New aspects regarding the WAXD investigation of spheroidal cast iron inclusions.

Pencea, Ion ; Stefanescu, Doru Michael ; Ruxanda, Roxana 等

1. INTRODUCTION

The cast iron functional characteristics depend on carbon inclusion (CI) morphology (Stefanescu, 1988). The morphology of CI is strongly related to the CI atomic structure (Pencea, 1997). The wide majority of metallurgists uses the term graphite for CI and do not make any distinction between crystalline graphite and turbostratique carbon/graphite that is more appropriate to CI structure (Cojocaru-Filipciuc, 2007; Riposan et. all., 2007; Double&Hellawel, 1995). The authors try to prove that the most ordered CI in cast iron called spherulitic graphite (SG) have a turbostratique structure. By the authors knowledge, the SGCI could not be of polycrystalline graphite nature but, more likely, of turbostratique one. In this sense, there were investigated by Wide Angle X-ray Diffraction (WAXD) technique powder samples of SGCIs obtained by dissolving a magnesium treated cast iron matrix. The WAXD technique has the advantages to be sensitive only to the crystalline or long-range order and to be the cheapest among the similar techniques. There were estimated the apparent average diameter of turbostratique quasicristal into SGCI and the average number of graphene into a quasicristal.

The further researches will address the structure of compacted (vermicular) carbon inclusions from gray cast iron. Based on these data and the foreseen one we intend to review the mechanism of the CI nucleation and growth and the influence of S, P etc on CI morphology.

2. THEORETICAL CONSIDERATIONS

The structure of a solid carbon material depends mostly on C atom hybridization (Pencea, 2009; Skaland, 2005). When [sp.sub.2] and [sp.sup.3] C atoms simultaneously aggregate, then, depending on [sp.sup.2]/[sp.sup.3] ratio could result: diamond like carbon, turbostratique carbon, ionic carbon or even amorphous carbon (Pencea, 1997; Pencea, 2009). Because there is little knowledge on C atom hybridization states into cast iron melt then it is quite impossible to predict the structure of CI from growth mechanism point of view. The issue of CI structure is complicated by the interference of S, P, lanthanide etc into CI growth mechanism (Cojocaru-Filipciuc, 2007; Riposan et. all, 2007). From thermodynamically point of view, it is quite probable that most C atoms could be in [sp.sup.2] hybridization state and a significant ratio in [sp.sup.3] state into melted iron at about 1350[degrees]C. The foreigner atoms as S, P, Ca etc could make covalent bonds with [sp.sup.3] C atoms and stabilize their [sp.sup.3] state. Stabilization of [sp.sup.3] C favorizes the turbostratique structure. One of the agreed CIs growth mechanisms is shown in fig. 1.

If a nucleus consists of 2-3 planar lattices, then the lattices are almost sure bonded together by foreign atoms at the edges, and they will probably grow preferentially as plate-like inclusions (fig. 1, b). If S, P, O atoms are eliminated from the melt then C atoms will nucleate and grow as nodule inclusions (fig. 1, c). Other three of the most considered SG growth mechanisms are schematically illustrated in fig 2. (Alp et. all.2002)

[FIGURE 1 OMITTED]

[FIGURE 2 OMITTED]

All the above CI growth mechanisms and many others are consistent with turbostratique structure but not with that of polycrystalline graphite.

3. EXPERIMENTAL

A magnesium treated cast iron bar was machined to produce metallic chips. The elemental composition of the bar was determined by SDAR-OES and is given in Table 1.

The CIs were extracted by dissolving the iron matrices in HCl based solutions. Three different extracting procedures were applied to get powder by-products (Table 2).

The WAXD investigation of SGCI specimens were done with a Philips PW 4280 diffractometer operated at U=40 keV, I=35mA, using Ni filtered [Cu.sub.K[alpha]1] radiation (wavelength [lambda] = 0.154 nm). Besides the three specimens of SGCI (Table 2), it was WAXD investigated other samples as polycrystalline graphite, metallurgical coke and other reference materials to get comparative data. Some of these materials where investigated using [Mo.sub.K[alpha]1] radiation ([lambda] = 0.071 nm) (see fig. 4).

4. RESULT AND DISCUSSIONS

All the SGs diffractograms should be considered related to the diffractogram of polycrystalline graphite (fig. 4). The peaks having Miller indices (101), (110) and (112) attest the crystalline 3D order.

[FIGURE 4 OMITTED]

The diffractogram in fig. 5 shows that SG 1 specimen contains mainly carbon and small amount of ferrite and iron oxides. The same, it shows that, besides the tough dissolving process undertaken by the SG 1 specimen, the ferrite is steel present in a significant amount.

[FIGURE 5 OMITTED]

The diffractogram of SG 2 (Fig. 6) shows that the matrix dissolving in 90v% HCl -10 v% HF solution is more effective than the dissolving in HCl 100v%. The same the diffractogram in fig. 6 depicts well defined (002) and (004) peaks of graphite.

[FIGURE 6 OMITTED]

The SG 3 diffractogram (tig. 7) shows an increased content of iron oxides when comparing the intensities of the correspondent peaks from SG 1 and SG 2 diffractograms. The SG1 diffractogram shows less iron oxide content than SG 2 and SG 3. If the dissolving conditions are correlated with phase content of the sample it can be concluded that during iron matrix dissolving in boiling HCl v100% the iron oxides formation is inhibited and the ferrite content is reduced. The dissolving process in 90v% HCl -10 v% HF solution produces a small amount of iron oxides while dissolving in 50v% HCl -50 v%. The distance between (002) planes, [d.sub.(002)], and the apparent height of the grapheme stack, [L.sub.(002)], were calculated as (Guinier, 1963):

[2d.sub.(002)] * sin([theta]) = [lambda] (1)

[L.sub.(002)]=0.9.[lambda]/([beta].cos([theta])) (2)

where [theta] is the Bragg angle, p is the (002) peak width at half height

[FIGURE 7 OMITTED]

The average number of grapheme Nc into a turbostratique quasicristal is Nc=[L.sub.(002)]/[d.sub.(002)]. The values of [d.sub.(002)](A), [L.sub.(002)] ([Angstrom]) and of Nc for SGs samples are given in Table 3. The relative uncertainties associated to [L.sub.(002)], UL(95%), and to Nc, [U.sub.N](95%), were estimated according to SR EN 13005:2005 with 0.05 degree of significance e.g. k=2. The UL(95%), values given in Table are less than 10% while UN(95%) are about 3% that is quite acceptable for WAXD technique.

5. CONCLUSIONS

The dissolving approach should be improved to reduce or even entirely elimination of the ferrite and iron oxides contents. WAXD data show clearly that carbon inclusions from SGCI are not polycrystalline graphite. Thought, the CIs have turbostratique structures that have surprisingly [d.sub.(002)] values close to graphite one e.g. to 3.34 [Angstrom]. Accepting that carbon inclusions are not polycrystalline graphite the growth mechanisms based on crystallographic considerations should be then revised.

We consider that the atomistic growth mechanisms of SGCI should be based on [sp.sup.2] and [sp.sup.3] C cluster aggregation and further growth. In this direction it is necessary, at least, to clarify: the elemental composition of the carbon inclusions and of the surrounding areas; the C atom hybridization in molten cast iron; etc.

As a final conclusion, the researches on this subject brought more knowledge on carbon inclusion structure but, most important, perhaps, how to conduct a further research on this topic.

6. REFERENCES

Alp, T., F. Yilmaz F., Wazzan, A.A. (2002). Microstructural characteristics-property relationships in cast iron, The 6-th Saudi Engineering Conference, KFUPM, Dhahran, December Vol. 5. 281-286

Cojocaru-Filipciuc, V. (2007). Spheroidisation of graphite in cast iron-theoretical aspects, Politehnium Publishing House, Iasi

Double D.D., Hellawel A. (1995). The nucleation and growth of graphite: the modification of cast iron, Acta Metall Mater. 43: 2435-2442

Guinier A., (1963). "X-ray diffraction on imperfect crystals", Ed. Willey, New York

Pence I. (1997). Contribution to the structure characterization of different bulk carbon materials, PhD Thesis, Univ. Polytechnic of Bucharest

Pencea I., F.Barca F., Sfat C. E, Ghiban, B. (2009). SEM, WAXD and EDP-XRFS characterisation of new sorts of coke, Metalurgia International, v XIVno.9, p.6773, ISSN 1582-2214

Skaland, T. (2005). Nucleation mechanisms in Ductile Iron, Proc. Of AFS Cast Iron Inoculation Conf., Sept. Illinois, 2930

Stefanescu D.M. (1988). ASM Handbook, v15, Casting Am. Soc. Metals, Metals Park, Ohio, USA

Tab. 1. The main chemical composition ot the dissolved cast
iron

Element C Si Mn Mg S P
c (wt%) 3.87 2.73 0.20 0.03 0.012 0.014

Tab. 2. The main characteristics of the extracted powder
procedure

 Sample Dissolving solution
No. code

1 SG 1 100 v% HCl, boiled at 100[degrees]C
2 SG 2 90v% HCl + 10 v% HF, boiled at 100[degrees]C
3 SG 3 50v%HCl + 50v% [H.sub.2]O, boiled at 100[degrees]C

Tab. 3. Structural data for SGCI samples

Sample [d.sub.(002)] [U.sub.L] [U.sub.N]
code ([Angstrom]) L([Angstrom]) Nc (95%) (95%)

SG 1 3.341 326.4 98 8.8 3
SG 2 3.344 317.8 95 8.6 3
SG 3 3.347 288.9 86 7.8 2