
values for all are predetermined.
By way of an example, let us find the solution for the case, when
and, consequently,
a1 =-- a =al. 0 I
I, S
a ..a ==a ?
2 2,S 2, 0
(VIII, 59)
(yin, 60)
This assumption can be substantiated in a certain measure by the sim-
ilarity of a number of sulfur and oxygen properties. The possibility,
therefore, is not excluded that if not the absolute values of these parti-
cles, then at least the relationship of their interaction forces with both
phases, are approaching each other.
Upon this assumption, equation of the expressions for v given in
formulas (VIII, 53) and (VIII, 58) permits reducing the exponential con-
taimingrz sa2, and determining the exponential including/1 s withouto(i.
Upon exclusion of this exponential from equation (VIII, 53), and after a
series of modifications, we find the unknown relationship between the rate
and the phase compositions
1. 0 101 (s2?)
k 2' 0 (02?) (SI ? k
s k s k
2.S I. S
V ?
I (S i+ k1'0 10i1 ? l(S2?) k2. (02?
)(
1,s s
1) With reference to the possible values of 47( see L257.
-486-
(v111, 61)
6
4
?
4
?
?
Certain Conclusions Resulting from
the Kinetic Equation.
Supposing the concentration of oxygen in metal is constant for each
given temperature. Then equation (VIII, 61) may be rewritten as follows
? la
V ?
(a1 x)
(VIII, 62)
Here x = (82), whereas the remaining quantities are grouped within
the parameters a, al and b.
It follows from expression (VIII, 62) that v drops with increasing 34
Moreover, it is easy to see that not only
but also
(VIII: 63)
(VIII, 64)
In other words, the fall of rate v. of the desUlphurization process is
retarded with the growth of concentration x of sulfur in slag.
Let us deal now with the second characteristic. From equation (VIII,
61) it transpires that the initial rate which complies to the condition
(S2-) = x =0, (VIII, 65)
may be represented by the formula
? a kstS1
=
0
ad 0
Ie
I is] kl
lb]l
I. S
(k yt,
2.0 ? (?2_)
k
2. S
(VIII, 66)
It may be seen from the above that the rate of the direct reaction in-
creases not only with the contents of sulfur in metal, but also with the
concentration of oxygen ions in slag, i.e., with its basicity. This deduc-
tion is valid for all cases, with the exception of the one when 0(1 ma
0 ( i.e. 0(2 = 1).
The third peculiarity refers to the tangent of the curve v(x). Its
expression is produced by differentiation of equation (VIII, 62)
-487-
Declassified in Part - Sanitized Copy Approved for Release @ 50-Yr 2014/05/01 : CIA-RDP81-01043R003400070002-1
Declassified in Part - Sanitized Copy Approved for Release ? 50-Yr 2014/05/01 : CIA-RDP81-01043R003400070002-1
[alb a2a b (1 ?a2) .11
(egi +x) -Fai
At the initial instant of time
1 k 1 / k V
, ks ta2 ISI + ?k-1..1) 10111 ( 2.81k
alb + ao 1.8 2. 0
V.v..? ,M =
all-f-a,
kkt. o (oil (02?)a.
a,
1, S
(VIII, 67)
(VIII, 68)
This relationship shows that the quantity (-4.40) is inversely pro-
portional to (021. In other words, the drop of the process rate at the
initial moment is the greater, the less the concentration of oxygen ions
in Slag, i.e., the lower its basicity. Let it be noted that this conclu-
sion is true when 0(2 is not equal to zero (0?1 0! 1).
The fourth singularity consists in the following. From expressionk
(VIII, 66) -- if the small addend in the denominator which contains LO/
be disregarded -- it transpires that the rate constant of the direct reac-
tion (at the initial moment) may be represented, considering equations
(VIII, 5.4) and (VIII, 56), in the following aspect
k,ks yi. 2. s k`1 =
2, 0 1. S 2,0 ?
2,S ?
(VIII, (.9)
Whence, and from expression (VIII, 51), it follows that the seeming
activation energy of the direct reaction is composed of the energies of
activation E1,5 - of the sulfur transfer from metal to slag, and E20 - of
the oxygen ions migrating in the opposite direction, or more precisely
(VIII, 70)
E, = o2 E s + E2. o ?
Since 1/40 depends on the stability of the bond of 02- ions with Slag,
and it, in turn, is lower the greater the basicity, then it may be expected
that the activation energy of the desalftrization process will decrease
with growing basicity of Slag.
-488-
?
?
?
?
4
4
2. DESULFURIZATION REACTION.
During the smelting of ferrous metals with charge materials, a certain
proportion of salfur, as a rule, is introduced, which lowers the quality
of production. Consequently, a necessity arises in a number of cases for
measures to be taken which would assure the elimination of mann,. They
are applied either during the metal smelting process or through ex-furnace
processing.
The first method is universally applied in steel production by basic
processes. in cast-iron smelting, ex-furnace procedures are also used
sometimes to remove sulfur in addition to the proportion eliminated into
slag during the smelting cycle. Listed among these methods should be the
treatment of east-iron in the ladle by means of various reagents extract-
ing sulfur and often containing salts of alkali metals.
In the process of metallurgical smelting the metal interacts with slag
and furnace gases as a result of which an exchange of sulfur takes place
between these three phases.
According to the opinion of a number of investigators, sulfur in ap-
preciable quantities passes into metal from furnace gases when sulfurous
fuel is being used. or this reason it seems to be interesting to consider
the interaction of metal both with the gaseous phase and with slag.
INTERACTION OF IRON WITH GASES CONTAINING SULFUR.
The reactions of gaseous 1125 and 502 with liquid metal are not only
of technological importance. They are broadly-used in the study of the
properties of Fe-S melts. This refers particularly to the activity of
sulfur dissolved in metal. The application of the method based on the
determination of the dependence of the equilibrium composition of the melt
on the temperature and the partial pressures of the reacting gases is most
widespread.
Reaction with Hydrogen Sulfide.
TO determine the activity of sulfur in iron the equilibrium of reac-
tion
SFe + H2. (9m4= H2S? (VIII, 71)
is investigated.
The selection of this reaction is based on the relative ease with
which equilibrium can be established, on the reversibility of the reaction,
and the commensurability of the concentrations of dIl substances which per-
mits the production of fairly accurate results by routine analytical methods.
In spite of this, the data of different investigators disagree to a
considerable extent, which fact should be attributed to measurement errors.
Some authors L267; Lg17, for example, failed to take measures to eliminate
thermal diffusion which modifies substantially the composition of gas near
-489-
Declassified in Part - Sanitized Copy Approved for Release ? 50-Yr 2014/05/01 : CIA-RDP81-01043R003400070002-1
Declassified in Part - Sanitized Copy Approved for Release ? 50-Yr 2014/05/01 : CIA-RDP81-01043R003400070002-1
the metal surface.
Britzke and Kapustinsky L287 were the first to demonstrate that the
effect of thermal diffusion is especially noticeable in the case of gas
mixtures possessing different densities. This, in particular, concerns
the mixture of H2 and H28 in which an appreciable separation of components
may ensue in consequence of thermal diffusion. Additional errors arise
L267 in view of the inhomogeneity of the temperature field, manifested dur-
ing heat-treatment of metal in the induction furnace.
The narrowness of the temperature interval contributes to the errors
1,267 because it reduces the reliability of the obtained equilibrium con-
stant versus temperature relationship. Also responsible for the discrep-
ancies is the existence in metal of appreciable concentrations of silicon
(2-3%). The presence of the latter considerably increases the activity
of sulfur.
Some authors 2'307in the study of reaction equilibrium (VIII, 71)
applied the method of the jet, adopted earlier in the work by BritZke and
Kapustinsky L287. The data obtained thereby refer mostly to the highly
sulfurous alloys (up to 37% 8) and embrace a fairly wide temperature
range (from 860 to 1530?C).
As may be seen from fig. 171, the results pertaining to pure FeS are
Close to those established earlier by Britike and Kapustinaky. Moreover,
the temperature dependence of the equilibrium constant proposed for this
composition
K ig Pir,s 3025
+ 0,873
PH,
doe's not differ practically from Britzke-Kapustinskyl s equation
t,s
Ig K ? lg
PH,
3070
+ 0,940,
7'
(VIII, 72)
(vIII, 73)
although the latter was obtained for a narrower temperature interval.
-490-
1
?
3.
4
?2,50
9,s 8,57,54j
Fig. 171. Effect of temperature upon the equilibrium composition
of the gaseous phase (H2, 1128) at varying contents of
sulfur in metal as per different authors' data:
1 - FeS 2)7; 2 - FeS/17; 3 - 28.9% s_L3P7; 4 - 36.48% S L307;
5 - 30% L307; 6 - L6 ; 7- 1% L307 -
Leeend; A) Thermodynamic calculation.
As regards the relationship to alloy composition, it appeared, accord-
ing to these data 2307, moreover that the ideal law of mass action is
operative when sulfur contents does not exceed 4%,
Z267
ig K _ ig PH,S N?
P11, ArFeS
3025
+0,766.
T
(VIII, 74)
The above relationship is more correct than the one suggested earlier
Ig I<
4500 + 1,853 .
(VIII, 75)
Finally, it was established L307 that both for solid and liquid Fe-S
alloys the isotherms reveal a horizontal section over an interval from 6
to 24% S (fig. 172).
In more recent works special measures were taken to eliminate thermal
diffusion. Thus, in the case of induction furnace!, special gas heaters
were used with argon being introduced into gas 2317. In other investiga-
tions /.327 hermetically sealed furnaces were utilized with spiral carbon
heaters assuring a sufficiently even temperature field. In addition to
this, the gas mixture was sparged through metal.
-491-
Declassified in Part - Sanitized Copy Approved for Release ? 50-Yr 2014/05/01 : CIA-RDP81-01043R003400070002-1
Declassified in Part - Sanitized Copy Approved for Release ? 50-Yr 2014/05/01 : CIA-RDP81-01043R003400070002-1
a
S
80
40
1417,1T
0,41
go go 20,0 NS]
Fig. 172. Effect of temperature and sulfur contents on the
equilibrium composition of the gaseous phase in
Fe1S-H2-H2S systems.
Let us dwell for a while on the experimental data bl7pertaining to
the temperatures 1530, 1610, and 1725?C and the sulfar contents ranging
from 0.4 to 4.8%.
Minor corrections for thermal dissociation of hydrogen sulphide were
usually introduced here into the initial compositions of the gaseous phase
as per equation
4740
lg
r
?72
Ha r
2,582. (VIII, 76)
Thereafter, a quantity similar to the equilibrium constant of reaction
(VIII, 71) was calculated
K' =.
I 14121% sl
(viii, 77)
Graphs were then plotted of lg K1 against 4, at L%07= const. The
values of were thereupon determined by interpolation for 1600?C, on the
basis of which the isotherm recorded in fig. 173 was plotted. The straight-
line relationship between NS and ggwas maintained approximately-up
to 0.6 - 0.8% A. Deviations therefrom indicate that the activity of sulfur
-492-
?
in the melt is not equal to its concentration, i.e., that this solution is
imperfect.
Fig. 173. Equilibrium composition of the gaseous phase in
system Fe2S - H2 - 11.28 versus sulfur contents in
melt (1600?C).
Therefore, the expression for the equilibrium constant should assume
the following aspect
whereupon
=
PH,S
==
PH, aS PH,1% SI "f,i Is
(viii, 78)
1g 7s == -- Ig . (VIII, 7/)
The dependence of the activity coefficient of sulfur )1's upon its con-
centrftion can be determined, for instance, by means of extrapolation of
the Ki values with respect to the point where [is] = 0. Since )(s = 1 here,
1
Ki = El according to equation (VIII,'?). Knowing the value of Ki and K,
for any contents of sulfur, it is possible to determine I's which corresponds
to the given MI
2 J 110,51
Fig. 174. Effect of tenperature and the concentration of
sulfur on its activity coefficient.
-493-
Declassified in Part - Sanitized Copy Approved for Release ? 50-Yr 2014/05/01 : CIA-RDP81-01043R003400070002-1
Declassified in Part - Sanitized Copy Approved for Release ? 50-Yr 2014/05/01 : CIA-RDP81-01043R003400070002-1
In other words, Ys represents the relationship of the tangent at any
given point of isotherm p (0) to its value for (P) = 0, whence
PH,
lgys = Ig tg a ? Ig tg cro
(VIII, Elio)
The values of Xs for various /Wand T determined in this manner are
recorded in fig. 174. As to the temperature dependence of the equilibrium
constant Ki (fig. 175) of reaction (VIII, 71), and the variation of standard
free energy .64, according to these data, their aspect is as follows
IgK1= ?
2150
A rs. a = 9840 + 6,54 T .
Whereupon, for the process
1 e
?2 = S(Fe) 7
2 94L'S)
(VIII, 81)
(VIII, 82)
(VIII, 83)
upon utilization of equation (VIII, 76) for hydrogen sulfide dissociation,
we obtain
A F,. 6 = ?31 520 + 5,27T. (VIII, 84)
It differs considerably from the equations derived on the basis of
investigations conducted without elimination of thermal diffusion Z297, L267
A ? 42 526 -I- 1 1 ,01 T ,
A F;. ? 42410 4- 10,35 T .
?494?
?
1
?
p-
?
42 40 7cvp
Fig. 175. Temperature dependence of the equilibrium constant
for reaction (VIII, 71).
Thus, experiments conducted with greater precision reveal that the
influence of temperature on the distribution of sulfur between metal and
gas is smaller (formula VIII, 84), than it was previously thought jormulas
(VIII, 85) and (VIII, 86)_2%
Reaction with Sulfurous Gas.
In the oxidizing atmosphere of an open?hearth furnace sulfur is to be
found almost entirely in the form of SO2. In view of this fact certain
investigators 2337, 4347 were inclined to place the responsibility for the
transfer of sulfur from gas to metal on, eaction
..Swo d- 02 = SO2 (30_,. (VIII, 87)
However, this is not so. Actually, by combining equation (VIII, 84)
with a similar expression
for the reaction
we find
whereupon
A F s = ? 86 380 + 17,30T
S2 (9?c -1- 02 f5.4= '3,6) ,
A rs. 87 = ? 54 860 + 12,03 T ,
?495?
Declassified in Part - Sanitized Copy Approved for Release @ 50-Yr 2014/05/01 : CIA-RDP81-01043R003400070002-1
Declassified in Part - Sanitized Copy Approved for Release ? 50-Yr 2014/05/01 : CIA-RDP81-01043R003400070002-1
Pso,
ig KS, 87 = ig
PO, as
11991
==2,63.
(VIII, 91)
Calculations made with the aid of equation (VIII, 91) disclosed that
the equilibrium constant K5187 for reaction (VIII, 87) is great at all
temperatures important for metallurgists. 4 drops with rising tempera-
ture, but even at 2000?K it amounts to 103*-w15. This points out to the con-
siderable disinclination of sulfur to pass from SO2 into metal according to
reaction (VIII, 87).
Moreover, as emphasized by Karnaukhov/357, the presence of oxygen in
the atmosphere causes oxidation of sulfides and a transfer of sulfur from
metal to gas in the form of SO2, i.e., a flow of reaction (VIII, 87) from
left
tween
(
This gives
AF = 1000 + 13,17T,
218. _
in, R. = Pso , ,
2,88.
s? 93 b as a6 7'
(Ian, 96)
(v1111 97)
It follows from equation (VIII, 97) that the equilibrium constant of
reaction (VIII, 93) is a small quantity which depends little on temperature.
This fact points to a relatively powerful solvent action of iron with regard
to SO2 which can be ascertained, for example, from the following activity
values of sulfur and oxygen for reaction (VIII, 93) at 1800?K (4293"::f0.001),
determined in the assumption that
to right. In the presence of slag, there develops in it a reaction be-
the anions of sulfur (sulfides) and the cations of trivalent iron
S2- + 6 Fe3+ -11-: 2 02- 6 Fe2+ SO (VIII, 92)
?si ; (it) ? (st (si ) 2 (346) I
(1,4=O0
P cdPri. 1,0
SO,
a S= a0 31,6
0,1
10,0
0,01
3,16
which also leads to the removal of sulfur in the form of SO2.
A substantial role (see L367) in combining sulfur with metal
played by another reaction, namely that of SO2 dissolution
S(Fe) -I- 2 0(re, = S02(90) .
It was stated in chapter VII that the interaction
10
(VIII, 93)
of s92 with
proceeds through dissociation of sulfurous gas at the interface
with their subsequent dissolution in metal. The development of
(VIII, 93) from right to left is confirmed here by the validity
root law
3
% SI =m V-P-C)-, ?
copper
into atoms
reaction
of the cube
With regard to SO2 interaction with iron, the change of the standard
free energy and the equilibrium constant for reaction (VIII, 93) may be
determined t- combining equation (VIII, 90) with an analogous equation for
the solution of oxygen in iron Z3V.
02(./c4 m=2C)(m,
AF?0.94..--55860--1,14,r.
-496-
(vizi, 94)
(wily 95)
A
Thus, in the case under consideration the equilibrium conditions are
favorable for the absorption of sulfur by metal from 802. This conclusion
is borne out by the increased sulfur contents in steel smelted with sul-
furous fuel.
In conclusion, a few wordy concerning the rate of iron interaction
with gases containing sulfur. The number of investigations confined to
this matter is scarce. Let us review the results obtained by Karamzin
1387 who studied the kinetics of sulfur absorption by solid iron from an
atmosphere of SO2 and H28 at 800-1000?C.
It transpired that the rate of the process in dealing with H2S, is
approximately twice as high as in the case of 802. In the latter instance,
there forms apparently a compact blanket of Fe0-FeS, through which diffu-
sion proceeds slower than through a layer of FeS (see part I of the
present monograph, page 406).
Further, according to Karamsin, the absorption rate of the sulfur
bearing gases grows in direct proportion to their partial pressures. This
accords with the assumption concerning the great significance of the dif-
fusion processes during the period of charge smelting. It should finally
be noted that the rate of the process develops sharply upon reaching the
melting point of the oxide-sulfide scale and the partial exposure of the
metal. In seems that in these conditions, before the formation of a pro-
tective slag layer while the scale is molten, the most intensive absorption
of sulfur by metal takes place.
-497-
Declassified in Part - Sanitized Copy Approved for Release @ 50-Yr 2014/05/01 : CIA-RDP81-01043R003400070002-1
Declassified in Part - Sanitized Copy Approved for Release ? 50-Yr 2014/05/01 : CIA-RDP81-01043R003400070002-1
Effect of Elements Dissolved in Iron.
The introduction of carbon into metal is known to cause a decrease of
sulfur solubility in it. In particular, Karamsin L.397 provides a few data
LA7' according to which carbonization of iron increases the ratio of sulfur
contents in slag to that in metal, i.e., (% S) : 5g, approximately by
4 - 5 times (see also L417).
Silicon and partially phosphorus also reduce the solubility of sulfur
in iron. Certain authors L427 believe that impurities can be broken down
into three groups in accordance with the character of their action.
To the first of these belong the elements whose affinity with sulfur
is less than that of iron. These substances, in their opinion, do not
produce any substantial influence on the solubility of sulfur in iron.
They refer Cu, Co, Ni, W, and others to this group of impurities. It
should be noted, however, that the dissociation pressure of 0u23 at high
temperatures becomes noticeably lower than that of FeS, while at tempera-
tures upwards of 135000 it is lower even than that of MnS.
The second group incorporates the metals producing sulfides more
stable than FeS. They reduce sulfur solubility. Classified here are Mn,
Al, Cr, Ti, Zr, etc.
Finally, the third group includes elements which interact powerfully
with iron forming with it various compounds. According to the authors of
this classification, their addition reduces the concentration of mfree"
iron, causing thereby a drop in sulfur solubility. This group covers Si,
C, and P.
Suggestions were even made LAR' in this connection for the computa-
tion of sulfur activity in the presence of Si, C, and PI according to
formula
S I 1% S],
s
- (Fe's')
in which gee is the concentration of ufreen iron.
Carbon and Silicon.
(VIII, 98)
The effect of Si and C on the activity of S was experimentally studied
also by way of equilibrium investigation of metal with gas, containing
H2 and 1125. It turned out L447, in particular, for carbon-saturated iron
that the activity of sulfur is double that registered in pure metal. As
to other carbon concentrations, corresponding experimental data L457 may
be found in fig. 176. It provides plots for S concentration versus C
contents for a number of temperatures and compositions of the gaseous
phase.
Knowing these relationships and the value of K1 (see equation VIII, 75)
it is easy to find the activity coefficients of .Ys for various carbon
concentrations. The relationship between Ys and Ns are shown in fig. 177.
-498-
6
.?
Moreover, a melt containing 1% S and 99% Fe was taken for the standard
state of sulfur in which as = Ns. The curve plotted in fig. 177 shows the
growth of the sulfur activity coefficient with increasing carbon contents.
A second curve is incorporated therein which illustrates an analogous in-
fluence of silicon Z327.
SimPar results were obtained by other authors 1.467, whose data are
recorded in fig. 178. In this case I's refers to the binary system Fe - S,
while )t refers to a ternary system, i.e., to Fe-C--S, or Fe-Si-S alloys.
In full harmony with the curves in fig. 177, the addition of silicon in-
creases Y more powerfully than the introduction of carbon, if the compos-
ition is recorded in atom %.
As to the influence of sulfur concentration, its growth leads to a
drop of the activity coefficient of S both in the presence and in the
absence of C and Si.
The attempts to provide a theoretical explanation for the action of
carbon and silicon additions can be broken down into two groups. One of
them considers the bonding of iron by additions to constitute the main
cause, the second holds it to be due to the reduction of the number of vacant
sites.
Fig. 176. Effect of temperature, gaseous phase composition,
and the contents of carbon in its alloy with iron upon
the concentration of sulfur in the latter.
-499-
Declassified in Part - Sanitized Copy Approved for Release @ 50-Yr 2014/05/01 : CIA-RDP81-01043R003400070002-1
Declassified in Part - Sanitized Copy Approved for Release ? 50-Yr 2014/05/01 : CIA-RDP81-01043R003400070002-1
Fig. 177. Effect of silicon and carbon concentrations on the
activity coefficient of sulfur in their alloys with
iron.
To the extent that in low carbon concentrations each atomic percent
of carbon increases if approximately by 6%, a purely chemical interpreta-
tion must assume that iron combines with the admixture into a Fe6C com-
pound. Similar and highly improbable formulas of chemical compounds must
be conjectured also for silicon 2477.
More realistic, it seems, is the other version which may be reduced
to the following z487. Carbon has a. greater affinity with iron than
sulfur. The atoms of the latter are, therefore, being driven off by carbon
from the environment of iron. From the ratio of C and Fe radii, it follows
that each C atom can be surrounded by six atoms of Fe. This means that
its appearance in the melt reduces approximately by 6 the number of the
possible vacancies to be occupied by sulfur. To put it in another way,
each atomic percent of carbon decreases the equilibrium contents of sulfur
by 6%, approximately.
0,2
211 6 8 (YMICI
Fig. 178. Effect of silicon and carbon on the activity coefficient
of sulfur in their alloys with iron.
The second interpretation departs from the fact that the number of
vacant sites to be occupied by sulfur and carbon is always equal to one
quarter of the total number of iron atoms (see chapter VII, page 428).
-500-
4,-
4
Assuming further that carbon and sulfur distribute themselves on an equal
basis among the available sites, one finds
-
1. - 5(N c+ N
(VIII, 99)
As may be seen from table 23, the above equation is in satisfactory
agreement with the experimental data b27, at least up to 3% C.
Table 23. Comparison of the observed Ys values in Fe - C - S alloys
at 1600?C with those calculated according to formula (VIII, 99).
PH2S . 103
PH2
Contents in the alloy, %
EXperiment
Calculation
2,44
1,70
0,59
1,75
1,73
2,42
2,75
0,42
2,57
2,57
2,43
0,01
(1,97
-1
-1
2,46
2,82
0,41
2,62
2,62
2,49
0,98
(1,73
1,42
1,39
2,52
1,97
0,55
1,97
1,89
2,55
3,84
0,29
4,00
5,12
4,93
2,20
1,10
1,95
2,03
5,12
4,02
0,57
4,20
6,95
Yet, the assumption of the equivalency of sulfur and carbon can hardly
be substantiated. Neither the particle sizes, nor the mode of the exist-
ence of these elements in metal, or their affinity with iron -- nothing
seems to provide any grounds for such an assumption. If the non-equivalency
of carbon and oxyten -- consisting in that the first dissolves by way of
replacement, whereas the other by way of intrusion 2'507.-- is to be consid-
ered true, then this should be even more so with respect to sulfur and
carbon inasmuch as the size of the S atom is larger than that of the 0 atom.
Furthermore, to the extent that the same basis of equality is acknowl-
edged for the distribution of C and Si L517, one should expect a similar
influence of Si and C admixtures upon the activity coefficient of sulfur.
Experiments, however, do not support this theory.
Az it transpires from figs. 177 and 178, after conversion into atomic
percents, Ys grows faster upon silicon introduction than after the addi-
tion of carbon.
Finally, it follows directly from the proposed explanation that the
sum total of carbon and sulfur atomic fractions in saturated solutions
should be equal to one fifth, i.e., Nc + Ns ve 0.2. This rule is obeyed
2:527only in a few cases, but by far not always.
The temperature and the composition of the gaseous phase affect the
-501-
Declassified in Part - Sanitized Copy Approved for Release ? 50-Yr 2014/05/01 : CIA-RDP81-01043R003400070002-1
I
51
i
1
Declassified in Part - Sanitized Copy Approved for Release ? 50-Yr 2014/05/01 : CIA-RDP81-01043R003400070002-1
total concentration of sulfur and carbon. Thus, for examples, with the
partial pressure ratio of H2 and H25 (p1 /2s : p = 2.5) being constant, the
S and C contents varied with temperature in the following manner 2327: at
12000C there was 4.35% C and 0.27% S in metal, whereas at 16000C - 5.3% C
and 0.16% S. At 1600?C and at the same concentration of carbon (5.30% C -
saturated solution) the sulfur contents in iron increases fr2m 0.16 to 0.48%
together with growing PHS : PH from 2.63-10-3 to 7.52 ? 10--/.
2 2
Thus, a correct interpretation of the effect of C and Si on the con-
tents of S in iron must be based on the non-equivalency of the sulfur and
admixture atoms both with regard to their affinity with Fe, and in respect
of their location in the solution.
Copper.
A study of the effect of copper admixtures on the activity of sulfur
in iron was conducted at a temperature of 1550?C in A1203 (99%) crucibles
which contained practically no silica at all 2537. The results obtained
are recorded in fig. 179, from which it may be seen that introduction of
Cu reduces as greater, the greater the contents of copper in the melt.
In explaining the results produced, the authors 2537proceed from the
assumption that the bond of sulfur with copper is stronger than with iron.
Hence, the introduction of copper into the melt forces the sulfur atoms to
accumulate round copper atoms. The interaction energy of sulfur with the
melt increases, while its activity declines.
The following considerations are set forth in support of the surmise
concerning the greater bond of sulfur with iron. Since Fe - Cu and Fe - FeS
binary alloys are prone to separate into two immiscible liquids, one may
assume that there appear in them isolated groups which, in the first in-
stance, are composed of copper, and in the second case, of FeS.
In the ternary system Fe - Cu - St where copper atoms attract the atoms
of sulfur and form stable bonds, these two types of groups unite with each
other. At the same time the tendency to isolation, and consequently also
to separation, augments.
Indeed, the fusibility diagram of the Fe-Cu-S system is characteristic
for its extended region of separation into two liquids, one of which is
rich in copper and sulfur, while the other manifests a high concentration
of iron.
-502-
?
?
?
Fig. 179. Equilibrium composition of gas in H2-S-Fe-Cu system versus
concentration of copper and sulfur in alloy:
1 - FeS alloy; 2 - Fe-Cu-8 alloy with Ou:S ri..1.3; 3 - same with
Ou:S
INTERACTION OF METAL WITH SLAG.
As mentioned earlier, the furnace gases which contain sulfur are
responsible for its transfer into metal. For this reason, exposed metal
is gradimily being enriched by sulfur. Slags capable of extracting sulfur
are ordinarily used to decrease its contents in metal. A distribution of
sulfur occurs thereby between two immiscible phases in conformity with
Shilov's law. It is natural that the effect of sulfur elimination becomes
a function of the temperature and the composition of metal and slag.
It has been long since established that successful removal of sulfur
from metal into slag can be achieved through the use of highly basic slags
rich in calcium oxide. An assumption was made in this connection .L47 to
the effect that desulfurization follows the reaction
FeSthvic) + CaO(51) = CaS(51) Fe0(51, .
(viii, 100)
It was initially believed that liquid Slag does not deviate too much
from ideal solution. While considering the metal phase as a dilute solu-
tion of FeS in Fe, it was customary to express the equilibrium constant
for reaction (VIII, 100) not through activities, but through concentrations
of reagents.
Thus, for instance, for slags rich in CaO and containing a small pro-
portion of FeO and CaS, its expression was recorded L267 inthe following
aspect
-503-
Declassified in Part - Sanitized Copy Approved for Release ? 50-Yr 2014/05/01 : CIA-RDP81-01043R003400070002-1
Declassified in Part - Sanitized Copy Approved for Release ? 50-Yr 2014/05/01 : CIA-RDP81-01043R003400070002-1
!i
1,
NraS Nre0
= (VIII, 101)
Ncao 17 Fes
, i '
where N = mole fractions of the corresponding components.
It appeared thereby that the temperature dependence can be satisfactor-
ily described by equation
Ig Ks
2047
By a similar method, for reactions
0,337. (VIII, 102)
FeSimet, Mnotsh Fed,si MnS,51, ,
FeSomt, ? Mg0(1, = Fe0,51, MgS(si ,
the following equations were obtained
Nm?,; NFeu
1g K; .1g ' =
Nm?0 1% FeS1
4234
- ?0,271 ,
1g K< = Iggti NFet) 7530
-0,337 .
N FySI
sAg0
(VIII, 103)
(v1111 104)
(VIII, 105)
(VIII, 106)
It followed from these that at 1600?C, KA = 3.7 ? 10-21 Kg = 2.9
nt
Ks = 4.6 ? 10-5, i.e., that the desulfurizing effect of MnO, and even more
so in the case of MgO, is s,W1 in comparison with the effect produced by
CaO.
However, with the development of the molecular theory of alags, it
became common practice to introduce miler fractions of "free" lime into the
expression for equilibrium constant Ks instead of total CO concentrations.
The methods used to calculate this quantity varied depending on the char-
acter of the assumed compounds and the degree of their dissociation.
In basic slags, in particular, the presence of almost undissociated
molecules 2Ca0 ? Si02 and 3Ca0 ? P205 was presumed to be possible.
Then
-504-
4
4
4
?
dr.
6
NCa0 !free) = NCii0 -F Nmgo Nmno -- 2 Ns,0 - 3 N
PO. ?
(VIII, 107)
Yet, the processing of the experimental data of a number of investi-
gators with the aid of these equations, according to Samarin, Schwartzmann
and Temkin AO, produced highly fluctuating and sharply divergent values
K. In this connection attempts were made to find other chemical compounds
S
for which calculations could assure the constancy of Ks.
Thus, for example, in order to review the results of 212 experiments
it became necessary to assume the existence 2:557of the following molecules:
(2CaO?Si02)2; (Ca0.8102)2; 4CaO.P205; 0a0.Fe203; 4Ca0.P2C?Cara; 0a0.111203.
The presence of such, molecules in liquid slag demands special proofs.
Moreover, in crystal chemistry the dimers of calcium ortho- and metasilicates
are totally unknown. On the contrary, the structure of these bodies
(isolated SiOi- tetrahedra and endless chains of them) would not permit
for polymerication to occur without a change of composition.
Without dwelling any longer on similar attempts, we shall remark that
a strong impulse towards the solution of the problem of sulfur equilibrium
was given through the application of the ionic theory of slags. As is well
known, equilibrium in a heterogeneous system (independently from the forms
of its components' existence in different phases) demands compliance both
with the law of mass action and with Shilov's law of distribution.
Consequently, alongside with the equation for the equilibrium constant
of the reaction of desulfurization
ar.jg 1.e
_
"FeS. Ft- 'CaO
valid also are the expressions
arcs
arcs. I .
L?
aho)
al?e0 re
In passing over to the ionic form
aCaS = acii?2+ ? ac.,0 ac.124- ? a02?.
ares art.24- ? (1.1a? ; are24-? a02-
(VIII, 108)
(VIII, 109)
and upon replacement of the activities of FeS and FeO in the metal phase,
in view of the smallness of their concentrations, by weight percentage of
S and 0, we obtain
-505-
Declassified in Part - Sanitized Copy Approved for Release ? 50-Yr 2014/05/01 : CIA-RDP81-01043R003400070002-1
Declassified in Part - Sanitized Copy Approved for Release ? 50-Yr 2014/05/01 : CIA-RDP81-01043R003400070002-1
_
? s
1 0 ?
a ?..7.2_
l% Sl
five:64-*//02?
i% 01
? 1% q)
K s =
? Ex, si ?
On the basis of the theAry of perfect ionic solutions, Samarin,
Schwartzmann, and Temkin DA/ assumed the activities to be equal to ionic
fractions in these equations, inferring that
where
Then
II
()2
11S,
= ; a?7- =
En_
_n -
fl 02? n (..so ? n ?2 n
mgo 4 "1,.0 so,
s ns2_1% 01
?
?2 nsio)1%S1?
(11 CaO MgO nre0
(VIII, 113)
(viii, 114)
(vizi, 115)
Knowing from the analytical composition of the slag the number of
moles n1 of the corresponding components (CaO, CaS, MgO, FeO, 8102) and
the weight percentage of S and 0 in metal at the moment when these two phases
were in equilibrium, they computed the values of Ks. It transpired that
the latter remain more or less constant so long as the silica contents in
slag does not exceed 20-25%. On the other hand, the values of Ls and Lo,
calculated in the same manner, begin to change appreciably at substantially
smaller concentrations of SiO2 (7-10%).
This was also established by Heinninnn L567 and other investigators
L5/7. As mentioned earlier (also see L27), this fact is due to the accept-
ance of the hypothesis of the equivalence of all the ions of the same sign,
i.e., a consequence of the premises based on the theory of perfect solu-
tions. Upon the appearance in the melt of noticeable concentrations of
silicon or aluminum-oxides and other complex anions this assumption, ob-
viously, loses its validity.
In addition to the ionic fractions, it becomes necessary to introduce
-506-
?
?
4
?
?
?
?
?
into the expressions for Ks, Ls and Lo also the coefficients of activity
y andy , which indicate the degree of deviation of the real Slag
$2- v 04-
from perfect ionic solution. Then
Ls
L0
NI.Le+ ? A'_ 71:,24- ? 7s2-
1% S1
NFe2+ ? NO2- ? T1?:2+ ? 10'2-
_
Ks--
No- ? 70-0-
? I%SL
1% 01
Ns2? ? 7s2? ? 1 nu 01
At the same time the quantity i 2 =)/ , ? Y may be, for
FeO Fe* 02-
example, determined from one of the previously recorded empirical equations
(see chapter V, page 339),in particular, on the basis of the formula pro-
posed by Samarin and Schluartzmann L587
1g02- 1,53 N510.41- ?0,17.
(VIII, 119)
As maybe seen from fig. 180, this relationship remains valid also
when instead of N 1 the sum is taken of the ionic fractions of all
SiOw"
I+
complex anions (3i0147 P0-1
4 4
in non-calcareous slags.
As concerns
A103) and when Ca0 is replaced by MnO, i.e.,
3
FeS Fe2
s2-
able constance of KSI it is possible to assume that -2( does not differ
FeS
very much from )( . Then the variations of )/ 2 with S102 may, in
FeO FeS
the first approximation, be encompassed by the same empirical equation,
in particular, by expression (VIII, 119).
, bearing in mind the consider-
It should be borne in mind, however, that the applicability of all
these empirical corrections must be limited by 30% contents of 3102
(N- ==0.9), since the changeover to acid slags is accompanied by the
Si
formation of other, more complex anions, whose influence is apparently far
more difficult to evaluate.
-507-
Declassified in Part - Sanitized Copy Approved for Release ? 50-Yr 2014/05/01 : CIA-RDP81-01043R003400070002-1
Declassified in Part - Sanitized Copy Approved for Release ? 50-Yr 2014/05/01 : CIA-RDP81-01043R003400070002-1
0,2 gy6? 9,61.414-
Fig. 180. Dependence of the activity coefficient of the oxygen
anion (Y02....) on the aggregate contents of complex anions
(INSiO4- =4 111503- NA103-) bluz
3
Let it be mentioned in conclusion that the considerable absolute
values of 1'2 and r2 have induced certain authors to assume the
FeO FeS
existence in liquid slags of molecules of oxides and of their compounds.
The considerations exposed in Chapter V seem to indicate that this fact
speaks rather in favor of a sharply expressed micro-inhomogeneity of such
melts, and of the existence in them of cybotactic regions enriched by
various sorts of ion pairs.
Upon these general remarks, it appears to be opportune to pass over
to a detailed description of the influences produced by the phase compo-
sition3 and temperature upon the processes of desulfurization.
INFLUENCE OF SLAG BASICITY
Of the four basic oxides CaO, MgO, FeO, and MnO, which are most com-
monly found in Slags, it is important to single out FeO to be subsequently
more carefully analyzed, since only FeO and eS are noticeably soluble in
iron.
By way of a quantitative characteristic of the process, let us take
the ratio of sulfur concentrations in Slag and in metal
(S)
(VIII, 120)
and call it the index of desulfurization. Since the contents of sulfur in
slag is, ordinarily, small, its weight and atomic concentrations are in
direct proportion to each other.
-508-
?
In view of this fact, the index under consideration may be represented
according to the ionic theory by the expression
N.....
? = L
sl ? s
l'e 2 ? Fe' ti"
(viii, 121)
If the concentration of Fe2+ in Slag and the temperature are given,
then LI will depend mainly on the product of the activity coefficients of
Iron and sulfur ions. The desulfurization index turns out to be inversely
proportional to this product
C,;
iFc2 1 ? 7:42?
const const
?
1-Fes
(VIII, 122)
As indicated earlier, the increase of the contents of SiO2 and other
substances (2205, A120 and so on) which yield complex anions, augments
substantially /(see, for instance, equation (VIII, 119)2. On the
other hand, the growth of the concentration of bases, for instance CaO,
must decreaseFeS strongly, i.e., increasing the desulfurization index.
Numegous tests Ake known to support the foregoing Z597, &07c 26i7,
L.E137, L.64/1L657, ZW. The data produced by Samarin and Teodorovich Z677
in tests run in an arc furnace are recorded here by way of illustration.
These results are represented in fig. 181, which shows that in spite of a
considerable spread of points," there appears to be a clearcut parallelism
here between the percentage of CO and L.
Li
90
.50 ?
JO
1,9
Fig. 181. Effect of calcium oxidercontents in Slag upon the
desulfurization index L5.
A. - Fusion in the order of growth of Ncao
-509-
Declassified in Part - Sanitized Copy Approved for Release ? 50-Yr 2014/05/01 : CIA-RDP81-01043R003400070002-1
Declassified in Part - Sanitized Copy Approved for Release ? 50-Yr 2014/05/01 : CIA-RDP81-01043R003400070002-1
A similar relationship for a variety of temperatures (1550 and 1675?C)
was obtained by other authors Z657in a high-frequency furnace with a
rotating crucible and a device for hermetic sealing, which assured the runs
being made in an atmosphere containing 85.5% N2, 2.7% CO, 1.4% CO2 and
0.4% 02. In all test series in which the metal was either oxidized or
reduced by silicon, or simultaneously carbonized, the growth of the desul-
furization index was registered with waxing CaO concentration (see, for
example, fig. 182).
Analogous observations_pertaining to the effect of CaO upon L7 were
also recorded by Pavlov AO/ who made a survey of the production experience
of one of our Southern metallurgical plants which used sulfurous coke.
From fig. 183 it may be seen that the contents of sulfur in cast-iron falls
regularly with the increase of slag basicity, or more precisely, of the
CaO-8102 ratio. In order to determine the effect of partial CaO replace-
ment by other basic oxides (MgO, MnO, and so on), the index of sulfur dis-
tribution can be more conveniently calculated from expression (VIII, 118).
4
2
N - 7 2 -
L. .
s 1%sl 81% 01 isz-
/6 7.5"C
Wee
Iii
?.5-
/met
625?e
as 1.0
Fig. 182. Desulfurization
Index L versus the degree of
slag basicity (R = CaO:Si02) and
temperature, upon reduction
through admixture of 0.2-05% of
silicon.
[AS]
11,18
Ca
40 42 /// 02
(VIII, 123)
Fig. 183. Effect of slag basicity
on the contents of sulfur in cast-
iron.
If te percentage of oxygen in metal is given, while the concentra-
tion of 0 - anions is maintained constant, then the introduction of
Mg2+ and mh2+ cations instead of Ca2+ should affect only the relationship
-510-
CI(
?
4.
of the activity coefficients. Inasmuch as the radii of the Mg2410.74
and Mn2*(0.91R) ions are :smaller than that of Ca2*(1.062), their inter-
action energy with 02- and S2- anions is greater. The emergence of these
cations in the melt will cause a decline of both activity coefficients.
Yet, since the size of the S2- anion (1.74 2) is greater than that of
02- (1.32.2),)/O2- will decrease more than Y . In other words, the
S'-
replacement of MgO, or MnO, for Ca0 may produce a drop in L.
It is difficult to say in advance how considerable this drop is.
Should one judge it by analogy with the variation of Ypes )'FeO ensu-
ing upon the admixture of 8i02, one might expect Lls to be practically con-
stant. On the contrary, should one proceed from equations (VIII, 102),
(VIII, 105) and (VIII, 106), then one could assume that there is a pos-
sible drop of Lts. Aetna-11y, by comparing expressions (VIII, 101) and (VIII,
118) we find that
N
K; = Ks Fe0 1o2
Lye 03 .._
(vIII, 124)
When the percentages of FeO in Slag and of oxygen in metal are fixed,
the magnitude of Ks depends only on the ratio of the activity coefficients.
Since equations (VIII, 102), (VIII, 105) and (VIII, 106) produce sharply
differing Kb values, the ratio )/2 / Y should be assumed to vary
0 S"-
strongly with the change from Ca? to MnO, and then to MgO.
However, the questionable accuracy of the cited equations dor not
lend confidence to the conclusion concerning the fall of Ks and Ls upon
the replacement of Ca21- cations by those of Mn2+ and mg21-. The available
published material fails to produce unambiguous indications to this effect
and speaks rather more in favor of a gm 11 variation of L.
Thus, according to Rulla and Hess L692; MgO admixtures from 4 to
6.8% decrease L. Further increase of MgO contents, though somewhat aug-
menting Lh, still causes it to remain smeller than in the absence of MgO.
However, it is true that in slags which contain 54% CO, where Si02:Ca0 =
= 0.45, Lh also increases with growing MgO concentration. But the authors
are inclined to ascribe this to the liquefying action of MgO, i.e., be-
lieving that these data do not pertain to the state of equilibrium.
Furthermore, according to Kochin 287, experimental runs with siderite
containing highly magnesian gangue, revealed that a considerable decline
in the desulfurizing capacity of slag is to be observed beginning with 18%
MgO (see also ZW ).
-511-
Declassified in Part - Sanitized Copy Approved for Release ? 50-Yr 2014/05/01 : CIA-RDP81-01043R003400070002-1
Declassified in Part - Sanitized Copy Approved for Release ? 50-Yr 2014/05/01 : CIA-RDP81-01043R003400070002-1
Yet capverse conclusions should also be registered. For example,
Pavlov Z68/ as well as Semiki_707observe that a partial replacement of
CaO by MgO does not affect desulfurization. The same transpires also from
the results obtained by other authors t717.
As a matter of fact, analysis L557 of a great number of published
data indicates that the index of sulfur distribution depends mainly on
the basicity of slag R. The latter is understood to mean the difference
between the aggregate molar concentrations of the basic (CaO, MgO, MnO)
and acid (8i02, P205, A1203, Fe203) oxides. In this conjunction FeO and
CaF2 are often considered as neutral substances.
Fig. 184 shows a straight-line which illustrates the growth of Lis
with R. for heats run with slags containing no MnO, while the points refer
to tests in which the Mn0 percentage in slag varied from 16 to 39%. As
may be seen from this diagram, the points are located close to the
straight-line, i.e., the effect of MnO is equivalent to that of CaO and
MgO.
Karmazini 397 also believes that the desulfurization index is
primarily determined by the basicity of slag R. To calculate the value of
L he deems it possible to employ the following empirical expression
(s) = 1,5 + 15R, (VIII, 125)
s 11'0 s1
where
R = 0,0178 (CaO) + 0,0141 (MnO) 0,025(Mg0)-0,333(Si02)? (VIII, 126)
?0,0282(P20)-0,0196(A1,03)? 0,0062 (Fe203)-
In the last equation the concentrations of the components are given in
percentages by weight. Formulas (VIII, 125) and (VIII, 126) are valid
according to the author's data for the R's located within -.1.0 to 1- 1.6.
They reflect the almost complete equivalence of the desulfurizing action
of CaO, MgO, and MnO.
0,0 420 030
Fig. 184. Effect of manganous oxide (from 16-39% - points) on
Lh at different basicity of slag R
Solid line is the standard relating to the slags free from MnO.
-512.
alk
?
?
?
?
It should be noted that the empirical relationship of L; versus R
which is sometimes represented in the form L557of
I
(VIII, 127) = a -I- h IN cal ,V \I ,1?0-2N - 4 Np.os -- A I 1?, 0,?
- 2N101 + bl? ,
may, with certapa allowances, as shown by Yessin, be produced from equa-
tion (VIII, 123).
Indeed, the ionic fraction of oxygen No2_ in a sufficiently basic
slag Z187 can be expressed by molar concentrations of the components
No2? --- Airco ? Ncan Nmgo N,1110 - 2 ?
? 3 Apo. ? NA% fal* N1.0) R
(VIII, 128)
approximately in the same vay as It, particularly so, if the composition
of the chemical compoundS1) is assumed to be the same in both cases.
This is due to the fact that both contents values (R and N62-) of
basic oxides are recorded with the plus sign (sources of 02- ions), where-
as the concentrations of the acid oxides (bonding 02- during the formation
of anions) appear with the minus sign. Only the concentration of mangan-
ous oxide is added to the quantity R since it remained unaccounted for in
the latter and because it also represents a source of oxygen ions.
In this manner, the following expression is derived from (VIII, 123)
I702?
Ls ---= Ks -
- ? -- (N Fe? R) a+ bR,
1%01 7,2
(nil, 129)
which, like formula (VIII, 127), includes two augends, one of which con-
tains R while the other does not.
As to the common variable factor
Ks 702
-- 1
1%01 7,2 -
it will be relatively constant only when the variation of the degree of
1) In equations (VIII, 127) and (VIII, 128) the chemical compound
composition is different for the phosphates and the aluminntes. In this
case the statement remains valid if the contents of P205 and A1203 is
small in slag.
-513-
Declassified in Part - Sanitized Copy Approved for Release ? 50-Yr 2014/05/01 : CIA-RDP81-01043R003400070002-1
Declassified in Part - Sanitized Copy Approved for Release ? 50-Yr 2014/05/01 : CIA-RDP81-01043R003400070002-1
metal oxidation, le.e, 2P7, is small, since the ratio Thr
0 Yin
basic slags changes little with their composition.
Apart from this, one should also bear in mind that the addend a in-
11.?,
corporates the ratio =EY- which can be considered as more or less con-
stant for slags containing less than 10% of acid oxides. With the slag
composition approaching that of orthosilicate, N2 will drop less rapidly
0
than it follows from equation (VIII, 128) in view of the detachment of o2
during the partial dissociation of the complex anions. In consequence
thereof, quantity a will inc/ude still another augend increasing with the
drop of basicity and partly compensating for the dimishing FeO/L%07 ratio.
An attempt was made recentlytgto determine the ratio of a CaO4Cati
in slag on the basis of a study of the equilibrium
CaS(scce H20(5.9 = CaOcs.imi+ H2S(r.s? (VIII, 130)
for which it was established that
A F? = I5650-0,87T. (VIII, 131)
According to equation (VIII, 118) this relationship is equal to
'CaO a02 K 1%sl
a a
CaS S'
S l%
oi
With a given L%07 it may serve as a measure of the desulfurizing
power of slags.
In conjunction with other thermodynamic data it was disclosed that
the ratio acas/acao augments with sulfur concentration in Slag and fails
with its growing basicity. The acas calculated from these results varies
with the contents of sulfur in slag according to the curve plotted in
fig. 185.
At the same time the value of aCaS =1 corresponds approximately to
calcium sulfide saturated slag.
-514-
?
?
4
'
4
?
3 a /6' (4/05]
Fig. 185. Dependence of calcium sulfide activity in blast-
furnace slags of varying basicity (R = CaO: 8102)
upon sulfur concentration in them.
Let it be noted in conclusion that relatively recently various 'ways
have been proposed for the treatment of liquid cast-irons by solid calcium
oxide with the addition of coke fines in special rotary kilns Z737. It
is supposed that the following reaction develops here
2 S(Fel 2 CaOcum -= 2 (aS(s04-1- SiO2 /or)),
(VIII, 132)
as a resat of which no liquid slag is formed, since calcium sulfide and
silica are being absorbed by the surface layers of lump lime.
EFFECT OF FERROUS OXIDE.
Reference is repeatedly made in the literature to the effect that
additions of ferrous o4sie reuce considerably the index of desulfuriza-
tion 4 (see 207, ,61,/, &2/, L747 and others). According to the molecu-
lar hypothesis this fact is usually explained 207 by the inverse process
of the homogeneous reaction in slag
FeSist Ca0(51, = Fe0(31, CaS131)
due to increasing FeO concentration.
(VIII, 133)
From the viewpoint of the ionic theory this reaction becomes sense-
less and changes into identity
S1 -1- (Y2-- ()2? s2?
tst) ish '
(VIII, 1331)
The reality of the (VIII, 133) process apparently can neither be sub-
stantiated by a reference to the micro-inhomogeneity of slags, i.e., to
the existence in the melt of regions rich either in highly interacting
Fe2+ and 02- ions, or in more wepkly combined Ca2+ and S2-ions.
The fact of the matter is that 01lshanskyL757 had experimentally
-515-
Declassified in Part - Sanitized Copy Approved for Release @ 50-Yr 2014/05/01 : CIA-RDP81-01043R003400070002-1
Declassified in Part - Sanitized Copy Approved for Release ? 50-Yr 2014/05/01 : CIA-RDP81-01043R003400070002-1
established a lowering of FeS solubility with the increase of CaO and
Al2O3contents in orthosilicate melts. The presence of MgO, on the other
hand, was practically-untraceable therein.
He attributes this to the fact that FeS, as a substance with a con-
siderable share of metallic bonds, is contained in "ion-electron" regions
which are rich in FeO. If the latter be partially substituted in silicate
melts by CaO, then the volume of these regions would decrease with a con-
sequent drop of FeS solubility. The absence of any effect upon admixture
of MgO is due to the closeness of the Fe2+ and Mg2+ cation radii.
Thus it is hardly possible to speak of the existence of cybotactic
groups with a predominant contents either of Fe,-,4 and 02- ions, or Ca24"
and S2- ions, at least as far as this concerns Slags containing a suf-
ficient quantity- of silica.
Furthermore, had ferrous oxide been fully identical to the basic
oxides specified earlier, its effect on L would have been relatively as
small as that resulting from the replacement of 0a0 by MgO. The effective
action of FeO is related to the previously mentioned fact that only Fe2T
cations are capable of liquidating the continuous accumulation of charges
arising during the passage of 02- anions from slag into metal.
To put it differently, the increase of FeO contents, when the basic-
ity (or Fe2? at a given 04- concentration) is fixed, augments the degree
of steel oxidation, that is leads to the growth of j07. According to
equation (VIII, 123) this results in a drop of the desulfurization index.
Analogous results may be obtained also upon application of expression
(VIII, 121). The increase of NFe2+ causes here a direct reduction of Is&
because the growth of the concentration of Fe24' ions contributes to a
greater transfer of S2- from slag to metal.
The foregoing is illustrated by the data contained in table 24, which
show L657 that the desulfurization index Lb diminishes with increasing
oxidation of metal and declining CaO:S102 ratio.
-516-
?
4
1
Table 24. Effectl) of metal reduction upon the index of desulfurization
Lb at 1600?0
Oxidized iron I Iron reduced by silicon I Iron reduced by
silicon and carbon
'.CaO
gi:T;
0,478
0,917
1,335
1,668
0,200
0,356
0,665
0,811
I
0,0705'
0.0380,
0,02871
0,0310
C;t0
Ls
CaO
Ls
s Sit),
It. -.StO,
0.381
0,227
0,0269 0,655
1,274
0.953
0,437
0,0250 1,090
7,123
1,257
2,043 0.0357 1,418
40.30
1.584
3,680 0.05381 1,823
197,0
1) It should be borne in mind that the introduction of Si and C
during reduction augments the activity coefficient of sulfur in metal
which also contributes to an increase of Lb.
Similar results are recorded by Samarin L767, who indicates that Lb
amounts to 3,4 in oxidizing Slags when the ratio of CaO to the sum of
MnO + FeO is equal to 0.88, whereas for Slags of the reduction period, when
this ratio varies from 30 to 39, the index of desulfurization is consid-
erably higher and attains 23 and even 49.
Various empirical formulas Also speak in favor of a decrease of Is;
with growing FeO concentration in Slag, for instance, L697
or Z617
L ' 1.3% Si) and Slag with
the above, do not contain a constant/f3 .
Further, it is easy to obtain from equation (VIII,
for the index of desulfurization
== (S) -L. A/ fl.
S 10/0 sl N L (1 IV
FeS %. ye) mn ? Lrts).
S
(VIII, 1W)
37=39% 8102, unlike
164) an expression
(VIII, 171)
It follows thereby that the value of L must grow together with Nmn.
As may be seen from fig. 190, the experiment L627 confirms this to
be so for diverse temperatures (from 1500 to 1900?C). Moreover, the
equilibrium constant for reaction (VIII, 147), in the first approximation,
is equal to the ratio LL3/14eSs Since Kmn > 1 and drops with the rise
of temperature, IIrdn s > 14es, while the difference (rithas - 1,Pes) diminishes
also with T. It follows therefore that the slope of the straight-lines
must increase with the lowering of temperature. The foregoing is in agree-
ment with experimental data (see fig. 190).
Finally, the results of the investigations L427 recorded in table 25
evidence that the product of gidge7by 2%'.97 for steel is approximately
double that for cast iron, and grows in both eases with temperature.
Table 25. Magnitude of the product /J ,%87 for oast-iron and steel
at different tem eratures
Temperature, 0C
1425
1480
1540
1600
1650
/Ail Os7
cast-iron steel
0.59
(1.74
D,g3
1,17
1.48
?527-
2,0
2,5
J2
Declassified in Part - Sanitized Copy Approved for Release @ 50-Yr 2014/05/01 : CIA-RDP81-01043R003400070002-1
Declassified in Part - Sanitized Copy Approved for Release ? 50-Yr 2014/05/01 : CIA-RDP81-01043R003400070002-1
These facts can also be easily explained by means of equation (VIII,
164). Indeed, it follows from the latter that
Nmn Ns -
NMn
7F [
(1 -Ns) NM
Ics 72
111S 712:eS
MuS L IS
IL
(VIII, 172)
It results therefrom that the smaller value of N N in cast-iron is
due to the presence in it of silicon and carbon which considerably increase
the activity coefficient of sulfur.
441 48 42 48 1% mni
Fig. 190. Effect of the temperature and manganese contents
in metal on the index of desulfurization L.
Independent observations L37 indicate that s for cast-iron may
(with one and same weight concentration of sulfur) be even as much as 5
times greater than in the case of steel. Furthermore, since L3 is
greater than LFes, and drops faster with growing temperature, the use of
equation (VIII, 172) helps to explain the experimentally observed increase
of the value of NiaNs product with rising temperature.
Thus the ionic theory, departing from the fact that manganese is less
noble than iron, is capable of explaining the regularities determined by
experiment which are related to its influence on the completeness of cast-
iron and steel desulfurization.
Let it be mentioned in conclusion that certain authors emphasize the
accelerating effect produced by manganese upon the process of sulfur
elimination from metal. Thus, for instance Karnaukhov Z607 believes that
manganese, while exerting no direct influence on equilibrium, catalyses
the reactions forming the intermediate compound MnS. Sch-mntically the
process is assumed to develop as follows
-528--
?
?
?
FeS(met) Mnone.t) = MnSontt) + Fetmet
MnSonet)= MnS(si)
MnS(51)+Ca0(51)=Mn0(si. + CaSish
MnO(si + Conit, = Mn(, + CO(24.0
+ CaOisi) + CaSis f)+ CO,.
(VIII, 173)
A similar point of view was expressed by Zamoruev L9371 Panfilov 2947,
Danilov /957, and other authors /96/.
Danilov, in particular, in substantiating his considerations refers to
fig. 191, from which it transpires that an increase of manganese concentra-
tion by 0.1% accelerates the rate of desulfurization, approximately, by
0.002% per hour.
v, rob
17,00,1,5
0,0041,5
431-0,J1 0,14-0,15 ,U?17
[V0/1147J
Fig. 191. Rate of sulfur removal versus manganese contents in metal.
At the present time there are no reasons whatsoever to deny either the
catalytic power of manganese, or the latterls favorable influence upon the
equilibrium distribution of sulfur between cast-iron and slag. From the
ionic theory viewpoint the summary equation of Karnadkhavis reaction
Soe?et, + 04t2i. = St2--1) + C015,4 (VIII, 174)
emphasizes only the fact that carbon is a more powerful deoxidizer of
metal than manganese. What it is then that constitutes the accelerative
effect of the latter - remains still unknown.
EFFECT OF SILICA AND ALUMINA.
As mentioned previously, the addition of SiO2 and A1203 as well as
as other oxides forming poly-atomic anions leads to an increase of ?I 2 ,
FeS
i.e., to a drop of the desulfurization index. In addition to the material
-529-
Declassified in Part - Sanitized Copy Approved for Release ? 50-Yr 2014/05/01 : CIA-RDP81-01043R003400070002-1
Declassified in Part - Sanitized Copy Approved for Release ? 50-Yr 2014/05/01 : CIA-RDP81-01043R003400070002-1
described before relative to the general influence of such oxides, let us
revise certain data pertaining to the effects produced by them individually.
Thus, the effect of silica can be seen quite Clearly-from the diagram
in fig. 192, which provides the lines of equal values of L827. The dia-
gram shows that the increase of 6102, if the relationship of Ca? and FeO is
given, causes a drop of L. According to Karmazin /397, ferrous Slags are
affected hy silica in a similar way.
Fig. 192. Desulfurization index versus Slag composition.
This fact, however, calls for additional explanation. The matter of
the fact is that the activity of ferrous oxide in FeC-Si02 Slags (in con-
trast to the more complex ones, such as, e.g., Fe0-CaO-8i02) is about
equal to its mole fraction &77, /987.
In this connection, it may seem to appear that YF , and consequently
also :Fesy
are equal to unity in this case, and that Lrdoes not change
X
upon 6102 addition. However, this is not so because for slags containing
n1 moles of FeO and n2 moles of 8102 -- i.e., consisting of n1 Fe24.,
(n1 - 2n2) 02-, and n2 SiO2 -- the mole fraction of FeO
NFv() =
111
? 1
n UI n, ns
n2
(VIII, 175)
is greater and it varies with growing 5102 less rapidly than the product
of ionic fractions
n.--2n2 ,
N F c2 ? N 02? =
Ill? n2
_530-
n2
(VIII, 176)
N
It follows directly therefrom that , 2equal to
FeS
Ve0
?
NFe2+ ? NO2-
9
efreS24 7%0 N Fe2+aPe.?NO2?
(VIII, 177)
is greater than unity and increases alongside with the contents of silica.
In other words, Ls in this ease must decrease with rising concentration of
SiO2.
Without dwelling any longer on similar data, we shall only remark
that sometimes a positive influence of S102 is also to be obeerved. Hort-
ever, this Always takes place in the absence of equilibrium and is attribut-
able to the decrease in Slag viscosity resulting from the addition of S102.
In consequence thereof, the process develops more rapidly although the
possible depth of desulfurization decreases.
IV
C 4-1 43 o 1
Al d g 1:1 I
:1:1
sg 0 ?
e- P
-,15 0 '0 0 -.4
-o
-.-1
a It -P xt
0
0
o
I
-e- -,..4 .-r-i 4) 5
9?
1--I
a
e
a VI
06. -43 43
I-4
-o- / A .
N
/
? 0 In V
-0
/ te II,
/ - cv
.../."?r-f CV bs)
./ :::0
--------
-
c?^ a
-617-
Declassified in Part - Sanitized Copy Approved for Release ? 50-Yr 2014/05/01 : CIA-RDP81-01043R003400070002-1
Declassified in Part - Sanitized Copy Approved for Release ? 50-Yr 2014/05/01 : CIA-RDP81-01043R003400070002-1
1
f,
In spite of the insufficient precision of the relationships, the
experimental materials still permit speaking of the variability of ac-
tivity coeffioients lr and lr . Moreover, one of them indubitably
0
grows with carbon concentration, whereas the other apparently drops.
ACTUAL CA
ON AND OXYGEN CONCENTRATIONS IN METAL BATH
The question of the actual carbon and oxygen contents in the bath
of open-hearth furnaces is of substantial importance. Even on the basis
of average data pertaining to the variation of their contents in the
prooess of melting, it appears to be possible to make qualitative con-
clusions as to the inter-relationships between the rate constants of the
individual stages of the process.
In their work Z1607 Stark and Chelishohev prooessed the data of 21
industrial heats in one and the same furnace. Samples of metal were
taken during the run which allowed not only the determination of the con-
tents of carbon and oxygen, but also the establishment of the tendencies
in their variations (rise or drop).
The results of their investigation are represented in fig. 236 in
the form of a solid curve above which are located points (1) correspond-
ing to the sinking tendencies of the oxygen concentration, and below (2)
those showing its tendency to rise. Thus, the recorded curve reflects
the relationship between the contents of C and 0 in steel for a stabi-
lized process of decarbonization. It is situated above the dotted
curve which corresponds to equilibrium conditions &697.
The comparison consequently shows that the actual contents of
oxygen in the bath is greater than at equilibrium
104...et [0]ev?1.
Similar results were also produoed by Stark and Filipov L1857 in
a small (8 kg) induction furnace. It became evident thereby that in
runs under slag peroxidation is somewhat less than when the metal is
'bares in which case, incidentally, the percentage of oxygen fluctuates
powerfully.
The same conclusions were reached by Samarin, Polyakov, and
Sohwartamann &797., who analyzed the data observed by a number of inves-
tigators. As may be seen from fig. 237, the actual oxygen concentra-
tions in metal (region B) are higher than those at equilibrium with
carbon (ourve A) and lower than those at equilibrium with slag (accord-
ing to different data either region C, or region A ). Analogous con-
clusions were also made by other authors L1867.
The data recorded here and numerous other observations indicate
that some one of the stages develops in the metal bath relatively
slowly and limits the process as a whole.
-618-
e
4
b001
r/
42 0,6 li.ef (%t1
Fig. 237. Relationship between 4% Fe7 and Z% C7 in metal
in the condi ions of:
A -- equilibriaa of o7 with LC7; B industrial heats; C and
A- equilibrium of metal with slag.
A form of relationship between the concentrations of C and 0,
other than (VIII, 308) considered above, may serve as an additional
confirmation of the fact that the equilibrim of reaction
C 0 = CO
(met ov.et,
fails to stabilize.
An illustration of the foregoing is provided in fig. 238 plotted by
Oyks, Maksimov, and Kaluzhsky. It displays curves reflecting the varia-
tion of the equilibrium (a) and actual ( S) values of mt = &,C7 e.07
With growing carbon contents. Fig. 238 shows that the actual value of
mt is greater than that of the state of equilibrium and that this in-.
equation increases with rising carbon concentration.
Analytic expressions were proposed for curve ( 6 ). According to
one group of investigators Z1867 its equation has the following aspect
= 0,0124 + O,050[% CIct . (11III? 318)
A linear relationship similar to the above is advanced by other
authors /1877, Z.1887
[C]f.i.c.t [Fe0]..c.t = 0,012 + O,087[% CJ
ECI [Fc0] = 0,0146 -4- 0,032 [%
(VIII, 319)
(VIII, 320)
All these equations re-emphasize the fact that boiling metal is
peroxidized, i.e., that it contains more oxygen than follows from the
conditions of equilibrium.
-619..
Declassified in Part - Sanitized Copy Approved for Release ? 50-Yr 2014/05/01 : CIA-RDP81-01043R003400070002-1
Declassified in Part - Sanitized Copy Approved for Release ? 50-Yr 2014/05/01 : CIA-RDP81-01043R003400070002-1
Which one of the transformations developing in the metal bath then
proceeds insufficiently rapidly? Is it constituted by the chemical re-
action between carbon and oxygen? Apparently not.
Fig. 238. Dependence of the product C7 Z% 07 upon C7 in
equilibrium conditions (a) an in conditions oT indus-
trial heats ( ) according to the experimental data of:
1 -- Oyks and co-authors; 2 -- Stark and Chelishohev; 3 -- Oyks;
4 -- Ageyevo
Indeed, the rate of the prooess of carbon burn-out increased hundreds
of times when other methods were used assuring more intensive oxygen sup-
ply into the metal bath (bessemerizing, blowing of metal with oxygen or
air in open-hearth or eleotric furnaces).
Consequently had the ohemical reaction been the limiting stage, then
it would have been reasonable to expect a far greater peroxidation of
metal to take place, i.e., a very sharp divergence of the curves reflect-
ing the actual and the equilibrium values of mt.
Yet this is not son This is evidenced, in particular, by the loca-
tion of curves (a) and (() in fig. 239, which refers to oxygen-blowing
of metal &897. In spite of the great quantity of oxygen absorbed by the
bath, its assimilation proceeds rapidly enough. Peroxidation (the diver-
gence of plots a and 1 ) remains approximately the same as in the case
of a normal open-hearth furnace run.
Actually, as may be seen from fig. 239, concentrations &eqfact
before, after, and during oxygen-blowing do not differ one from another
and correspond to the limit characteristic for active boil (curve a).
Many other similar data pertaining to the assimilation of oxygen are
reported in the book by Yatsunskaya and Starovioh &757.
-620-
?
?
4
r',1)
?
g 2
(70C]
Fig. 239. Relationship between L% 07 and 4% C7 in metal at
equilibrium (a) and during blowing in converter (8'):
1 -- samples taken prior to blowing; 2 -- same after blowing;
3 -- during blowing.
Analogous results were obtained by Maksimov &907 during a study of
the Bessemer process. Here (fig. 240), as earlier, curve (2) of actual
oxygen oontents is to be found above curve (1) pertaining to equilibrium
with oarbon, and below curve (3) which refers to equilibrium with slag.
Fig. 240. Z.% 07 versus Lyn c7 in the conditions of equilibrium
dissolved carbon and oxygen (1), during Bessemer blowing
and at equilibrium of metal with slag (3).
of
(2).
Further, it appears that in the case under review the magnitude of
..h.
ml grows with carbon contents slower than in the instance of open-hearth
melting. The curves for m'0 and m1
c7- = 0.3%4 At greater carbon conte1111: intersect when concentration
open-hearth metal peroxidizes
more than Bessemer metal; at lower carbon contents the process is re-
versed. According to Maksimov, this is due to the fact that the forma-
tion of CO bubbles is greatly impeded in the open-hearth furnace. This
is not so in the Bessemer converter.
-62:1-
Declassified in Part - Sanitized Copy Approved for Release ? 50-Yr 2014/05/01 : CIA-RDP81-01043R003400070002-1
Declassified in Part - Sanitized Copy Approved for Release ? 50-Yr 2014/05/01 : CIA-RDP81-01043R003400070002-1
The foregoing justifies the belief that the chemical reaction stage
is not the limiting stage of the decarbonization process.
In summing up, it may be seen that the average actual contents of
oxygen /d7
faat mayvary within the limits from equilibrium with carbon
Lb7c to equilibrium with slag MIA . In oases conducive to the forma-
tion, growth, and elimination of carbon monoxide bubbles, the concentra-
tion afact tends to approach Lo7c. Such is the ease when the contents
of carbon is small (up to O), when the quantity of the forming CO is
not large, and there is a sufficient number of centers of generation for
the formation of the bubbles.
In the case of metal blowing, the question of bubble formation does
not arise and the question of supplying them with carbon becomes more
important. Hence, the approximation of concentration afaot to ac
becomes more explicit here with increased carbon concentrations.
It should be borne in mind that the equilibrium contents of oxygen
LOTC corresponds to pressures exceeding 1 atm. As was repeatedly men-
tioned by Dobrokhotov, Andreyev, Karmazin, and other authors, the pressure
of the melt layer and the forces of surface tension must be taken into
account. In the opinion of a number of authors, total pressure in the
bubble amounts to 1.5 - 3.0 atm.
In the converse case, when the formation of CO is strongly inhibited,
the actual concentration of oxygen begins to grow and in a few special
oases it may attain the magnitude of Z0731. Nevertheless, as a rule, it
is smaller, i.e., the inequation ZO7fact < LO7
fa continues to be ob-
served. Most frequently this takes place when Z.% c7> 0.0, i.e., when
the quantity of oxygen consumed every second is sufficiently large.
In the opinion of Karmazin and Pukhnarevich 6.917, the elimination
of carbon monoxide is strongly inhibited at the initial periods of melt-
ing, when not only the contents of carbon is relatively large, but the
superheating of metal is insufficient, its viscosity is high, and the
bassing surfaoe small. In these circumstances peroxidation becomes so
extensive that it leads to red-shortness of steel.
LOCATION OF THE REACTION FRONT
As mentioned earlier, the molecules of CO do not exist in liquid
iron (see also L.1927). Consequently, reaction (VIII, 288) ?met + Suet
= COgas is not homogeneous but heterogeneous. It develops at the
boundary of metal with bubbles. In view of this, the reaotion front is
located where there are bubbles.
If these form at the furnace bottom and then make their way through
metal into the furnace atmosphere, then reaction (VIII, 288) will take
place at the hearth and within the volume of the entire bath. On the
contrary, if the bubbles emerge only at the slag-metal interface, then
the whole front of the reaction will ooncentrate there.
-622-
The possibility of bottom and surface ebulition was first substan-
tiated by Andreyev L1937 both by means of a study of the reaction of soda
solutions with aoet o acid (model), as well as by way of visual obser-
vations of boiling metal.
"iWhile conducting laboratory runs in an induction furnace with a
thin layer of slag (10 mm), he noticed the formation of bubbles at the
metal-slag interface. They are directly distinguishable by a light spot
appearing at the surface of slag.
? Apart from these relatively small bubbles, larger ones form near
the bottom and penetrate through the layer of metal and slag.
Similar observations (with the aid of binoculars) were made by him
also in the open-hearth furnace bath. Here too two types of bubbles
evolve: large - bottom ones, and small - surface bubbles. The latter
are to be registered after the charge of ore and they persist the longer,
the lower the temperature of the bath. With consumption of ore and in-
creasing temperature, the surface boil weakens and finally subsides al-
together.
Depending on the conditions of the process (composition of slag and
metal, their temperature, the presence and the nature of the non-metallic
inclusions, the condition of the hearth and batters, etc.), different
types of ebulition may therefore occur: - bottom boil only, surface boil
only, or both together.
DISTRIBUTION OF THE CONCENTRATIONS OF ELRMENTS
IN DEPTH IN THE BATH.
4
Data pertaining to the different locations of the reaction front may
be obtained not only on the basis of visual observations of the frequency
in the appearance of bubbles, their size and the energy of their rising,
but also by means of a study of the distribution of concentrations of
different elements (H, C, 0, and so on) along the height of the bath.
The experimental difficulties here are sufficiently great, consist-
ing not only in the requirement for simultaneous sampling of metal from
different levels of the bath but also in that it is necessary to prevent
subsequent interaction of the metal components in order to fix its in-
itial composition.
Investigation into the distribution of carbon along the height of
the bath of an open-hearth furnace shows that the degree of inhomogeneity
a=
is small.
Moo-g? ol; ? 100%
0,ver
(VIII, 321)
Thus, for instance, Chelishohev /1947 succeeded in establishing the
fact that the oontents of carbon in siTmples taken simultaneously from
three levels (500, 700, and 900 mm below the slag layer) differed only
Declassified in Part - Sanitized Copy Approved for Release ? 50-Yr 2014/05/01 : CIA-RDP81-01043R003400070002-1
-623-
Declassified in Part - Sanitized Copy Approved for Release ? 50-Yr 2014/05/01 : CIA-RDP81-01043R003400070002-1
by 0.01- 0.02, and rarely by 0.05%, i.e., that it is almost the same.
A similar constancy of concentration was also established for manganese.
These results are analogous to those published earlier by Karmazin
Z1957; according to whom the oontents of carbon and manganese during the
Fot run of furnace does not change in height, which fact points to in-
tensive mixing in the bath caused by the evolving gases (table 39).
Table 39
Mn and C concentrations at the upper and lower levels of the bath
Near the slag
At the depth of 1300 - 1500 mm
0,16
0.15
0,16
0,19
0.22
0.37
0,36
0,42
0,46
81,40
0,54
0,60
Mn
0,15
0,15
0,14
0,18
0,19
0,37
0,36
0.41
0,41
0,40
0,48
0,57
The minor dependence of carbon concentration on the height of the
slag layer was also noted by other authors 6.967, Z1977.
In more recent works, it became possible however not only to confirm
the small difference in carbon contents of AN in height, but also to
expose the regularity of 41 N variation in the course of melting.
Thus, for instance, gyks, Maksimov and Kaluzhsky &987, who studied
the variation of concentrations of elements in the process of melting
took samples simultaneously from three levels (100, 350, 600 mm. under
the slag-metal interface). They succeeded in establishing that at rela-
tively low temperatures carbon content in the upper layers is lower than
at the bottom of the furnace. This difference levels down with heating
and becomes inverse during the hot run of the furnace - carbon content
in the lower sections of the bath turns out to be inferior to that in
the upper layers.
A far greater difference in concentration along the cross-section of
the metal layer is to be observed with regard to oxygen &997, L2007,
987, 2017, 6.547. According to Yavoysky 2027, the degree 0r irregu -
arity Tn oxygen distribution is many times that of carbon and depends
on the period of melting. It reaches a maximum at the end of ebulition
with ore and at the beginning of the clear boil. Irregularity is greater
during this period in 350-ton furnaces than in those with a capacity of
100 and 185 tons. Towards the end of clear boiling it drops to o(. =
40-50% and becomes closer for different capacity furnaces.
Fig. 241 illustrates the change in the character of oxygen distri-
bution in depth during the transition from one melting period to another.
It turns out that maximum oxygen concentration is to be registered in
-624-
?
?
?
different plaoes - near the slag, or close to the bottom, or , finally,
in the intermediate levels. This can be seen particularly clearly from
fig. 242, where aotual and equilibrium contents of oxygen are compared.
There is no doubt that suoh oharaoter of oxygen distribution, as
well as that of carbon, is caused by the location of the reaction front.
Indeed, the presenoe of surface boiling is indicated by the fact
that in a number of oases the difference A Z07 between observed and
equilibrium oonoentrations near the slag-metal interface is smaller than
at the bottom and within the volume of the bath. This also is suggested
by the fact that the least contents of carbon is to be registered in
the layer situated immediately under slag. In brief, it transpires
from these data that the reaction takes place in the proximity of the
slag-metal interface. It is precisely here that the greatest drop in
carbon and oxygen oonoentrations and their approximation to equilibrium
values is to be observed.
In a converse case, when the contents of carbon and deviation of
oxygen concentration from equilibrium prove to be minimum, the reaction
front should be located close to or even at the bottom.
A B
D
tyin, C, %
41 EN
47
4
45
4.1
42 !
4/
a
Y=0 41.=0,41 Kvar
1 I 1i i
ci mi
1 I 1
I 48,75 1 Igni
%.fe"
, , wmteofa Zlir -I/ ;I 796'
1 Ingne
(V
Wild-M1PetegiSHE? 3 6,10
gam
opiAlorawarAllwas, loiaartreriTi
m
re
0,02.15
lJc'
22,9
Ar Will Mr
Fig. 241. Distribution of oxygen in depth of bath and change of
its concentration in a heat run in 100-ton capacity furnace during
chromium reduction process.
A -- Addition of ore and bauxite; B -- Additions of bauxite, lime and
ferromanganese; C -- Tapping of slag; D Reduction and addition of
ferroohromium; E --Melting; F Distance from the bottom, mm; G --
2 hrs 301; H -- 4 hrs 201; I -- 5 hrs 101; J -- Bottom layer; K
Time, hrs-min.
-625-
?11.1.0
Declassified in Part - Sanitized Copy Approved for Release ? 50-Yr 2014/05/01 : CIA-RDP81-01043R003400070002-1
40/0R
Declassified in Part - Sanitized Copy Approved for Release ? 50-Yr 2014/05/01 : CIA-RDP81-01043R003400070002-1
t=1.52.5C: l'I/2%:
MS Vr= e.-fi 70,141rV5'l fur nee e
ROW
4,092/
g6v/.70 Q,10 6.0 .000
A
magog1/
? 0 oil hr:Ifein furnace
ei 8888/5
IA, ----RAW
_fg/ 1 0001)
yop gip stye llsg Nal
0,0041
130
00005
087.51
/10
A
o/i 90.41
101.
t-i.a5t;e=09% ;
Ye=82%hr,agn furnace
aost 'gem
000.15 ? mos
Of,
449 850 1/59 1(100
A
0,02/ eye ?Wd111,100/? 0012
=.i '/q'; 00grf arai.ac
0,01, 101. Reg! '
./71/ Oft 190 .950
A
t=1010?1":1'-0/2%;
012 089 ow [0] 00541 i/c=4/870/hr Min furnace
A (01,,Y00,0
AS 3/0 4,70 610 7.98.958
A
Op?i ag.fa? 0059 mg, 400
(01?
/50 vou 550 Ng NO
1=100 ?;
=0.9 % ;
--.004 J/0/11r;
asin furnace
001/2
Fig. 242. Oxygen concentrations - in equilibrium with carbon
6:17 and as observed in the bath LO7H:
P'
a -- during sub-slag carbon oxidation; 6 -- during bottom
volume oxidation of oarbon; o -- during bottom boil of low-
carbon metal.
A -- Distance from the bottom, :um.
Data pertaining to the distribution of hydrogen and nitrogen serve
as an additional confirmation of the foregoing. Indeed it is reasonable
to expect that during surface boil the bubbles of carbon monoxide will
outgas the metal far less than during bottom ebulition when they are
penetrating the entire volume of the bath.
As shown by Yavoyskyls investigations &027, as well as those of
other authors, the variations in N2 and H2 contents along the height of
the bath are considerable and exceed those of oxygen concentrations.
Depending on the melting conditions, three types of distribution are to
be registered here. In one of them (A) the concentration of dissolved
gas grows from the bottom to the surface of metal, in the second (B) it
-- on the contrary -- decreases, while in the third case (C) - it changes
irregularly along the height.
-626-
4
Type A distribution is characteristic for such periods of melting
when the total oxygen contents in bath increases with time. Type B, on
the contrary, corresponds to the conditions of general de-gasification
of metal. It is probable that in the case of A there develops a surface
boil which does not assure de-gassing and does not impede the passage of
H2 from the furnace gases through slag into metal. Hence, the gradient
of concentrations peouliar to the process of H2 and N2 absorption by the
bath is here preserved.
On the other hand, the B and C type distributions are oharactteris-
tic for bottom ebulition during which the CO bubbles intensively A re-
move N2 and H2.
The fluctuation limits of the degree of distribution irregularity
for different elements in furnaces with different capacities (017are
recorded in table 40.
Table 40
Degree of Distribution Irregularity of Elements
Furnace
capacity,
m.
Irregularity, %, for elements:
0
330
1/35
1-8
0-20
1-15
6-27
5-18
7-25
3-17
4-12
20-60
1- 1 f)
7-23
i--60
3m-80
13-51
111-80
18-85
12-46
10-50
38-130
59-151
2u-110
BOTTOM AND SUBFACE BOILING
The recorded material confirms Andreyev/s observations and indicates
that there exists a possibility for the reaction front to be located either
at the metal-slag interface (surface boil), or at the interface with the
hearth bed (bottom boil) and at the surface of the emerging bubbles.
As proved by tests, the surface boil characterized by the formation
of a great number of small bubbles, develops the more intensely the lower
the temperature of metal. In this case the metal is cold, viscous, and
insuffioiently super-saturated in the bottom layers with respect to the
reaction of burn-out.
Surface boil is also facilitated by increased carbon contents and
greater oxidizing capacity of slag. It is enhanced by poor penetrability
of slag, i.e., large quantity of solid particles close to the surface
boundary.
Stark and Chelishchev L1587, 1.1597, &607are even apt to accept
the one-sided assumption that the reaction front concentrates mainly at
1 The degassing character of the boil was proved by Yavoysky also by
special tests on bath sparging by nitrogen, carbon dioxide and dry air
L2027.
Declassified in Part - Sanitized Copy Approved for Release ? 50-Yr 2014/05/01 : CIA-RDP81-01043R003400070002-1
-627-
Declassified in Part - Sanitized Copy Approved for Release ? 50-Yr 2014/05/01 : CIA-RDP81-01043R003400070002-1
the boundary of liquid phases. In their judgment, only a small propor-
tion of the oxygen which passes from slag actually penetrates deep into
the metal.
The following is cited in support of this theory.
In the first place, the absence of a strict interrelation between
the variations of carbon and oxygen concentrations in time and along the
height is emphasized. However, this fact points only to the co4lexity
of the combustion process and contributes nothing to the theory concern-
ing the location of the reaction front at the metal-slag interface.
Secondly, an assumption is made that the contents of carbon and
oxygen within the metal volume approaches equilibrium. Yet, it is not
always so, as may be seen, for example, from fig. 242.
Thirdly, the favorable influence is underlined of the slag-metal
interface on the development of the reaction. Yet, for the latter the
Influence of the other boundary, namely that of metal with the hearth
bed, is not less if not more favorable.
Finally, total reaction, according to the data of these authors,
proceeds with heat absorption and this is in perfect agreement with
the regularity of the growth of decarbonization rate with increasing
bath temperature as observed in practice. This conolusion is also
wrong, since almost any reaction, whether exothermic or endothermic,
accelerates with growing temperature.
Thus the arguments produced by Stark and Chelishchev are insuf-
ficient for the negation of bottom boil. The latter takes place most
frequently after intensive heating of metal, at moderate carbon concen-
tration and a favorable condition of the hearth bottom.
As a rule, it is accompanied also by the development of the reac-
tion in volume, i.e., at the surface of the CO bubbles. This is evi-
denced by the 'deoxidizing* effect of carbon dioxide blown through the
metal, as well as by some of the curves pertaining to oxygen distribu-
tion along the cross-section of the bath.
One of the decisive arguments corroborating the existence of bottom
ebulition is constituted by the slowing down of the decarbonization
process in proportion to the slagging of the hearth bed.
KINETICS OF EEACTION BETWEEN CARBON AND OXYGEN
DISSOLVED IN METAL
The practical impossibility of spontaneous formation of CO bubbles
in steel which does not contain sonims will be demonstrated later in
the discussion of the bubble-metal equilibrium. Here the assumption is
made that the bubbles have already emerged and the discussion is con-
fined solely to the growth of the existing bubble-nucleus.
There are grounds to believe that the intensity of the total process
of decarbonization is limited by the rates of interphase crossings at
-628-
the boundaries of metal with slag and with the bubble. Most essential
among them are the inequations referred to earlier.
10 41 > Aka'? (VIII, 322)
They bear wltmess to the fact that steel is -- on the one hand --
peroxidized with regard to carbon dissolved in it, and -- on the other
hand -- insufficiently oxidized with respect to slag. These inter-
relationships become comprehensible if one assumes that the stages
limiting the flow of the process are constituted by heterogeneous
processes at the boundaries of metal with bubble and with slag.
As to the stage of homogeneous chemical transformation
Co?ez, 0(?.b)= CO1
to which reference has been made by many investigators (for instance
&037, 42047, L.2067), its existence is highly questionable. The absence
of CO m1
o eoules in liquid steel speaks against this theory.
It was repeatedly indicated that the chemical reaction of carbon
oxidation bears a heterogeneous character as it develops at the metal-
bubble boundary, and is tightly connected with the event of carbon mon-
oxide desorption. In other words, the atoms of carbon and oxygen dis-
solved in iron form a CO molecule only at the bubble interfaoe, where-
upon it passes immediately into the gaseous phase.
The fact that CO molecules are absent in metal was disregarded for
a long time by many investigators Z.1987, L427, &067. Attempts were
therefore made by them to aocomplish a kinetic analysis of decarboniza-
tion process in the assumption of the development of a homogeneous reac-
tion of CO formation. In a number of cases, it was surmised that the
intensity of the process is limited by the stage of CO molecule diffu-
sion towards the surface of the growing bubble.
This theory will not be discussed below as totally groundless.
Greater attention is due to the theories based on the surface character
of the reaction.
CABBON OXIDATION BY AIR
According to Sohwartzmann, Samarin, and Temkin el.517, &067, the
oxidation of carbon, sulfur, silicon, and manganese oxygen has a dif-
fusion character. Its intensity is determined by the rate at which these
elements are supplied from the metal volume to its surface.
The kinetic characteristics of metal decarbonization as defined by
them are close to those of the Bessemer process. The rate constant of
the latter, calculated from induOrial data and related to the unit of
specific area, is equal to 2910 Q., whereas the constant determiped in
experimental runs in a small induction furnace amounts to 30?10 Q.
Declassified in Part - Sanitized Copy Approved for Release ? 50-Yr 2014/05/01 : CIA-RDP81-01043R003400070002-1
-629-
Declassified in Part - Sanitized Copy Approved for Release ? 50-Yr 2014/05/01 : CIA-RDP81-01043R003400070002-1
1-J
The above coincidence indicates the related character of the
processes and emphasizes only the fact that they are both heterogeneous.
In other words, the formation of CO molecules occurs here not in the
volume of metal but at its interface with gas. This surface process
requires the simultaneous presence of carbon and oxygen atoms at the
interphase boundary. Moreover, direct interaction of carbon particles
with the oxygen molecule is hardly probable. The solid bonds between
the atoms of the latter must be considerably weakened beforehand. This
takes place during oham-adsorption of oxygen on iron.
In connection with the foregoing, it seems to be doubtful indeed
that there should not appear an extremely thin layer of liquid iron ox-
ides at the surface of the metal. It must isolate iron to some degree
or other from immediate contact with gas. Neither the oiroulation of
metal, nor -- and even less so -- an attempt to clear its surface by
means of a rod, can prevent it from being practically immediately recoated
by a fresh film of iron oxides1. The same, apparently, occurs during
blowing.
These circumstances, it seems, draw the process in question closer
to that of carbon combustion in open-hearth furnace. Even the equation
proposed by Sohwartzmann, Samarin, and Temkin coincides in its form with
the one derived in 1901 for the rate of decarbonization which was sub-
sequently confirmed by the data resulting from Amelting by ore process
carried out at the Nadezhdinsky plant Z077.
It is true that in the case of air oxidation the integumentary layer
of "slag" is very thin. Local fractures of the oxide film and direct
evolution of CO into the furnace atmosphere are possible. However, re-
gardless of this, the C and 0 atoms present at the surface are the par-
ticipants in the reaction.
If one assumes the rate of oxygen chem-adsorption to be very con-
siderable, then the process may be interpreted as that of reduction of
the Fex07 ronomoleoular 'protective 1' film by the atoms of carbon. Ao-
tunny, experiments show L2087that the rate of interaction of liquid
oast-iron with molten iron oxides (FeO and Fe304) conforms to the equa-
tion of the first order. It is true though that these measurements can-
not be considered as reliable.
Be that as it may, the essential fact here is still that in the
surface reaction zone the concentration of oxygen is many times higher
than that of carbon. Hence, an equation of the first order may be ob-
tained not only because of the insufficient rate of carbon diffusion,
but also in consequence of the deceleration of the chemical act itself
(VIII, 2881)
?(met.surf) 4. ?(met. surf) = CO(gas)
Indeed, the reaction velocity v will then be determined solely by
carbon sinoe the quantity (0)sur is praotically oonstant
f.
See also Part I, page 395.
(VIII, 323)
kicConsl
Moreover a number of contradict the diffusion character of
the process of oxidation by air of carbon contained in metal.
Listed among them should be the high coefficient of carbon diffu-
sion in iron, the negligible irregularity of carbon distribution in the
bath of open-hearth furnace, and the weak dependence of the rate of ox-
idation on the concattration of carbon. It is not proportional to the
value of Z.07.
Experiments with air and oxygen blowing of metal show L1757 that
the rate changes relatively slowly with carbon contents, for example, in
proportion to log Zgor according to a more complex law (fig. 243).
Sohwarzmann also seems to be inclined to accept this theory at the pre-
sent time.
Fig. 243. Decarbonization velocity versus carbon contents in
metal.
The weak dependence of the reaction velooity v upon z.c7 is apparently
attributable to the fact that carbon constitutes a strong surface-active
component in iron. A proportionality between the values of LC7surf and
LC7exists only when the percentages of carbon in the volume are small.
As a rule, the value of zp7surf. changes far less rapidly than that of
Z?7 at medium volume concentrations of carbon.
Thus, experiments on carbon oxidation by air and oxygen indicate
that this process has no diffusion character and is, apparently, deter-
minable by the progress of heterogeneous reaction (VIII, 2889.
The law temperature dependenct of the velocity of oxidation calls
for additional experimental proofs'.
1 According to Z087, the rate of liquid oast-iron reaction with iron
oxides is very sensitive to temperature (activation energy 37,000 -
43,000 oal/Mol). These data also call for verification (variable inter-
face area between the phases, rise of the oxide temperature during heat-
ing, etc.).
-631-
-630..
Declassified in Part - Sanitized Copy Approved for Release ? 50-Yr 2014/05/01 : CIA-RDP81-01043R003400070002-1
Declassified in Part - Sanitized Copy Approved for Release ? 50-Yr 2014/05/01 : CIA-RDP81-01043R003400070002-1
This is also required because the rise of temperature in induction
furnaces is achieved by increasing strength of current. This, in turn,
provokes intensive stirring of metal, i.e., facilitating carbon diffu-
sion. Hence, even if the intensity of the process were limited by dif-
fusion, it still would be reasonable to expect a fairly strong accelera-
tion of interaction with growing temperature (and, consequently, also
of the bath mixing).
ON THE POSSIBILITY OF OTHER REACTIONS
According to Schwa.rtzmann, the burn-out of carbon contained in metal
under the influence of oxidizing slags is subordinated to the same rela-
tionships as those governing its oxidation in air. In both oases, a pro-
portionality is observed between the velocity of the process and the con-
centration of carbon. Moreover the rate constants also prove to be ap-
proaching eaoh other. These circumstances induced him to assume again
that the limiting process is constituted by carbon diffusion towards the
reaction front.
Observations of carbon monoxide bubbles indicate that the reaction
front is situated within the metal near the hearth bottom and not at the
interface with slag. By what means then can rapid supply of oxygen be
assured there if it is known that the coefficient of oxygen diffusion is
appreciably lower than that of carbon diffusion? A very dubious surmise
is furnished, by way of an answer to this question, to the effect that
carbon monoxide penetrates from slag to the hearth through the capillar-
ies of the refractory lining.
This prooess could hardly be faster than oxygen diffusion in metal,
powerfully mixed by evolving CO.
The fact is also disregarded that the passage of 02- and Fe2+ ions
from slag to steel is a relatively slow process and it is assumed that ,
the following reactions develop near the bottom
Fe0(51)+C(Act) = Fe(rnav)-1-00(.9..st
CO 4-Fe0 =.Fe -*CC12.(9.4),
CO(, 04.x)
CO2. (5,0 + C(Hiet) 2 CO(9vse .
(VIII, 2881)
(VIII, 324)
(VIII, 325)
Analogous conjectures pertaining to the pattern of the process de-
veloping at the bubble surface were also expressed before Z097. More-
overt, it was assumed that along with reaction (VIII, 288)
C 0 ? CO
%et (m.15 (bidi
1 Major errors were made in study p097. Thus, for instance, it is as-
sumed that by dividing the volume o a gram-molecule of gas by Avogadro's
number it is possible to determine the size of an individual particle.
As a result of such erroneous calculations, an Wicredibly large radius of
CO molecules was produced which is equal to 38 A. It is obvious that
postulates based on such foundations are groundless and require no dis-
oussion.
4.
there develop two other reactions according to equations
C016.4 + 0 (met =
CO2 +Comm = 2 COtb,.4) .
(VIII, 3241)
(VIII, 325)
However, such assumptions seem hardly justified. The slow progress
of reaction (VIII, 325), which hampers the study of equilibrium between
liquid steel and gaseous CO2, is well known. At any rate, the velocity
of this interaction is considerably lower than that of reactions (VIII,
324) and (VIII, 324').
But one should expect then that the bubbles which form in the bath
oontain almost pure carbon dioxide. This, however, is not the case. The
presence of an appreciable content of CO2 in gas given off by the bath
is caused by another easily developing process, namely, by the reduction
of ferrous oxide in slag by carbon monoxide.
The flaw of reactions (VIII, 324) and (VIII, 324') is also impeded
by condltions unfavorable for their equilibrium with metal. Actually,
at 1600 C the gaseous phase at equilibrium may contain only 1% CO2 when
the oonoentrations of carbon are inferior to 0.2%. At higher carbon
contents, the possible concentration of CO2 becomes even maaller.
The experiments conducted by Andreyev and Yavoysky with pure carbon
dioxide blowing through metal are also in conflict with the noticeable
development of the additionakreactions advanced in Schwartzmann's study
and in work Z067. About 3 at' 002 were introduced into an electric fur-
nace with 10-tcn capacity. The result was that the degree of metal oxi-
dation not only failed to grow, but actually dropped somewhat (on the
average of 0.015% FeO) while 0033 kg of oxygen were released from the
bath.
Thus blowing with pure carbon dioxide does not contribute to the
development of reaction (VIII, 325) from left to right and of interac-
tions (VIII, 324) and (VIII, 324') from right to left. It increases only
the area of the metal-bubble interface assuring thereby a broader front
for the progress of reaction (VIII, 288).
The foregoing as a whole permits one to think that the principal
role in the process of metal decarbonization is played by interaction
(VIII, 288). As to reactions (VIII, 325), (VIII, 324) and
these develop apparently on a very limited scale.
ON QUANTITATIVE RELATIONSHIPS
(VIII, 3241),
Many attempts have been made to determine the analytic dependence
of the rate of carbon burn-out upon the principal conditions of the
process development (Andreyev, Maksimov, Chuyko, Loerber, Larsen, and
others). Yet they oan hardly be considered successful. The fact is that
the process under consideration is extremely complex from the point of
view of kinetics. Its intensity is limited not by one but by several
stages. Moreover, not all of them are defined with due conclusiveness.
-63a..
Declassified in Part - Sanitized Copy Approved for Release ? 50-Yr 2014/05/01 : CIA-RDP81-01043R003400070002-1
Declassified in Part - Sanitized Copy Approved for Release ? 50-Yr 2014/05/01 : CIA-RDP81-01043R003400070002-1
?
On the basis of the afore-said, it is possible to cite the follow-
ing stages which inhibit carbon burn-out. First of all, Ws is caned
by diffusion of ions within slag, then by the passage of 0'- and Fe4
from slag to metal and further by the heterogeneous reaction between
carbon and oxygen.
Along with the mentioned stages, it seems to be necessary to note
the slaw diffusion of oxygen towards the reaction front inside the metal.
This is evident from the fact of the very uneven distribution of oxygen
along the height of the bath. In addition to this, one should also con-
sider another fact, namely that the coefficient of oxygen diffusion is
appreciably lower than that for carbon.
Finally, a very important stage which could bring the process as a
whole to a complete standstill is constituted by the event of bubble
formation to which we shall duly refer later on.
It is obviously difficult to embrace in one equation such a large
number of factors, the more so because their number may vary depending
on concrete conditions. By virtue of this fact, all efforts to derive
such a relationship were accompanied by certain simplifications of the
problem. The surface character of the carbon-oxygen reaction wr more
oilln disregarded, as well as the slaw-down in the passage by 0 and
Fe ions of the metal-slag boundary.
In view of this, instead of surface carbon and oxygen conoentrations,
volume concentrations of the same were introduced into velocity equations,
i.e., zp7
and Z07 were considered to be equal respectively to
q7 and 611/. On the other hand, the equilibrium contents of oxygen with
ag was substituted for its actual contents at the slag boundary, where-
by LO7 was identified with 6)7 . All this makes the pro- ,\
fact surf. slag
posed equations so approximative and singular that there is hardly any
sense in dwelling on them in greater detail.
In conclusion, we shall consider only one regularity observed in
the tests. From fig. 238, it follows that with diminishing carbon con-
centration in metal the product re, pertaining to the factual contents
of 0 and C in metal, tends to equilibrium.
At first glance, this appears to be very strange since this creates
the impression that the process reaches equilibrium (accelerates) more
rapidly when the concentration of one of the reacting substances is
dropping. Nevertheless, this is not so, and the observed regularity can
be explained rather easily.
With the conditions established, the velocity of the decarboniza-
tion process as a whole is equal to the intensity of each of the suc-
cessive stages and, in particular, that of the reaction between carbon
and oxygen
V = V., = k1 [Cl [.01 k2PCO. (huh) ?
-634-
(VIII, 326)
?
4
At small contents of carbon and oxygen heir surface concentrations
become proportional to volume concentrations'
V = kik' [C] [01# k2Pc0 (644.1v ?
(VIII, 327)
By replacing pco in the expression for equilibrium conditions, we
obtain
whereupon
= k ([C] [C][0]..v.zo ,
= ICI [Olfazt_
+1.
[C) (Olp k IC] [0] p
(VIII, 328)
(VIII, 329)
Whenever the concentration of carbon shows a tendency to vanish, the
reaction rate also approaches zero. Nevertheless, the product
[C1[O]i=m
still continues to remain constant according to the law of mass action.
Then the experimentally observed regularity results directly from equa-
tion (VIII, 329)
I
/71 ICI - ?
0
EQUILIBRIUM OF BUBBLE AND METAL
PURE CARBON MONOXIDE
(VIII, 330)
The equilibrium of carbon monoxide with liquid steel was discussed
earlier for the case when the phase boundary is flat. The matter be-
comes somewhat more complicated when the CO bubble is in equilibrium
with metal. In this case, its dimensions and its location within the
bath should be taken into account.
Actually the pressure of carbon monoxide inside the bubble, as men-
tioned repeatedly by a number of authors, in particular by Dobrokhotov
and Andreyev, is determined by equation
wherein Pa_
hi and ri
2 a
PCO = Pa + 4(1 )1(51 ) h(re) itmet ?r
(VIII, 331)
atmospheric pressure;
height of the column and specific gravity of slag and
steel;
surface tension; r = radius of bubble.
1 This apparently explains the fact that equation (VIII,
agreeing with experimental values when the reaction rates
L2107, loses its validity when the contents of carbon are
-635-
327) while
are small
elevated /1757.
Declassified in Part - Sanitized Copy Approved for Release ? 50-Yr 2014/05/01 : CIA-RDP81-01043R003400070002-1
Declassified in Part - Sanitized Copy Approved for Release ? 50-Yr 2014/05/01 : CIA-RDP81-01043R003400070002-1
By combining the first three addends, we have
Pco= Po+ ?r ?
(VIII, 332)
Formula (VIII, 632) permits the calculation of the CO pressure in
a bubble which has a definite size and is located at any given level in
the bath. The knowledge of poo value is important for the clarification
of the question concerning the thermodynamic stability of the bubble.
As is well known, the latter depends on the sign of 41 F, while for a
general case
F - RT (In n a ? In K3)
VIII, 333)
When the aotivities of carbon and oxygen are given, equation (VIII,
333) assumes a more simplified aspect
A F RT (In Pco.
(VIII, 334)
Whereupon it follows that the bubble and metal reach a state of
equilibrium when
In pco. r In Pco...
If the phase interface is flat, the equilibrium
tion (VIII, 288) may be recorded as follows
Pcno.-
1 PCO.r ?
According to equation (VIII, 334) this means that AF > 0 and the
bubble is unstable. Its size will diminish as a result of the passage
of gaseous CO into metal in the form of C and 0 atoms as per reaction
(VIII, 288). The latter must proceed from right to left which will
cause a still greater decrease of r, the growth of AF > 0, and, in the
final count, this will lead to the solution in metal of the entire avail-
able carbon monoxide, resulting in the disappearance of the bubble.
On the contrary, if the bubble's size is greater than critical, i.e.,
if r > ror' then equation (VIII, 338) would be upset because in p CO,r
would become smaller than in p? . Since in this case A F < 0, a
CO,r
spontaneous progress of reaction (VIII, 288) from left to right becomes
possible. The bubble radius will grow, assuring the successful develop-
ment of decarbonization.
The foregoing is illustrated in fig. 244. Here, for different r,
the values are given of pro which correspond to mechanical equilibrium
as per formula (VIII, 3327 and to chemical stability according to formu-
la (VIII, 337). The point of intersection of these curves corresponds
to the critical parameters of the bubble.
Fig. 244. Effect of the bubble site on poo, corresponding to
mechanical (poo,r) and chemical (40,r) equilibrium
with metal.
-637-
Declassified in Part - Sanitized Copy Approved for Release ? 50-Yr 2014/05/01 : CIA-RDP81-01043R003400070002-1
Declassified in Part - Sanitized Copy Approved for Release ? 50-Yr 2014/05/01 : CIA-RDP81-01043R003400070002-1
MIXTURE OF TWO GASES
On the basis of somewhat different considerations, it is possible
to come to similar conclusions with regard to the existence, under given
conditions, of bubbles of a stable size.
The change of free energy during the formation of a bubble is equal
to its increment caused by the passage of particles from one phase into
another (from chemical potentials of C and 0 to the potential of CO),
plus the work needed for the formation of a new surface
A F = (
&co- !LC ? 1L0) a + a le, .
Here the addend 41Tr2 is assumed to be equal to
(VIII, 339)
an2/.
So long as the difference between the chemical potentials is posi-
tive, the magnitude of A F > 0, and the probability of the formation
of a bubble (or the number of such bubbles)
A P
N = A e-rf
(VIII, 340)
will be the lesser, the greater the number of CO particles in it, i.e.,
the greater the value of n.
In the other hand, if }x. 44Le -A" AL0, then LiF will initially in-
crease with growing n, but will begin to fall later (fig. 245). The
maximum of 41 F corresponds to the size of bubble which can also be
called critical. Indeed, its further growth could proceed spontaneously
since the growth of n in excess of nor causes a decrease of A F.
Recently Nesis and Frenkel Z:2117, proceeding along these lines, ex-
amined a more complex problem when the bubble is formed by two gases,
namely by vapors of the solvent and the dissolved substance. Having
designated the number of molecules of the first and second component in
the bubble by n! and nll, and their chemical potentials in both phases
by jut1, ',1Li', IA., and i.4.11 they composed an expression for the change
/ 2 2
of free energy
AF - 61:2 - n' + n" + 4
(VIII, 341)
Assuming that the liquid composition remains constant, while replac-
ing the chemical potentials of the components in the bubble by means of
equations
we find
n'
}L2 V2,0 kT In
n' n"
n
112= 1L2,0 kTin
n' H'p
-638-
(VIII, 342)
(VIII, 343)
F = (
o !L't) n' 4 (!,72.? - ) n" + kT (n' In
n" In e )4- tive3 .
n' n'
By equating further
47t r2 a a (n' rzn)1
and designating
we obtain
n'
n' a'
p...2. 0 ? 111:I == A 7
13,2. 0 B,
A F = - A n' - B n" -V (n' n"
kT In - n" In .
F1' 1- it' n"
(VIII, 344)
(VIII, 345)
(VIII, 346)
(VIII, 347)
(VIII, 348)
This formula is represented in fig. 246 by a surface with a double
curvature. Its plane projection may be composed of three fields. In
the conditions of the first (I), where
and
F >0
On' (VIII, 349)
the bubble is unstable and dissolves in the liquid phase.
On the contrary, for field III, where
F 900. Once a
bubble is formed here, it will always be leaving a nucleus behind upon
take-off. This removes the obstacles referred to above to the passage
of carbon and oxygen atoms into the gaseous phase.
Moreover, the solid surface not wetted by metal (0.> 900) does not
only preserve the nuclei left behind by bubbles, but also facilitates
the emergence of new ones. This is due to the fact that the curvature
radius of the lenticular bubbles may attain considerable values even if
the volume of gas is very small. Consequently, in this case the capil-
lary pressure decreases. In other words, a far smaller amount of mole-
cules may produce a nucleus of critical size here than in the case of a
spherical bubble. All this contributes substantially to the process of
nuclei formation.
In addition to this, the selective adsorption of gas molpules on
the surface of the solid body not moistened by liquid (0 > 90 ) will also
facilitate the emergence of bubbles. It is especially strong for an un-
even contact surface with a highly broken relie:.
Indeed, the change of free energy 4 F during the formation of a
gaseous phase nucleus upon a solid surface
AF=1/3(a S. S --a25Sjzs) ? (VIII, 359)
s9
Assuming as a first approximation that the surface of the nucleus
is a section of the sphere, upon introduction of angle e (see equation
/VIII, 354/) we obtain
0
AF? S (1 + cos? ctg2 --2).
3 R-9 -119
(VIII, 360)
From this expression it transpires that with the growth of angle e,
i.e., with increasing adhesion between gas and the solid body, the
parenthesized quantity will diminish. Hence, if the surface area Si_g
is fixed, the variation of free energy Zi.F will drop. This implies
an increasing probability of heterophase fluctuation assuring the forma-
tion of a stable bubble. The reduction of surface tension of metal
should lead to the same result.
A different picture arises when the solid body is wetted by liquid
(0 4: 90?). In this case there appear relatively small almost spherical
bubbles which leave no nuclei on the surface upon take-off. This is to
be attributed to the fact that for e 4: 90? the work needed for the
breakaway of the entire bubble is smaller than that required for its
partial separation. In addition to this, the low effectiveness of such
smooth surfaces is also due to the considerable curvature of the par-
ticles which increases substantially the equilibrium pressure n
-646-
Declassified in Part - Sanitized Copy Approved for Release
?
?
EFFECT OF WETTING
Let us review a few experimental data which illustrate the fore-
going. First of all, it was repeatedly established that the volume V
of the separating bubble in greater, the less the solid body moistened
by liquid. In other words the value of V grows together with the rim
angle 9.
a
Fig. 250. Diagrams of photographs of bubbles with different
rim angles 43 before separation from the electrode:
a -- e 18o b 9 = 75o o -- 9 = 97?(magnified 6
times).
Fig. 250 reproduces photographs &21.7 of bubbles before take-off
from an electrode for 9 = 18, 75 and 97 . As may be gathered from the
diagram, their size increases with A.
The conditions and the peculiarities of the boil of liquid heated
from underneath a horizontal surfaoe was studied with the aid of films
:
2227. Fig. 251 confirms that, depending on the wettability of the
lining, the shape and the site of the bubbles change as described above*
It was established that the separation of the first type of bubble
(0 > 900) occurs as a result of necking-off while a ready nucleus re-
mains on the surface of the solid body. This was not observed in the
case of moistened surfaces, i.e., when the angle 8 4:: 900 &237.
It is interesting to note that the growth of bubbles on a heated
solid surface proceeds faster than after take-off. This may be explained
by lesser superheating (and, consequently, also supersaturation) of the
liquid volume as comparee with the surface through which heat is being
transmitted.
Similar observations were made by Andreyev Z1937, L2247, t2257 con-
cerning the existence of three types of bubbles -- as illustrated in
fig. 251 -- depending on the wettability of the underlying surface.
Mile carefully pouring off the aqueous solutions of acetic acid and soda,
he established, in addition to the above, that in new glass vessels with
well-fused walls no evolution of gas was to be registered. On the other
hand, upon introduction of coarse partioles or lumps not wettable by
the solutions -- as well as in the case when the walls of the vessel
were eroded -- a powerful evolution of CO2 was observed. Gassing de-
veloped at the surface of the solid body.
50-Yr 2014/05/01: CIA-RDP81-01043R003400070002-1
-647.
\it
Declassified in Part - Sanitized Copy Approved for Release ? 50-Yr 2014/05/01 : CIA-RDP81-01043R003400070002-1
-.1??=30.111M,T11.
'//e/
If .///
Fig. 251. Dependence of the vapor bubble on the mettability
of the solid surface.
I -- not wettable; II -- poorly wettable; III -- wettable.
It is important to emphasize two facts here - first, intensification
of ebulition on non-wettable surfaces (0 > 900), the causes of which
were considered above, and second, facilitation of bubble formation by
coarse bodies which is to be discussed presently.
As is known, the intensity of boiling of open-hearth baths depends
substantially on the condition of the hearth bed, and primarily on the
degree to which it is slagged /1567.
Of course, the drop of the carbon burn-out rate in proportion to the
impregnation of the magnesite bed with slag is a complex phenomenon (a
change occurs in the composition, the porosity of metal, etc.). Never-
theless, the fact is noteworthy that the wettability of the hearth by
metal increases simultaneously and parallelly with this.
Similar observations were made by Sapiro '.226.7*, L.2277, Levin L2287
and by other investigators. According to the ormer, the rim angle e,
for a bead of metal (0.7-1.14% C and 13.0-0025% Mn) placed on a slagged
and a clean magnesite body amounted, respectively, to
o o o o
before slagging . . . . 128o 118o 125o 108o
after slagging ? ? ? ? 63 84 82 69
i.e., it dropped considerably during slagging.
Analogous results were produced by Leving for metals of different
composition. He found, for example, that the average value of angle 8'
for Armco-iron on unslagged magnesite comprised 126'. The value of 0'
dropped to 104? after the lining was covered with ore or oxidizing slag,
or even with reduction period slags.
Andreyev furthermore believes that the intensive bubbling develop-
ing around a metal rod dipped into the bath is to be attributed to the
fact that it is poorly wetted by steel. This is possibly due to the ab-
sence of a direct contact between the atoms of Fe caused by the presence
of scale.
-648-
4
?
The boil subsides when the rod becomes hot.
Thus the conditions for the formation of bubbles on flat poorly
wetted surfaces are more favorable than in the volume of metal. Apart
from this, an exceptionally high influence on the probability of bubble
formation is produced by the relief of the solid surface.
EFFECT OF COARSENESS
How important is the role of porous and coarse surfaces for the
process of decarbonization may be gathered, for example, from the follow-
ing tests &297. If during smelting of steel in a high-frequency induc-
tion furnace, through increase of power, a circulation is produced at
which liquid slag isolates the metal from the walls of the crucible, then
the evolution of gas slows down abruptly while the contents of carbon may
become tens or hundreds of times greater than equilibrium.
Upon re-establishment of contact between steel and the refractory
lining, resulting either from a reduction of the power input, or through
destruction of the separating slag layer with the point of a quartz rod,
intensive decarbonization may again be achieved.
The importanoe of the relief of the hearth bottom consists in that
its coarseness changes the angle of wetting, while porosity is respon-
sible for the preservation of the gaseous phase nuclei.
While speaking of coarseness, it seems worthwhile to draw a line of
distinotion between the macroscopic rim angle 0 and the micro-rim angle
Bo. There exists the following relationship between them L2307 which was
theoretically substantiated by Deryagin /2317
cos0==kcos0 =. --cm?0*
S
(VIII, 361)
Here S and So are the true and the apparent surfaces of the
solid body;
k - is the coefficient of coarseness.
Sometimes use is made of expression
1
k ?
cos ";$
(VIII, 362)
in which cos ,8 is the mean value of the cosine of the steepness of the
micro-relief.
Deryagin also specified the limits of the applicability of relation-
ship 361). It is valid when the height and the distanee between
the neighboring ridges of the micro-relief are small in comparison with
the capillary constant and the radii of the meniscus macro-steepness near
the walls. It is important too that the sphere of molecular action
should possess a radius smaller than that of the micro-relief steepness,
and that inequation k cos 904(.1 be observed.
-649-
Declassified in Part - Sanitized Copy Approved for Release ? 50-Yr 2014/05/01 : CIA-RDP81-01043R003400070002-1
Declassified in Part - Sanitized Copy Approved for Release ? 50-Yr 2014/05/01 : CIA-RDP81-01043R003400070002-1
It follows from equation (VIII, 361) that the rim angle increases
with growing coarseness, i.e., that wettability decreases. This, in turn,
facilitates the formation of bubbles.
The relationship between the rim angle and the coarseness of the
lining has been confirmed experimentally by a number of authors 11937,
/12287, Z.2327, 12337 both for low and for high temperatures. Table 42,
by may of illustration, incorporates the data produced by Sapiro 12267
for steel.
Table 42
Dependence of the rim angle on the coarseness of the surface.
Contained in steel, %
Rim angle 9, with magnesite surface
Mn
coarse
smooth
1,14
0,13
150--I53?
139-135?
0,15
0,50
162-17(1?
1244
0,16
0,50
151-157?
I21-126a
It should finally be noted that in a number of cases the sharp ribs
(ridges) of the coarse surface may appreciably reduce supersaturation
required for the formation of bubble nuclei &937, 1227, Z:2277.
POBES WETTED BY LIQUID METAL
When the liquid is in contact with a solid body the pores of which
contain at least a mall quantity of gas, the formation of bubbles is
greatly facilitated. When wettability is perfect, the metal surface will
not only prove to be concave but its curvature radius r will be equal to
that of the pore rn (fig. 252). In other words, the gas contained within
the pore volume constitutes a nucleus of a bubble which will be stable
if rn r
or
Thus, in case of perfect wettability of the surface by metal, only
those pores will become active whose radii are greater or equal to the
radius critical for the bubble. This, of course, refers only to the pores
not entirely filled with steel. If, for some reason or other, there is
no gas in them at all, then such pores, regardless of their sizes, will
remain inactive.
Conditions change gamewhat when the lining is wetted by metal im-
perfectly (0 9 4: 90 ). In this case
r
n.xp
(VIII, 363)
i.e., the critical size of the pores increases with declining wettability
(with the decrease of angle e).
?650.
* 1 0
? 4
I
The maximum pressures of gas in the bubble take place apparently,
until it still remains in the pore volume. After its emergence into the
bath volume, the conditions of growth become more favorable.
?
Fig. 252. Diagram of a pore partially Fig. 253. Diagram of the growth
filled with the liquid wet- of a bubble in a pore
ting the solid body. wetted by liquid.
Moreover, the take-off of the bubble will occur at the top of the
pore (fig. 25$). However, unlike the situation which develops in the
case of a smooth surface where a breakaway at 6 900 leads to the liqui-
dation of the nucleus, here there will remain a small amount of gas in
the pore which will facilitate the subsequent formation of bubbles.
PORES NOT WETTED BY LIQUID METAL
In those calms when a porous solid body is not wetted by liquid
metal, the surface of metal in the pore will be bulging towards the gas-
eous phase (0 > 900). Such curvature assures better conditions for gas
evolution than those characteristic for a smooth surfaoe.
When 6 900 the gas pressure in the pore
2a
Pr = Pext.?r
(VIII, 364)
is smaller than external pressure (Pext) For this reason in a super-
saturated system there will develop a spontaneous growth of the gas
volume in the pore. The liquid will be forced out from the capillary
by gas.
Upon reaching the opening of the pore (fig. 254) the contact sur-
face will become concave (III) instead of convex (I, II) with respect
to gas, while retaining the same rim angle 6. Moreover, according to
equation (VIII, 364) the changeover from position (I) to position (II)
may occur under the influence of capillary forces even when gas pres-
sure in the pore is inferior to external pressure. This is not the case
when position (II) changes into position (III), since here the inter-
face is flat at the initial moment.
Evidently, in this case, the growth of the bubble and the passage
of the contact surface into a position with a stable rim angle (6 > 900)
Declassified in Part - Sanitized Copy Approved for Release ? 50-Yr 2014/05/01 : CIA-RDP81-01043R003400070002-1
-651-
Declassified in Part - Sanitized Copy Approved for Release ? 50-Yr 2014/05/01 : CIA-RDP81-01043R003400070002-1
are possible only when the metal is supersaturated with carbon and oxy-
gen to a degree which corresponds to flat interface.
Fig. 254. Diagram of the growth of bubble in a pore not wetted
by liquid.
Thus for unwettable linings all the pores unfilled by metal may
serve as centers of degassing. If steel, at the beginning of reduction,
was not supersaturated with gas, then the latter may .dissolve while
passing from the pores to metal. Steel will fill in the capillaries if
their radius satisfies the following condition
2 7
rn > cos 0 .
Pext
(VIII, 365)
As a result of this, such pores may became deactivated, i.e., they
will not participate in the formation of bubbles during subsequent super-
saturation of metal.
The growth of a bubble formed over a pqe proceeds in the condi-
tions of the existence of a constant angle 84 by means of gradual expan-
sion of the contact perimeter of three phases (gas-hearth bottom-metal).
The take-off of the bubble when 0 > 90? occurs in the same manner as
in the case of a flat surface, i.e., only the upper part of the bubble
separates. The balance -- its stem -- remains fixed to the solid body
and serves as a center for the accumulation of new portions of CO.
In conclusion, it is important to state that both cases considered
(8 4: 900 and e >90?) are limiting. The quantity fi does not remain
constant during the melting process. Thus, for instance, according to
Sapiro &267 for freshly fritted lining, the porosity of which on the
average approaches 6 - 7% Z23576 the rim angle is obtuse, varying, as
stated earlier, from 108 to 128 . Here the overwhelming majority of
pores is active.
As the slagging of the hearth bottom proceeds the angle el decreases
and drops to values considerably lower than 900. In this case only those
capillaries the radii of which satisfy equation (VIII, 363) become ac-
tive. In view of this, the evolution of gas becomes inhibited and the
process of decarbonization slaws down.
1 For information on the hysteresis of the rim angle, see D.937, /2317
and L2347.
-652-
4
Together with the decrease of angle Of, i.e., along with diminish-
ing activity of pores in the process of melting, their number also
changes. According to Andreyev and Matyukha L.2257* in the acid process
decarbonization accelerates not only because of improved conditions of
metal heating. The intensity of bubbling often rises also in conse-
quence of the loosening of tne upper layer of the hearth bottom as a
result of reaction
Si020.441 2C(0,et.)=Si(met) + 2
Kramarov, in studying the behavior of metal in a crucible with acid
lining, revealed a similar phenomenon. As ebulition develops, there forms
in the surface layer of the crucible a network of capillaries which con-
stitute the centers of gas evolution.
He noticed at the same time that the frequency of the emersion of
bubbles is very low (5-8 per second) whereas the number of active pores
is great (up to 2000). This seems to imply that in tne conditions of
Kramarov's tests, the evolution of gas was limited not by the formation
of nuclei, but by the growth of bubbles.
SURFACE-ACTIVE ADDITIONS AND CAVITATION
Additions of surface-active substances may exert a strong influence
both on surface tension of metal and on the interphase energy at its
interface with the hearth. This causes a change of the rim angle e and
affects the formation and growth of bubbles. In view of this the com-
position of metal assumes considerable importance, i.e, the presence in
it of substances which powerfully reduce surface tension.
Problems pertaining to the wettability of refractory materials by
molten ferrous metals were thoroughly studied by Levin L2287. He came
to the conclusion that "the rim angle of a given alloy Ls greater with
the material with which it is least apt to react and with which it pro-
duces a system closer to equilibrium."
Voluminous factual material relative to the liquid non-ferrous
metal--fused salt systems is provided in the book by Belyayev and Zhem-
chuzhina L2347.
The influence of surface-active substances upon the intensification
of heat exchange presents an interesting technological phenomenon.
Lozhkin and co-authors &367, r7'
237found that the coefficient of
heat transfer from the walls to boi. ing mercury is small because of poor
wettability and the formation of a vapor layer. To enhance gas evolu-
tion, they introduced small quantities of capillary active metals
(potassium and sodium) into mercury. These additions reduced surface
tension and clearing the contact surface of metal improved the wetting
of the walls by mercury. As a result of this, the heat exchange coeffi-
cient increased tens of times.
-653-
Declassified in Part - Sanitized Copy Approved for Release ? 50 Yr 2014/05/01 ? CIA-RDP81
Declassified in Part - Sanitized Copy Approved for Release ? 50-Yr 2014/05/01 : CIA-RDP81-01043R003400070002-1
k
.4f,AficOre
Fig. 255. Successive stages of the formation of cavitation bubble
at the nozzle surface (diagram).
On the other hand, the introduction of oleic acid into aqueous solu-
tions makes metal surfaces unwettable which leads to the "spreading of
bubbles along the heater and a changeover to 'film" boiling characterized
by low coefficient of heat exchange L2387.
In conclusion, let us note still another possible source of gas
phase nuolei formation. It is known that local stresses may develop as
a result of rapid flaw of liquid near solid bodies oscillating at high
frequencies. If these stresses attain the ultimate strength of the li-
quid, then raptures appear in it and the so-called cavitation bubbles
arise L2397%
-654-
I
Fig. 255 provides a diagram based on a cinematographic film show-
ing the development of a cavitation bubble formed during the flow of
water issued through a rectangular nozzle at a rate of 15-30 mimic. It
can be seen therefrom that this bubble originates at the narrowest sec-
tion of the nozzle and moves without leaving the surface in the direc-
tion of the flow of water until it finally disappears.
Fig. 256 represents schematioally the cavitation zone around a
screw propeller. A spiral track can be distinguished here which is
composed of bubbles carried away by the stream of liquid.
Fig. 256. Cavitation zone around a screw propeller (diagram).
It is still difficult to say what the significance of cavitation is
for the process of decarbonization. One should think that, in spite
of the powerful stirring of the open-hearth bath during boiling, it is
hardly capable of giving rise to cavitation bubbles in appreciable quan-
titiesl. However, the possibility is not excluded that cavitation plays
a certain role in the Bessemer process.
FORMATION OF BUBBLES AT THE METAL-SLAG INTERFACE AND FROTHING2
As mentioned previously, boiling during decarbonization may develop
not only at the interface of metal with the hearth bed (bottom boil), but
also at the contact surface with slag (surfaoe boil). In the latter case
the frothing of slag arises fairly often and this complicates very much
the normal development of the melting process.
FORMATION OF BUBBLES
In the opinion of Andreyev iy7i
3the appearance of bubbles at the
metal-slag interfaoe is facilita ed by the fact that the work required
for nuclei formation is considerably smaller here than in the volume of
1 It should, nevertheless, be expected that cavitation phenomena develop
with greater effectiveness in supersaturated solutions than in unsatu-
rated ones.
2 This section is written in collaboration with S. I. Popeli.
-655-
Declassified in Part - Sanitized Copy Approved for Release ? 50-Yr 2014/05/01 : CIA-RDP81-01043R003400070002-1
Declassified in Part - Sanitized Copy Approved for Release ? 50-Yr 2014/05/01 : CIA-RDP81-01043R003400070002-1
metal. Apart from this, one should expect a greater supersaturation of
steel in these regions (closeness to the oxidizing slag, lower total
pressure, and so on). It is also enhanced by carbon and oxygen capil-
lary activity whioh increases their concentrations at the contact sur-
face i1857, L2187, L2407.
Similar views are also held by other investigators. Thus, accord-
ing to Sapiro 24l7 the variation of free energy during the formation of
a bubble nucleus of critical size at the metal-slag interface sey be ex-
pressed by equation
AF= (am, A Sol+ 009A Ssi3 St,,,,)
--I S 1 + cos 0, + ctg2 -2-1)
3 tris 2
(VIII, 366)
Here Cri and 41Si stand for corresponding interphase tensions and
the increments of contact surfaces; 8s-g is the rim angle as shown in
fig. 257.
Since the surface tension of steel (crm-g ) is ordinarily 2.5 - 3.5
times that of slag (ors_g), the interface area 1-3 noticeably exceeds
that of 1-2 (see fig. 267). In other words, the emergence of a nucleus
here is facilitated by the fact that the greater part of the surface re-
quires less work during its formation.
Besides, the critical size of the bubble in this case must be ap-
cr
proximately ( ) 17:30 times smaller than in the metal volume. Fi-
s-g
nally, the presenc6 of solid unsintered slag particles at the boundary
with metal also creates favorable conditions for the formation of bubbles.
nag ?
'cf,?A?????.u,?., Metal /2)."...,.
? ".??? ????./ ' ?
? - ? ? ? ?
Fig. 257. Diagram of bubble at the slag-metal interface.
FROTHING
The frothing of slags is a highly undesirable phenomenon. It im-
pairs the transfer of heat from the flame to the bath, retards the process,
and leads to superheating of the furnace roof and ends.
-656-
1
?
A'
4
Andreyev L1937 distinguishes two types of slag froth - volume and
surface. The first is formed in slags with increased viscosity where the
rate of the emersion of bubbles is slower than the speed of their pene-
tration into slag. Surface frothing is characteristic for fluid slags
which possess, however, high surface viscosity. The latter hinders the
running off of the film and prolongs the life of the bubbles.
In Sapirols judgment L2417, Z2427, 2437, &447, frothing is facil-
itated by initial low surface tension of slag, its subsequent decrease
due to adsorption of capillary-active substances, and the mechanical
strength of the bubble film. He further believes that the envelope of
the bubble consists of two parts -- the inside part formed during ad-
sorption of substances dissolved in metal, and the external film formed
at the slag-gas boundary. The higher is the viscosity of slag, the slower
the run-off of the envelope and the more stable the bubble.
Sapiro states that two types of slag are subject to frothing
either homogeneous ferrous slags or highly basic raw slags.
FACTORS DETERAINING THE STABILITY OF FROTH
Since the rate of bubble ingress into slag is determined by the
process of decarbonization, then -- in order to prevent frothing -- it
is necessary to reduce the period of gas presence in the slag layer. In
particular, the rising speed of gas bubbles should be increased. It de-
pends on the size of the bubbles, the viscosity of the medium, and other
factors.
If the bubble was generated at the furnace bed, then it expands in
rising both due to diminishing pressure and to absorption of gas from
the bath. On the contrary, the bubbles formed at metal-slag interface
are quite small and ascend slowly.
In view of the foregoing, surfaoe boiling constitutes one of the
factors which contribute to frothing. This is also facilitated by in-
creased viscosity of slags. It reduces, for instance, the rising speed
of the bubbles. Its significance was noted by Grum-Grzhimaylo &97,
Schenk L.617, and by other authors &137, L2457.
Testscarried out with alcohol and water revealed 42467 that adsorp-
tion of surface-active substances retards the movement of bubbles in
solutions. A similar phenomenon is conceivable also in slags.
Alongside the viscosity of bubble motion, a substantial role is also
played by the stability of its envelope. It is determined by a number
of factors, the most important among which are - splitting pressure,
sensitivity of adsorption to concentration changes, mechanical strength
and surface separation.
Talmud and Bresler Z2477 with the aid of a movie camera (1 shot
per 1/40 sec.) succeeded in establishing that in clear water a bubble
explodes immediately on reaching the surface. Yet, if the bubble is pro-
tected by a layer of soap then it remains intact on the surface for a
period up to 10 sec.
-657-
Declassified in Part - Sanitized Copy Approved for Release ? 50-Yr 2014/05/01 : CIA-RDP81-01043R003400070002-1
Declassified in Part - Sanitized Copy Approved for Release ? 50-Yr 2014/05/01 : CIA-RDP81-01043R003400070002-1
Yerchikovsky 2.187*arrived at the conclusion that a rapid merging
of bubbles coming Irrom opposite directions takes place not only in pure
liquids but also in certain diluted solutions (8.1% phenol and 0.9% KC1
in H20).
Analogous results were obtained by Venstrem and Rebinder Z097, as
well as by other investigators /2507.
Thus, a more or less stable foam is produced by solutions and not
by pure liquids. The capillary-active substances which are to be found
in the solutions, and which are responsible for the reduction of sur-
face tension, increase the stability of froth*
SPLITTING PRESSURE
Deryagin and co-workers disclosed that solid surfaces pressed against
each other separate after moistening by liquid.
Such splitting pressure manifests itself also when small particles
covered by adsorption layers, as well as bubbles, approach each other.
It comes into_display usually at distances around 10-6 am, attains a
maximum at 10 7 to 10-8 cm, whereupon it begin:3 to subside.
Tatiyevskaya L2517, /2527 succeeded in determining the splitting
pressure developing when air bubbles draw together in water. In the pre-
sence of molecules of aliphatic acids, accumulated on the surface of the
film, the latter becomes more stable. This is partly attributable to
splitting pressure which prevents the bubbles from merging causing froth
stability.
To the extent that slag and steel are not single component liquids
but constitute solutions, adsorption phenomena should be expected to
occur at their boundaries. As a mattqT_ of fact, experiments show /2527
that a number of cations (Na4., K4., Ca'', etc.) and anions (P01-,
z-
SixOy and so on) are capillary-active at the slag-gas interface. Their
accumulation on the outer surface of the bubble envelope may prevent the
bubbles from merging because of the electrostatic repulsion of the like
ions.
Moreover, it is possible that splitting pressure also comes into
display near the slag surface in impeding the bubbles to cross the boun-
dary and thereby provoking frothing.
SENSITIVITY OF SURFACE TENSION TO THE CHANGE OF
CONCENTRATION.
The bubble bursts in consequence of continuous diminishing of the
envelope thickness. The latter is caused through imbibition of liquid
into the sections with greater thickness (rims of the film), and also
due to its run-off under the influence of the force of gravity &567.
With the film getting ever thinner, its surface increases while the con-
oentration of capillary-active substances diminishes and becomes uneven.
In consequence thereof, flows of liquid towards the thin spots of the
-658-
film develop in its surface layer. They are inverse to those which re-
duce the thickness of the film and this prolongs the life of the bub-
ble.
Obviously, the greater the change of surface tension with concen-
tration, the more intensive the inverse liquid flaws are and the more
stable the bubble becomes. An additional confirmation thereof is con-
stituted by the fact that solutions of capillary-active substances froth
most intensively when the surface layer is insufficiently saturated
2547, i.e., in cases when the value of 0- varies strongly with concen-
ration.
A similar effect is to be expected to take place also in slags. In-
deed, such substances like 5102 or P2O5 reduce appreciably the surface
tension of ferrous oxide. In particular, the addition of 0.5% P205
causes cr to drop from 650 to 510 erg/cm2, i.e., by 15%. In other
words, adsorption may produce a considerable difference in surface ten-
sions in the narrow and wide sections of the film and increase thereby
the stability of the bubble.
MECHANICAL STRENGTH
Tests show that films which form the bubble envelope in aqueous
solutions possess a definite viscosity, elasticity and shearing strength
2557, Z25671, L2577. Moreover, soluble capillary-active substances pro-
uoe adsorption layers the mechanical properties of which do not differ
from those of a layer of pure water. The converse is apparent for hydro-
philic colloids which have a smaller surface activity and form gel-like
layers with high mechanical strength.
Thus according to Bebinder and Trapeznikov Z2557, the film on a 3%
solution of saponin behaves as a two-dimensional solid body. Intrinsic
w ? to it are elastic properties and a mensurable tensile strength.
't 4
4
The mechanical properties of layers formed by substances insoluble
in water vary depending on their nature.
In certain cases, they do not differ from the properties of a clear
water layer, in other instances there appear films with considerable sur-
face viscosity which produce plastic and even brittle fractures (see,
e.g., &587, L2597).
This type of bubble stabilization is, no doubt, characteristic also
for slag froth. Thus, Semik /2607 revealed increased viscosity of the
surface layer in blast furnace slags. He demonstrated that the critical
shearing stress in slag at rest is greater than in slag with rapidly
ruptured surface layer. In the latter case, the fracture lines of the
film are discernible upon solidification. They have a different colora-
tion than the rest of the slag surface.
The high mechanical strength and even the brittleness of slag froth
may sometimes be registered during the melting process. It is apparently
due to the increased contents of 5i02 (or CaO) in the surface layer and
-659-.
Declassified in Part - Sanitized Copy Approved for Release ? 50-Yr 2014/05/01 : CIA-RDP81-01043R003400070002-1
Declassified in Part - Sanitized Copy Approved for Release ? 50-Yr 2014/05/01 : CIA-RDP81-01043R003400070002-1
to the presence of small particles of lime, magnesia, and other sub-
stances, wetted by slag.
SURFACE SEPARATION
Teitelbaum ?617, &627, &637, in studying the solutions of organic
liquids with similar surface tensions and viscosity, revealed a sharp
maximum of froth stability for medium concentrations. In the absence of
a common explanation for this fact, he advanced the theory of surface
separation of such solutions.
It is assumed that the surface emulsoid whioh forms thereby increases
sharply the viscosity of the film, stabilizing the froth.
On the basis of its existence, Teitelbaum explains the positive tem-
perature coeffioient of surface tension of the solutions which is to be
registered at times (see page 307).
It should be noted however that Yerchikovsky ;2487 reached opposite
conclusions. In his opinion, surface separation siiould reduce consider-
ably the stability of froth.
Be that as it may, the possibility of surface emulsoid formation in
liquid slags still seems to be quite conceivable.
The faot of the existence of volume separation speaks in favor of
this theory. Volume separation, for example, in Me0 - Si02 systems be-
gins to develop at concentrations of silica which are the smaller the
smaller is the radius of the Me24'cation.
EFFECT OF INDIVIDUAL SLAG COMPONENTS
The available published communications on the influence of slag com-
position upon froth stability are often contradictory in character. Thus,
Andreyev 41937 believes that A1203 and CaF2 prolong the life of a bub-
ble. On the other hand, in slags containing CO, FeO, and 5i02 it is
short.
Umrikhin q617 also observes that in heats run in basic open-hearth
furnaces with s ags containing a high percentage of A1203 the stability
of froth and the period of its formation increase.
Revebtsov and Rybakov Z2657, on the contrary, came to the conclu-
sion that in operations with &luminous slags no frothing is to be ob-
served.
According to Sapiro Z243.7, Z2447 non-viscous slags with high
contents of FeO and 5i02 Troth most frequently. In addition to this,
the period of effervescent ore boil extends considerably at high concne-
trations of Ca? + MgO (%i 50%) and low concentrations of A1203(2 -
The addition of small portions of lime to acid slags reduces the dura-
tion of foaming.
-660.
4
4
?
As attested by certain data, p667, admixture of finely ground iron
ore to a slag composition of 36% CaO, 22% Si02 , 10% MnO and 11% FeO
covering metal containing up to 44 C resulted in abundant evolution of
gas not accompanied by frothing. Addition of Ca2S3.04, borax, and
fluorite produced foaming which lasted for more than 10 min.
Finally observations show that addition of tar and mazut to gaseous
fuel reduces the frothing of slag in open-hearth furnaces. Some au-
thors /2677 erroneously connect this with the increase of reduoive pro-
perties of the gas jet, others /2687-with the growth of its caloricity.
Sapiro &437 ascribes this effea to the mechanical destruction of the
foam by the stream of the denser jet.
All the foregoing evidences that this problem is still insufficiently
known and that further systematic investigation of the causes of slag
frothing is required.
EMERSION OF BUBBLES
The emersion of bubbles concludes the process of decarbonization.
At the same time it fulfills a number of very important functions.
First of all among them should be cited the develop&ent of the re-
aotion front in the bath volume. The rising bubbles play the role of
centers of gas evolution. They grow at the expense of the molecules of
carbon monoxide they contain, as well as of the gases (N2, H2, etc.)
dissolved in metal, and serve thus as active degasifiers of steel. Fur-
thermore, the bubbles passing through slag react with iron oxides re-
ducing them to metal or to ions of the lowest valence.
They produce an exceptionally important hydrodynamic influence on
the bath by causing powerful mixing. This leads to a sharp increase of
the coefficients of mass and heat transfer and also to the ejection of
drops of metal into slag and into the atmosphere of the furnace. The
latter fact produces a considerable effect on the saturation of steel
with gases.
The efficiency of bath mixing may serve as a certain characteris-
tic of these processes. As shown by Kooho /2697 and Levin ?.2707, the
calculation of mixing efficiency made ,by Schenk is inoorrec . It does
not take into account the effect of the height on the work produced by
the bubble. An assumption was made that the latter"s volume is con-
stant and that the work required for its passage into the gas phase is
entirely absorbed by the bath.
Kocho improved the accuracy of this computation, having taken into
consideration both the hydrostatic pressure and the expansion of the
rising bubble. He also analyzed the case of the bubble's growth caused
by the passage of CO molecules from the metal. At the same time, it
was assumed that the volume of the bubble increases linearly with the
height of its ascent.
Further elaboration for greater precision was a000mplished by Levin,
Declassified in Part - Sanitized Copy Approved for Release ? 50-Yr 2014/05/01 : CIA-RDP81-01043R003400070002-1
-661.
Declassified in Part - Sanitized Copy Approved for Release ? 50-Yr 2014/05/01 : CIA-RDP81-01043R003400070002-1
who took the site of bubble formation into account and based his cal-
culations on a more correct regularity of the growth of bubble volumes
while making an allowance for surface carbon oxidation.
The work of Pomerantsev and gyrkin Z2717 which deals with the as-
cent of bubbles presents a certain interest in that it provides a fur-
ther detailization of the picture. They have established that a steady
rate of ascent is being achieved here practically immediately.
Frank-Kamenetsky &727suoceeded in calculating the discoid shape
of the rising bubble. Proceeding from the balance of the forces of
surface tension and resistance to motion, he developed the following
equation for velocity
4 ea
vr.=
Here a = coefficient of resistance;
g = gravity acceleration;
= specific gravities.
(VIII, 367)
In conclusion, let us dwell on the ejection of beads of metal by
bubbles. Andreyev studied this phenomenon on a model by blowing air
through mercury 6.937% He found that the bubbles exploded immediately
on a clear metal surface. This process decelerated, however, after the
surface of mercury was covered with a layer of water. Moreover, upon
the egress of the bubble from water, a rupture of the metallic film was
to be observed together with the dispersion of mercury in the form of
small drops. The replacement of water by transformer oil, produoing
a more stable foam, facilitated the passage through its layer of a
greater number of bubbles.
Assuming that the quantity of metal carried away into the atmos-
phere of the furnace corresponds to 0 of the lifting power of bubbles,
that only of them actually reaches the gas-slag interface, and that
the rate of carbon burn-out comprises 0.006% per minute, Andreyev de-
termined that in a 100-ton capacity furnace around 500 kg of metal are
being ejected into the gas phase per minute.
Medzhibozhsky investigated this process in industrial conditions.
He reports that the number of metal beads contained in slag grows regu-
larly with its diminishing viscosity and the increasing thickness of its
layer (fig. 258) as well as with the rise of the carbon burn-out rate
(fig. 259).
Further it became evident that the distribution of the beads along
the height of the slag layer during intensive boiling is -- as one should
have expected -- uneven. Their number is greater below (5.95%) and
smaller at the top (0.81% at fluidity of 40 mm). Towards the end of
ebulition this unevenness levels down.
It was established by chemical analysis that the contents of oar-
bon in the beads of steel in slag is about half its contents in the
-663-
V
bath. Moreover, the concentration of C is lower in the bigger beads
than in the smaller ones. This is probably due to the fact that the
former originated from the bursting film, whereas the latter formed as
a result of the ejection of drops of a more compact type.
Fig. 258. Effect of slag fluidity
(1, ma) and the thickness of slag
layer (a - 135 mm, & - 80 mm) on
the contents of metal beads in slag
(g, %)
On the basis of Stokes, formula
Medzhibozhsky proposed a formula for
beads ejected per second
where
Pel' h81,
q=
Fig. 259. Contents of metal
beads in slag (q, %) vs. carbon
burn-out rate.
and the conditions of steady process,
the calculation of the number of
3,81 kPsi /le
q = rate of emergence of beads in relation to the
weight of slag, % per minute;
k their contents, %
&
quantity in % to the weight Of metal, thickness of
7,
the layer, and viscosity of slag;
rk = mean radius of bead.
(VIII, 368)
In particular, with k = 0.5%, ho =
poises, rk = 0.015 cm.' Psl an Ve
q = 0.1% per minute. The value produced
that calculated by Andreyev.
8 am., 71 si = 0.2 - 0.3
= 0.2% per hour, we have
is considerably smaller than
Ark() of interest is the conclusion to the effect that the beads of
metal introduce a relatively small quantity of oxygen into the bath,
namely 0.005% per hour, while a far greater amount penetrates from slag
- 0.48% per hour.
In summing up, one should emphasize once again the fact that the
process of carbon burn-out is very complex and far too insufficiently
studied. Its detailed and systematic investigation is essential.
Declassified in Part - Sanitized Copy Approved for Release ? 50-Yr 2014/05/01 : CIA-RDP81-01043R003400070002-1
-663-
Declassified in Part - Sanitized Copy Approved for Release ? 50-Yr 2014/05/01 : CIA-RDP81-01043R003400070002-1
BIBLIOGRAPHY TO CHAPTER VIII
1. V. A. Vanyukov
2. O. A. Yessin
3. L. Chang and J. Chipman
4. O. A. Yessin and L. K.
Gavrilov
5. J. Chipman
6. O. A. Yessin and L. K.
Gavrilov
7. O. A. Yessin, L. K. Gav-
rilov and B. M. Lepinskikh
8. V. P. Yelyutin and B. E.
Levin
9. F. Koerber & W. Oelsen
10. P. V. Geld, N. V. Zaimskikh
N. N. Serebrennikov and
U. P. Nikitin
11. O. A. Yessin, L. K. Gay-
rilov & N. A. Vatolin
12. V. P. Yelyutin, U. A. Pav-
lov and B. E. Levin
13. K. P. Ranadin
14. O. A. Yessin and I. G.
Sryvalin
15. A. N. Volt sky
16. O. A. Tessin, Yu. P.
Nikitin & S. I. Popelf
- 'Zhurnal Russkogo metallurgicheskogo
obshchestvas (Journal-of tRe Russian
Metallurgical Society), 2, 123, 19120
- (Electrolytic Nature of Liquid Slags)
Elektroliticheskaya priroda shidkikh
shlakov, ed. Technical House of The
Ural Polytechnical Institute, Sverd-
lovka, 1946.
- *Metals Technology,' October, 1946.
? "Izvestiya Akademiyi nauk SSSR,"
Otdeleniye tekhnicheskikh natik,(Bu1-
1etin of the USSR Academy of Sciences,
Department of Technical Sciences,)
7, 1040, 1950.
- *Discussions of the Faraday Society,*
4, 4, 1948.
? "Izvestiya Akademiyi nauk SSSR,"
Otdeleniye tekhnicheskikh nauk, (Bul-
letin of the USSR AcaTemy or Sciences,
Department of Technical Sciences),
7, 1040, 1950.
- *foklady Akademiyi nauk SSSR (Re-
ports of the USSR Academy of Sci-
ences), 88, 5, 1953.
- "Stal" (Steel), 9-10, 554, 1946.
? Nitteilungen K.AUlhelm Institut fuer
Eisenforshung,' 18, 89, 1936.
^ *Zhurnal prikladnoi khimiyi" (Journal
of Applied Chemistry), 25, 687,
1952.
- *Doklady Akademiyi nauk SSSR" (Re-
ports of the USSR Academy of Sci-
ences), 85, 87, 1952.
? Ferrosplavy (Ferrous Alloys), 424-425,
Metallurgizdat, 1951.
- Elektroliticheskiy pertnos v metalli-
cheskikh zhidkikh i tvrdikh rastvorakh
(Electrolytic Transfer in Liquid and
Solid Metal Solutions) Proceedings of
the N. E. Zhukovsky Academy, ed. 167,
17, 1947.
^ "Zhurnal fizicheskoy khimiyi* (Journal
of Physical Chemistry) 25, 1505, 1951;
26, 371, 1952.
? Osnovy teoriyi metallurgicheskikh
playok (Rudiments of the theory of In -
&atrial Heats), Metallurgizdat, 1943.
- *Doklady Akademiyi nauk BSSE* (F0 -
ports of the USSR Academy of Sci-
ences), as, 431, 1952.
-664?
!1?., S'
? rc,
1
17. W. U. Filbrook, K. M.Gold-
man, M. M. Hesel
18. A. M. Samarin, L. A.
Schwartzmann and M. I.
Temkin
19. K. Balajiva, A. Quarrel
and Vajvagupta
20. Ya. I. Gerassimov
21. P. Antipin et al.
22. 0. A. Tessin and A. I.
Okunev
23. A. N. Frumkin
24. A. N. Frumkin
25. O. A. Tessin
26.
27.
28.
29.
30.
31.
32.
33.
J. Chipman and Ta Li
K. Jellinek and G.
Sakovsky
E. W. Britzke and A. F.
Kapustinsky
D. White and H. Skelly
E. Maurer, G. Hammer and
H. Moebius
C. W. Cherman, H. Elvan-
der and J. Chipman
J. P. Morris and R. C.
Buehl
C. H. Harty
34. G. A. Jontech
35. M. M. Karnaukhov
36. W. Koch and K. Fink
37. M. W. Dastur and J.
Chipman
^ "Problemy sovremennoi metalurgi i"
(Problems of Contemporary Metallurgy),
3, 3 GIIL, Moscow, 1952.
- 'Thurnal fizioheskoy khimiyi" (Jour-
nal of Physical Chemistry) 20, 111,
1946; "Acta Physicochimica USSR,*
20, 111, 1945.
^ *JOUrnal of Iron and Steel Institute,"
153, 115, 1946; 155, 562, 1947.
? "Uspekhi Khimiyi"TIchievements of
Chemistry), 14, 289, 1948.
- Elektrokhimiya resplavlennykh soley
(Electrochemistry of Fused Salts),
part I, ONTI, 1937.
- "Izvestiya Akademiyi nauk SSSR,"
Otdeleniye tekhnicheskikh nauk, (Bul-
letin of the USSR Academy of Sciences,
Department of Technical Sciences),
10, 1472, 1952.
? "Murnal fizicheskoy khimiyi" (Jour-
nal of Physical Chemistry) 10, 568,
1937; 24, 244, 1950.
^ Trudy 2-qy konferentsiyi po korroziyi
metalov (Proceedings of the 2nd Con-
reThriFE on Metal Corrosion), ed, of the
USSR Academy of Soiences, vol. 1, 1940.
- "Zhurnal fizioheskoy khimiyi" (Jour-
nal of Physical Chemistry) 14, 717,
1940.
*Transactions of the American Society
for Metals," 25, 435, 1937.
- *Zeitschrift fuer anorganische Chemie,
142, 1, 1925.
- Ina, 194, 323, 1930.
^ "Journal of the Iron and
Steel Insti-
tute, * 155, 201, 1947.
"Archly fuer das Eisenhuettenwesen,*
16, 159, 1942-1943.
- *Journal of Metals," 183, 334, 1950.
- rbid, 188, 317, 1950.
? *Transactions of the American Institute
of Mining and Metallurgical Engineers,"
76, 26, 1929.
^ *Stahl mid Eisen,' 31, 911, 1924.
? Metallurgiya stali rietallurgy of Steel),
part II, p. 126, -1934.
^ "Archly fuer das Eisenhuettenwesen,'
22, 371, 1951.
- "Transactions of the American Insti-
tute of Mining and Metallurgical En-
gineers,* 185, 441, 1949; *Journal of
Metals,' August 1949.
?665-.
Declassified in Part - Sanitized Copy Approved for Release ? 50-Yr 2014/05/01 CIA-RDP81-01043R003400070002-1
Declassified in Part - Sanitized Copy Approved for Release ? 50-Yr 2014/05/01 : CIA-RDP81-01043R003400070002-1
38. V. I. Karmazin
39. V. I. Karmazin
40.
41.
42.
43.
44.
45.
46.
47.
48.
49.
H. Wentrup
S. O. Lifschitz
K. Trubin and G. Oyks
B. M. Larsen
J. A. Kitchener, J. O'M
Bockris and A. Liberman
J. P. Morris and A. I.
Williams
C. Sherman
J. Chipman
L. S. Darken
A. A. Samarin and
L. A. Sohwartzmann
50. A.
L.
51. A.
L.
52. A.
53. T.
54. B.
55. N.
A. Samarin and
A. Schwartzmann
A. Samarin and
A. Schwartsmann
Norro and S. Lindqvist
Rosengvist and E. Cox
Osann
Grant & J. Chipman
56. A. S. Heinman
570 O. A. Yessin and
V. A. Kozheurov
58. A. U. Samarin and
L. A. Sohwartzmann
59. V. E. Grum-Grzhimaylo
60. M. M. Karnaukhov
- Trudy instituta chernoy metallurgiyi
Aka emiyi nauk USSR, (Proceedings of
the Institute for the Meallurgy of
Ferrous Metals of the Ukrainian Academy
of Sciences) 3, 25, 1949.
^ "Stalf* (Steel), 8, 9, 139; 5-6, 24,
? "Archly fuer das Eisenhuettonvesen,"
13, 535, 1936.
? Teoriya i praktika metallur iyi (Theory
and Practices of Metallurgy) 9, 9, 1937.
- Metallurgiya stali (MetalluriF of
Steel), Metallurgizdat, 1951.
in the book "Open-Hearth Steel Process,"
Metallurgizdat, 1947.
? "Discussions of the Faraday Society"
4, 39, 1948.
- Transactions of the American Society
for Metals," 41, 1421, 1949.
"Journal of larals," 188, 1349, 19500
? in the book "Open-Hear-ER' Steol Process,'
Matallurgizdat, 1947.
? "Journal of Metals,' 188, 1349, 19500
"Isvestiya Akademiyi k SSSR,"
Otdeleniye tekhnicheskikh nauk, (Bul-
rarh of the USSR Academy of Soiences,
Department of Technical Sciences),
3, 407, 1951.
^ Mid, 8, 1231, 1949.
- Ibid, 11, 1696, 1960.
^ 'Jernkontorets Annaler," 130, (3), 118,
1946.
- "'Journal of Metals," 188, 1959, 1950.
"Stahl und Eisen,* 287751, 1908.
^ "Metals Technology,v- T. P., 1988,
April 1946.
^ "Izvestiya Akademi i nauk SSSR,"
e eni e tekhnic es i nauk, (Bul-
letin of the USSR Academy of Sciences,
Department of Technical Scienoes),
10, 1439, 1946.
"Thurnal prikladnoi khrmiyi" (Journal
of Applied Chemistry), 21, 1'65, 1948.
^ "Izvestiya Akademi i nauk SSSR,"
Otdeleniye tekhnic eskikh nauk, (Bu1s,
letin of the USSR Academy of Soienoes,
Department of Technical Sciences),
9, 1457, 1948.
^ P7oizvodstvo stali (Production of
Steel) GI, 1931.
? Metallurgiya stali (Metallurgy of Steel)
Metallurgizdat 1933.1934.
61, G. Schenk
62. P. Bardenheuer and
W. Geller
63. E. Maurer & W. Bischof
64.
65.
66.
G. Manterfield
F. Scott & T. Joseph
A. M. Samarin and
F. P. Edneral
67. A. M. Samarin and
O. K. Teodorovich
68. M. A. Pavlov
69. I. V. Bulla and B. A.
Hess
70. I. P. Semik
71. W. Holbrook and
T. Joseph
72. T. Rosenqvist
73. B. Kalling, C. Danielson
and O. prance
74. C. Bettendorff & I. Warg
75. Ya. I. Olshansky
76. A. M. Samarin
77. A. M. Danilov
78. D. Fox
79. B. Osann
80. S. Sohleioher
81. O. Meyer and
J. Herrisen
82.
83. G. Derge and
S. Marshall
- Fizicheska a khimi a metallur icheskikh
processov, ? sioa emis ry o
Metallurgical Processes), part II,
GONTIU, 1936.
- Iiitteilungen K. - Wilhelm! Institut
fuer Eisenforschung," 16, 77, 1934.
- "Blast-Furnace and Steel' Plant,"
410, 1936.
^ tallurgy," 9, 110, 1939.
- 'Metals and AlThys,' 5, 745, 1942.
- Proizvodstvo i obrabAka stali, Sbornik
trudov Mosokovs ogo instituta stall
(Production and Processing of Steel,
Proceedings of the Moscow Institute of
Steel), 29, 64, Metallurgizdat, 1950.
- Issledovaniye protsessa desulfurizat-
siyi v dugovykh pechakh (Investiga-
tion of the Desulfurilation Process
in Arc Furnaces), Moscow, VNITOM, 1940.
^ Metallurgi a chuguna (Metallurgy of
Cast-Iron j, part II, Metallurgizdat,
1949.
^ "Domes,* 8, 23, 1934.
^ "Stall*, Steel, 3, 203, 1947 and others.
? "Transactions orthe American Institute
of Mining and Metallurgical Engineers,"
120, 99, 1936.
^ ITFOblemy metallurgi (Problems of
Metallurgy), 3, 11, 952; "Journal of
Metals," 3, (/), 535, 1951.
- *Journal -Of Metals," 3, 732, 1951.
? *Stahl und Eisen,' 2, 25, 1932.
- Trudy instituta georogicheskikh nauk,
Akademiyi nauk SSSR, vyp. 137, petro-
grafioheskaya seriya, Transactions
of the Institute of Geological Sci-
ences of the USSR Academy of Sciences,
ed. 137, Petrographic series) 40, 33,
1951.
? Elektrometallurgiya (Electrometallurgy),
195-216, Matallurgizdat, 1943.
? 'Stall" (Steel), 4, 104, 1945.
"Iron Age," 9, 8371909.
? Ibid, 47, 11T, 1927.
und Eisen," 41, 357, 1921.
- "Archiv fuer das Eisenhuettenwesen,"
7, 665, 1934.
- gartenovskoye proizvodstvo stali
715-3-e-in_aair-producton,etal-
lurgizdat, 1947.
? "Problemy soveremennoy metallurgi i"
(Problems of Modern Metallurgy),
29, 1952.
-667-
Declassified in Part - Sanitized Copy Approved for Release ? 50-Yr 2014/05/01 : CIA-RDP81-01043R003400070002-1
Declassified in Part - Sanitized Copy Approved for Release ? 50-Yr 2014/05/01 : CIA-RDP81-01043R003400070002-1
84. V. Heike
85. W. Oelsen
86. G. Evans
87. O. Meyer and H.
Schulte
88. J. Chipman, J. B. Gero
and T. B. Winkler
89. F. Koerber and W. Oelsen
90. N. N. Dobrokhotov
91. S. L. Levin and
V. F. Lyade
92. G. A. Volovik
93. V. M. Zamornyev
94. M. I. Panfilov
95. A. M. Danilov
96. L. S. Darken and B. M.
Larsen
97. C. R. Taylor and
J. Chipman
98. R. Schumann and P. Ensio
99. V. G. Voskoboinikov
100. V. I. Karmazin.
101. A. S. Mikulinsky and
Ya. M. Umova
102. F. Koerber and W. Oelsen
103. C. S. Pease
104. S. I. Popell, O. A. Yessin,-
? ? _
? "Zhurnal Russkogo metallurgicheskogo
ohshohestva" (Journal of the Russian
Metallurgical Society), part II, 366,
1931.
^ 'Stahl und Eisen," 44, 1212, 1934.
? *Metals Technology,w-september, 1938.
? "Archiv fuer das Eisenhuettenwesen,"
11, 187, 1934.
? lgournal of Metals," 188, 341, 1950.
"Kitteilungen K-Wilhelm Institut fuer
Eisenforsohung" 15, 271, 1933.
"Teoriya i praktrEa metallurgiyi"
(Theory and Practices of Metallurgy),
9, 35, 1936.
- Ibid, 10-11, 37, 1939.
Nauchnye trudy Dnepropetrovskogo metal-
lurgicheskogo institute, (Proceedings
of the Dnepropetrovsk Metallurgical
Institute), ed. 13, 65, 1948.
Sovremenwye metody proizvodstva vyso-
kosortnoy stali, (Modern Methods of
High-Quality Steel Production), ONTI,
1936.
"Stall* (Steel), 1-2, 23, 1940.
Ibid, 4, 104, 1945.
wYFWntaCstions of the American Institute
of Mining and Metallurgical Engineers,"
150, 87, 1942.
^ Ibid, 154, 228, 1943.
SRI
105. P. Heragymenko and
H. Speight
106. A. M. Samarin, L. A.
Schwartzmann
107. Ya. S. Umansky, B. N.
Finkelstein and M. E.
Blanter
108. B. V. Stark, Ye V. Chel-
ishohev and A. Kazachkov
?111
"Journal of Metals," March, 1951.
*Stall" (Steel), 7-8, 243, 1945.
"Teoriya i praktika metallurgiyi"
tTheory and Practices of MeTallurgy),
9, 35, 1936.
lifetallure (Metallurgist), 10, 1934.
'Stahl und Eisen,* 34, 905, 1938.
"Journal of Metals,173, 313, 1951.
"boklady Akademiyi nea SSSR" (Reports
of the USSR Academy of Sciences),
75, 227, 1950.
lgournal of the Iron and Steel Insti-
tute" 166, 169, 289, 1950.
^ "Mturnal fizioheskoy khimiyi" (Journal
of Physical Chemistry) 22, 565, 1948.
- Fizioheskiye osnovy metiTiovedeniya
(Physical Principles of Metallography),
Metallurgizdat, 1949.
- "Izvesti a Akademiyi nauk SSSR," Otdel-
eniye te hnioheskikh natik, (Bulletin
of the USSR Academy of Sciences, Dept.
of Technical Sciences), 11, 1789, 1951.
-668-
109. See e.g. E. Dipschlag
110. F. Schoenwalder
111. L. M. Lindeman
112. L. Chang and K. Goldman
113. G. Derge, W. Filbrook,
and K. Goldman
114. T. Rosenqvist
115. T. B. Winkler and
J. Chipman
116. L. A. Schwartzmann
and P. L. Gruzin
117. V. A. Kamensky and Ye.
Abrosimov
118. N. A. Konstantinov
119. M. Hansen
120. J. Berak
121. Ye. A. Poray-Kosohitz
122. C. H. Herty
123. W. Krings and H.
Skhaokman
124. Bischof and E.
Maurer
125. Y. K. Zea
126.
127.
128.
T. Firely
V. K. Gorin
P. V. Umrikhin and
N. I. Kokarev
129. P. V. Umrikhin
130. MoCanoe
131. T. B. Winkler and
J. Chipman
132, P. V. Unrikhim
- "Domennyi protases* (Blast Furnace
Process) GNTIU, 1953.
- "Stahl und Eisen" 37, 949, 1933.
- "thurnal Russkogo fiziko-khimicheskogo
obshchestva" (Journal of the Russian
Physic? -Chemical Society), II, 1-2,
61, 1916.
- "Metals Technology," 15, 4, 1948.
- "Journal of Metals," ITS8, 9, 1111,
1950.
Ibid, 188, 11, 1336, 1950.
- lfferals Technology," T. P. 1987,
April 1946.
? Problemy metallovedeniya i fiziki
metalov, (Problems of Metallography
and the Physics of Metals ) vol. 3,
237, Matallurgizdat, 1952.
D. - Trudy Moscovskogo instituta stali
(Proceedings of the Moscow Institute
of Steel), vol. 12, 167, 1939.
- "thurnal Russkogo fiziko-khimioheskogo
obshohestvaw (Journal of the Russian
Alysico-ohemical Society), 41, 1220,
1909.
^ Struktura binarmjkh splavov (Structure
of Binary Alloys), part II, Metallur-
gizdat, 1941.
- "Archiv fuer dass Eisenhuettenwesen,"
131, 1951; ref erat "Problemy sovremen-
nay metallurgiyi" (Tech. -Paper "Prob-
lems of Modern Metallurgy), 3, 81, 1952.
? "Uspekhi Khimiyi" (Achievements of
Chemistry), 13, 115, 1944.
^ *Transactions of the American Institute
of Mining and Metallurgical Engineers,"
73, 1107, 1926.
^ neitsohrift fuer anorganische und
algemeine Chanties' 213, 161, 1933.
- "Arohiv fuer das Eisenhuettenwesen,"
6, 415, 1932/t.
- TJournal of the Iron and Steel Indus-
try," 151, 459, 1945.
- TEld, IZIT, 161, 1947.
- Will" (Steel), 5-6, 150, 1944.
- Osonvy skorostey martenovksoy plavki
(Rudiments re. Rates of Open-Hearth
Smelting), Metallurgizdat, 1947.
- 'Stall" (Steel), 7, 596, 1947.
- "15UFhal of the Iron and Steel Indus-
try," 331, 1938.
- "WetairTeohnology," T.P. 1987,
April 1946.
- Trudy Urallsko o politekhnicheskogo
institute. Imeni S. M. Korova (Transac-
tions of the Ural Po4rech. Institute
of S. M. Koriv), vol. 22, 5, 1945.
-669-
Declassified in Part - Sanitized Copy Approved for Release ? 50 -Yr 2014/05/01 ? CIA-RDP
-01 043R00340007nnn9
Declassified in Part - Sanitized Copy Approved for Release ? 50-Yr 2014/05/01 : CIA-RDP81-01043R003400070002-1
133. L. S. Einstein and
A. I. Nestsrov
134. P. V. Umrikhin
135. E. Maurer and
V. Bischof
136. V. A. Kozheurov
137. S. T. Rostovtsev
138. Yu. P. Nikitin, O. A.
Yessin and S. I. Popel
139. L. O. Sokolovsky
140. Ye. Kostyuchenko
141. L. Darken and R. Gurry
142. J. Chipman and L. Chang
143. V. A. Kozheurov
144. J. nit.
145. M. Ya. Medzhibozhsky
146. P. V. Unrikhin
147. P. Ya. Ageyev
148. C. Taylor and J.
Chipman
? "Uraltsksiya metallurgi e (Metallurgy
of the Urals), 6, 1936.
- Netallu (Metallurgist), 3, 1938.
'- Jburnal of the Iron and SteR In-
stitute,* 132, 13, 1935.
? *Zhurnal prikladnoi khimi i' (Jour-
nal of Applied dhamiitry , 23, 233,
1950.
- Teoriya metallurgioheskikh protsessov
(-Theory of MetallurgloaI-Processes),
287, Metallurgizdat, 1945.
? *Doklad Akadamiyi nauk SSSR" (Re-
ports oi the USSR Academy of Sciences),
87, 5, 813, 1952.
- 1Vestnik mashinostr eniya" (The Ma-
chinebuilding Herald), 2, 30, 1947.
0 metode ozdorovleniya 7letalla shlakom
MA fiziko-ikolloidno-khimioheskcy
?macre, methd oreta1i1ity
Improvement by Slags on the Basis of
ghysioal and Collidal Chemistry),
of the Ukraine, 1936.
- 'Journal of the American Chemical
Society," 68, 798, 1946.
? "Journal of Metals," February 1949.
? *Zhurnal fizicheskoy khimiyi" (Journal
of Physical Chemistry) 25, 694, 19510
? 'Journal of the Iron and-Steel Insti-
149. M. Dastur and J. Chipman -
150. N. L. Bowen, J. F. Scheirer
and E. Posnjak
151. L. A. Sohwartgmann, A. A. -
Samarin and M. I. Temkin
152. L. S. *bakov
153. P. Heragymenko
154. G. Leiber
tube," 148, 579, 1943.
osk laboratoriya' (Plant Labors.-
tory'), 115, 1948.
Sbornik statey *Proizvodstvo stali"
(Collection of articles 'Steel Produo-
tion'), 6, Metallurgizdat, 1952.
Trudy Urallskogo industriyallnogo
stituta (Transactions of the Ural In-
dustrial Institute), vol. 22, 17, 1945.
'Transactions of the American Institute
of Mining and Metallurgical Engineers,*
154, 228, 1943.
TETa, 185, 441, 1949.
'American Journal of Science,' 26, 1939,
1933.
"Zhurnal fizicheskoy khimiyi" (Journal
of Physical Chemistry) 21, 1027, 1947.
Trudy Urallskogo politeMicheskogo
institute, ibornik statey ("Proiz-
vodstvo Transactions of the
Ural-PoIytechnioal Institute, collec-
tion of articles 'Steel Production"),
62, 1952.
*Journal of the Iron and Steel Insti-
tute," 157, 515, 1947.
*Stahl urartieen," 9, 237, 1937.
-670-
A
155. N. N. Dobrokhotov
156. N. N. Dobrokhotov
157. P. Ya. Ageyev
158. B. V. Stark and
Ye. V. Chelishohev
159. B. V. Stark and
Ye. V. Chelishohev
160. B. V. Stark and
Ye. V. Chelishohav
161. C. H. Herty
162. P. Ya. Ageyev
163. P. Ya. Ageyev
164. Kintzel and Eagon
166. B. M. Larsen
166. J. Chipman
167. J. Chipman and
A. Samarin
168. S. Marshall and
J. Chipman
169. H. Vaoher and E.
Hamilton
170. H. Vacher
171. G. Phragmen and
H. Kalling
172. Le Chatelier
173. A. A. Baykov
174. S. L. Levin
- "Teoriya i praktika metallurgiyi"
(Theory and Practices of Metallurgy),
9, 35, 1936.
- Mmremennaya tekhnologiya vy lavki
stali va martenovgkikh pechakh,
izdan-
iy. Akademiyi nauk USSR (Catemporary
rooasses of Steel Smelting in Open-
Hearth Furnaces, pub. the Ukrainean
SSR Academy of Sciences), 1951.
Thud Uraliskogo industri alfnogo
institute (Transactions o the Ural
indistrial Institute), vol. 22, 1945.
? *Metallurg' (Metallurgist), 7, 17, 1939.
Trudy Moscovskogo institute stall. (Pro-
ceedings of the Moscow Institute of
Steel), 19, 1941.
? Ibid, 30, 1951.
- !Wining and Metallurgical Investiga-
tion Bulletin,* 68, 1934.
? 4-aya nauchno-teUhicheska a konferent-
siva Leningradskogo p1itekhxiioheskogo
institute (4th Scientifio-Teohnical
Conference of the Leningrad Polytech-
nical Institute) pub. Len. Polyteoh.
Inst. (LPI) 1947.
? Yezhegodnik NITO metallurgov (Yearbook
of the Soientifio-Research Teohnical
Society), Leningrad, 1947.
- 'Transactions of the A.I.M.M.E.,"
"Iron and Steel Division," 304, 1929.
- *Metals and Alloys," 1, 763, 1930.
- "Transactions of the 17.I.M.M.E.,"
Open-Hearth Proceedings, 110, 1937.
? 'Novosti inostrannoy literatu " (News
in Foreign Literature), 7, 1937.
- "Transactions of the American Society
for Metals," 30, 695, 1942.
*Trensactions of the A.I.M.M.E.,' 95,
124, 1931; 'Stahl lind Eisen," 1033?,
1931.
^ *Bureau of Standards Journal Research,"
II, 541, 1933.
- *Jernkontorets Annaler," 123, 199,
1939.
? "Revue de Metallurgie,* 9, 673, 1912.
Sobraniye sochineniy, (Collection of
Morks), pub. USSR Academy of Sciences,
vol. II, 382, 1948.
^ Nauohnye trudy Dnepropetrovskogo
lurgloheikogo institute (Proceedings
of the Dnepropetrovsk Metallurgical
Institute), ed. II. 68, 1940.
Declassified in Part - Sanitized Copy Approved for Release ? 50-Yr 2014/05/01 : CIA-RDP81-01043R003400070002-1
Declassified in Part - Sanitized Copy Approved for Release ? 50-Yr 2014/05/01 : CIA-RDP81-01043R003400070002-1
175. O. I. Yatsunskaya and
M. N. Starovich
176. B. V. Stark
177. L. S. Darken
178. M. I. Temkin and
L. A. Schwartzmann
179. A. M. Samarin, A. Yu.
Polyakov and L. A.
Schwartzmann
180. E. Shale
181. F. Koerber and W.
Oelsen
182. A. M. Samarin and
L. A. Sohwartzmann
183. H. Tielmann and
M. Wiemmer
184. P. L. Gruzin, Yu. V.
Kornev and G. V. Kurdy-
MOV
185. B. V. Stark and S. I.
Filippov
186. C. Fetters and J. Chipman
187. B. M. Larsen
188, W. Filbrook
189. J. Slotman and
F. Lawnsberry
190. Yu. M. Maksomov
191. V. I. Karmazin and
G. P. Pukhnarevich
192. O. A. Yessin and
P. V. Geld
193. I. A. Andreyev
- Primeneni e kisloroda v martenovskom
ilirolzvods ve (Use of Oxygen in Open-
Hearth Process), Metallurgizdat, 1952.
? "Izvestiya Akddemiyi nauk SSSR,* Otdel-
emye e ? es nau , Bu e n
of the USSR Academy of Solencea, Dept.
of Technical Sciences), 5, 655, 19480
- Netals Technology," T.P. 1163, 2, 1940.
? "Zhurnal fizicheskoy khimiji. (Journal
of Physical Chemistry) 23, 755, 1949.
^ "Izvesti a Akademi i naliESSSR,* Otdel-
enfie tekhnioheslcikh nauk, (Bulletin
of the USSR Academy of-Soiences, Dept.
of Technical Sciences), 12, 1639, 1947.
- "Problemy soveremennoy metallurgi is
(Problems of Modern etallurgy)?
44, 1952.
- Nitteilungen K.-Wilhelm Institut fuer
Eisenforschung," 17, 39, 1935.
*Izvestiya Akademiyi nauk SSSR,"
Otdel-
eniye tekhnioheskikh nauk, (-Bulletin
of the USSR Academy of Sciences, Dept.
of Technical Sciences), 8, 1231, 1949;
6, 891, 1949.
- Itahl und Eisen," 10, 389, 1927.
^ "Doklady Akademi i nauk SSSR" (Reports
of the USSR Academy of Sciences), 80,
49, 1951.
? "Izvesti a Akademi i nauk SSSR, Otdel-
eniye tekhnichesicikh nauk, (Bulletin
of the USSR Academy of Sciences, Dept.
of Technical Sciences), 3, 413, 19490
^ *Transaotions of the A.I.V.M.E.,* 140,
1/0, 1940.
- Ibid, 145, 67, 1941.
- IsTa, Ii, 136, 1940.
- 1173n Age," 159, 42, 1947.
- Trudy Moscovskogo institute. stali
(Proceedings of the Moscow Institute
of Steel), ed. 29, 17, Metallurgizdat,
1950.
^ "Teoriya i praktika metallurgi ai*
(Theory and Practices of Metallurgy),
1, 28, 1940.
^ itspekhi Khimiyi" (Achievements of
Chemistry), 22, 62, 1953.
^ Trudy TsentrMnogo nauchno-issle-
dovatellskogo institute. NKTP (Proceed-
ings of the Central Scientific Re-
search of the Peoples Commissariat of
Fuel Industry), 2, 17, 1945.
-672-
194. E. V. Chelishchev
195. V. I. Karmazin
196. W. Alberts
197. V. S. Koch?
198. G. N. Oyks, Yu. M.
Maksimov and Ye. A.
KaluzhinskY
199. L. P. Vladimorov and
A. N. Zotenko
200. V. I. Yavoysky and
B. A. Pupyrev
201. B. Kalling
202. V. I. Yavoysky
203. A. Field
204. E. Jette
205. G. Schenk, W. Riess,
and E. O. Bruegeman
206. N. M. Chuyko
207. E. Deloisy
2080 T. Dancy
2090 A. Jay
210. W. Mackenzie
211. E. I. Nesis and Ya.
Frankel
2120 0. N. Smirnov and
B. F. Ormont
- Iketallurf" (Metallurgist) 3, 11, 1940.
? "Teoriya 1 praktika metallurgiyi"
(Theory and Practices of Metallurgy),
10-11, 15, 1939.
^ 'NET und Eisen,' 117, 1931.
? *Stall' (Steel), 8, 698, 1947.
- TFOTMoscovskogrinstituta stali
(Proceedings of the Moscow Institute
of Steel), ed. 28, 46, 1949.
- "Teoriya i praktika metallurgi i"
(Theory and Practices of Metallurgy),
10-11, 1939.
a^ lrfiElil" (Steel), 12, 1948.
- "Jernkotorets Annaler," 3, 1937.
- Gazy v vannakh staleplavillrykh peohey
(Gases in the Baths of Steel Smelting
Furnaces) Metallurgizdat, 1952.
- "Trans. of the A.I.M.M?E:," 90, 23,
1930.
- *Trans. of the A.I.M.M.E.," 91, 80,
1931.
? 'Stahl und Eisen,' 38, 562, 1932.
- 'Stall" (Steel), 4, 1941, and other
x7E17.
? "Revue de Metallurgie," 369, 1926, (as
under 84) II, 5, 642, 19777
"Problemy soveremennoy metallurgiyi
(Problems of Modern Metallurgy), ....,
28, 1952.
^ "Journal of the Iron & Steel Insti-
tute," 157, 167, 1947.
? *Ertiirfron and Steel Institute,"
36, 1946.
I. - 12hurna1 tekhnichesk
213. H. Becker, F. Garders,
and G. Kornfeld
214. O. D. Hvolson
215. Ya. B. Aron and Ya. I.
Frankel
216. Ya. B. Aron
217. G. M. Chick?
Declassified in Part - Sanitized Copy Approved for Release @ 50-Yr 2014/05/01 : CIA-RDP81-01043R003400070002-1
fiziki" (Journal
of Industri1Physiosj T50O, 1952.
^ "Doklady Akademi i nauk SSSR" (Reports
of the USSR Academy of Sciences), 82,
751, 19520
"Archiv fuer das Eisenhuettenwesen,"
20, 363, 1949.
- Kursk fiziki (Manual on Physics), vol.
III, 504, 1923.
- "Zhurnal ekelerimentallnoy i teoreti-
oheskoy fizisi* (Journal of Experi-
mental and Theoretical Physics) 19,
807, 1949.
? *Zhurnal fizicheskoy khimiyi* (Journal
of Physical Chemistry) 24, 82, 1950.
cap]ic,Inilchrez- Ranchet1=_?ov (Cal-
culation of drop-shaped reservoirs),
Moscow-Leningrad, 1951.
-673..
218. S. I. Pope)., O. A. Yessin -
and Yu. P. Nikitin
219. E. I. Nesis
220. Ya. I. Frenkel
221. A. N. Frumkin, V. S. Bag-
otsky, Z. A. Yoffa and
B. N. Kabanov
222. M. Jacob
223. M. A. Mikheyev
224. I. A. Andreyev
225. I. A. Andreyev and
L. C. Matykha
226. S. I. Sapiro
227. S. I. Sapiro
228. A. M. Levin
229. F. Koerber and
W. Oelsen
230. R. Wenzel
231. B. V. Deryagin
232. P. A. Rebinder, N.
Kalinovskaya, N. Lipets
et al.
233. Sh. G. Dzhadzhgava
234. A. I. Belyayev and E. A.
Zhemchuzhina
235. A. S. Berezhnoy
Declassified in Part - Sanitized Copy Approved for Release
"Doklady Akademiyi nauk SSSR" (Re-
ports of the USSR Academy of Soiences),
83, 253, 1952.
^ vihurnal tekhnicheskoy fiziki" (Jour-
nal of Indtstrial Physics) fF, 1506,
1952.
- Zhurnal eksperimentaltnoy i teoretiches -
koy fiziki Journal of Experimental and
Theoretical Physics) 18, 659, 1948.
^ Kinetika elektrodynkh protsessov, -
daniye Moskovskogo Gosudarstvennogo
universiteta (Kinetics of the Electrode
Processes, pub. by Moscow State Uni-
versity), 222, 1952.
^ "Zeitschrift des Vereines Deutscher
Ingenieurs (V.D.I.)? 76, 1161, 1932.
? Osnovi tezloperedachi (Principles of
Heat Transfer) Energoisdat, 1947.
- "Syulletininauchno-issledovatellskikh
(:ueil of the itific-Research
Studies of the Dzerzhinsky Works), 1934.
^ "Urallskaya metallur iya" (Metallurgy
of the Urals) 9, 21, 939; 9, 12, 1940.
? "Stal", (Steer), 3, 395, 047; 7-8,
449, 1946.
? Trudy Stalinskogo oblastnogo otdelenia
NITOM, (Transaarons of the StaIinsk
BETga Division of the Scientific-Re-
search and Tech. Society of Metallur-
gists), 1, 21, 1949; 8, 35, 1950, etc.
? Sbornik Trudov DneprOlietrovskogo metal -
lurgioheskogo institute (Collection of
Studies of the Dnepropetrovsk Metallur-
gical Institute) issue 28, 80, 1952.
- *Mitteilungen K.-Wilhelm Institut fuer
Eisenforschung.," 17, 39, 1935.
- 'Industrial and Engineering Chemistry,'
28, 988, 1936.
- rffoklady Akademiyi nauk SSSR" (Reports
of the USSR Academy of Sciemcies), 51,
357, 1946.
- Issledovani a v oblasti poverkhnostnikh
yavIeni 4, Investigations in the Region
of Surface Phenomena ONTI, 1936.
- "Doklady Akademiyi nauk SSSR" (Reports
of the USSR Academy of Sciences), 70,
417, 1950.
- Poverkhnostnyye yavleni a v metallur-
gioheakikh protsessakh Surface Pheno-
mena in Metallurgical 'Processes),
Metallurgizdat, 1952.
- "Stalt" (Steel), I, 1948.
-674-
Declassified in Part - Sanitized Copy Approved for Rel
?
? 50-Yr 2014/05/01
: CIA-RDP81-01043R003400070002-1
236.
A. N. Lozhkin and
I. R. Izraelit
"Zhurnal tekhnichesk fiziki" (Journal
of Industrial Physic), 9, 21/4, 1939.
237.
A. N. Lozhkin and -
Ibid, 8, 1872, 1938.
P. I. Krol
238.
A. L. Jacob and -
Industrial and Engineering Chemistry,"
L. C. Bischman
40, 1360, 1948.
239.
M. Kornfeld -
Wiugostl iTroohnostf zhidkosti
to.
(Elasticity and Strength of Liquid),
75-81, GTTI, 1951.
240.
I. A. Andreyev
Trudy TsNII NKTP (Studies of the Central
Scientific-Research Institute of the
Peoples Commissariat for Fuel Industry),
24, 5, 1945.
241.
S. I. Sapiro
YFUay Stalinskogo oblastnogo otdeleniya
NITOM, (Transactions of the Stalinsk
Oblast Division of the Scientific-Re-
search and Tech. Society of Metallur-
gists), 1, 21, 1949.
242.
S. I. Sapiro
"Stall", ?(Steel), 3, 20, 1945.
243.
S. I. Sapiro
Ibid, 7-8, 449, 19T6.
244.
S. I. Sapiro
TER, 37395, 1947.
245.
T. Ondoxin
Trans.?of the A.I.M.M.E., Open-Hearth
Proceedings, 304, 1949.
246.
V. G. Levioh
Fiziko-khimichesk a gidrodymamika
(Physico-chemicalhydrodynamics) pub.
USSR Academy of Sciences, 1952.
247.
D. L. Tamud and
Poverkhnostnyie yavlen4a (Surface
S. E. Brasier
Phenomenaj GTTI, 1934.
248.
G. O. Yerchikovcky
Obrazovaniye floatatsion pe (For-
mation of Flotation Froth) GO I, 1939.
249.
E. K. Venstrem and
P. A. Rebinder
"Zhurnal fizicheskoy khimiyi" (Journal
of Physical Chemistry) 7, 754, 1931.
250.
G. Holmes
Rukovodstvo k laboratorriym rabotam po
imi i-(Hand book for Lib fork in
Chemistry) 81, ONTI, 1936.
251.
B. V. Deryagin
"Kolloidnyi zhurnal" (Colloid Journal),
6, 291, 1940.
252.
B. V. Deryagin
l'Priroda" (Nature) 2, 23, 1943.
253.
P. A. Rebinder and
K. A. Pospelor
in the book by Clayton 'Emulsions',
GIIL, 1950.
254.
W. Clayton
Emylsiyi (EMulsions), GIIL, 1950.
255.
P. A. Rebinder and
A. A. Trapeznikov
"Zhurnal fizicheakoy khimiyi" (Journal
of Physical Chemistry) 12, 573, 19380
2560
A. A. Trapeznikov
Ibid, 12, 673, 1938.
257.
D. L. Talmud, S. D.
SUkhovollskaya and
Prochnostt adsorbtsionnilch sloyev,
Prilozheniye k zhurnalu "Tsveniye metal-
/
A
N. M. Lubman
ly"(Stability of Adsorption Layers,
Appendix to the journal "Non-Ferrous
Metals") Moscow, 1930.
258.
P. A. Rebinder -
'Izves#ya Akademiyi nauk SSSR, Otdel-
eniye tekhnicheskikh nauk, (Bulletin of
the USSR Academy of Sciences, Dept. of
Technical Sciences), 5, 639, 1936.
50 -Yr
/0 .
-
1-01043Rnmannn
Declassified in Part - Sanitized Copy Approved for Release ? 50-Yr 2014/05/01 : CIA-RDP81-01043R003400070002-1
f
t-
259. S. G. Mokrushin
et al.
260. I. P. Samik
261. B. Ya. Teitelbaum
et al.
262. B. Ya. Teitelbaum et al
265. B. Ya. Teitelbaum et al
264. P. V. Umrikhin
265. V. P. Revebtsov and
L. S. Rybakar
266. G. Davis
267. B. M. Larsen
268. V. O. Filbrook
269. V. S. Koch?
270. S. L. Levin-
271. V. V. Pomerantsev and
S. I. Syrkin
272. as cited by
S. S. Kutateladze
273. A. I. Kholodov
274. H. Flood and
C. Griotham
- "Kolloidnyi zhurnal" (Colloid Jour..
nal), 12, 448, 1950; *Zhurnal Prik-
ladnoi khimiyi* (Journiir-07Eigria
Chemistry) 26, 143, 1952.
- Vyazkostt zaakostei i killoidnykh
rastvorov (Viscosity of Liquids and
Colloid Solutions) pub. USSR Aoademy
of Sciences, 1941.
? "Izvestiya Kazanskogo filiyala Akademiyi
nauk," seriyo.khimicheska a (Bulletin
of the Kazan' Branch of the Academy of
Sciences, chemical series), 1, 26, 106,
1951.
? "Zhurnal fizicheskoy khimiyi* (Journal
or Physical Chemistry) 25, 911, 1043D
1951.
- 'Kolloidnyi zhurnal" (Colloid Journal),
12, 375,1.950; 14, 372, 1952; 14, g?2,
I752.
- "Urallska a Metallurgi a (Metallurgy
o the Urals), 1936.
- Issledovaniye osnovnogo martenovsko o
protsessa pri radote na glinozemistikh
shlkakh (Investigation of the Basic
517a=Teirth Process during Operations
with Alumina Slags), Sverdlovsk, 1939.
- "Trans. of the A.I.M.M.E." Open-Hearth
Proceedings, 198, 1949.
^ Ibid, 231, 1949.
- ITU, 234, 1949.
- 111W.11", (Steel), 2-3, 1945; Trudy Ural'
sEiTindustriyallnogo instituta (Trans.
of the Ural Industrial Institute), 26,
60, 1948.
^ Sbornik trudov Dnepropetrovsko o metal-
lurgicheskogo instftuta (Colleo ion of
studies of the Dnepropetrovsk Metal-
lurgioal Institute( 19, 1949.
- Proceedings of the TsiTI, 8, 1936.
- Teploperedacha pri condensatsiyi i
kipeniyi (Heat Transfer during Conden-
sation and Boiling) Leningrad, 1952.
- Sbornik state "Proizvodstvo
(Collection o articles "Steel Produc-
tion"), Metallurgizdat, 1952.
^ 'Journal of the Iron and Steel In-
stitute," 177, 61, 1952.
-676-
Declassified in Part - Sanitized Copy Approved for Release ? 50-Yr 2014/05/01 : CIA-RDP81-01043R003400070002-1