MICROSTRUCTURAL CHARACTERIZATION OF SPRAY FORMED

Ni-Al-Cr-C ALLOYS Aroldo Mourisco (1), Danilo B. Cunha (1), Helio Goldenstein (2), Claudio S. Kiminami (1), Claudemiro Bolfarini (1). (1)- UFSCar- Universidade Federal de São Carlos, DEMa- Dep. de Eng. de Materiais, P.O.Box 676, São Carlos - SP - 13565-905, Brazil. (2)- EPUSP- Escola Politécnica da Univ. de São Paulo, Dept. de Eng. Metalúrgica e de Materiais, Av. Prof. Mello Moraes, 2463, C. Univesitária – São Paulo – SP - 05508-900, Brazil.

pmour i s@ir i s .u f s car .b r ABSTRACT In this study three Ni-Al-Cr-C nickel based casting alloys with 0.5, 1.0 and 1.5wt% carbon content were processed by spray forming aiming to investigate the potential of achieving substantial microstructure refinement by the high cooling rate involved in this process. Some attempts were done in order to evaluate the wear resistance of these alloys by means of pin-on-disk tests. Two values for the gas to the metal flow rate ratio, GMR, were used ( 0.12 and 0.23) and nitrogen was used as the atomization gas. The overspray powders and the deposit were characterized by using optical and scanning electron microscopy. The high cooling rate resulted in a strong microstructural refinement, with carbides of about two order of magnitude smaller than those obtained in the conventionally cast materials, dispersed in a predominantly gamma-prime matrix. Higher GMR led to a more refined microstructure due to higher cooling rate imposed to the atomized droplets. The microstructure observed in the deposit could be correlated with that observed in the overspray powders, indicating transformations during deposition process. The atomized alloys having higher carbon content (1.5%) and consequently higher volumetric fractions of carbides, presented better wear resistance. Their results are similar or slightly better than those obtained for a conventionally cast high chromium white cast iron. Keywords: spray forming, nickel, rapid solidification, carbides, intermetallic INTRODUCTION A new family of wear and high temperature resistant nickel based cast alloys containing Ni3Al intermetallic compound, known as “White Cast Intermetallic Compounds, WCIC” has been introduced by Yoshimura and Goldenstein [1]. These alloys, containing a hard carbides dispersion in a gamma-prime matrix, were proposed as substitutes for refractory steels and conventional cobalt and nickel based wear resistant cast alloys for use at high temperatures. WCIC are basically Ni-Cr-Al-C alloys, with composition balanced in order to have a microstructure with predominance of ordered gamma-prime matrix with a dispersion of eutectic and in some cases also proeutectic carbides. [2]. The predominantly γ’ matrix takes advantage of the anomalous increase of flow stress with the temperature until a peak (anomalous flow stress peak phenomena that several intermetallic compounds with the L12 ordered structure exhibit) to improve its high temperature resistance. In order to increase γ’/γ’ grain boundaries ductility 0.06 wt.% B is usually added [3, 4]. This work aims to investigate the influence of rapid solidification on the microstructure and properties of WCIC alloys by using the spray forming process. In this process the atomized droplets are consolidated onto a substrate to a dense deposit. Rapid solidification allow generally a strong refinement of the microstructural scale and even in many ferrous alloys, an extension of the solubility of carbon and carbide formers on the austenite phase [5]. The grain size in spray formed deposits is determined by the conditions of the spray upon impact, the spatial distribution of solid particles after impact, the time for solidification and the cooling rate after solidification [6, 7].

EXPERIMENTAL PROCEDURE Three different compositons of WCIC nickel based alloy were studied, as presented in Table I. Baths of 4 kg of each composition were prepared using electrolytic nickel and chromium, commercially pure aluminum AA1100 and Al-5wt%Ti-1%wt%B master alloy. The alloys were melted in an induction furnace under argon atmosphere using alumina crucibles. For the spray forming processing the molten alloy was superheated up to 1600°C and atomized by nitrogen. Two values for the gas flow rate to the metal flow rate ratio, GMR, were used (0.12 and 0.23). Flight distance was adjusted to 360 mm. To compare the microstructure of the spray formed alloys with the conventionally cast alloys samples of the three composition were prepared by melting under argon atmosphere and casting in a ceramic mould (inner diameter 12mm, length 150mm). The characterization of the conventionally cast samples, deposits and the overspray powders (droplets which do not impact on the deposit top surface or/and bounce off impacting droplets or powder from the deposit surface) was performed by optical microscopy (OM) and scanning electron microscopy (SEM). Hardness tests were performed by using Vickers hardness with a load of 1kg. Wear resistant pin-on-disk tests were performed according to ASTM G99-90, under load of 515g using cylindrical samples against a rotating (53 rpm) disc. Abrasive papers, 200 and 600 mesh, were glued at the disc surface before each test and the mass loss of each sample was measured after abrasion test.

Table I. - Chemical composition of the three alloys Alloy %Ni %Al %Cr %C %B

1 and 4 82 9.44 8 0.5 0.06 2 and 5 82 8.94 8 1.0 0.06 3 and 6 77 9.44 12 1.5 0.06

It will be used, hereafter, the codes conv and atm for conventionally cast and atomized alloys respectively. Table 1 shows the chemical composition of alloys conv-1, 2 and 3., whose chemical composition are the same as for the atomized alloys : atm-1, atm-2 and atm-3, respectively. The alloys atm-1, -2 and -3 are corresponding to spray deposited alloys atomized under 1.0bar of gas pressure ( GMR = 0.23 ). The alloys atm-4, 5 and 6 present the same spray deposited alloys atomized under 0.5bar of gas pressure ( GMR = 0.12 ). RESULTS AND DISCUSSION The deposits presented a gaussian-like geometry with an outer diameter of about 200mm and a central height of approximately 60mm. Qualitatively it was possible to observe that the amount of overspray was inversely proportional with the carbon content of each compositon. This results can be explained by the decreasing of the solidus temperature with carbon content and the fixed pouring temperature used for all composition, i.e. the superheating temperature was higher for higher carbon content alloy. Figure 1 presents the microstructure of the spray formed alloys processed under high GMR (0.23) and containing 0.5%C (a), 1.0% C (b) and 1.5%C (c) and processed under low GMR (0.12) and containing 0.5%C (d), 1.0%C (e) and 1.5%C (f). Some previewed phases from the Ni-Cr-C phase diagram are graphite, Cr2C3, Cr7C3 e Cr23C6 and from the Ni-Al binary diagram are γ, γ’ and β. A previous work [4] presented some results of X-ray diffractometry for the same alloys which are being studied here and these results will be used hereinafter in order to help understanding the microstructures showed in figure 1. According to these results the carbides which are being observed in the microstructure are mainly Cr3C2. The black points and regions in the micrographs are regions of carbides extracted by an overetching, which was necessary to show the grain boundaries.

(a) 25 µm (b) 25 µm (c) 25 µm

(d) 25 µm (e) 25 µm (f) 25 µm Figure 1.- Optical micrographs of the deposit obtained under high GMR, and containing (a) 0.5%C, (b) 1.0%C and (c) 1.5%C, and the deposits obtained under low GMR, containing (d) 0.5%C, (e) 1.0%C and (f) 1.5%C. The grain sizes measured by the line-intercept method was 11±6 micrometers for both 0.5 and 1.0%C content deposits. We were not successful to measure the grain size of 1.5%C content

deposit. Comparing the microstructures for the same GMR, one can assert that the amount of acicular carbides rises as the carbon content increases. Examining the size of the carbides shown in figure 1 it is possible to assert that the carbides are smaller for the deposits obtained under higher GMR. This fact can be explained by the higher cooling rates prevailing by these processing conditions and then, favoring higher carbon content in solid solution. At the same time, higher GMR generally results in smaller droplets (smaller mass media diameter of the atomized droplets) and then, smaller droplets can, by higher cooling rate, generate smaller carbides. It can be observed as well in figure 1 that in alloys atomized under higher pressure (high GMR) the carbides present maximum size of about 5 µm. Otherwise, the alloys atomized under lower GMR have shown carbides slightly larger. In conventionally cast alloys, carbides of minimum size of about 4µm can as well be seen in the microstructure. However, the majority are very larger and carbides as big as 100µm can be found. In the X-ray diffractograms [4] four peaks are superposed and are related to γ or γ’ and the majority of the peaks observed for the conventionally cast alloys are also present in the diffractograms of atomized alloys, suggesting the presence of similar phases in the alloys obtained by both processes This could suggest that γ and γ’ are present in materials produced by both processes. On the other hand, some preliminary transmission electron microscopy observations revealed an extensive presence of typical diffracted spots of ordered structure, showing forbidden diffraction pattern, for the spray deposited materials. Therefore, the presence of γ’ in the matrix of the spray deposited materials was confirmed. Moreover, the presence of minor intensity peaks which are depicted only in the diffractograms of the atomized alloys, and identified as related to γ’ phase, also suggests a massive presence of this phase in the atomized deposits. The absence of these peaks in conventionally cast alloys could also be related to an effect of texture, which could hide such peak in that case. The peaks related to the Cr2C3 carbides are present only in the diffractograms of the spray formed deposits. Figure 2 shows the microstructure of a overspray powder (a-b-d, 1%C) and of the conventionally cast alloy (c, 1%C). Analyzing the same alloy composition (1%C), by the conventionally cast material it can be seen (fig.2-c) a γ’ shell around the carbides. It has been stated [1,2] that the small domains of ordered γ’, in a disordered γ matrix, are regions that solidifies as γ and had a transformation to ordered γ’ in solid state. There are also some regions where γ’ is associated to the proeutectic and eutectic carbides originating regions of haloes at around the carbides. It suggested that the reaction which forms γ’ could take part in the eutectic carbide formation It can be seen in figs.2-(a) the SEM picture of a fine particle of overspray. In fig.2-(b) it is shown a relatively coarse particle of overspray, and in fig.2-(d) a higher magnification of fig.2-(b). Comparing fig.2-(a) and fig.2-(b) one can observe the presence of very fine dendrites in the coarse particle and a near-cellular structure for the smaller particle. In both particles a net of fine carbides is present in the contours of dendrites or cells. Taking a higher magnification picture (fig.2-d) of the coarser particle, a very fine net of carbides can be observed. Comparing this fine net of carbides, with that of fig.2-(c), it can be observed approximately, the same carbide configuration in both cases, but remarking the significant difference between the magnifications (five times) of both pictures. Despite the similar carbide arrangement found in the overspray particle and in the same conventionally cast alloy, the already mentioned haloes around the carbides are not present in the overspray powder, suggesting a different way of solidification. In the same way the deposits obtained by spray forming do not present the mentioned haloes around the carbides.

(a) (b)

(c) (d) Figure 2.– SEM micrographs: (a) fine overspray particle having 1%C showing non-dendritic structure; (b) large overspray particle showing dendritic structure; (c) conventionally cast sample with 1%C showing large acicular carbides; (d) overspray particle with 1%C presenting near cellular structure, similar to the one present in (c) in different scales. (GMR=0.23, Etched) The pin-on-disk wear results are presented in table II. In order to compare with a well known wear resistant material, wear tests were also performed on a conventionally cast high-chromium white cast iron (2.98%C-19.7%Cr), here named WCI.

Table II.- Measures of mass loss during wear resistance tests at room temperature. Material Loss of mass (%) 600mesh Loss of mass (%) 220mesh HV(N/mm2)

WCI 0.0517 ± 0.0079 0.1282 ± 0.0052 647 ± 25 Atm-01 0.1286 ± 0.0138 0.3225 ± 0.0132 455 ± 13 Atm-02 0.1287 ± 0.0043 0.4083 ± 0.0641 467 ± 4 Atm-03 0.0745 ± 0.0127 0.4128 ± 0.0595 573 ± 5 Atm-04 0.0528 ± 0.0130 0.2860 ± 0.0441 474 ± 9 Atm-05 0.0648 ± 0.0063 0.2872 ± 0.0919 492 ± 8 Atm-06 0.0303 ± 0.0042 0.1629 ± 0.0195 506 ± 7

Analyzing the results using abrasive paper 600#, it can be observed similar, and even better results for the spray deposited alloys, than that of WCI. As expected, the atomized alloys having higher carbon content (Atm-3 and 6), and consequently higher volumetric fractions of carbides, present better wear resistance. Keeping in mind that these alloys have similar and even better wear resistance than WCI, it can lead us to assert that the wear mechanisms involved in both materials are similar, that means, the microfurrow and squeeze of material in ductile matrices. In WCI the ductile part of the matrix is the austenite.[9] Both wear micromechanisms above mentioned were also observed in preliminary examinations by SEM, of the spray deposited alloys. Analyzing the results using abrasive paper 220#, it can be observed the better results for WCI. Comparing the spray deposited alloys atomized under high GMR (Atm-1,2,3) and under low GMR (Atm-4,5,6), one can see better wear resistant results for the low GMR alloys, where the carbides are slightly larger than the ones of alloys atomized under higher pressure. It can suggest that larger carbides could better protect the ductile matrix against the wear damages of coarse abrasives (220#). Applying higher gas pressures for atomizing can increase the solidification speed in such a way that the carbon in these alloys can remain in solid solution, and it could explain why alloys with the same carbon content have different sizes of carbides leading to different hardness and different wear behavior. CONCLUSIONS The microstructure of the alloys processed by spray deposition, when compared with the same alloys processed by conventional casting, are very different; the main difference being the refinement of the carbides. However, the nature of the carbides is similar and Cr2C3 is the main carbide type found in spray formed and in conventionally cast alloys. Despite the lower values of Hardness when compared with high chromium cast iron, the spray formed Ni-based alloys here studied presented similar and for some processing conditions even better results of wear resistance by using pin-on-disk tests. ACKNOWLEDGMENTS The authors would like to thank to “FAPESP - Fundação de Amparo `a Pesquisa do Estado de São Paulo” (Projs. n. 2000/05893-3 and 2000/00873-4) for financial support; H.G., C.S.K and C.B. acknowledge partial support from “CNPq - Conselho Nacional de Desenvolvimento Científico e Tecnológico”. REFERENCES [1]- H.N, Yoshimura, H.Goldenstein, In: 51o CONGRESSO ANUAL DA ABM, 1996, Porto Alegre - RS -Brazil. Proceedings...1996, p. 287-299. [2]- Y.N. Silva, H.N. Yoshimura, H. Goldenstein In: 2o CONGRESSO INTERNACIONAL DE TECNOLOGIA METALÚRGICA E DE MATERIAIS, 1997, São Paulo-SP- Brazil- (Proceedings in cd-rom). [3]- J.H Westbrook., R.L. Fleischer, Intermetallic Compounds – Principles and Practice, John Wiley & Sons editions, v.1, ch..21, (1995). [4]- A. Mourisco, Y.N. Silva, H. Goldenstein, C.S. Kiminami, C. Bolfarini, EUROMAT-2001, 2001, Rimini - Italy (Proceedings in cd-rom). [5]- M. Boccalini Jr., H. Goldenstein, Int. Mat. Reviews, v.46, Nº2, pp 1-23, 2001. [6]- A.R.E. Singer, J. Institute of Metals, v.100, p.185-190, (1972). [7]- R.W. Evans, A.G. Leatham, R.G. Brooks, Powder Metall., v.28, p.13-20, 1985. [8]- M. M. Pariona, C. Bolfarini; R. J. dos Santos; C. S. Kiminami, J. of Mater. Proc. Tech., Vol. 102/1-3, pp.221-229, 2000. [9]- K. H. Zum Gahr, D. V. Doane; Met. Transactions A, v.11A, p.613-620, 1980.