By FRANK H U D SO N (Member) The non-ferrous foundry has m any imm ediate
problem s, m ore so perhaps than iron and steel foundries. Every engineering project is d e
veloped from a casting of one kind o r another and production, particularly of copper-base cast articles, constituted the first m etallurgical developm ent that the world has ever known.
Accordingly, the uninitiated m ight anticipate that brass foundries and ingot casting shops are m odel exponents and possibly originators of works metallurgy. Indirectly the last thought m ay be true, and if one undertook to trace the first application o f w orks m etallurgy it m ight conceivably be found to originate from a dis
pute, shall we say, ab o u t an unsatisfactory article.
So far as the direct application of metallurgy is concerned, the brassfounder, with few excep
tions, has shown little initiative. T here is some excuse for this as the very antiquity of the art prom otes habits and m ethods unreceptive to m odern developments, and it is not unusual to find that the whole fabric of bronze and gun- metal casting production has been built upon rule o f thum b m ethods w ithout the presence of guiding fundam ental facts. Such a state of affairs is, o f course, entirely wrong, and to-day, more than ever before, there is profitable scope and application for research of a practical nature.
F o r m any years the practical m an has been, and in some cases still is, labouring under en
tirely erroneous ideas as to how metals should be melted. U ntil quite recently it was con
sidered satisfactory practice to em ploy neutral or slightly reducing furnace atm ospheres using charcoal as a cover. Follow ing investigation into the effect of gases on copper and degasi
fication m ethods it is now known that such practices are incorrect and that the best results are obtained with oxidising conditions— a com plete reversal of the old ideas. It is also now appreciated that the com position and condition of the m etal being melted, in conjunction with alloying technique, bear an im portant relation to therm al reactions.
Practical application of these discoveries has led to pronounced im provem ent of casting quality in a wide variety of alloys, both as regards density and m echanical properties.
They have exploded the theory widely held am ongst practical men th at virgin m etal heats, say of 8 8 :1 0 :2 gunmetal, give inferior results to the use of remelted m etal such as ingots.
They provide inform ation as to why various brands of copper behaved differently when used for alloy production in the foundry, particularly so far as sand castings are concerned, and in
dicate means by which these differences m ay be corrected—a most im portant point at the present time.
D uring the past two years it has been the privilege of the author to come in close con
tact with m any brassfounders, large and small, antiquated and m odern, throughout the length and breadth of the British Isles, and it would appear that m ore efficient results would un
doubtedly be obtam ed if melting operations, in
T a b l e I .— Vapour Pressure of Zinc in the Industrial Brasses.*
Partial pressure of zinc vapour at various temperatures.
Composition. M elting poir t (liquidus). A pproxim ate ca stin g temperature. Boiling point.
t D eg. C.
(Vapour pressure Zn 760 mm.) Temp.
Deg. C.
Vapour pressure.
Mm. Hg.
Temp.
Deg. C. Vapour pressure.
Zinc 4 1 9 .5 0 .1 3 9 500 1 .27 918
60 : 40 900 160 1,040 600 1,070
65 : 35 930 170 1,070 595 1,100
70 : 30 955 150 1,100 540 1,145
80 : 20 1,010 85 1,150 265 1,300
90 : 10
(estim ated values)
1,055 20 1,200 80 1,600
* “ The Casting o f Brass In g o ts,” by Genders and B ailey, R esearch Monograph No. 3, British N on-Ferrous M etals Research A ssociation.
177 k
the full sense o f the term , were better u nder
stood. In the first place, it m ight be m entioned th a t it is no t essential to rely upon furnace control to prom ote oxidising conditions, as suit
able trea tm e n t can be given even in the presence o f reducing atm ospheres. F u rth e r
m ore, the degree o f oxidation required is not such as to cause abnorm al m etal loss. It would sim plify the problem to com pare the m elting operations entailed in connection with, say, three groups o f alloys, brass, gunm etal and the true copper-tin alloys such as the 9 0 :1 0 alloys and phosphor bronze, etc., and it is proposed to ad o p t such a course in this Paper, keeping all
conditions showed 150 per cent, increase in porosity. It is surm ised th a t the fac to r re
sponsible for this exceptional behaviour o f brass can be attributed to the high v apour pressure o f the m olten alloy as show n in T able I.
It is well know n th a t liquids under certain conditions tend to give off vapour, and this process o f tran sfo rm atio n is called evaporation.
A t constant tem perature the space above the liquid contains a definite am o u n t o f the vapour of th a t liquid, and this exerts a pressure know n as the vapour pressure. Liquids w hich readily evaporate give high v apour pressures; for exam ple, ether a t 20 deg. C. has a vapour
pres-T a b l e I I .— Z inc Loss o f IQ : 30 B rass under Reducing, N eutral and O xidising Surface Atm ospheres.*
A lloy. Treatm ent. Loss. Per cent.
70 ; 30 brass 800 deg. in air 0 .2 5
1,050 deg. in air 1 .7
1,050 deg. in nitrogen ( + 2 per cent, oxygen) 4 .7
1,050 deg. in hydrogen 1 2 .6
A lum inium brass ( 0 .2 per cent. 800 deg. in air 0 .2 7
Al) 1,050 deg. in air 0 .9 2
1,050 deg. in nitrogen (-(- 2 per cent, oxygen) 2 .5 2
1,050 deg. in hydrogen w ith flux 1 3 .8
Alum inium brass ( 2 .5 per cent. 800 deg. in air 0 .0 0 3
Al) 1,050 deg. in air 0 .0 0 7
1,050 deg. in nitrogen ( + 2 per cent, oxygen) 0 .0 7 1
1,050 deg. in hydrogen 1 5 .3
Silicon brass ( 2 .0 per cent. Si) 800 deg. in air 0 .1 2
1,050 deg. in air 0 .6 7
1,050 deg. in nitrogen ( + 2 per cent, oxygen) 6 .0
1,050 deg. in hydrogen 1 2 .8
Phosphorus brass (0 .0 5 per 800 deg. in air 0 .3 6
cent. P) 1,050 deg. in air 15.1
1,050 deg. in hydrogen 1 6 .4
‘ The C asting o f Brass In g o ts,” b y Genders and B ailey, R esearch M onograph N o. 3, British N on-Ferrous M etals Research A ssociation.
rem arks as simple and practical as possible so th a t they m ay be the m ore readily understood.
Brass M elting
T here is no doubt that hydrogen is the gas w hich causes m ost trouble in m elting operations, and the first p o int which the foundrym an should b ear in m ind is that the solubility o f hydrogen varies w ith different metals. F o r exam ple, the affinity of zinc-free bronze for this gas is infinitely greater than, say, 7 0 :3 0 nrass. It is now know n, in fact, that 7 0 :3 0 brass is rela
tively im m une to hydrogen unsoundness.
G enders and Bailey have shown in their book,
“ T he C asting o f Brass Ingots ” * a standard w ork o f reference on the subject, th a t the sound
ness o f sand-cast 7 0 :3 0 brass m elted in an atm osphere o f hydrogen was unaffected, whilst a 5 per cent, tin-bronze treated under sim ilar
* R e se a rc h M o n o g rap h No.
R esearch A ssociation. B ritish N o n -F e rro u s M etal
sure o f 442.2 mm. o f m ercury, whilst w ater at the sam e tem perature has a value of 17.41 mm. M olten copper at 1,320 deg. C. has a v apour pressure of only 0.001 mm. o f m ercury.
V apour pressure increases w ith rise o f tem pera
ture and a liquid boils when its tem perature is such that its vapour pressure equals the pres
sure o f the atm osphere, nam ely, 760 m m . of m ercury. T he solubility o f gas in a liquid, p ro viding it does not react chem ically w ith that liquid, decreases with increasing vapour pres
sure o f the liquid, becom ing zero at the boiling point. W ater contains air and if one watches the effect o f heat on w ater in a glass vessel, this air can be observed com ing out, long before boiling point is reached, and it is all expelled on boiling, and, so long as the w ater is boiling, no air can be re-dissolved.
A sim ilar analogy exists in connection with m olten m etals. As the vap o u r pressure rises 178
the solubility of gas becomes increasingly less are practically unaffected by variations in m elt
ing conditions so far as gas absorption is con
cerned. In other w ords, it does not m atter whether one employs melting furnaces giving re
ducing, neutral or oxidising atmospheres. It is im portant to note, however, that the evolution o f zinc vapour should take place in order to obtain satisfactory results. Care m ust there
fore be taken to ensure adequate superheating relatively fool-proof so far as melting technique
’is concerned.
In view o f brasses being im m une to hydrogen pick-up, it m ust also be evident that the initial gas content o f the virgin metals em ployed for m aking up the alloy is of little moment. F or example, cathode copper, which contains appre
ciable quantities o f hydrogen, can be readily used for brass production although special precautions w ould have to be taken if such appears incorrect as the above investigators have proved that reducing conditions lead to
highest zinc loss. T he experim ental m ethod utilised consisted in heating small quantities of alloy in a tube furnace through w hich either in reducing metallic oxides should be particu
larly noted as its action in this direction is of considerable im portance in connection with alloys such as gunm etal and bronzes, to be dis phosphorus that a phosphorus-alum inium brass behaves as if phosphorus were absent. are obtained with oxidising furnace atmospheres, but this does not apply when even a trace of
All the com mercial gunmetals absorb hydrogen, and the degree of absorption depends on several
x - 2
factors. F ro m w hat has been said in regard to furnace atm osphere, w hether reducing, neutral o r oxidising. M elting tim e, superheat tem pera m entioned th at industrial hydrogen is principally m ade by passing steam over coke at a high tem perature.) R educing conditions assist this re
action, but it tends to be reversible in the presence o f carbon dioxide or oxygen, hence the m odern tendency for utilising oxidising melting conditions as fa r as possible. oxidising gases in furnace atm ospheres, either as oxygen o r probably as carbon dioxide, limits elasticity fo r general needs although extremely useful fo r specialised w ork. F o r all general purposes, either oil or forced-draft coke-fired furnaces are h ard to beat, providing a plentiful air supply is available. In fact, w ith proper control, uniform fuel and judicious selection of charge m aterials, correct m elting conditions can be so m aintained as to require no additional oxidising technique.
U nfortunately, the present conditions in most non-ferrous foundries in this country preclude this ideal being readily obtained, and some de
gasification treatm ent will usually be found necessary. Coke-fired n atu ra l-d ra ft furnaces as a class are inferior to oil o r forced-draft units in view o f the inadequacy or wide variation in the air supply available and the consequent difficulty o f obtaining oxidising conditions.
W ith these types of furnaces the addition of oxidising agents becom es m ore o r less essential.
Coal-fired, gas an d electric arc furnaces, unless handled with skill and special precautions taken, are liable to give gassy m etal, and degasification oxidising atm osphere during m elting the presence o f furnace hydrogen is m inimised. A still more
C.—0.15 p e r c e n t. Ox y g e n. D.—0.32 p e r c e n t. Ox y g e n. Fi g. 2 .— M i c r o s t r u c t u r e o f C h i l l - C a s t C o p p e r C o n t a i n i n g V a r i o u s O x i d e C o n t e n t s .
x 100. E l e c t r o l y t i c a l l y P o l i s h e d .
Oxygen in ca st copper can be read ily d etecte d an d th e a m o u n t p resen t estim ated u n d e r th e m icroscope a t 75 to 100 m agnifications.
Specim ens should be ta k e n a t least £ in. aw ay fro m a ca st o r “ set ” face as th e oxide c o n ten t in surface areas is u sually m u c h higher th a n in o th e r p a rts. E x a m in a tio n should be con d u cted a fte r polishing a n d lig h t etch in g w ith a fr e s h ly p r e p a re d 10 per cent, aqueous so lu tio n o f a m m o n iu m p ersu lp h ate. Sm all p ercentages o f oxygen begin to show in cast copper as a th in g rain b o u n d ary o f copper-cuprous oxide eu tectic , as show n b y th e above photom icrographs, g rad u ally th ick en in g a n d filling u p th e grains u n til ap p ro x im a te ly 0.39 p er ce n t, oxygen is reached w h en th e s tru c tu re consists en tirely o f eu tectic. A bove th is c o n te n t spangles o f free p rim ary cuprous oxide
a p p e a r as'sh o w n in m icro g rap h 3D.
181 n 3
oxide is present, and m etallic copper w hich is redissolved. Similar reactions take place with nickel oxide.
It is thus evident th a t the presence o f copper or nickel oxides in an active form in their re
spective alloys either as a result o f melting conditions, com position of m etallic charge or as a definite addition, provides a practical means
A .— 0.05 p e r c e n t . O x y g e n .
of overcom ing troubles due to dissolved gas.
In fact, m any founders have already found that degasification conducted by oxide additions is m ore suited to routine foundry purposes than furnace control.
These m ay be utilised with the initial charge or added after the m etal is molten and super
heated. Black cupric oxide, added in an
B.— 0.10 p e r c e n t . O x y g e n .
a m o u n t equivalent to between 1 and 2 per cent, o f the charge w eight (as well as certain p ro prietary fluxes) is a satisfactory addition m aterial for incorporating w ith the charge, w hilst red cuprous oxide can be utilised for m olten m etal
t
A .— 0 .0 5 p e r c e n t. Ox y g e n.
o f the m etal after m elting has an added ad vantage inasm uch as crucible attack by the flux is negligible. T he use o f oxidising slags in the m anner suggested is particularly valuable when melting m ust be conducted under reducing
con-B .— 0 .1 5 p e r c e n t. Ox y g e n.
I '
.
\
••• •
V "
V A i
/ ■ r P p
• ■ V
T.--, c „>
/IP
■ r - K r 'c 'y / p p p \
mm
- ' i ’ ■■ >-x
i .■w
• ■ . .
' ■ v - • ,
C .— 0 3 2 p e r c e n t. Ox y g e n. D.—0.50 p e r c e n t. Ox y g e n.
F i g 3 .— M i c r o s t r u c t u r e o f C h i l l - C a s t C o p p e r C o n t a i n i n g V a r i o u s O x i d e C o n t e n t s . x 5 0 0 . E l e c t r o l y t i c a l l y P o l i s h e d .
degasification. In this latter case around 0.2 to 0.5 per cent, should be sufficient, and this is added and stirred into the metal, preferably in the furnace, a few m inutes before pouring.
It is, o f course, im portant that these oxide additions should be perfectly dry. T reatm ent
ditions. It m ust also, o f course, be appreciated that gunm etals treated with oxidising m edia must be properly deoxidised, preferably with phos
phorus, before being poured into castings. Phos
phorus, as was shown in the case o f brass, has the property of reducing nearly all the m etallic
oxides except th a t of alum inium ,* and such an addition will effectively remove the oxides present in gunm etal if the oxidising treatm ent level, but rapidly increases with oxygen contents below 0.05 per cent. “ Best Selected ” copper ingots favoured by the non-ferrous founder con
tain between 0.05 and 0.10 per cent, oxygen, and there is no doubt that this m aterially assists casting production by tending to prevent gas absorption and constitutes a probable reason for the popularity of the grade. Fig. 2 illustrates the m icrostructure of cast copper containing various oxide contents. There are, however, many this recently. A certain foundry producing Adm iralty gunm etal castings had made these test-bar requirem ents, obtaining about 11.0 tons per sq. in. tensile. These low results were principally due to gas pick-up, probably caused by melting under slightly reducing conditions in conjunction with a lower initial oxide content in the charge. The higher oxide content in the previously used “ Best Selected ” copper was just sufficient to prevent the absorption of hydrogen.
The trouble was eliminated by adding a small
som ewhat higher than those previously obtained when using “ Best Selected ” copper. The de type of copper contains appreciable quantities of hydrogen, and in the absence of oxide is liable to give erratic results. Melting under oxidising conditions alone may not provide enough oxygen to meet requirem ents, but satisfactory results can always be obtained by the addition of cupric or cuprous oxide during melting. In fact, in these days when foundrym en as a whole must work economically, the use of oxide addi
tions to gunm etal is a practice which can be thoroughly recommended to give uniform re
sults regardless of the raw m aterials employed. that m aterial carrying hygroscopic corrosion p ro ducts may accentuate gas troubles due to the following points should receive consideration: —
(1) Melting atm ospheres begin to become of im portance, and oxidising conditions should be obtained wherever possible.
(2) The raw m aterial basis of the metal
183 N 4
charge has an effect on the results obtained, and quality production is facilitated by the presence o f oxides in or on copper o r its alloys.
(3) T he use of direct oxide addition in con
junction with phosphorus deoxidation affords a convenient way o f obtaining uniform m elt m ethods already described, in conjunction with judicious use of zinc and phosphorus, is essen
tial if the best results are to be obtained, especially so far as sand castings are concerned.
C ertain precautions are, however, necessary in the presence of appreciable am ounts of phos resistance to corrosion is required. P hosphor and gear bronzes are further m odifications of been actually substantiated in service. P erson
ally, it is felt th a t the presence of under 1 per phosphorus. Similarly, larger am ounts of phos
phorus, such as are present in phosphor-bronze, phosphorus. W hen appreciable phosphorus is present during the m elting-dow n perod, prob assists degasification with less phosphorus loss.
In the m elting of special alloys fo r sand castings such as nickel silver, which, unlike
brass, is highly susceptible to gas reactions, cupro-nickel and M onel, etc., it can be truly claim ed that the use o f oxidising melting treat
ments follow ed by adequate deoxidation have been prim ary factors in establishing satisfactory foundry technique. castings careful control of melting operations is not so im portant. to learn about non-ferrous m elting conditions, and he had described in the Paper one o r two possessed the suitable physical characteristics for producing good castings. In the production of special high-m elting-point alloys in the brass foundry, such as M onel, etc., brass sands should never be employed. Ironfoundry sands carrying coal dust were called for. Brass foundries had received initial orders for these special alloys, and naturally had used Mansfield sand. T he sand from blemish. This particular foundry obtained very good results by pouring from about 10 ingates on each side, and using two big down- runners fed from two pouring ladles. T o obtain the best results on plate castings, one should use dry-sand methods, and cast horizontally, with as m any gates as one could get down one side o r (if necessary) both sides. F or bronze cast
ings, risers should be avoided if at all possible.
Melting Losses
M r. A. Ph illips rem arked th a t in the table showing the melting losses under oxygen, nitro
gen and hydrogen, no test had been taken on the stressed that alloys containing alum inium could have the oxides reduced by phosphorus ad d i losses under reducing and oxidising conditions, suggested th at the only reason that the nitrogen test was om itted was because there was no difference between oxidising and reducing con
ditions. It was no t w orth while doing a test on the phosphorus brass with the neutral atm osphere, because quite obviously the result would be the same. plunged below the metal surface.
He did not know o f the existence as yet of excessive additions w ould m ake it ineffective.
Mr. Ph illips said that for some time the firm with which he was associated had observed the practice stressed in the P aper o f having oxidising conditions in melting. H e was fully in agreem ent w ith the au th o r; oxidising a t
m ospheres were quite satisfactory for the
brasses and the bronzes, but w ith alum inium w hich revealed very favourable aspects.
Mr. Hud so n pointed ou t th a t both sodium
was using 0.2 per cent, zinc as deoxidiser and the best results were obtained. H e believed that Mr. H udson had recom m ended phosphorus. D id any of the phosphorus rem ain in the metal, and did it im prove the strength o r otherwise ?
Mr. Hud so n replied that the am ount w hich rem ained was practically undetectable by normal
Mr. Hud so n replied that the am ount w hich rem ained was practically undetectable by normal