August, 1923
ILVD CXTRIAL A X D EAVGIATEERIAVG CHEIMISTR Y
for wooci, wall board, etc. By treatment with caustic soda solution and washing, it may be converted into a wood-flour substitute. The concentrated discharge liquor from which the free furfural has been expelled contains, among other substances, all the potassium originally present in the cobs. As the potassium content of the cobs is more than 0.5 per cent, under favorable conditions recovery might be profitable.
GENERAL The plant here described has a capacity of over 100 pounds of furfural per 24-hour day, if run on a continuous basis. Its features are capable of ready enlargement to a commercial scale. The type of digesting equipment employed is similar to that used in garbage reduction plants. Large capacity may be attained by increasing the size and number of units. A continuous column still, large enough to handle all the furfural produced in a plant, can be constructed, although two columns would be preferable in a very large plant. In the experimental plant, the operations were performed by two chemists and a helper. A plant of several times
829
the size could be opzrated with little additional labor. The labor item, however, will be the main factor in the cost of producing furfural except with very large installations. The process as developed is simple and commercially feasible. Not only corncobs, but many other pentosan-containing vegetable wastes, such as oat hulls, rice hulls, bagasse, etc., may constitute the raw material. Some of these products give very good yields of furfural, but none of them have proved equal to corncobs. ACKKOWLEDGMENT Acknowledgment is made to the E. B. Badger h Sons Company, the J. P. Devine Company, and the Vulcan Copper and Supply Company, for their cooperation in designing the column still; also to the Kutstown Foundry and Machine Company for valuable suggestions in connection with the digesting equipment. A. C. Fieldner, of the United States Bureau of Mines, furnished data on the fuel value of the cellulosic residue. The Office of Development Work of the Bureau of Chemistry has given much assistance in planning, designing, and drafting.
Nonmetallic Inclusions in Hypereutectoid Steel’ By E. G. Mahin and G. B. Wilson P U R D U EUNIVERSITY, LAFAYETTE,IND
Methods already outlined in earlier papers f o r studying solubility effects upon ferrite separation were applied to the observation of these effects upon cementite separation in hypereutectoid steels. The results indicate that at least silicon, phosphorus, titanium, chromium, nicliyl, aluminium, and copper, singly or conjointly, are e#ectioc in causing premature cementite separation if these elements are themselves segregated in the steel. It would appear also that any stresses that m a y be set u p by forcing the insert into the body steel under pressure can have no effect of this kind.
The undesirable effects of nonmetallic inclusions cannot be removed by thermal treatment, since such inclusions furnish a n inexhaustible supply of contaminating material. The only remedy seems to be in refining the processes of manufacture of steel to’the highest possible degree, so that steel m a y be produced which has a minimum of these objectionable bodies, and so that this minimum of material m a y be as finely divided and as widely disseminated as possible.
H E influence of foreign inclusions of a nonmetallic character upon the localization of ferrite in cooling hypoeutectoid steel has received much attention from various investigators. I n four earlier papers from this laboratory,2 this influence has been discussed from the standpoint of solubility effects, and support hasd been given to the view that not only is total insolubility of inclusions theoretically impossible, but that the existing slight solubility of such bodies is an important factor in the determination of the localization of ferrite. If this theory is correct, a similar explanation should be indicated for the localization of cementite by inclusions. We are here dealing with the opposite branch of the transformation curve, in a system in which cooling austenite becomes saturated with cementite upon entering the transformation range. Similar conditions should obtain in this case. ’ The austenite becomes saturated first a t .Acm. Supersaturation becomes acute upon further cooling, and it is finally relieved, a t Arcm, by local rejection of the solute. Kormally and in austenite of uniform composition, cementite might bth expected to form at boundaries and cleavage faces of austenite grains, and this is usually the case. But here, as with ferrite in hypoeutectoid steels, slight differences in carbon content and in cementite solubility-the latter due
to localization of dissolved impurities-may have an influence in the determination of the locus of cementite in network formation. In the last paper cited above2 i t was pointed out that ferrite segregation from these causes is not so noticeable or important in steels of only slightly hypoeutectoid composition, because of the narrowness of the transformation range for such steels, the entire movement of excess ferrite, across and out from the austenite grain, taking place in a comparatively short period of time. A similar statement applies to hypereutectoid steels, and, since the majority of commercial steels of this character contain carbon sufficient to form only a comparatively thin network, and rarely isolated grains except where cementite has coagulated, it follows that in most cases co-segregation of cementite and inclusions is not particularly striking or important. However, there seems little room for doubt that such co-segregation is a fact. Many metallographists have noted this,3and a careful examination of such steels will usually reveal case;, such as those illustrated in Fig. 1. Here, as in other similar steels, inclusions are found to be either directly in the cementite network or at an appreciable distance from it and well within the pearlite grain. That is, the directional influence of the inclusion is not sufficiently important to shift the locus of the envelope of cementite through any considerable distance, but if there is any tendency for this envelope to form in the
T
Received March 16, 1923. Presented before the Division of Industrial and Engineering Chemistry at the 65th Meeting of the American Chemical Society, New Haven, Conn , April 2 t o 7, 1923. THIS JOURNAL, 11, 739 (1919); 12, 1090, 1096 (1920); Chem. Met. Eng , 27, 980 (1922)
3 Brueil, J. Ivon Steel l n s t (London), 74, 57 (1907); Stead, I b i d , 91, 140 (1915); Levy, Iron S l e d I n s t . (London), Carnegie Schol. Mem., 3, 260 (1911).
830
Vol. 15, No. 8
I-VDUSTRIAL A N D ENGl 'NEERING CHEMISTRY
Pie. 3---S n ~ AnS I'm 2, Bxciirr Sonrus PICRATE A n r c w .
x
100
immediate viciuity of the inclusion, the hitier will he found directly in the envelope itself. There are relatively few exceptions tu this rule, hut this is a matter that appears not to have been fully appreciated hy most metallographists who have discussed t.hc question of inclusion influence. The occurrence of an inclusion in the body of a pearlite grain should not he regarded as exceptional unless it is found near the edge of such a grain. In a geueral way it may be said that the center of the austenite grain which generated the pear1it.e grain is highest in carbon in hypo-, and lowest in carbon in hypereutectoid steels, unless these have been normalized. The disturbing influence of the inclusion is then opposed by the natural tendency of ferrite or cementite to form at graiu boundaries, and it. is only when the inclusion is somewhere in the vicinity of such houudary that its influence might he expected to give any striking results.
PROCEDURE I n the effort to determine the relative importance of solubility effects upon cementite segrexatiou, the writers have used the general method as developed in their laboratory for studying similar effects in hypoeutectoid steels. The introduetion of inclusions of the general nature of those naturally occurring in commercial steel has not appeared to he feasihle, lor reasons discussed in thc first, paper of this series. Clean, unoxidizrd rods of alloys and .specid steels wcre driven
Frc. P-PHOSPWOPUS SIEIJI.(asrow) rs AYP~~XIIUTI$CIOIE. BODY. Two Iioiins A T !loo'. x ill0
into freshly drilled holes of hyperrutcctnid steel. The composite specimen was then heated for varying periods at temperatures above the transformation rarige for the body steel. After furnace cooling, sections were polished and examined to determine the character of contact. They were then etched and reexamined. The first series of metals used as inserts consistmi of specially made ~ t e e l s . ~For use in this connection these steels possrssed the disadvantage that they were hypoeutectoid in carbon content and, in consequence, they invariahly absorbed carbon from t,he hypereutectoid body steel. However, the thermal treatments were continued long enough to permit more carbon to move into the zone thus impoverished, so that hypereutectoid composition was maintained in this region. Diffusion of the excess impurity of the insert into the surrounding metal might be expected to occur, as usual. The body steel for these experiments contained the following percentages of the respective elements: carbon, 1.2; manganese, 0.30; phosphorus, 0.02; sulfur, 0.03; silicon, 0.20. In the illustrations the material of the insert oecnpies the lower portion in each ease.
RESULTS EFFECT OF SILICON-The
etching treatment suitable for revealing the structure of carbon stecl does not definitely etch the 4 per cent silicon steel used h e x . A narrow hand of cemen-
4
T ~ I JS O U ~ N A L la, , io90 (ISZO).
INDUSTRIAL A N D ELVGINEERING CHEMISTRY
August, 1923
FIG.7-Snlan.
AB
Frc. R . EXCEPTSoomla P ~ C R AATTACK. TB X 250
lite almost entirely surrounds the insert, lying just outside the line of contact (Fig. 2). T h a t this is cementite, and not ferrite formed by decarburization by the hypoeutectoid insert, is evident from the fact that the cementite network of the body steel makes direct connection with this band. Sodium picrate confirms this conclusion (Fig. 3). EFFECT OF P ~ o s ~ ~ o n u sl.OG - A per cent phosphorus steel, heated in contact with the body steel for 2 hours, produced results similar to those obtained with silicon steel, as shown in Fig. 4. Here again the bright ring joins the cementite nctwork, which proves its hypereutectoid character. Longer heating entirely removes this effect through extensive dissemination of the previously localized phosphorus. After 13 hours of heating at 900' to 1000" C. it becomes almost impassible to find the line of contact between the two pieces. Carbon distribution has become uniform and there is now no sharp dividing line between high and low-phasphorus areas, so that cementite separation is not now locally affected. EFFECT OF TITANIUM-Ordinary thermal treatment of a composite specimen containing a titanium steel, as the insert, showed no effect of the latter upon cementite separation. After heating for 17 hours a t 900" to 1000' C., slight cementite localization rras noticeable (Fig. 5 ) . This lack oJ pronounced effect is therefore apparently due to the slowness with which titanium migrates in the heated steel. EFFECT OF CHROM~UM AND NICKEL-An insert of chrome1 (an alloy of chromium and nickel) was used. Figs. 6 and 7, showing the heat-treated specimen a s etched with nitric acid and sodium picrate, respectively, indicate the action of these two metals, when used together in these proportions. As was noted in the discussion of the action of silicon, the cementite network dimctly joins the band surrounding the insert.
831
Fia.X-ALUMZI. (aerow) IN HYPGP.SUTECIOIDSTBBI.. T x r a r l r ~Houns ~ AT 9 00'. X 500
EFFECTOF A L ~ M I A I AAD ~ M NiCKEl,-An i n s x t of alumcl, a nickel-aluminium alloy, war used. Fig. 8 shows a broad band of unetched material. ioining the cementite n!atcs of thc pearlite grains as wrll as the-network without. Sddium picrate colors the carbide plates of the pearlite grains slightly and the network more decply, as usual. Thc bright border material is not colored appreciably in 20 minutes application of the boiling picratr solution, the red line of the cemcutite network runnins into and joining this. One might conclude from this that the broad, bright band is either ( a ) austenite, retained in the cooled system through the action of nickel i n retarding the transfomation, or ( b ) a high nickel (and alumiilium) solution of nickel or nickel carbidc in cementite, the nickcl inhibiting the action of sodium picrate. The first assumption is rendered unncceptahle by the conclusion that if austenite has remained in the zone immediately adjacent to the contact line. it sliould shade into martensite and troostite in the low-nickel areas. Instead of this, the band appears to be continuous with carbide plates of pearlite grains. Hypothcsis ( b ) has thcrefor? been provisionally adopted. This view is further strcngthened by the fact that a shorter thermal treatment gives a narrow band which is colored by sodium oicrate. as shown in Fir. 9.
treatment of the composite specimen gave a very interesting result. Apparently on account of the unusual ability of this alloy to absorb carbon, the insert has exerted a pronounced decarburizina effect upon the body niece. T h e movement of
832
IXIXCSTRIAL A S D ENGl'NEERING CHE.llISTR Y
steel, causing premature ferrite separation as usual, so that during cooling through the. transformation range a perfectly even and definite ring of entirely decarburized iron-nickelcapper alloy is produced, outside of which lies a still wider, but less definite, diminishing zone of hypoeutectoid steel. Apparently, the inner band of carbonless alloy marks the limit of advance oi the invading nickelkopper, while the ferrite-pearlite z o k lying outside this is the result only of carboil absorption by the insert. Still farther out this merges with a zone oi eutcc-
r01. 13, No. 8
have,,,een causq$,@rtiy by the resulting pressure should be considered. Theoretifk considerations, as well as expe&mXntal results, lead believe t,hat such pressure had no effect of this kind. separation of both ferrite and cementite fmm. cobling: austenite is accompanied by pronounced dilatation. The effect of pressure tipon the system would therefore necessarily be to binder or retard, rather than to aid, the separation of these excess solutes in the zone about the insert. In order to make a practical test of the question, inserts were made of the same material as that 'of the body and the composites were treated in the way already described for other specimens. In some of these cases a l i g h t tendency toward such separation was noticed, but ill the most carefully cleaned pieces it was cxtremcly difficult to find the line of contact between insert and body after the heating, the pieces being perfectly welded together. It may be observed also, in this connection, that the long heating that was applied to all these specimens shonld have given ample opportunity for all internal stresses to be relieved, as well as for the disappearance of all essential differmces in grain size, the latt,er possibly being due to the presence of a narrow ring of deformed metal in contact with the insert. That these differences did disappear in all caws is evident from an inspection of the photomicrograplis.
CoxcLusIom The logical condusion from all this work appears to be of tteel, excluding the one special case nf i:omposition, segregation of any dissolved constituent will h a w an effect upon carbon distribution in all but quenched steel, if indeed this case is to be excepted. This will apply to the ordinary four e~emcntsoccurring in small quantity in all carbon steel, as well as to metallic or nonmetallio elements intentionally added in the manufacture of special steels. It will apply also t,o tliosc sources of con-
FIO. IO-FrYPI:RI'I:IECI"," 'I'W"
STEEL
wlrw
i i < > O U SAT
C""I*UTAY
!,Ill,?
x
INSBRT
(eeLo$r).
loll
st,ant contamination known as noninetallie inclusions, and witli this addition-that while ordinary segregation of totally dissolved elements can be nearly or entirely cured by t.herma1 treat,rnent and, to some extent, by forging. no amount of either forging or t.hermal treatment that is possiblc iti practice will serve'to cure the evil effects of the nonmetallic inclusion, since tho relatively large amount of matcrial so concentrated and the relatively small solubility of this mat,erial in t h e solid solution surrounding it, make its exhaustion by diffusion a practical impossibility.
