JOURNAL OF RESEARCH of the National Bureau of Standards-A. Physics and Chemi ~ t ry Vol. 64A, No. 1, January- February 1960 Nitriding Phenomena in Titanium and the 6Al-4V Titanium Alloy
1. R. Cuthill, W. D. Hayes, and R. E. Seebold I
(September 21 , 1959)
::\ i t riding unalloyed Litan ium in pu rifi ed n iLrogen at J,800° F produced a ulliformly t hick ca~e that co nsisted of fi ve d istinct zones. The sam e Lr eatm ent applied Lo a 6AI-4V t i tanium alloy res ulted in a t hinner niLride cas£' t hat co nsi ted of t hree zones and elongn,ted nitride grains t hat penetrated into the co re at approximately 45 deg rees to the spec imen ;.; urface. The aluminum appears Lo be responsible for Lhe format ion of the elongaLe d grains. Tlwse grains, in t um, appear to be rC'sponsible for the nitriding havi ng a m or e adverse effect 0 11 tlw t oughness of Lhe alloy than of 1IH' unalloyed titanium, as indicated by preliminary ilTlpact test resu lts. The nitride case 0 11 Lhe t itanium appears to i ncr ease in t hickness with inerrase in nit riding tintC withouL limi t. The nitride case ex hibits a hardlH?ss equivalcnt to abollL 77 R ockll'cll CaL t he su rface down to almost 50 R ockwell CaL the inte rface.
1. Introduction em britLlement of t he Litanium may be induced during lJitriding in ammonia [6]. Tlierefore, in Lhe Titanium is knolnl to han: a s trong affinity Jor iniLial phase of this currenL program, it was decided ni Lrogen and to form a ni tride case of high hardn ess Lo nitride in purified nitrogen in order to avoid when heaLed in the presen ce of nitrogen or ammonia. any poten tial complications involving hydrogen However, th el"(' has been liLtle previous inve Liga embriLtlemell t . tion of Lh e rffecL of m etallic alloyino- clem ents on t he 2. Materials morphology of Lit e lliLride phase or phases gen erated in the course of t he nitriding treatment. This p aper The Armou r Research FOLlndaLioll [5] i nvestigftted is con cerned primarily with the efl' ect of aluminum Lhe effect of v arious allo~r addiLions on the n itriding and vanfld ium . The morphology of Lhe nitride case cltaracLeris tics of LiLa nium, and reporLed th at vana produced in commercially pure LiLanium is com dium and boron wer(' LIJe only elem enLs found to pared with the nitride formations produced in a res ulL in an increased n i Lr iding rale and incr eased tiLanium 6AI-4V alloy. The work was carried ouL in ultimftte hftl'ciness. TIle 6AI-4V alloy, which r e the course of a broader program of m aLerials in ('e Il L l~r h as ("ome inLo widespread usc, therefore, vest igaLion for the N a Iry Bureau of Aerona u Lics appeared Lo be of s ufficient theoreLic al as well as regarding an applicaLion where good r es is Lance to pracLical inLerest Lo warranL inclusion in t he program galling and weal' beLween s urfaces in sliding contact along wi lh C'ommel"C'ial1~r pure t i Lanium. is r equired , as well as hi gll surface h ardness, light The ch emical composiLio ll s a nd tell sile properties weigh L, good saIL water COlTO iOll resis Lallce, flnd of both maLerials, as fLlrnis hed by t he supplier, arc good m achinability . W yatt and Gr an t's [1]2 ex g iven in Lable l. Toe m aLcrial was specified as p eriments indicated LhaL the nitriding of ti tanium double arc melLed "aircra[L quali ty" O.050-in. thick did produce a s urface with saLisfacLory properLi es s heet, 1\0. 2 fini sh . r equired for Lh e mftintenance of surfaces in sliding contact. T AB I. E 1. Composition and properties of t-itanium and 6Al-4 V :\1any procedures for surface hardening LitaniulD alloy ' have b ee n Lried b~r previous investigators, includi ng carburizing, ni triding, and anodiz ing [2 , 3, 4, 51. I t Ti!.anil!m 16.\ 1-4 \ ' alloy h as been generally concluded that ni triding is th e Y i:'Ici strelloth: mos L promising m ethod for producing a h ard , wear L onzilucl illal _ p~i :: 4.
I Physical chemist at thp ).;ava l R esC'urch Laboratory, \\'ashington . D .C. 2 Figures in hrackets ind icate thr litr raLu re I'cfpren ccs at the end of this paper. • D ata furnished b y manufacturer.
119 3 . Experimental Procedures " Water-pumped" grade nitrogen was passed through a liquid oxygen cold trap to remove water Specimens. approximately ~{6 in. by 2H in. , were and then over titanium turnings at 1,8 00 ° F to re <' ut from the 0.050-in . thi ck sheet; all of the specimens move oxygen before entering the nitriding chamber . of each material were cu t from the same sheet. The In spite of its strong affinity for nitrogen , titanium edges were milled so as to permit the accurate deter was used in the getter furnace because of its superi mination of the cross section of specimens uaed for ority in removing oxygen. Fortunately, the oxide unnotehed Charpy-type impact tests. Preparatory reaction is more rapid than the nitride reaction, but, t o nitriding, the specimens were pi ckled for 15 min in more importan t, the titan ium nitride that is formed the followiu g standard pickling solution [7] at 80° F: even tually rever ts to th e oxide. However, to insure high efficiency in oxygen removal, the charge of H N 0 3 47 parts by volume titanium turnings was replaced before every run. I-IF 2 par ts b~- volume Gas leaving thc nitriding cll ft mber was passed H 20 .51 parts by volume through a mercury bubble trap and then exhausted Upon removal from the pickling solu tion th e speci through a wet-test gas meter, which permitted the m ens were thoroughly washed in tap water and air absolute pressure in the nitriding chamber to be dried. maintained anywhere between 1 and 2 atm . How Specimens were nitrided at 1,800° F for 4, 16, 24, ever, tests early in the progl'J,m revealed no signific:1nt 48, 96 , or 168 hI', to ob tain data on case thickness difference in nitriding ch aracter istics res ulting from versus time. Figure 1 is a drawing of the nitriding varying the pressure over this en tire range. There furnace. The ni triding chamber was an Incond fore, the bulk of the specimens were nitridcd with a tube wi th water-cooled ends. TJlC specimens were pressure of 2 in. of ll1('}'cury above atmospheri c in held in th e cool "one at tlte rear of t he ni triding the nitriding chamber. chamber until the chamber with its purifi ed nitroge n More t han 300 spec imens were nitrided for 48 hI' atmosphere was brought up to temperature. Each in the course of this program. To facilit a te n itriding specimen was then individually drawn in to the hot of this many specimens a much larger chamber was zone by magnetic attraction acting on t he iron bob also used . This larger chamber was also of Incond attached to t he specimen by means of a fin e wire. and the same type of gas purifying system was used, E ach specimen, at the end of its preseribed time, but of correspondingly larger capacity. The speci was drawn into t he cool zone at the front of the mens were sealed in the chamber, and, with the puri chamber b:--- the sam e meelJauism . fied nitrogen fl owing through the system, th e chamber
r------44" ------~ 9 f ------r------25f ------r~------9 " IfIt------~------24 " ______---, IL %,00 BRASS OUTLET l 14 TU BE "", I ~~~-r~~~~~~~~~~~~~~ TRAN SI T E EN D PLATE S
WATER COOLING BRASS FLANGE B CO ILS \ END PLATE 5.5. DISC RADIAT ION SHIELDS
2," OD X 32X WA LL SPLIT PORCELA IN INCOLOY T UB E RING SEAL. EA CH EN D
. 3foD x f WALL 16 GA . NICHROME :ll. WOUND ON PORC ELAIN TUBE ;f MANDREL. 40 F1 I WINDING. 3 WINDINGS, 6TURNS EA CH . ~' I
I RON WEIGHTS ( 6 mm PY REX TUBE
NITRIDING FURNACE
\.
F I G U RE 1. Schematic diagram oj appamtus J OI' nitridi ng specimens f ol' dijJel'e nt lengths of time without opening chambe'.
120 was then rolled into Lbe [-urnace after the latter was The immediatelv evident diA'cl"C'll ce betwecn the up to temperature. At t he end of the prescribed metul and the "all oy, considerirw only th e gross time, the chamber was rolled out of the furnace and characteristics, is that the metal forms a uniformly allowed to cool to room temperature before the at thick case with a sharp line of demarcation between mosphere was turned off and the chamber opened. the case and the core, whereas Lhe all o ~r exhibit elongated nitride grains penetrating as far as the 4. Results and Discussion ce nter of the core, given sufficient time, and with the lon a axes at approximately 45 dcg to the surface. 4.1. Microstructure These
&2
,I -1
FIGUH.E 2. Titanium nitrided in purified nitrogen at 1,8000 F FIG URE 3. 6 1/·417 titanium alloy nitrided in purifi ed nitrogen for the lengths of time indicated. at 1,8000 fi' fm' the lengths of ti me indicated. E tched in 1 part J-IF (48%) , J2 parts U N 0 3 (cone), and 87 parts H 20 . X 200 Samc etehant as for fi gurc 2. X 200. 121 matrix. These elongated grains penetrating in to traverse across the case. The change in etching j the core might be expected to act as stress raisers characteristics is not unexpected because the TiK and result in lower impact values . Preliminary structure is retained from the stoichiometric compo results do indicate that nitriding has a more adverse sition down to a deficiency in nitrogen corresponding I effect on the impact properties of the allo~T than to the composition, TiNo.d8]. It might be mentioned of the unalloyed titanium. here that to retain the surface film on specimens I The hardness of the surface film on both materials prepared for metallographic examination, it was . was equivalent to about Rockwell C 77, and decreased necessary to insert sheet plastic between the speci- . to approximately C 50 at the interface between the mcns before placing the pack of specimens in the case and the core. The elongated grains in the alloy mounting press for moulding in plastic in the con also exhibited hardnesses equivalent to about Rock ventional manner. Also, it was preferable that the well C 50, in contrast to only about C 35 in the polishing be done on a wax lap. adjacent matrix. At the interface bctwee n the case and the core Upon closer examination the " uniform case" on of the Ullalloyed ti tanium there is a "fri ll ge" of the unalloyed titanium (fig . 2) is seen to consist of a each of the ni tride grains of the case protruding over surface film covering a thicker or principal layer. the prominen t boundary of the case into t he region This principal layer is approA'imately 0.004 -in. thick of the core. That this fringe is actually an integral 1 and the surface film is about 0.0005-in . thick on part of the respective grains of the case is r evealed specimens that have been nitrided for 48 hr. There by polarized light, as will be seen by comparing t he is an abrupt change in etching characteristics at tilc upper and lower photographs in figure 5. Polarized middle of the layer. Nitriding for a longer time light (lower photograph) also reveals that the case is brings out a sharply defined transition band between composed 01' grains of different orientations but the upper and lower halves of the layer, as may be generally with each grain extending completel~ - " seen in figure 4 which also shows a microhardness through tho principal layer of th e case and i.neludillg I the "fringe" at the interface with the core. However, V' _I the thin surface film is seen to be composed of grains pe separate from those of the principal layer . Figure 6 - ~ -1650 is an interference pa,ttern of the same region shown in 6gure 5, obtained with a Zeiss ver tical interferometer -1425 microscope. The "fringe" of the "hour glass" shaped .' grain is revealed to be below the portion of the grain 1210 making up the principal layer. Note that the scratches, which are valleys in the specimen surface, provide a m eans of identifying which are high and which are low areas in the interferometer patterns. 784 Wyatt and Grant [3] co ncluded that a maximum case of approximately 0.004 in. was attained when unalloyed titanium was nitrided in ammonia. The 784 time to produce this maximum was a function of the nitriding temperature. If nitriding were COll tinued beyond the time required to produce the 659 maximum thickness, the case was reported to actually decrease in thickness. They concluded that the mechanism responsible for the decrease in 593 case thickness was the decomposition of the TiN structure at the interface between the case and the ~ 538 core and the nitrogen diffusing into the core at a faster rate than nitrogen diffused through the TiK 303 case to form new case. However, in the present investigation, no evidence for a maximum was found. As shown in figure 7, the thickness of the nitrided case on titanium increased with an in crease 232 in time at 1,800° F ; a thickness in excess of 0.008 ill. was obtained by nitriding 168 hr. Also it is secn that the plotted points fall reasonably well on a straight line, which indicates that the increase in 205 total case thickness with time is a parabolic rela tionship.
FIGU RE 4. Jl!Iicrost rtlcture and Vicker·s miCl"ohardness indenta The growth of the elongated nitride grains in to tion (50-g load), and corresponding Vicker-s hardness ntlm the matrix of the 6Al- 4V titanium alloy at 45° with r bers, across the case on a titanium specimen n itrided 168 hr the surface in bot h directions (fig. 3), appears to be at 1,8000 F in purified nitrogen. associated with the cube planes of a beta phase H ardness range corres ponds to Rockwell C 77 to Rockwell C 12. Same etchant as for fi gure 2. X 350. matrix at 1,800° F. vVith an electron-probe micro- 122 :analyzer, vanadium content in these elongated nitride grain s was fou lld to be only one-half of that in t he matrix:. Furthermore, thero was no observ able vanadium composition grad ient in the matrix in the immediate vicinity of t he nitride grains. Determinations wer e made on 18 o( the elonga ted nitride grains in a specimen which had been nitrideel for 48 hr. On a sample of pure vanadium, 480 counts per secolld were recorded, and on a sample of pure titanium 66 cps were r ecorded when the G eiger coun tor was et on t iiO VKa wavelength . Correcting for this background , 22 to 24 cps were
Onc fri ngc= 11.8X lO-6 inch d iffcrencc in elevation.
7
6 c vi (J) ILl z 5 :.£
J <.l .... J: I- 4 • ILl (J) 2 • o ~O~---2~--~4----~6-----Le----~----~--~14 FIGURE 5. Case and interface between case and core in titanium -J NITRIDING TI ME', hr nitrided 48 hr- at 1,8000 F in purified nitrogen. Same etchant as fo r figm e2. X 500. Upper: white light. Lower: polarized F IG lJ RE 7. Case thickness in titanium versus square 1'00t of li ght. nitriding time at 1,8000 F. 123 obtained recording the VKa wavelength on the Using the relationship between the free energy, tJ. F, nitride grains and 44 to 48 cps on the matrix. and the equilibrium constant, K , The growth of these elongated nitride grains is obviously a function of nitriding time. Nitrogen 10gK= and aluminum are strong alpha stabilizers in titanium 2.30 RT but vanadium is a strong beta stabilizer. The fol lowing calculation of eq uilibrium constants suggests The equilibrium constants for tb(fonnation of VN, that the aluminum pla~' s a prominent role in the AIN, and TiN, at 1,800 ° F, are, formation of the elongated nitride grains. From Kubaschewski and Evans [9], the heats of formation, !:IF, are: 2V + N2 ---72(VN) !:IF = - 83,300 + 39.7 T cal/mole (a) = - 3:3, 500 cal/mole at 982 0 C (1,800 0 F ), 1 K 53 ,800 9 ..·3~{ og A!N= + 2.303 RT 2AI + N 2---7 2 (AIN) !:IF =- ]21 ,000+ 44.5 T (b) = - 65 ,200 cal/mole at 982 0 C (1 ,800 0 F ), 104,750 log K TlN=+ 2.303 HT 18.22 2Ti+ K 2-?2(TiN) !:IF =-160,500 + 44.4 T (c) =-104,750 cal/mole at 982 0 C (1,800 0 F ). These resul ts impl~~ that the ni trogen would com bine preferrntially with the aluminum in respect to However, to give a more realistic comparison, the vanadium, an d support the expel'imen talrcs ult that data for alumin um ancl. vanadium should be correctcd the aluminum must be largely r esponsible for the for a 4 and a 6 percent solution, respectivel.,-, of df'velopment of the elol1gated nitride grain s which r alumin um and vanadi um. in titan ium. As a "bes t do not occur in the absence of both the aluminum and approximation," in lieu of the necessar~' experimental the vanadium. Furthermore, both the thermo data on these systems, an ideal solution is assumed d~· lI amie calcul ations and the electron probe results and the free energy of dissolu tion for the correspond indicate that the vanadium is not responsibl e for the ing atomic percent con cen trations, i.e., 10.2 atomic formation of these nitride grains. Obviously, a percen t aluminum in titanium, and 3.8 atomi c per determination of the alumi.llum con ten t of the cent vanadium in titanium, is computcd from the clon$'ated graills is needed to confirm or refute this relationship CO Jlciusioll. Such a determination will be made when !:l FM= RTln N a vacuum spectrometer-equipped electron probe, now where, under construction, becomes available. !:lFM= tll e free e n e rg~r of mixing cal/mole of solute. 4.2. X-Ray Diffraction Results The expression on the righ t is really th e en tropy of mixing but because the heat of X -ray diffraction patterns were obtained of the solu tion is zero for an ideal solution, the surface of the titanium and 6AI-4V alloy before and en tropy term constitutes the en tire free after ni triding, and again after the ni tride surface J energy. layer had bee n polished oft'. " R = the gas constant, 1.987 cal/mole/deg Un-ni trided titanium exhibited the expected alpha titanium pattern except that the (002) reflection was T = the absolu te temperature, 0 K N = mole fraction of solute equally as in teJJ se as the (011 ), the most intense line as reported in the AST.M Index. This is probably due to preferred orientation as a resul t of rolling. At 1,800 0 F, one obtains for, AI---7 [10.2 a/o Alhh !:IF= -5,700 cal/mole The pattern from the nitrided surfaee revealed the expected Ti.N pattern, except that the most in tense and for, line was also the (002) reflection instead of the (220) reflec tion as given in the ASTM Index, and also an V ---7[3.8 a.jo V]n !:IF= - 8,200 cal/mole. extra line corresponding to a "d" spacing of 1. 52 A. This was the most in tense line in the diagram. After Combining these results with the values given for polishing off most of the nitride layer, this extra line the original reactions, (a) and (b), gives had not diminished in int en s i t~r although the TiN had been reduced drastically and the alpha titanium 2[Vhl+ N 2 -?2(VN), !:I F=- 17 ,100 (a/) pattern was beginning to appear. The alloy, before nitriding, exhibited the expected 2[Alhl+ N z-?2(Al), !:IF= - 53,800. (b /) ti tanium "d" spacings, but the relative in tensi ties 124 L differ appreciably from those nonna ll ~ ' expected from core of Lhe allo.,· at 45 d og to t llO surface, appar e ll tl~ ' the pure alpha t itanium pattern . There was no act as tres raiser causin g a significant reduction indication of beta phase . T he extra lin e corres pond in the impact trength of the 6Al-4V alloy. ing to it "d" spacing or 1 .52 A occu rred also in the pattem of the ni trided all oy su d ace and in the pat 6. References tern of the alloy after the nitrided surface had been polished off. 1] J . L. WyaLL and N. J . G ra nt, N it rid ing improves t itanium properties, Iron A3e, pp. 124· to 127 (J a n. 28, 1954) . 5 . Summary and Conclusions [2] A. J . Gri est, P . IE. ;VIoo rehead, P. D . Frost, a nd J . H . J acks n, Surrace ha rdening of titanium by carburizing a nd induction heat treatment, ASM T rans. 46, 257 1 Nitridi ng of unalloyed ti tanium in purified Ili trogen (1954). ~ produced a sharply defined and uniformly th icl,;: case [3] J . L . Wyatt and N. J. Grant, N itriding o f titanium ,,·ith which increased in thickness with increase in ni trid ing a mmonia, ASM Trans. 46, 540 (1954). [4] J. L. Wyatt a nd N . J . Grant, N it riding produces better time and apparentl~ ' would co ntinue to increase ha rd case on titanium, Iron Age, p. 112 (J a n. 14, 1954). witllOut limit. Five distin ct zones could be iden tified. [5] R . 'vV. H a nzel and V. Pulsifer, Surface hardenin g of t ita There is no evidence of a maximum case thickness nium with metall oid element, June 1,1952, to May 3 1, being reached followed by a decrease in case thickness 1953 Rept. (A rmour Research Foundation to Water town Arsena l on Co ntract No. DA- ll- 022- 0RD- 289) . with continued nitriding, as reported by other in - [6] 'E . L. White, P . D . Miller, a nd H.. S. Peoples, Ant igalling .., vestigators who nitrided titaniulll in ammonia. coatin gs and fabricants for t itanium, Battell e TML The same treatment applied to a 6Al-4V alloy of llept. No. 34, p . 10 (Feb. 24, J 956) . t itanium resulted in a thinner nitride case that con [7] Metals Handbook, ]954 Suppl. , Am. Soc. Metals, p. 83 (1954) . sisted of three zones and in elongated nitride graills [8] P. E llrlich, Z. Anorg. Che;n. 259, 1 (1949). t hat penetrated into the core at approxim ately 45° [9] O. Kubaschewski and E. LL. Evans, M etallurgica l wi th the surface of the specimen in both directions. Therm ochemi t ry (J ohn 'Wi ley &; SO il S, lew York, The aluminum i indicated to be primarily responsi .Y., 1956) . ble for the format ion of the elongated nitride grains. :/ The elo ngated ni tride grain s, penetrating into the WASHINGTON, D.C. (P APER G4Al-33) I, ''" ~ 125 I