Metallographic Manifestations of the Air Oxidation

Vol. 105, No. 2

General

In the preceding discussion, the ox idat ion k inet ics at temperatures greater than 350~ have been ap- p rox imated to parabo l ic and l inear rate equations. A compar ison of ox idat ion and Mg evaporat ion rates demonst ra ted that the oxide film offered res istance to ox idat ion at al l stages of exposure. Meta l lograph ic and x - ray examinat ions have shown that deplet ion of Mg by select ive ox idat ion resu l ted in fo rmat ion of A1 inclusion in the oxide. These inclusions, in turn, caused a gray to b lack d iscolorat ion of the surface. Many of the v iews which have been ex- pressed must remain speculat ive because chemica l composit ions and st ructures of the oxides in the composite film, and determinat ions of Mg act iv i t ies and dif fusivit ies for the a l loy and oxides are re- qui red to e luc idate the ox idat ion and Mg evapora - t ion behavior .

Acknowledgments The author wishes to express his thanks to former

col leagues, especia l ly J. S. K i rka ldy and H. P. Godard, at A lumin ium Laborator ies , for he lpfu l dis- cussions.

Manuscript received Nov. 26, 1956.

Any discussion of this paper wil l appear in a Dis- cussion Section to be published in the December 1958 JOURNAL.

OXIDATION OF AN A1-3% Mg ALLOY 71

REFERENCES 1. S. Dobinski and M. Niesluchowski, Nature, 144, 510

(1939). 2. L. De Brouck~re, J. Inst. Metals, 71, 131 (1945). 3. R. Eboral l and C. E. Ransley, ibid., 71, 525 (1945). 4. O. Kubaschewski and H. Ebert, Metall]orschung, 2,

232 (1947). 5. M. Whitaker and A. R. Heath, J. Inst. Metals, 82,

107 (1953). 6. A. J. Swain, ibid., 80, 125 (1951). 7. P. T. Stroup, "Controlled Atmospheres," A.S.M.

(1941). 8. E. A. Smith, Jr., Light Metal Age, 12, 24 (October,

1954). 9. W. W. Smeltzer, This Journal, 193, 209 (1956).

10. M. S. Hunter and P. Fowle, ibid., 103, 483 (1956). 11. T. E. Leontis and F. N. Rhines, Trans. Am. Inst.

Mining Met. Engrs., 166, 265 (1946). 12. I. A. Makolkin, Zhur. Priklad. Khim., 24, 460 (1951). 13. E. A. Gulbransen, Rev. Sci. Instr., 15, 201 (1944). 14. U. R. Evans, Trans. Am. Inst. Mining Met. Engrs.,

166, 292 (1946). 15. J. Loriers, Compt. rend., 231, 522 (1950). 16. W. W. Webb, J. T. Norton and C. Wagner, This

Journal, 103, 107 (1956). 17. E. A. Gulbransen, Trans. Electrochem. Soc., 87, 589

(1945). 18. C. Wagner, This Journal, 1{}3, 571 (1956). I9. R. M. Barrer, "Diffusion in and Through Solids,"

Cambridge University Press, p. 19 (1951). 20. R. D. Guminski and R. A. Hines, Pr ivate communi-

cation.

Metallographic Manifestations of the Air Oxidation

of Tantalum at 7S0~

Robert Bakish 1

Sprague Electric Company, North Adams, Massachusetts

ABSTRACT

The crystal lographic and structural factors involved in the air oxidation of tantalum at 750~ are presented. The role of the {100} planes in the oxidation process is discussed, and a tentat ive mechanism for the conversion of metal to oxide under these conditions is proposed.

The ox idat ion of Ta has been studied by a number of invest igators w i th emphas is on the type and s t ructure of the ox ide fo rmed (1-4) or on the k inet - ics of ox idat ion (5-9) . Work on the so lubi l i ty of oxygen in Ta has also been repor ted (10). In al l these invest igat ions, the nature of the ox ide-meta l inter face has been complete ly d isregarded, and no ment ion of the c rys ta l lograph ic dependence of the tanta lum oxygen react ion has been made.

Pure Ta (Ta 99.9 + Fe 0.03 max and C 0.03 max) suppl ied by Fanstee l Meta l lurg ica l Corporat ion was used for this invest igat ion.

Experimental Procedure

I t has been shown that tanta lum oxide exists in a stable modif icat ion in the temperature range 650 ~ 1300~ (3). The tanta lum oxide used in the most recent s t ructure determinat ion was prepared by

1Present address: Ciba Limited, Rare Metals Division, Basle, Switzerland.

ox idat ion of the meta l at 700~ (4). In order to avoid poss ib le compl icat ions int roduced by work ing w i th an unknown oxide, an ox idat ion temperature was selected on the basis of the above references. A l l specimens in this invest igat ion were subjected to 45-min ox idat ion in air at 750~ fo l lowed by quenching in air at room temperature .

X - ray powder technique was used to ident i fy the oxide formed. The single crysta ls used in this s tudy were grown by s tandard stra in anneal technique and thei r or ientat ions determined using Gren inger ' s method (11). The ident i f icat ion of the p lane of ox idat ion was carr ied out as fol lows: two surfaces wi th known angu lar re lat ions were ground on crystals of known or ientat ion, and the angles made by the p lanes of p re fer red ox idat ion wi th the common edge were measured. This in format ion was then p lot ted s tereograph ica l ly and ana lyzed using a s tandard cubic project ion, in accordance wi th s tandard procedures (12).

) unless CC License in place (see abstract). ecsdl.org/site/terms_use address. Redistribution subject to ECS terms of use (see 129.81.226.78Downloaded on 2015-05-16 to IP

72 JOURNAL OF THE ELECTROCHEMICAL SOCIETY

Standard metal lographic techniques were applied in this study, and a solution consisting of H~SO4 (96%), HNO~ (70%), and HF (48%) in ratio (2: 1: 1) was used as the etching reagent.

In oxygen embritt led Ta, both inter- and trans- crystal l ine fracture can be observed. Intergranular failure is predominant in relatively small grain size material (about 0.I ram). The oxide distribution on grain surfaces obtained by fracturing small grained metal were examined in order to gain additional information regarding the sites of preferred oxidation. These observations were made on freshly frac- tured wires without any subsequent surface preparation.

Experimental Results Air oxidation of Ta leads to surface oxidation and

crystal lographical ly dependent internal oxidation. The latter proceeds at higher rates along certain planes and directions in the Ta lattice. The oxide- metal interface in polycrystal l ine Ta oxidized in air at 750~ is shown in Fig. 1. Observe here the surface oxide and the spear- l ike or ientat ion-dependent platelets of the oxide. Note that no preferential oxidation exists at the grain boundary. This type of interface is typical of the large number of oxide- metal interfaces examined and is unaffected by longer oxidation treatments which only moves this interface into the metal.

X - ray study by application of powder technique indicates that the oxide formed under these conditions is Ta20~, of the same var iety as commercial ly produced mater ia l of this formula. The structure of this oxide has recently been reevaluated (4). X - ray lines characteristic of the oxide pattern were also detected after removal of the surface oxide in specimens which, by metal lographic examination, were shown to have internal preferential oxidation.

The oxide formed on polycrystal l ine Ta is of very fine part icle size, as indicated by diffuse x - ray pat- terns. This oxide has a high degree of porosity which can be detected on careful microscopic examination of the polished oxide. The same type of polycrystal l ine oxide seems to be formed on single crystal surfaces.

X - ray analysis of the internal oxidation occurring in single crystals shows that, within the accuracy

February 1958

Fig. 2. Details of early stage internal oxidation. 2000X before reduction for publication.

of the measurements of about 2 ~ , the planes of internal oxidation are paral lel to the (100) planes of the bcc lattice. On successive lapping and polishing of single crystal surfaces normal to the oxidized surface, and in several cases normal to the ~100~ directions, it was found that quite often the solid platelet profile of (100) planes was substituted at a greater depth by a bead profile and eventual ly with discrete sites of oxidation. See Fig. 2 for the appearance of (100} traces at some depth from the oxide metal interface. (Observe both "bead" profile and discrete oxidation sites.) These oxide platelets do not grow appreciably in thickness normal to the {100) planes after they reach a thickness of about 0.002 mm (see Fig. 3). Heavier oxidation is characterized by the presence of a greater number of more narrowly spaced platelets. The porous oxide is but a thorough oxidation of the volume of the metal by oxidation along {100} planes. F igure 4, which is a taper section of an oxidized surface of a single crystal, supports this contention. Observe the islands of porous oxide, the high density of platelets in their immediate vicinity, and the relat ively smaller density of platelets as the distance from the oxidized surface is increased. That this appears to be the nature of oxidation seems to be also supported by examination of polished oxide surfaces

Fig. 1. Oxide metal interface, cross section normal to Fig. 3. Traces of the {100} planes. 2000X before reduction oxidized surface. 2000X before reduction for publication, for publication.


Vol. 105, No. 2 AIR OXIDAT ION OF TANTALUM AT 750~ 73

F,g. 4. Taper seclion of oxidized single crystal surface. 500X before reduction for publication.

Fig. 7. Fractograph showing sites of nucleation of oxygen metal reaction on a grain surface. 500X before reduction for publication.

Fig. 5. Polished oxide surface (polarized light). 500X before reduction for publication.

under po lar ized l ight (see Fig. 5). This revea ls a fine s t ructure wi th e lements para l le l to the traces of the {100} p lanes of the substrate.

Another aspect in wh ich the c rys ta l lograph ic dependence of ox idat ion mani fests itself is the for - mat ion of steps bound by {100} subst rate p lanes on the ox ide-meta l inter face of single crystals. F igure 6 shows the appearance of such an inter face as v iewed on a (112) plane. When such single crysta ls were subjected to oxidat ion, it was possible to observe that the direct ions in the {100} p lanes were the direct ions of a h igher rate of oxidat ion. This observat ion and the occurrence of the {100} step format ion, tend to show that the ~100~ direct ions

Fig. 6. Crystallographic (100} steps on oxide metal interface plane of photomicrograph parallel to (112). 2000X before reduct,on for publication.

in the {100} p lanes are the direct ions of p re fer red oxidat ion.

In addi t ion to ox idat ion w i th in the grains, evi - dence of in terna l ox idat ion was found on gra in sur - faces. This in format ion was obta ined by micro - scopic examinat ion of f ractured, fine grained, ox i - d ized Ta wire. F igure 7 shows a d is t r ibut ion typ ica l of those observed.

Discussion of Results

The f indings of this invest igat ion show the pres - ence of a set of h igh act iv i ty p lanes in the (100} p lanes of the Ta latt ice.

Two a l ternat ives are offered as possible exp lana- t ions of the h igh act iv i ty of the {100} p lanes: it can be due e i ther to the inherent high act iv i ty of this p lane; or it can be the resul t of the presence of favorab ly or iented imperfect ions which could act as short c i rcui t ing paths for diffusion and oxidat ion. It is on these imperfect ions that the meta l -ox ide re - act ion appears to nucleate. No pre ferent ia l gra in boundary ox idat ion was observed in this study. The ox idat ion process seems to proceed by nuc leat ion of ox idat ion at h igh ly local ized sites along the traces of the {100} planes, leading to the growth of oxide p late lets para l le l to the {100} planes. These p late lets grow in th ickness unt i l they reach about 0.002 mm. Fur ther ox idat ion proceeds by nuc leat ion and the growth of add i t iona l p late lets unt i l a l l the meta l is consumed in the react ion. The high act iv i ty usua l ly associated wi th gra in boundar ies in ox idat ion re - act ions is not apparent f rom examinat ion of meta l - ox ide inter faces in this study. No sat is factory ex - p lanat ion for this behav ior can be advanced at pres - ent. Ox ide-meta l react ion, however , was found to nuc leate at d iscrete sites on gra in surfaces, and discrete oxide part ic les can be seen on observat ion of these surfaces. The exact nature of the sites which act as nuc leat ing centers for the ox ide-meta l react ion is not known. Their appearance and d is t r i - but ions suggest the poss ib i l i ty that they are dis loca- tions. This, however , cannot be asserted at present.

Conclusions

On the basis of this study, it is proposed that ox idat ion of Ta proceeds by pre ferent ia l ox idat ion


74 JOURNAL OF THE ELECTROCHEMICAL SOCIETY February 1958

along the {100} planes in the ~100D direction. The complete conversion of the metal to oxide is effected by the nucleat ion and growth of {100} plates, which eventua l ly fill the vo lume of the metal. The ind i - v idual plates do not exceed 0.002 mm in thickness.

On examinat ion of tapered sections, the Ta~O~ shows clearly the crystal lographic dependence of its growth process and the or ientat ion of the traces of the {i00} planes of the substrate. Latt ice imperfec- tions, bel ieved to be dislocations, seem to act as sites for the nucleat ion of the meta l -ox ide reaction.

Acknowledgment The author wishes to acknowledge the assistance

of D. Rogers and N. Harv in in the course of this work. Thanks are also due to W. Bernard for read- ing the manuscr ipt .

Manuscript received Apri l 22, 1957. This paper was prepared for delivery before the Buffalo Meeting, Oct. 6-10, 1957.

Any discussion of this paper will appear in a Dis- cussion Section to be published in the December 1958 JOURNAL.

REFERENCES 1. G. Brauer, Z. anorg, u. allgem. Chem., 248, 1 (1941). 2. S. Lagergren and A. Magneli, Acta. Chem. Scand.,

6, 444 (1952). 3. R. J. Wasilewski, J. Am. Chem. Soc., 75, 1001 (1953). 4. L. K. Frevel and H. W. Rinn, Anal. Chem., 27, 1329

(1955). 5. E. Gulbransen and K. F. Andrews, This Journal, 99,

6 (1949). 6. E. Gulbransen and K. F. Andrews, Trans. Am. Inst.

Mining Met. Engrs., 186, 586 (1950). 7. R. C. Petersen, et al., ibid., 200, 1038 (1954). 8. J. T. Waber, J. Chem. Phys., 20, 734 (1952). 9. J. T. Waber, et al., This Journal, 99, 121 (1952).

10. E. Gebhardt and H. Preisendanz, Plansee Proceed- ings, 254 (I956).

11. A. B. Greninger, Trans. Am. Inst. Mining Met. Engrs., 17, 61 (1935).

12. C. S. Barrett, Structure of Metals, 40 (1952).

Cathodic Reduction of Oxide Films on Iron

II. Determination of a-FeO and FeO

K. H. Buob, A. F. Beck, and M. Cohen

National Research Council, Ottawa, Ontario, Canada

ABSTRACT

A study of the cathodic reduction of a-Fe~O~ on iron has been made with the object of using the technique for the accurate determination of both the a-Fe~O~ and Fe~O4 formed during the oxidation of iron. The films were examined after oxidation and after various stages of cathodic reduction. Ex- amination was made by weight change, electron diffraction, x-ray diffraction, and chemical analysis of the films. Some evidence for the existence of a thin layer of ~-Fe~O~ between the layers of Fe~O~ and a-Fe~O~ was found. A standard pattern for Fe:~O4 was obtained.

When Fe is oxidized in oxygen at temperatures up to about 450~ a duplex film of Fe~O4 next to the metal and a-Fe~O~ is formed. The over-a l l growth of the oxide film should be related to the rates of growth of these two layers. In a previous paper (1) there was described a technique for the cathodic reduction of a-FelOn. In this paper the appl icat ion of this technique to the determinat ion of a-Fe~O~ and Fe~O4 is described.

The method proposed was to weigh the oxidized specimen and to reduce it cathodical ly unt i l the po- tent ia l of the Fe indicated a change in the cathodic process. The specimen was reweighed and the Fe~O~ determined by difference. The Fe~O, was calculated from the total weight gained and the weight of FelOn.

To check on the val id i ty of this method, reflection electron diffraction identif ication of the oxide was made dur ing various stages of the cathodic reduc- t ion process. F i lms were str ipped for identif ication by both diffraction and chemical analysis, and the effect of cathodic reduct ion on magnet i te films was determined.

Experimental Preparation of oxidized specimens.--The speci-

mens measur ing 5 x 1 cm with a handle 2.5 x 0.2 cm were cut from rolled Armco I ron sheet 0.150 mm thick. The Fe contained 0.113% total impurit ies. The specimens were degreased in benzene, wiped dry with Kleenex, stored in a desiccator for 1 hr, and weighed on a microbalance to x 10 4 g. They were then supported in a quartz tube over which a


Documents

Metallographic Manifestations of the Air Oxidation