AN ATOM-PROBE STUDY OF Fe-Ti ALLOYS

The ToF atom-probe FIM was used to study surface segregation of Fe-Ti alloys. Ti segregates to the surface upon annealing at or above 650°C. The degree of segregation depends on the crystallographic orientation, 100% at the (110) and (111), 70% at the (100) and £50% at the (121) plane. It is found that the presence of H2(D2) during annealing enhances surface segregation of Ti and that oxygen deters Ti segregation. An interesting observation is that when a small amount of oxygen exists during annealing in H2 (VL0^Torr ) the thick Ti02 layer(up to a few hundreds layers) is formed on the surface. Titanium is known to trap O2, H2, N2 and C in Fe-based alloys and to improve mechanical strength at higher temperatures by forming Ti clusters when a small amount is added. This is to study the microscopic aspect of the behavior of Ti in Fe-Ti alloys. The focusing-type time-of-flight atom-probe developed in our laboratory was used for this purpose. Our ToF atom-probe has a number of unique *_ tures compared to the others currently in operation. One of the most important is that we can detect all the particles being aimed with 100% detection efficiency, which is highly commendable for the study of surface and interface^). The thermal behavior of Ti in Fe1.48wt%, 0.62wt% Ti alloys was studied by annealing the specimen tip in vacuum, H2(D2) and 02 at the temperature range of 550 to 900°C. (1)•Appealing in vacuum. We previously reported in our preliminary study that Ti segregates to the surface upon heating the Fe-0.29wt% Ti tip(). A more thorough investigation was carried out here accumulating a large quantity of data as a function of annealing temperature, time and crystallographic orientation. When Fe-Ti tip was annealed, a FI image (Fig.l) showed new features distinctly different from that of clean Fe-Ti image, which is very similar to that of pure iron. Fig.1(a), (b) and (c) were taken with P(H2) = lxlO Torr and at 30K by gradually evaporating the surface layers after annealing at 900°C for lOsec. Fine spots forming the inner circle at the (100) and (211) planes and random spots inside the (110) are never observed in the case of pure Fe and are thought to be Ti atoms. Article published online by EDP Sciences and available at http://dx.doi.org/10.1051/jphyscol:1984969 C9-418 JOURNAL DE PHYSIQUE Our atom-probe study supports this view, revealing rich Ti on the surfacecorresponding to those fine image spots. A typical data is shown in Fig.2,where Ti, Fe and their compounds were plotted in the order of detection at the(100) plane. Strong segregation of Ti atoms is evident. It is striking to seethat a large portion of Ti detected at the surface layers is in the form of itsoxides(Ti0 and Ti02). From the net plane size determined by layer-by-layerevaporation, Ti concentration can be plotted as a depth profile, as is shown inFig.3. This figure shows the depth profile of Ti at the (110), (111) and (121) inaddition to the (100) condition. Close to 100% segregation was observed at the(110) and (111) planes while the (121) plane yielded only 50%. In those threecases, Ti concentration decreases rapidly. However the segregation at the (100)plane showed interestingly different characteristics. Ti segregation does notreach 100% at the topmost layer and only 60-70%, but its decay is rather gradualover several layers. According to bond-breaking model, segregation should be mostpronounced at the loosely packed net plane. Our data does not completely agreewith its prediction. We will discuss this later in conjuaction with the influenceof 02 present during the annealing. The effect of the annealing temperature israther simple. Fig.4 shows that in the range of 650"-900°C the depth profile ofthe segregation is essentially independent of the annealing temperature decayingexponentially although small differences are noticed at the second layer.Enrichment of Ti at the surface is almost always followed by a few tens of Tidepleted layers where no Ti atoms or very small number of Ti is detected. We haveat present time no quantitative information on the relationship between Tisegregation and its depletion.We now consider the fact that a large percentage of Ti's segregated at thesurface is detected as Ti0 or Ti02 while Ti's from the bulk are in the elementalform. This is so even the background pressure during the annealing is 610-loTorr range and partial pressure of 02 is monitored by a quadrupole mass spectro-meter is 610-11 Torr. Titanium oxides were found most abandantly at the (100)surface, approximately 80% of Ti at the surface or selvedge is in the oxide formwhile the other surface show 650% in the oxide form. This can be understood bythe fact that the (100) plane is loosely packed and the formation energy of Ti0and Ti02 is, thus, small.(2) Annealing in the presence of oxygen.When annealed in ~(0?)2.10-4 Torr at 900°C, a thick FeO layer was formed on thesurface (Fig.5). surface concentration of Ti is not very high unlike the caseof vacuum annealing and Ti atom tend to form fair size ($15 atoms) clusters bytrapping 02. This is evident in Fig.6 where more than 16 atoms were evaporatedand detected at once by a single trigger pulse. This kind of irregular evapo-ration rarely occured under the normal condition and is safely taken as evidenceof sudden detachment of an entire cluster. Indeed its chemical composition isvery much different from that of the bulk as shown in Fig.7. A large fractionof them are Ti and TiO. Some of the signals corresponding to m/n = 16 are not~i3+and are 0+, most likely the dissociative products of Ti0 and Ti02. Whenoxygen pressure is in the range of Torr, pure Fe and titanium oxides (Ti02and TiO) are intermixed to form a disordered structure over several tens ofsurface lavers (Fin.8)(3) ~nnealin~in the presence of H z (D2).When a Fe-Ti tip was annealed in P(H~)Q~o-~Torr at 650°C, we noticed slightenhancement of surface segregation but essentially same as that of vacuum annealing.However if a small amount of 02 is present in this case, a drastic changedeveloped. A thick Ti0 layers was formed, whose thickness depends on theannealing duration. Fig.9 shows a typical example obtained by annealing for60 min at 900°C. Over 45 layers Ti layers, mainly Ti0 (2.75%) and the rest,Ti (2.20%) and Ti02 (2.5%), were formed with virtually no Fe in this region. ThenTi concentration suddenly dropped off and pure Fe surface emerges. We believethat the effect of hydrogen in this case is minimal, but criticalinenhancing thesegregation rate of Ti to the surface. Now that the segregation rate of Ti isfast enough compared to the diffusion rate of oxygen into the bulk, Titanium atomsat the surface should trap oxygen (external oxidation) and prevent diffusion ofoxygen into the bulk (internal oxidation). The theory proposed by C.Wagner givesqualitatively a good agreement with our experimental observation. References(1) T.Sakurai, T.Hashizume and A.Jimbo, Appl.Phys.Lett., 46, 38 (1984)(2) H.W.Pickering, Y.Kuk and T.Sakurai, Appl.Phys.Lett., 36, 902 (1980)(3) T.Sakurai, R.J.Culbertson, A.J.Melmed, Surf.Sci.Lett., 78, L221 (1978)(4) C.Wagner, Z.Electrochem., 63, 772 (1959) Fig.l FI micrographs of a Fe-1.48 wt% Ti tip after annealingat 900°C for 10 sec.Immediately after annealing,Large flat net plabes appear atthe (100) direction(a) . Gradualfield evaporation produces moreflat planes (b) and (c). Fineimage spots forming inner circlesand random spots surrounding themare thought to be Ti.