By R. C. TU C K ER , M.A. (Member)
It is the a u th o r’s intention to portray the present know ledge o f heat-resisting cast irons, with special reference to chrom ium cast irons.
He will only touch upon some of the highly- alloyed cast irons, Ni-Resist, Nicrosilal, Silal and C ralfer, because these are the results of investigations carried out by large research organisations, and there is a constant flow of reports concerning these alloys em anating from the interested organisations. There is no central industrial o r research organisation devoted to the study and com m ercial application of chrom ium alloys as in the case o f nickel and m olybdenum alloys. The result is that published d ata tend to have an “ interested ” bias, and it will do these alloys no harm for someone to point ou t in as unbiassed a m anner as pos
sible the results w hich can be obtained by the use o f chrom ium alone.
The sim plest type o f heat-resisting cast iron available to-day is “ hem atite,” w hich loose but traditional phrase covers a wide range of struc
tures and types. By weight, it is the most im portant heat-resisting cast iron and in spite of the long period over which it has been used in heat-resisting conditions there are still m any problem s associated with its use.
The Ingots C om m ittee o f the Iron and Steel Institute has, as all cast-iron metallurgists know, published some very interesting reports on the life o f “ hem atite ” cast-iron ingot m oulds, and their conclusions are roughly as follow s:—
(a) It is only possible to determ ine the factors affecting the life of ingot m oulds by the careful statistical analysis of a large num ber o f carefully controlled service tests.
(b) T he presence o f 10 per cent, o f steel in the cupola charge can have a beneficial effect in som e cases.
(c) T he presence o f up to 0.3 per cent, of phosphorus has a beneficial effect and is
not m ore com m on because of the widespread practice of introducing scrap moulds into the acid open-hearth process.
It is very likely that no such com plete analysis o f service results has been attem pted in any other type o f service. M ost of the service conditions of heat-resisting cast iron are not under strict control and service records unless of a statistical nature are distinctly mis
leading in m any cases.
“ H em atite ” cast iron does no t resist oxida
tion, is very prone to “ growth ” and is m echanically weak on “ static ” tests. Its only good quality is its ability to resist m ajor crack
ing caused by heat shock, and there seems to be distinct evidence that the worse it behaves when tested in the ordinary ways, the better will be its resistance to heat shock. The present m ethod of testing for heat-shock resistance is to w ater-quench 1-in. dia. test-pieces from 650 deg. C. a num ber of times. This test, like m any other laboratory tests, has m any lim ita
tions because it is difficult to introduce the factors of thickness and shape, and, in any case, alm ost any hem atite cast iron will resist this, test for a large num ber of quenches.
It has been found, however, in studying ingot moulds (loc. cit.) that certain m icrostructures are desirable for the avoidance of m ajor crack
ing (Fig. 1). T he coarse graphite is surrounded by ferrite with only small am ounts o f pearlite and other impurities, and this is perhaps the best test that can be applied in the absence o f service records.
Pig-Casting-Machine Moulds
Some years ago the author encountered som e surprising results in the m anufacture o f pig- casting-machine moulds, and a description of the case will, no doubt, raise a sym pathetic feeling in the minds of many foundrym em
T ra d itio n said th a t such m oulds should be m ade in p u re W est C oast hem atite. T he Ingot- m oulds Sub-com m ittee and private investiga
tion w ere beginning to p o int to the use o f some steel and phosphoric iron as indicated above.
Boldly, these were introduced and tw o sets of m oulds to the sam e design w ent through the shops a t the same tim e an d were installed in
discrim inately on two new m achines. In a few m onths, from one m achine, cam e a serious com plaint o f m ajor cracking, from the other a satisfactory report. Visits to the m achines and careful investigation showed th a t the ser
vice conditions were widely different. The replace moulds for the cracked m oulds were m ade in pure West C oast hem atite, and the com plaints ceased for a time. L ater, however,
(c) One m ould of (a) plus 1 per cent, nickel.
(d) One m ould o f (a) plus 1 per cent, copper.
These were limed and repeatedly filled with m olten foundry iron, and cooled with a hose as nearly as possible as in practice. T hey were rotated and tipped, “ stickers ” w ere rem oved with a crow bar and the underside was hosed for a sim ilar period. T hey were then re-limed and refilled w ith m olten m etal as soon as pos
sible. T he two alloyed m oulds show ed early cracks a t the lip which gradually increased in size. T he o ther m oulds showed no sign of m ajor cracking fo r 200 casts, except for som e very sm all cracks on the lips, afte r 100 casts, w hich appeared unchanged afte r 200 tests.
T he m oulds w ere then placed in general use for casting spare m etal in the foundry, and after several m onths w ithout lim ing o r quench
ing gradually failed by growth an d the tearing out o f lum ps from the bottom by “ stickers.”
This test reproduced as nearly as possible
ser-Fi g. 1.— Mi c r o s t r u c t u r e o f Go o d He m a t it e. x 500.
several m ore orders were received, indicating a continuance o f rapid failure and sim ilar lack o f satisfactory service from other suppliers!
Obviously, if the m oulds are of uniform quality, and the m achine starts with all new moulds, there will com e a period of very heavy replace
m ents followed by an apparent im provem ent, followed again by a period of fairly heavy replacem ents, the intensity of these periods gradually dying down to a fairly uniform wastage.
A heavy mild-steel fram e m ounted on tru n nions was m ade and a set o f six moulds m ounted on it: —
(a) Tw o in West C oast hem atite o f an approved analysis.
(b) Tw o o f a 10 per cent, mild-steel mix w ith E ast C oast hem atite and o f alm ost identical chemical analysis.
Fi g. 2 .— Lo w- Te m p e r a t u r e Ca r b o n i s a t i o n Re t o r t Ca s t i n g s i n Lo w- Ch r o m i u m Ir o n.
vice conditions in a norm al pig-casting m achine and agreed with the results obtained in the com plaint-free m achine.
T he o ther m achine was operating in condi
tions which are com paratively new to the pig- iron industry, and these conditions lead to rapid failure o f alm ost any type of cast-iron mould.
This has been confirm ed by two o ther cases.
This exam ple shows how departure fro m prece
dent can bring a host o f recrim inations ab o u t the ears of the foundry technical staff from cus
tom er, sales and draw ing office if it coincides with a change in service conditions.
T he im provem ents to be expected by d ep a rt
ing from precedent are probably small, and of the order o f a 10 to 15 per cent, increase in life at the m ost, and the au th o r is now firm ly co n vinced th a t the introduction o f steel and phos
phorus into an established good hem atite m ix
ture is a step in the w rong direction where heat shock is probable.
Recently, the B.C.I.R.A. tried their titanium - carbon dioxide m ethod of refining the graphite in these irons, and a set of ingot m oulds was cast. Some im provem ent in life was noticed, but the test should be done in m oulds that are prone to periods o f m ajor cracking, i.e., m ould designs th a t are on the lim it of safety, and the results o f such tests are aw aited with interest.
T he refining o f the graphite should im prove the resistance to growth.
High-Duty Cast Iron
The next group o f cast irons used fo r heat resistance is synonym ous with the group known as high-duty cast irons and m ade usually for their m echanical properties. O f these the earliest was “ L anz Perlit.” They have a rather m ore stable pearlite and finer graphite than ordinary foundry iron. These two factors en
abled experim enters (e.g., D awson) to dem on
strate their superior growth resistance at 450 to 550 deg. C. A t higher tem peratures, how ever, and for interm ittent service the difference is very small. They generally do no t resist to scaling or heat shock, and on short-tim e tensile tests their strength is only m aintained up to 500 deg. C. These irons contain no carbide stabi
lisers or special strengthening elements and are not very im portant from this point of view.
Low-Alloy Cast Irons
The third group consists of low-alloy cast irons designed fo r specific purposes, and the influence of special elements begins to be felt.
These low-alloy irons are not resistant to heat shock and are only very slightly m ore scale re
sistant than ordinary engineering cast irons (q.v.). The properties which m ust be sought are, therefore, m achinability, good foundry p ro perties, grow th resistance, toughness and strength at operating tem peratures.
The author has found, from practical experi
ence, that these low-alloy irons depend for their properties on very close control of the struc
ture and analysis, and a com bined carbon of 0.8 to 1.3 per cent, is aim ed at in m ost cases.
It is only possible to control the structure by the use of m odern m ethods, as the old m ethods of chrom ium addition were uncertain. Much has been w ritten ab o u t the difficulty o f dis
solving high-carbon ferro-chrom e in cast iron, and it is true that, with dull iron, hard spots are often encountered. The problem is not one of melting the ferro-chrom e, but is one o f dis
solving it, and the proper operation of the cupola gives m etal hot enough to dissolve com pletely the high m elting carbides o f 4 to 6 per cent, carbon ferro-chrom e.
T he addition o f chrom ium can be accom plished in various ways. L um p ferro-chrom ium in the cupola often spreads to other charges and is usually unsuitable except for melts o f some tons collected in one ladle and well mixed.
F erro-chrom e briquettes are m uch m ore satis
factory, and a trial carried out on a 10-ton
Fi g. 3.— Mi c r o s t r u c t u r e o f Lo w- Ch r o m i u m Ca s t Ir o n, x 2 5 0 .
casting when these briquettes first became avail
able gave the following results:—
M etal in ladle. Chromium. Per cent.
3 tons 0 .6 0
5 tons 0 .6 2
7 tons 0 .6 3
10 tons 0 .6 2
Sought 0 .6 0 to 0 .7 0
The cheaper variety a t 4 to 6 per cent, car
bon is quite satisfactory, and there is no need to purchase the low -carbon m aterial. C rushed
4 to 6 per cent, carbon ferro-chrom e can be added to the cupola stream in am ounts up to
0 .6 per cent. C r when treating small quantities of standard metal. The ladles should be weighed, but it is nearly always sufficient to exercise proper control on ladle lining and to estim ate the quantity of m etal by eye. I f closer control of ladle additions is necessary, then pow dered ferro-chrom e is an unsatisfac
tory m aterial (as is pow dered ferro-silicon or ferro-m olybdenum ), because some of the pow der is invariably blown away from the falling stream by air currents.
The m ost accurate and reproducible m ethod o f chrom ium addition is by the use o f m odern
ex o th erm ic m ixtures. T he m ost satisfactory so f a r tried by the au th o r is of C anadian origin, a n d is m arketed under the nam e o f C hrom X.
This m aterial is in the form o f 10-lb. bricks,
■each containing 5 lbs. o f chrom ium and giving practically 100 per cent, recovery. They con- -sist of chrom e ore m ixed w ith carbonaceous a n d exotherm ic m aterial and are usually added to the ladle. They are deliquescent and m ust be dried before use, unless a ho t ladle is avail
able; in the latter case the m aterial is placed in th e hot ladle and the action starts im m ediately.
T he m aterial glows white hot and dissolves com pletely in the m etal when it is tapped on to it.
Fi g. 4 .— Ch e m i c a l Re t o r t s Ca s t in Lo w- Ch r o m i u m Ca s t Ir o n.
T h e slag is alkaline and gum m y and should be T ried up by lim e unless teapot ladles o r a mixer be used. T he slight inconveniences of drying the bricks and dealing with the slag are hardly noticed in a w ell-run foundry and are well balanced by the certainty o f the results obtained.
It is fa r easier to control chrom ium by this m ethod than by any other simple and inexpen
sive m ethod. W hen added to the cupola charge, the bricks rem ain in their correct charge an d tests carried out by the au th o r indicate that the chrom ium lost in this way is less than 5 per cent., as com pared with 15 to 20 per cent, with
¡briquettes.
In order to see how m uch chrom ium can be added by the use o f C hrom X, a severe test was carried ou t in the follow ing way. A 10-ton ladleful o f low -chrom ium cast iron (0.62 per cent.) was well m ixed by “ rodding ” and sam pled by m eans of a hand ladle for a chill test. A b o u t 20 lbs. o f m etal was left in the hand ladle and C hrom X broken up was added to give a m axim um o f 10 per cent, o f chrom ium . T he slag and com pound were skim m ed off before the action ceased and a test-bar could still be poured from the metal. In fact, fluidity seemed to be unim paired. T he chrom ium in the test-bar was 6 per cent. In a repeat test under m ore favourable conditions, ab o u t 20 lbs.
o f hot m etal (free from chrom ium ) were poured on to 4 lbs. C hrom X ( = 10 per cent. Cr), reacting in a hot hand ladle and the reaction allowed to go to com pletion. The slag was held back and a test-bar cast. T he analysis showed 10 per cent, chrom ium .
Types of Low-Chromium Irons
T he various types o f low -chrom ium cast irons used for heat-resisting purposes are as follow s:
(a) Low -tem perature carbonisation retort castings operating at 700 deg. C. These castings, show n in Fig. 2, are 1 in. thick and 10 ft.
long and m ust be m achined on each flange.
T here w ere occasional com plaints of hard corners, bu t when the m achine shops became used to them and a suitable technique was adopted, no fu rth er troubles were experienced.
C arbide-tipped tools could no t be used because of holes cored in each flange to take bolts.
The castings have proved very successful in service and only “ grew ” i in. in length in two years, except on one plate w here p oor con
ditions had led to direct flame im pingem ent.
A fter re-m achining to size the castings went back into service for a fu rth e r cam paign.
T he m etal used was o f the follow ing average com position: T.C , 3.25; Si, 1.35; M n, 0.6; S, 0.1; P, 0.3; and C r, 0.6 per cent.
These castings would be a m uch easier p ro position to-day because of the greater control over the chrom ium additions, although the con
trol was m ade sufficiently good fo r the job by the use of rapid chill tests an d adjustm ent of the m etal in the ladle by soft iron. T he struc
ture o f one o f the castings, w hich is shown in Fig. 3, is a carbide netw ork with pearlite and m edium graphite, and h ard spots were encoun
tered on m achining (com bined carbon is 0.86 per cent.).
T he next problem was the m anufacture of chem ical retorts fo r use a t 850 deg. C. Two o f these 9-ton retorts (3 in. thick) are shown in Fig. 4, and a sim ilar com position (T.C, 3.25;
Si. 1.2; M n, 0.6; S, 0.1; P, 0.1; and C r, 0.65 per cent.) was adopted with due regard for
ïh e extra thickness and increased danger of cracking.
These castings have stood up very well in service and the tem perature has been raised to 950 deg. C. w ithout decreasing the life.
They are too large to test easily and, as no w asters have been m ade as yet, it is only possible to judge the initial structure by the
Fi g. 5 .— Ne e d l e Ex c h a n g e El e m e n t s Ca s t in 30 p e r c e n t. Ch r o m i u m Al l o y.
ap p e aran c e of the head or ingate. There is a m inor carbide netw ork and fairly coarse graphite. A fter service the structure is very sim ilar to that shown in Fig. 12, showing that after the carbide has been spheroidised, there is little fu rth er change.
Heat-Exchange Elements
A m ore difficult problem is encountered in th e m anufacture of the well-known Newton needle heat-exchange elements, because the thin sections and needles (Fig. 5) dem and very fluid metal. A part from any m etallurgical re
quirem ents, it is only possible to cast these elements w ith few rejects if the body of the casting be grey. It is necessary therefore to determ ine the best com position and structure which will not only meet the service require
ments, bu t give little trouble in the foundry.
Again, chrom ium cast irons have proved the m ost satisfactory all-round m aterial, but over 1 per cent, chrom ium is necessary, with silicon raised to give the desired structure. A t one tim e it was thought that a slight m odification o f the previous com positions would be ideal, b u t if the cross-section is too small to allow the full developm ent o f the carbide netw ork found so useful in the thicker castings and if the castings are m ade soft enough to be grey with th e low -chrom e : low-silicon alloy irons, there is great danger of the scaling and growth being greater than the corresponding unalloyed iron.
T h e introduction o f a high-chrom e : high- silicon iron gives less scaling and much less growth. T he body of these castings may
■operate up to 650 deg. C. and the needles in
the gas stream are up to 850 deg. C. at thetr tips.
The tests on which these com positions were standardised were carried ou t some years ago and modified slightly to aid the production of a high percentage of good castings. R ecent tests have shown, however, th a t the m odifica
tion has led to slight im provem ent in growth resistance.
Testing for Elevated Temperature Use A few words about the testing procedure for scaling, growth and strength at high tem pera
tures are necessary in order to show why some of the au th o r’s results are not in accord with earlier work.
G row th Tests.— O utibridge carried ou t tests on test-bars held in cast-iron pipes plugged w ith clay. C arpenter and Rugan heated their bars to 900 deg. C. in a cast-iron muffle, enclosed in a brick muffle to avoid direct flame. D onald
son heated his bars to 800 deg. C. for 8-hour periods, and all these experim enters had trouble with scale and in fact m easured growth plus scale.
F ortunately, m ost of their irons grew so much at these tem peratures that the results were not greatly affected and their results are mostly correct. W hen, however, as in D onaldson’s early w ork,1 a test-bar containing chrom ium was tested, 0.28 per cent, shrinkage was noted. The au th o r’s earlier tests also used machined bars and included the scale, but above 650 deg. C.
the results were very erratic and at 750 deg. C.
the results were very erratic and at 750 deg. C.