O N BLAST FURNACE SLAGS AND T H E F U S I B I L I T Y O F SILICATES.
BY AUGKST~ J. ROSSI. (Concluded.) An examination of a few specimens of natural silicates occnrring as ordinary minerals and of which the composition is very often reproduced artificially in blast furnaces, will corroborate the preceding statement that, in a slag, in R multiple silicate, t h e silicates of the bases RO and R,O, are of the same type. The following examples are taken from a table of silicates by Prof. Eggleston, of Colnmbia College, in which the old formula for silica, SiO,, is used : Orthoclase.*-0Ratioi
R,O : R,O, : SiO, 12 or
3Si0, (Trisilicate of Percy), or, written with Totnl-0 of SiO, : 0 of bases: : 24 : 8=12 : 4 = 3 : 1, a s e s p i - m i d silicate; the 0 ratio for each silicate is 3 : 1 in both formulE, as for the mineralitself. I t is the tri-silicate of the present time derived from H,Si,O, =2(H,O)3 SiO,. 0 ratio G : 2 x 3 : 1. Sodalile.*--l : 3 : 4. 3Na0, Si0,+3(81,0,, SiO,). (Tribasic or monosilicatc of Percy). Formula, with SiO,, 2Ka0 SiO,+ 2R,O33Si0,. Bibnsic silicnte. 0 ratio 1: 1 for each silicate, for the compound 12 : 1 2 = l : 1; derived from H 4 S i 0 4 = 2 ( H z 0 ) SiO,, a monosilicate or ortliosilicate with 0 ratio ; 2 : 2 = 1 : 1. Anorthi/e.---l : 3 : 4. 3Ca0, Si03+3(A1,0,, SiO,) the same as socialite, 0 of acid, : 0 of bases=3 : 3 = 1 : 1; or with SiO,, 2Ca0 SiO, +2A1,0, 3Si0, a bibasic silicate. 0 of acid : 0 RO, SiO,+AI,O,,
SiO,, 2RO 3SiOz+2R,0, OSiO,.
-
* Found in the United States. 307
308
BLAST P C R F S C E SLAGS, ETC.
of base=Z : 2 = G : G-1 : 1 for each silicate and for the compound. Epidote.-1 : 2 : 3 or 31 1. ‘2 ltOLCaU,48. 0s 0 ratio t! : 1 t h e quantity of Ii10 (magnesia for example) which replaces ~b certain quantity of lime bears a direct proportion to their saturation for SiO,, which, for this class of oxide NO, is as to t h a t of their equivalent Ill0 and HO, it isevident that the silicate will p i - e -
317
BLAST FURNACE SLAGS, ETC.
serve its type, the oxygeii &io qf the silica renini,ai?tg the same and being still double that of the base. This can be established in an absolutely general manner as follows : lo. Let Si0 , = a CaO =It6 =b be a regular type of silicate of any kind for which we know t h a t s 2 0 of silica =k, 0 of base. 0 of SiO, =- a. 0 of CaO = -b, N
15 and we have, as characteristic of the compound, s ?b
x k.
-85-
15
c 1
Let 116 first mplace any part b' of the weight b of lime by a weight b" of MgO, saturating the same qnantity of silica as b' of lime, thus b"magnesia : b' lime : : 20 equiv. of MgO or 40 atomic weight of MgO : ?S eq. of lime or 5 6 atomic weight of lime b" 0 5 = 5 : ; - , b"=-b', b' 7 the composition of the silicate mag be written then : SiO, = a CaO = b - b '
-__.
P,
5
MgO = b" = - b' Iv
the 0 of the silica in the new compound 8 Pb - a=- x K
-_
15
I
2
as above.
0 of bases
P
- ( b - b') 0 of lime l
2b' -2b7 - 2b' -+7 7
0 of bases : =
2
5b'
5
7
+ - x - 0 of MgO 2b
=
7
;11-?m b u t , if -- - 1 wc rclturii to the oi-ipinal silicate, this will 11apl)eii if n
= m.
it-.lm
m=l then 11
SiO,=a
111-2m 1: I)=
= = jii
b :tiid tliccompositioii becomes
5m
CnO = b
MgO = b-b' = b"= 0 it should be.
! 0 of
)
21)
acid - x k 4
=0
'2 b of bases - x k 11s 1
319
BLAST F U R N A C E SLAGS, ETC.
3'. I n the same manner if in the silicate
{ CaO-
sio2z wea replace
b'
of lime by a weight b" of tilnmina equivalent as saturation for silib' ca (see Table 111. a ) =1.631 We have for the new composition
i
SiO, = a CaO = b-b' Also, = b k b " 1.631
This number 1.631 has been obtained by the consideration t h a t for saturation of silica, for the same type (neutral) 1RO = 28 saturates 1Si0, = 30 or 3 R 0 = 84 saturate 8Si0, = 90, while 1A1,0, =51.50 saturates 3Si0, =90. Hence: 84 lime are equivalent to 51.50 81,0,, since they saturate the same quantity 90 of Si0 51.50 84 Hence : 1 lime= -A1,0, or 1A1,0, = --- 1.631 CaO. 84 51.50 84 1.631=61.50 b' b'x51.50 a n d we have : b" =-, The composition of the sili84 (5%) cate can then be written : SiO, -a CaO =b-b' A1,0, b' X 51.50
,.
i
84
I
2
1
r
51.50)
24
I b ' x - 1 x -= I 84 J 51.50 \ / \ \/ \ /'
0 of bases=- * (b-b')+ CaO
'"' 2b'
0 of AI10,
= 1
320
BLAST F U R N A C E SLAGS, ETC.
(2b
?b
I _ _ _
'!4b
I+
P
!b
2b
Zb
;
*
.
_ _ = _ _+ --
7 , J
i'; O of acid= --a
1
84
#
I
? i, =- x k
,
1 -5
2b r ' 1 o t a l O o f bases=----+
2b'
2b'
2b
*
n
u
1
1
i
- and ---
1
'2 b 0 of acid = 0 of bases - X
k as before, the type has not been
t-
changed. Let us calcnlate the formula of the compound in this case. Sb-Zb' 2b' 0 of It0 =---; 0 of X2O3 =c
1
1
s
26
Total 0 of a c i d = - a a 15 '?cj
Total 0 of bases=--
7
;
x k.
ci 0, 0 of
-< I + I
'Jb-zb' 1iO = -n 6
Zb'
O,, 0 of R e o 3 = ; i
.?b ?b Oofacid-xk=Oofbases--x
7 Proportioning t h e ' 0 of SiO, :
k.
7
1 (Ib-"')
(ab
0 of silicacombined witliRO=
-x k I -( ___ 7 J
2b'7
O o f silica combined with
Reo3
Zb I
and formula is : (' 2b-2b' 7
(2b-2b'
k-k
en-
Zb'
I-
I $
( 2b-2b' Oofacidk
7
1
1
-
I
7
=
321
BLAST FURNACE SLAGS, ETC.
0 of bases
o of acid k
[
2b'
J
[ 2b'
=O of base
-
1x
k
( Y J
The silicates in RO and R,O, are of the same type as k x 0 of base and of the same type as the compound itself. We could prove in the same manner that if the weight b" of Al,O, were not equivalent for saturation of silica to that of the lime i t replaces in the compbund, the type of the silicate, the oxygen ratio, is changed, b u t remains the same for the silicates in RO and E803as tlie new ratio of the componnd itself. Hence: Let a slag be expressed with SiO, or SiO,, as t h e symbol for silica ; let the equivalents or the atomic weights be used, the empirical or rational formulir, the nomenclature neutral, e. g., sesquiacid, sesquibasic- - - - -slag, or the more modern names derived from hypothetical radicals be employed : in all cases the oxygen ratio remains the same, as it should be. This ratio constitutes, so to speak, the individual feature of a silicate, it decides as to its greater or lesser fusibility, its possible coexistence with certain grades of iron, and, since the transformation of all t h e bases into their equivalent of lime does not modify i t in any manner, as we have just seen, we possess, then, in this method an absolutely exact and easy mode of comparison of two slags. No notation or symbol of any kind is required ; we have to deal only witli~figii~es representing the composition of tlie substance as given by the analysis itself ; the relative character of two silicates can be judged a t once, the tables of references furnishing all the other d a h pertaining to each case.
C'AI,CULATION OP Sr,Ac,.-This method of transformation of all the constituents into lime will be found very advantageous in calculating a slag. It .simplifies the work considerably, and from the beginning to the end the operations can be performed by t~ non-tcdinicnl person without using any symbol 01-formula.
Assume that we have position is :
ii
mixtur(f of ores. of which the average com-
We use a limestone of comp(raition :
1 ,yig9
'
L
Metallic iron, 50s
l(x1.25
6.l)O 1 And anthracite 1.15 1 coal with S ? S % ash, of whlch , 19.00, ash the compo- { 44.20 1 aition in per ct. -I of coal i f i 1011.35 I
80.00 I.
MgO ~
O,lo o.10
6.28%
We have decided to obtain a slag of such ~1 character that t h e fusibility will be about that of a se.spihctsic .sl/iy, t h a t is, if preferred, of one iii which 0 ratio is 1:3. Looking a t the table we see that, for such 5 type, 1 of lime satnr:itcis 0.7'14 of silica, or 1 of silica takes u p 1.400 of lime. Assuming a n y proper amount of coal per ton of ore srnelted :ind, in most cascs, 0.7.5 toii is all that 1s required, m e have a l l the datiL ncccssrtry for our calculatious. Transform ct2Z
t?tc aiiuI!lse.s iuto 7r'vic : 01th:.
Si0 =20.00$- - - ...- - .- -20.00
j
AI2b3=3..)0X1.G31 ==5.92$) 'l'lie ore ist siO, =20.00s = 12.11$ erinivzllen Ca0=3.10 CaO -12.11% irgo=z.cox 1.40 CaO per ton to >I I1 O,=O. 90 = .I5 J
2:;: 1
STOKE.
1
COAL.
__
SiO,=. - - - - - - - .- - - 3 . 3 5 g Si(),=. - - - - .- - - - - - _ _ -6.00 A l , 0 3 = l , l j ~ 1 . G 3 1 = : 1 . 8 7 ) 58.47%~i120,=9.73X1,ti31=4.45 10 (';iO=30. . _ - .-=30.00 Lime. ('aO- - __ ._ _ _ _ _ ..- -=0.10 0 Mgo - .- ..- .. - .- -=o. 14 I( e iygo=ig x 1.40 - =2G. (io
5
IIence, a s we use only 2 ton of coal per toii of ore, the coal used is e q u i v d e n t to 9 of t h e above analysis, o r : SiO2-- - -2.3'1 CaO- - - - 3 . 3 2 ---
SiO,=2O+S.~S=SiO, ---22.52 per ton of ore; the coal and ores are equivalent to CaO = 12.11+ 3.52 C:rO- -13.63. Since, to make the proper silicate, 1 of lime ; -
323
BLAST FURNACE SLAGS, ETC.
takes up 0.714 of silica, the 15.63 of lime in coal and ores will take up: 0.714 x 15.63 =11.16$ of silica, leaving as f 7 - e ~silica in t h e ore and coal 22.52 - 11.1G = 11.36 SiO, to saturate with limestone. T h e G$ of silica qf the stone will require, a t the rate of 1.400 lbs. lime per lb. of SiOp, G x 1.40 = 8.40 lime, leaving of .free lime or the equivalent in the limestone, 58.4'7 - 8.40 or 50.0'7.free Zinze. We have to saturate in coal and ores, 11.X free silica. A t t h e rate of ssturatton adopted, it will take: 11.3G x 1.40 lime = 15.91 lime; we have 50.07 free lime in 1 ton of limestone, we require only 15.91 of lime to strtttrnte f h e SiO, in coal and ores, hence, we need only per tot, qf'ore n i i d 2 t o i l c o d , 1,;. !I 1 --
s
0.J17
.j0. 0 T t o n of stone. Tlic c l ~ n r p iti'c' s thus: I ton of ore, 0.75 ton of coal, 0.317 ton of limestone and, as the ore contains 50% qf irow, w e repiire :
2 tone ore p e r f o i l o f p i g iiintle. 1.50 tons coal 0.634 ton stone The composition of the slag is : ,Cilicn in ore and coal per ton 1 of ore and per 3 ton of coal ( - - - - - - - - - - - - 520
i
'"'*
In stone GX0.317 ton-. - - - - - - - _ _ _ - - - - - - - 1.909 Iota1 SiO,--. _ _ - - - _. - - - _ _ _ - - - .24.-!28 Lime iu 3 ton c o d and 1ton ore_(per ton ore) - 1LGR I n stone 0.315 ton x 58.47$=- - - - - .- - _ _ - - - - - -18.59 Total lime- - - - - - - - .- - - - - - - - - - - - - -34.22 and composition of slag is : SiO, = 24.488 or reducing to a SiO, = 41.66 CaO = 34.8220 percentage : CaO = 58.34 I 1
--
--
58. G48
100.00
exactly a sespuibasic silicate. See Table V. Using the preceding charges of ores, stone and coal we should have every reason to expect a slag of the above composition or of one very close to it.
We have :idopted 14 toil coal p e r ton of pig. If it were found that this could be redoced it slionlcl be clone and tlic sing would hardly be modified i i i gc:ner:~l chariictcr IJJ this c h r g e . If greater accuracy were iiecessary the I)reccdiiig calculations could be made over ag:iiii with tlic i L L ' / ! * r h t r r ~ j c si i i coal; but, 1)r;t(:tically, it, is absolutely useless, t h c t :ish of cwal onteriiig, iis it niay be seen, as ii sinull pciw?iit:tge i n t o tile gerieral compositioii. \Vitli inferior cokes o r aiitliracite it bcootiies i i i i iinportarit factor n o t t o be qioylecterl but too qf'leii ~ y m m d . Coki.s w i t h 135 of :tsli are n o t aincommoii in ct:rtain lociilities. As ;rn es:rrnl)lo of tlie close coiiicitlence betwecn slags :tctually ~ 1 1 1 1 from kiiowii calc~ilitted cliiirges and the slag cleteimiiiied a priori wc quote the folloivirig slag ruii iii :L furiiace (io feet high, 1 G feet bosh, riiiiiiiiig o i l hot blast 850 to 900" 1.'. Prossure of blast 7+ lbs., hnierican ~ U ~ I I R C L 'anthriicite , coal. 'L'lie analyses of materials were as follows : stom.
Ores.
SiO, - - - - - _ _ -23.21 AI,O3. _ _ _ _ _ ._ _ 4.51 CRO _ . . - - _ _ 1.61 --_
COd.
9.90 3.85 2S.00 NgO _ _ _ _ . :1*11 _ _ _ _ l(j.00 _ Alkalies _ _ _ _ _ _ _ 2.67 ..__
Mn,O,
_ _ _ _ - - - traces
3.00 2.30 0.10 0. os
____ ____
..__
P205_ _ _ . _ 0.31 _ _ _ _ . _ _ _ --_.. . -_ -_ _ - _0.0s .--. __-_
s--
__-
-- --
Making the calculations proportionally to the quantity of the different materials charged, we find t h a t the slag contained : Metallic iron 46.46%.
Charges : 8cwt. l q Olb. 924 lbs. Silica - _.._ ____________ Alumina _ _ _ _ _ _ _ _ _ _ _ _ _ _ Lime _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ Magnesia _ _ _ _ _ _ _ _ _ _ _ _ _ _ --___ Alkalies __.._ Mang. oxide _ _ _ _ _ _ _ _ _ _ _
____
Total weight of slag
S e w Jersey Dolomite.
3cwt. l q 141bs. 378 1bs. Ores.
215.38 41.67 14.90 31.50
24.67 Traces
Stone.
37.42 14.66 105.84 60.48
___-
Lehigh Summit.
5-1-0 588 I bs.
Coal. 17.64 13.52 0.59
0.47
_--_-_-
Total.
270.44 lbs. 69.85 " 121.33 " 92.45 '' 24.67 ((
_ _ _ _ _ _ _ _ _ _ _ _ _ _ 578.74 Ibs.
____
-__---- - _ _ _
_--_
or ~n:tiiyslags Lvc 1i;tw also tlie analyses of the orcs, stoiie arid couls. :tnd tlic charges, b u t they d o not figurc in this paper. 111 tlw tixiisformatioil of the slags into lime, we have often stopped the operations a t tlic first decimal ; i t is ii snfiicient approximation for ;t study of sncli compounds.
B L A S T F U H N A C E SLAGS, ETC.
... ... ... ... ... ... .. .. .. .. .. .. ... .. .. .. .
. . . . . . . . . . ... .. .. .. . .. . . .. .. ... .. . . . . . i. :. : . : . : .. .: . . :. .. ....::::: : ,. ... : . .. .. .. : . ,. .. . . . . . . I . . .
. . . . :. .
; m : ; i : : :
~.,j&O"; A arbZZr%?e"i:vj 2
. . . . . ... ... ... ... ... . . . . .
. . . . ... ... ... ... ... . . . . .
328
HLBST F U R N A C E SLAGS, +:'l'('.
.. .. . . . . .. .. . . . .. .. .. . . . . . . . . . . . .
.. .. .. .. .. .. .. .. ... ,... . ... . .. . .. . .. . . . . .. .
:A,::::: 0, m4u2i~gni
&&%O+
. . ... ... , . .., ... io"
*a"
.. .. .. ..
... . ... .. .. . . ...
.. ... ...
... . ... .. .. . . .. .
.. .. ...
, . . .. . . . .. ,. ... ... , . ...
BLAST E U R N A C E SLAGS, ETC.
329
I
Id
t
2
I
Y
I
a -4
330
BLAST I'CRNAC'E S L A G b , ETC.
(1.) Edsken, Sweden.-Hot blast, 572" Farh.; height of furnace, 44' to 48' ; boshes, 7'4" ; blast, Ok.07 to Ok.105; average, 0.0875 ky. per o centimetre = 1.20 Ibs. per square inch ; very fluid slag. ( 2 . ) Dowlais furnace.-Ores, mixture of argillaceous carbonates and hematites ; iron run ; generally gray Yo. '2 and 1-0. 3.
(3.) Dowlais furnace. (4.) Westplialia slag, leek green, t h e quantity of FeO, lO.99$, is remarkably high ; vitreous and conchoidal fracture. ( 5 . ) Argillaceous ores, enamel like, bluish white, wood furnace, Sweden. (6.) Swedish furnace, wood a s combustible, green and vitreous slag ; spathic cltrbonates and brown hematites ; air veTy hot. ( 7 , ) Spiegeleisen. (8.) 110. Spiegeleiseii Swedish furnace. ( ! I . ) Hot blast, Dowlais. (10.) Cold blast, English. (11)-(12.) Argillaceous ores ; pisolithic ores and puddle cinders. (13)-( 14.) Aberdare U. furnaces, argillaceous ores, cold blast. (15)-(16,) dberdare furnaces, hot blast 315" C:. (L7.) Coke furnace, Tonlaw. (1%) Lake and bog ores, cold air 0.07 kilo. per em,, Tinspong, Sweden, charcoal furnace. (19.) Manganiferous ores containing 37.19% mang. oxide, Cleveland furnace. (20.) BessAges, I3elgium, hot blast 536" Farli. : pressure, P . 1 2 to Om.14. (21)-(2?-(?3.) I)o. (24)-(25.) Bei.ge Borbeck furnace, hematites and argillaceous ores, poor ; avcr:ig:e, 35$ iron ; coke furnace, 47 ft. high x 13'tJ" boshes ; crucible, 5':;" diam. ; t e m p e r a t u x of air, 180" C. ; pressure, 0.1Gm. mercury per square centimetre. (26). Austrian furnace, hematites and magnetites ; hot blast ; pressure, 0.OE9ni. mercury ; coke furnace, height 45'; b;oshes, 13'6"; hearth, 3'4": temperature blast, 150°C'. ; pressure in lbs., 2 3 to 3 lbs. ( 2 7 . ) Hnmm furnace slag, manganiferous ores ; manganese oxide in slag :!5.84$. (28.) 33.06% MnO in slag. (29.) Miisen furnace, Spiegeleisen 26.20$ MnO in slag. (30.) Miken slag, 6.80 FeO in dag, 540 MnO. (31.) Do., Spiegeleisen MnO in slag 33.40s. (32.1 Cold blast I-lamm furnace, slag containing 29.2Og MnO and 21.50 FeO. (33.) Terreiioire (France). (34.) Scotch iron. ( 3 5 . ) Do. ( 3 6 ) - ( 3 7 . ) Besseges (Belgium) furnaces, 20$ puddle cinders added to charges for slag, Xo. 37. (38.) English slag, Dowlais furnace ; average. (30.) Slag from Colorado furnace; H o t Spring and Calumet ores, mixture of
BLAST FURNACE SLAGS, ETC.
331
rich magnetites 58$ iron ; and rich hematites 4 5 s iron ; Bessemer ores 25$, puddle cinders containing 9.4$$ titanic acid added regularly to the charges. Iron No. 1 and No. 2 Foundry, very tough ; coke furnace 70 ft. high, 1S$ ashes. (40.) Magnetite ores from New Jersey, very rich, general average, 50 to 5 5 4 . Iron, with gangue of free silica, hornblende, iron mica, furnace 45 feet high. Anthracite. Pressure of blast 3+ lbs. per square inch; temperature, 750 to 800" F. (41.) Same ores, with a mixture of one-half Staten Island hematite, containing 1.25$ chromium, which went into the iron making it white and hard. (42.) Half Staten Island hematite, half New Jersey mag(43.) All New Jersey netites, chromium in the pig iron, l . l O $ . magnetites. (44.) Do. (45.) Do. ; ores very silicious at the time, containing as much as 40% hornblende and iron mica, the furnace being only 45 feet high, they did not have time to be reduced, got into the hearth, fused and passed out as slag, which explains the anomaly of such an acid slag in connection with a gray forge iron. (4G.) Staten Island hematites added to the charges, making bhe iron white; chromium in the iron, 1.20$. (47.) Regular N. J. hematites. (45)-(49)-(50.) Do. (51.) Same ores, but furnace GO feet high ; prkssure of blast, G to t lbs.; temperature, 800 to 900" Fahr. (52) to ($0.) Do. Average charges for Pig, 1.00. Ores,
New Jersey magnetites generally. Stone, 1.00. New Jersey dolomites. Coal, 1.50. Lehigh a n t h r a c i t e and Scranton. 2.00.
T o compare these slags we have to transform them all into lime, and, as the approximations of a type to " basicity " is more logically expressed by the quantity of lime, we shall tabulate only the quantity of lime (calculated in some cases to only one decimal) to which all of the basic elements of the slags are equivalent. As some of the results fall between two types, the slag of the type to which i t comes the nearest will determine the designation. An acid slag containing 31.80%CaO and the next type, the sesquiacid,
332
BLAST FURNACE bLh(;b, ETC.
38.35% CaO ; (see tables) any slag which contains a quantity of 38.''-
lime greater than the a v e r a g e - - - ~ ~ ~ _ l = 3 6 . 0 8 %
of Iime will be
tabulated as sesgzciacid, if less as an acid slag. Any slag containing a quantity of lime greater than 4.3.J1$, the average of 38.35 (sasquiacid) and 45.2s (noutral), will be tabulated as neutral, below this figure as s e s p i a c i d . A n ? slag containing more than 53.3ZCaO$, the average of 45.25% (neutral) and 55.34$ sesquibasic, will be tabulated as sesyuibosic, below that figure as neutral. Any slag containing more than Sl.;YSCaO, the average of 55.34 neutral and 65.12 (bibasic), will be tabulated as bibasic, below this sespihnsic. Any slag containing more than 6'9.42%CaO average of 65.12 (bibasic) and 73.70 (tribasic) will be tabulated as tribasic, below this as bibasic. We give in a special column the amount of lime contained in each type, the figures being followed by capital letters, the first letters of the denomination of each type. Thus : A stands for acid, S.4 for sesquiacid, S for neutral, SB for sesquibasic, B for bibasic, 1 ' ' for tribasic. I n the next column we have put down the preceding figures, representing the limit of lime a t which a slag has been classified under the type nest in basicity. This was done in case of a possible approximation to one type, leaving in doubt the character of the slag ; but, as it will be noticed, it was riot necessary in any case. Another column contains the character of the slag determined as explained. The nest gives the character of t h e pig iron accompanying each slag, and the last contains such observations as may explain certain anomalies observed in the type in connection with the grade of iron, or such other special data regarding certain slags which make them rather abnormal. For details we refer to notes, pages 69 to 71, and to the Table VI., which gives the composition in full.
TABLE 171.
0.99 FeO and 5.80 MnO in Slag. Jerv hot blast. MnO 25.64. :no %3.96. Spiegeleieen. .old blaet. 1
?iiddly,cinders n ,! the chiwgeu. 51;?3 Bibasic.
Grav,No. 1. Gray, No. 2. Gruy, No. 2. Gray,No. 2. Gray. 1766.:10 B. 18 54.40 158.34 S . B 53.31 .Seeqiiibasic. Light gray. Mottled. Mn033.77; Fe0=4.39. Cold blast. 69.42 h'ribaeic. Gray, No. 1. ,78.70 1'. 19 72..-X HnO; 8.76. 201 65.20 135:!2 B. 't1162.20 B. 22 til.30 " B. 23 85.10 " B. 24 58.80 5 8 ; 3 S . B S. B 25 61.30 26 63.50 fi5.12B. 27 5:hN 58;64 S. B MnO 25.84%. 281 5 M i O M n O 38.96. S.B 29 40.40 4 8 ; B A'. % n o 28.20. 301 47.70 K. MnO 5.4C%; FeO 6.80. 31 4B.M " N. MnO 98.40. 32 5X.B7 58.34 S. B Ej.31 Sesquibasic. White. MnO 29.20; FeO 21.50. Gray No. 1. 61.73 Bihabic. 98 tii.00 .ti5.12 B. Scoidh gray, No. 1. ,54 fi9.Xl '73.72'1'. B 69.42 T r i b a w . Scotch ray, No. 2. 3!! 66.54 ti5;!2 B. 6I6T3 Blbasjc. Bibasic. Gray, $0. 1. 86 66.40 B. Graphitic Iron. Bibaeic. 37 813.40 " B. Gray. No.1 and No. 22 38 60.00 58.34S. B 53.31 'Seequibasic. White and mottled Gray, Nos.1 and 2.' % Puddled cinder containing 9.47 61.73 IBibasic. 39 63.07 65.12B. SiO.. 5:j.81 Sesquibauic. White. 61.78 Bibasic. White. '/B Staten Island Hematite. Chromium in ironmlfiring it white. ' I Bibasic. White. % Sttrten Is!and Hematite. !.25 Chrom. in iron making it white. '/ 'Bibasic. Gra 43 62.00 " B. 41 60.00 5 8 ; F S . B 53.31 ISeequlbasjc. Whrte. 'I !Sesquibasic. Gray forge. 45 53.30 .S.B 45$ Silicates of iron Hornblende and Mica In orenotrrdncednnd increasin unduly YiO 61.73 IBibaaic. /White. 461 65.00 165.12 B. '/B State! Igand Hemati?;. Chromium In iron. 47 59.70 i56;?4 9. B.' 53;U Sesquibasic. White. 9. B. Seequibeuic. White. 48 87.70 " S.B.~ " Sesquibasic. White. 49 58.00 '' S.B. ' I Sesquibasic. White. 50 56.40 51 80.40 I ' S.B.1 'I Sesquibasic Gra No 2. I' SePquibasic: W d l e and mottled. 52 56.10 '' S B: d. j 61.73 ,Bibask. Gray. 9. B.1 5%?l 'Sesquibasic. Light ay for e. Sesqnibasic. Yottleyand wgite. Y. B. S. B.! ' I Seequibasic. Mottled. 9.B.i '' Sesquibauic. Mottled. Qray, No. Band No. 3 B. 61.73 Bibasic. Bibasic. Gray. B. Qray, No. B x . Bo 64.10 B. '' Bibasic. Gray, No.2. 61 62.35 B. '' Bibasic. 62 57.90 58.34S.B.53.31 Se8qulbneic. Gray. Gray,No. 3. 63 6'2.80 65;!2 B. 61;.(5 Bibasic. Bibasic. Gniy, No. 2. 64 83.50 B. 85 61.80 " B. , '' Bibasic. Gray. No. 3. 66 59.20 58;?4 8. B 53;?l SeEquibaP/c. No. 3 and Mottled. Seequibasic. Mottled. U7 58.50 h. B. 68 60.00 '' S. B./ I' Sesqn!baeic. No. 3, Gray. No. 3, Gray dark. 89 61.80 65;!2 B. 61,.?3 Bibaslc. Bibasic.
'( Bibasic. '' Bibaaic. '' jBibaeic.
!
I
.
1
~
I
~
::
1
I
/.
.
liefcrriiig to tlic t1at;t coiitaiiitvl i i i tlic l ~ i ~ c ~ v ~t i i~l i li( :l iwe ~ i see ~ that : When the slags were * ' m x t i * d . " wlricli Ii:t1)pened lire times: the iron was four times wliite :tii(1 oricc' i i i c i t t l c ~ d . \\'lien they were 'i.ce.sy~~i6~~.sl'c.'' wliicli I i ~ i p p ~ n c d 11.2 tiiiic.s> the iron was t w l v c tiriirs white : thrcLcb tirries I v l i i t c :ind tnottlecl : f o u r times mottled : twicc light gri'y :1!if1 inottlcil, :t11(1 c i c v e ~ itimes gray : tlie shades of gixy Iwiiig mostly Iiglit. When they we'i.e i * h i h i c , " which 1i:tpperied 30 t,irnes, the iron WILS three tiiiies white ; twicx g : l i l t 1 lllottletl gray, a l l d ? 3 timw grity. gr:tpliitic So. 1. S o . . So. '1" aitd q o . ?, the darker sii:des corrcsl)oiidiiig most p ~ i i c ~ i ~I?i t tl o slags I*e:icIiiiig very n w r tlie typical bihasic sliig o r r r e ! ~goiiig bc~yoiidit i i i i1iiaiitity of lime. ?''/I/; i/i/*ce ~ ; ~ / S C . Swlic.n t,lir: i i . o i i \vas \vliitc" i t \VU due invariitbly t o tlie 1)rcseilce of c:liroiiiiuiii III tlie pig : cliroriiium makiiig t h e pig iron white i t i i i l 1i:trtl wlien present in c:urt:tin qiiaiitities : the conditions of the fririi:icc ere iiot iiurmal .io far as tlie study of t y p of s h g s is c o ~ i c e r ~ i o ~ l . When they were * . i/.ihrtsic." which 1iiipi)enecl twice, t h e iroii wus h o . 1 grapliitic: i r o i i . When tile! ~ w r u (icid?'' wliicli li:tp1wiied oiic(1, tlic iroii w:is gray. lye are justified in coticlucliiig a t lcltst f r o m tlie examples. quoted, that, with :t nentral slag tlic iroii c;tn be c:sp~ctecl t o be n o r ni a11y w li i t c , ex ce p t i oi i a1I y mot tl ed , 1i-itli a slag npproaching :t sesqiii1):isic slag. accordiiig to the conditioiis of tlie furiiace, lieiglit, terii1)ei'attire :ind pi'essurc of blast, tlic iron can be expected t o ;iA't:c:t tlie lightest grades of gray or to lie niottlec1,bcing wliite oiily i i i specid conditions, arid a darkc: gray if the composition of the slag is decidedly sesquibasic. Irons riin with such slags arc very adraiitageous 3s forge irons. r 1 Ilie pig coritairis n greater perceiittage of iron rtnd less silicoii and carbon than grayer pigs, coiitlitions very fiivorable in tl:e puddling procc?ss, which is intended to eliniinnte all siibstaiices foieigii to t h e irnr!. K i t h a bibusic slag, in n o r m a l conditions of ores and charges, w e caii expect invltrisbly a gray pig from the darkest shades to a i
a '
335
BLAST F U R N A C E SLAGS, ETC.
No. 2. No. 1, No.
2xx and kish cinders, if the composition in lime reaches over the type of bibasic No. 2 x and No. 2, if not quite up to the percentage of lime of the latter, as the case may be. With a tribasic slag the iron will invariably be a No. 1 iron, Scotch gray (judging from the limited number of examples of such slags, which are not very economical). An acid slag, as could be expected from diminished fusibility when the proportion of silica goes above a certain limit, corresponds to gray iron. The higher grades of iron contain the most silicas and carbon (total carbon), corresponding therefore to higher temperatures in a furnace, a condition necessary to insure the reduction of the silica and the passage of silica into the pig, as well as the solution of the carbori in the molten iron ; they are accompanied by the more basic slags. Blast furnace practice corroborates then completely t h e direct experiments on the fusibility of silicates which showed that the more basic they are the less their fusibility, and that, beyond a certain limit of acid, the silicates become again more refractory. If objections are made to the denomination “Basic,” (‘Bibasic,” ‘* Neutral,” slags used in this discussion, we can say : Putting aside all considerations as to the molecular constitution of silicates and taking only as a guide the oxygen ratio, a?t element of comparison which the aiialysis.fiirnisAes in all crises. The silicates in which the ratio of oxygen of SiO, to oxygen of (Seequibasic.)
(Bibasic.)
the bases is coinprisod between 4 : 3 and 4 : 4 (1 : 1) accompany most invariably the higher grades of iron ; when this ratio falls between 4 : 2 (neutral 2 : 1)and 4 : 3 (sesquibasic), mottled iron and lighter grades of gray may be expected, darker as the ratio approaches 4 : 3. At 4 : 2 (2 : l), white iron and occasionally mottled iron can be depended upon. When this ratio reaches 1 :4 (acid), the fnsibility appears to diminish again, and as a consequence, the darkest grades arc obtained. At 4 : 6 (2 : 3) (tribasic) graphitic iron and kish cinder are the
33G
I1LAST FUENACE SLAGS, ETC.
rule. If these deductions are correct, and, the more numerous the observations the better their t r u t h could be ascertained, their importance is not to be overlooked. The failure to obtain a certain grade of irou with a given slag within certain limits would then be a warning t o an intelligent iron mitster to look in his ores or other materials charged in the furnace, for a disturbing element or for abnormal conditions of temperature, volume and pressure of blast. It would call for R thorough chemical analysis. T h c same ores frequently change in composition. When silica is present in a combined state in an iron ore, i t s hornblende or iron mica, these silicates of iron may melt before being reduced, especially in t i low furnace, and pass o u t in the slag (1.9 S I I C ~ and , unreduced completely chknging its character as calculated from a “gcneral analysis,” or ‘*texpected’’ from :L previous practice with the same ores. Suqh a circumstiince did present itself in one of tlie slags quoted, KO. 45, I t showed a t the analysis 50.10 silica and 4.8s FeO, corresponding, transformed into lime, to 8 slag containing 5 3 . 4 0 CaO, just about tlie limit, 53.31, to be called a sesquibasic slag (58.34) ; i t was tabulated so, but, strictly speaking, it was nearer a neutral slag, 48.28. W i t h such a slag a n iron o f lighter gray than gray forge iron was t o be expected. I f a darker one was obtained, it wiis certainly due to tlie fact that, omitting tlie silicate qf irou melted without reduction in a low furnace, and existing simply as an i?iert nzatter in tlie liquid slag, t h e slag had :I different saturatioii for silica than the apparent one. A more complete and thorougli examination of tlie slags quoted would disclose for several of them other nnomalies, more or less important, but it would carry us too far. We have intended to show, only in a general manner, tliat the relation between the chemical character of a slag and grade of iron is much closer than, we believe, it is generally admitted t o be, leaving far more complete and numerous observations to determinc die question, nut forgetting t h e wise maxims of Sir Lowthian Bell and his reserve : If it is proper to give an opinion on a subject which one has not thoroughly experimented.” T h e following Tables V I I I . aiid I X recapitulate practically tlie results of this examination. KO f o r m u l s or symbols are used i n ((
BLAST. FURNACE SLAGS, ETC.
337
i t ; the bases being called by the names they aye generally known by. It is, so to speak, a practical table allowing any one not familiar with technical chemistry to transform a slag of any composition into lime and also to find the oxygen ratio if desired. Comparing the results with those tabulated, the character of t h e slag can be ascertained and permit its fusibility and the iron rrlikely” to be expected in normal conditions of furnace, as well as its relation with another one. The Table IX. of equivalence includes all t h e bases likely to be met with in slags, some as constant elements, others as occmional ones in certain localities. Table VIII. giving also the saturation of lime for silica, or inversely, for different types of slags of different fusibility these figures can readily be used, as explained previously, for the calculation, apriori, of a slag from certain charges in ores, stone and coal of which t h e analyses may be given:
H* U ci
3 t: c;:
c
339
BLAST F U R N A C E SLAQS, ETC.
TABLEIX. Bases likely to be found in slags. ~~
Equivalence in Lime.
Oxygen in 1 lb. of Base.
= 1.631 Lime 1lb. Alumina 1 Ib. Magnesia = 1.400 " 1 Ib. Potash = 0.594 " 1 Ib. Soda = 0.903 " 1 lb. Iron Oxide (FeO) = 0.780 1 Ib. Manganese Oxide (MnO) = 0.7SO 1 Ib. Baryta = 0.365 1 lb. Copper Oxide (CuO) = 0.704 1 lb. Lithia = 1.866 1lb. Nickel Oxide (NiO) = 0.747 1 lb. Iron Sesquioxide (Fe,O,) = 1.050 1 lb. Chromium Sesquioxide(Cr,O,) = 1.100 Iron and Chromium Sesquioxide, (Ferric and Chromic oxide,) t o be treated as Alumina. ((
0.466 0.400 0.170 0.258 0.222 0.225 0.104 0.201 0.550 0.213 0.300 0.314 0 of Silica: 0.533
