812 Analyst, September, 1979, Vol. 104, pp. 812-821 Identification and Determination of Titanium Sulphide and Carbosulphidle Compounds in Steel W. R. Bandi and George Krapf United States Steel Corporation, 125 Jamison Lane, Mowoevalle, Pa. 15146, USA As part of the experimental work to identify titanium compounds formed to control sulphide morphology in high-strength low-alloy steels, heats with special additions of titanium were melted. I t was determined that five titanium compounds could be distinguished from one another by differential thermal analysis - evolved gas analysis techniques. Further, by analysing residues obtained by using both the acid and the ester - halogen methods of dissolving the matrix, it was possible to quantiEy the results for the five titanium compounds.The results showed that every experimental lieat or experimental alloy examined contained Ti,C,S, and that one sample contained both Ti,C,S, and y-Ti,S. Another unidentified titanium sulphide was thought to be present in one of the steels. Keywords : Titanium compounds ; high-strength low-alloy steel ; dijferential thermal analysis ; evolved gas analysis ; ester - halogen extraction For several years a metallurgical investigation directed toward controlling the sulphide morphology of high-strength low-alloy (HSLA) steel has been in progress at the United States Steel Research Laboratory. I t was presumed that addition of titanium to the steels would result in a change in the sulphide morphology as the result of reaction of part of the sulphur with titanium rather than with manganese.By means of a recently developed chemical method,l it was possible to show tha.t the manganese(I1) sulphide content of the steel was reduced, but it was not possible to determine what titanium sulphide phases were present. 13ecause knowledge of the amount and type of titanium compounds precipitated in these steels was needed in order to understand the mechanical properties of the steel, work was initiated to determine the concentration and crystal structure of the titanium precipitates. The literature on titanium sulphide compounds found in steel is confusing and incomplete. Although sulphides are often studied by microscopic, diffraction and microprobe procedures, the separation and determination of sulphides by chemical methods has been ignored to a large extent.According to a recent survey of the literature,2 the most likely sulphide precipitates in titanium-bearing steel are hexagonal TiS (often written TiS,-,), y-Ti,S and Ti,C,S,. However, misidentification is common when microscope, microprobe and particu- larly X-ray diffraction procedures are used because the diffraction patterns of y-Ti,S and Ti,C,S2 are similar. Over the last 15 years, differential thermal analysis - evolved gas analysis (DTA - EGA) procedures have been developed for the determination, after isolation from the steel matrix, of approximately 35 second-phase compo~nds,~-6 some of which had not previously been identified. Because sulphur dioxide and carbon dioxide can be differentiated when residues isolated from steel are combusted in oxygen, it appeared possible to differentiate between y-Ti,S and Ti,C2S, by DTA - EGA procedures.Therefore, the chemical isolation and DTA - EGA identification and quantitative determination of these precipitates were investigated. This paper summarises this work and further describes the quantitative DTA - EGA deter- mination of titanium nitride and titanium carbide, and also two niobium carbides, which are present in some HSLA steels. Materials and Experimental Work The first two steels were experimental heats made for the purpose of identifying titanium compounds ; all alloying metals that form carbides and sulphides, except titanium, are present only in residual concentrations. Steels 3 and 4 were two HSLA experimental heats that were used in the study of sulphide shape control.Table I shows the partial composition of the steels used in this study.BAND1 AND KRAPF TABLE I 813 COMPOSITION OF SAMPLES (yo) Element c . . . . Mn . . . . Si . . . . Ti . . . . Mo . . . . A1 (total) . . Nb . . . . N, . . . . 0, . . . . . . . . . . . . . . . . . . . . . . 1 0.025 t 0 . 0 1 0.064 0.60 <0.01 <0.005 <0.002 0.001 N.D.* Steel No. 2 3 0.21 0.13 (0.01 1.49 0.064 0.024 0.53 0.053 <0.01 0.33 <0.005 0.042 (0.002 0.016 0.001 0.009 N.D.* 0.007 h 7 4 0.11 1.45 0.026 0.13 0.33 0.030 0.053 0.007 0.011 * N.D. = not determined. Residues from the steels were isolated in bromine - methanol ( l + 3 V / V ) , 10% m/V iodine in methanol solution, 10% hydrochloric acid and fuming perchloric acid. Of these isolation procedures the most useful is that with hydrochloric acid because it is easier to identify the presence of phases such as y-Ti,S and/or Ti4C,S, in the residues by chemical and X-ray diffrac- tion procedures.I t has also been established from past experience that the best DTA - EGA results are obtained when residues are isolated in hydrochloric acid because smaller amounts of interfering phases such as Fe,C, MnS, (MnTi)S and FeS will be present. Therefore, acid- isolated residues have higher concentrations of phases such as y-Ti,S, Ti,C,S, and TiC,Ny, and ultimately a smaller percentage of these phases can be detected in the steel because the DTA - EGA response is more sensitive, less complicated and more easily resolved into the individual components.Finally, because residues isolated in acid contain smaller amounts of amorphous carbon and Fe,C than those from most other isolation procedures, there is less tendency for rapid combustion of Fe,C. Such combustion can cause uncontrolled temperature changes in the DTA sample holder and lead to pre-ignition of other carbides or nitrides. Residues isolated in bromine - methanol, iodine - methanol and fuming perchloric acid were analysed for titanium, sulphur and niobium. The results were compared with similar values obtained on hydrochloric acid residues in order to gain qualitative and quantitative information about the phases present in the steel and to verify the qualitative interpretation of the DTA - EGA peaks. The microchemical analysis was also used to confirm the quantita- tiveness of the EGA results. As an example, Fe,Ti is insoluble in 10% hydrochloric acid, and if the amount of titanium dissolved in iodine - methanol is greater than that dissolved in the acid, it would indicate the possible presence of Fe,Ti in the steel.If more sulphur and titanium are present in an iodine - methanol residue than in a 10% hydrochloric acid residue, i t indicates the possible presence of (MnTi)S or hexagonal TiS in the steel. If the titanium phase is insoluble in fuming perchloric acid, it is probably TiO, or titanate. Before any attempt was made to isolate residues, the steel specimens were surface-ground to remove oxide scale and were then machined to produce the finest possible millings for treatment with bromine - methanol, 10% hydrochloric acid and perchloric acid.Acid- isolated residues were obtained by treating 2-4 g of steel with 10% hydrochloric acid at 38 & 2 "C. (For samples containing niobium, 10 g of tartaric acid were added to each 100 ml of 10% hydrochloric acid.) To prevent oxidation of the solution and precipitation, oxygen was removed from the dissolving reagent by bubbling argon through it, and an argon atmosphere was maintained during dissolution of the isolated residue. Filtration was accomplished by using a Millipore filtration apparatus and a 500-nm methylcellulose organic filter membrane. Details of the filtration and weighing of these residues have been published previously . 9 7-g To isolate (MnTi)S or hexagonal TiS, together with y-Ti,S and Ti,C,S2, solid pieces of steel ( 1 0 g or more) wcrc shaken for 3 h with 10% m/V iodine in methanol solution.The solid piece was removed from the iodine solution with the aid of a magnet, treated ultrasonically in methanol t o remove adhering residue and then re-weighed. The residue was filtered on a Gelman Alpha 8 membrane according to previously described pr~cedures.~*~-~ Titanium814 Analyst, V o l . 104 was determined by ignition of the residue, fusion with sodium hydrogen sulphate and applica- tion of the ASTM method for the photometric determination of titanium as peroxytitanic acid.10 Sulphur was determined by heating the residue with Escha's mixture and applying the grat-imetric barium sulphate method to the leached solution.ll To determine which sulphides could be isolated in bromine - methanol, fine drillings were treated with 5 ml of bromine and 15 ml of methanol per gram of sample.Filtration has been described previou~ly.~*~*~ Titanium and sulphur were determined in the bromine- isolated residues in the same manner as described above for the iodine - methanol residues. The presence of TiO, or titanate in the steels was determined by dissolving the steel in perchloric acid and vigorously fuming the resulting solution. After filtration and ignition, titanium was determined in the ignited contents by fusion with sodium hydrogen sulphate and determination of titanium photometrically as peroxytitanic acidlo in the acid-disolved fusion. Titanium in acid-isolated residues was determined by the photometric peroxytitanic acid method.10 Sulphur in the acid-isolated residues was determined by the combustion method.12 The DTA - EGA curves for the combustion of the isolated second phases were obtained with an instrument described in detail p r e v i ~ u s l y ~ , ~ , ~ ~ and summarised for this paper as follows.An R.L. Stone, Model 12EC,, instrument was modified to limit the gas volume and increase the sensitivity of the EGA system by substituting a specially built Inconel sample holder. This sample holder included an 8-min Inconel gas exit tube, which allowed the evolved gas to flow upwards out of the sample holder through the top of the furnace and into the EGA thermal conductivity cell. A hole €or the exit tube was drilled through the top of the furnace. The sample and reference material were placed on 6-mm platinum dishes situated on top of thermocouple rings. The insulated thermocouple wires were enclosed in a 5-mm Inconel tube, which was inserted donwards through the 8-mm exit tube into the sample compart- ment, and a gas seal outside the DTA furnace was obtained with an O-ring fitting. The total volume of the gas train was about 21 ml.The sensitivity was set so that the evolution of 5 ng of carbon dioxide in 10 min could be measured by thermal conductivity with an oxygen flow-rate of 3 ml min-l. Changes in the thermal conductivity were measured with a Gow-Mac TRIIIA thermal conductivity cell. A bridge current of 6 mA and a tempera- ture of 25 "C were maintained while recording the thermal conductivity changes on a 1-mV recorder. In some instances it was necessary to use a DuPont 310 curve resolver to deconvolute overlapping EGA carbon dioxide responses and determine the area associated with a specific carbide.In resolving the individual areas associated with a specific carbide, use was made of previously established knowledge of the EGA combustion temperature peaks for the specific carbides. Also, the appearance of the simultaneously recorded DTA response was used to interpret the EGA response. Attempts were also made to correlate the EGA response with the composition and heat treatment of the steel, the EGA temperature peaks of com- pounds identified by microchemical analysis of the acid-isolated residue, the EGA temperature peaks of compounds identified by X-ray diffraction examination of the acid-isolated residue, the solubility of phases in dilute warm hydrochloric acid and knowledge that minor amounts of Fe,C and amorphous carbon will contaminate acid-isolated residues.To establish the DTA - EGA responses caused by the decomposition of y-Ti,S and Ti,C,S,, 1-3 mg of the hydrochloric acid isolated residue were transferred into a 3 mm diameter platinum dish and placed on the platinel thermocouples. Alumina was used in the reference pan. The DTA thermogram was recorded on a 1 mV in-l recorder at a sensitivity of 0.33 "C in-l and a heating rate of 10 "C min-l. An initial EGA response was recorded without the sulphur dioxide trap in the evolved gas train so that the thermal conductivity changes caused by the evolution of both carbon dioxide and sulphur dioxide could be recorded. After a small sulphur dioxide trap (75 x 51 mm diameter plastic tube) containing man- ganese(1V) oxide had been placed in the EGA gas train, a second thermal conductivity for the carbon dioxide evolved was recorded.By comparison of the DTA - EGA responses, both sulphide and carbosulphide responses were discerned. The sulphur dioxide trap installed in the EGA exit gas train was small because the gas volume and gas-flow charac- teristics of the evolved gas would be changed by installation of a large sulphur dioxide trap and such a change would affect the shape, area, sensitivity and peak temperature of the EGA response. BAND1 AND KRAPF: IDENTIFICATION AND DETERMINATION O FSeptember, 1979 TITANIUM SULPHIDE AND CARBOSULPHIDE COMPOUNDS IN STEEL 815 The determination of TIN and N, as TiC,N, was accomplished by use of an automatic sampling gas chromatograph placed in the evolved gas train to separate nitrogen from oxygen. A description of this apparatus has been published p r e v i ~ u s l y .~ ~ ~ The same acid- isolated residue was used for the nitride determination. The EGA gas flow-rate was 3 ml min-l; helium (75 ml min-l) was used as the carrier gas in the gas-chromatographic column. The EGA sefisitivity was set so that as little as 0.0003~0 of nitrogen as TIN could be detected by using the same Gow-Mac thermal conductivity apparatus described for carbon dioxide detection. Results and Discussion Table I1 shows the X-ray diffraction identification of the precipitates in residues isolated with hydrochloric acid.The similarity in the diffraction patterns of Ti4C,S, and y-Ti,S prevents differentiation. The identification of TIN, TIC and NbC, particularly in the HSLA steels, was expected because past experience indicated that such precipitates should be present. The fact that no titanium - sulphur compounds were identified in the HSLA steels does not mean that none were present because small amounts of the precipitate could escape detection in the presence of large amounts of carbides, nitrides, amorphous carbon and oxides. TABLE I1 PRECIPITATES IDENTIFIED BY X-RAY DIFFRACTION Steel No. Component 1 y-Ti,S or Ti4C,S,, TiN 2 y-Ti,S or Ti4C,S,, TIC 3 4 TiN, Tic, plus some unidentified lines TiN, Tic, NbC, plus some unidentified lines The amount of titanium found in the residues isolated by treating the steels with various dissolution media is shown in Table 111.If it is assumed, for example, in steel 1 that the titanium compounds present are TIC and y-Ti,S, then about half of the titanium in the steel (0.30%) should be present in the hydrochloric acid residue. However, if it is assumed that Ti4C,S, instead of y-Ti,S is the primary precipitate, then less than one third of the titanium in the steel (0.18%) would have been precipitated. Table I11 shows that for steel 1, 0.14% of titanium was precipitated, which indicates Ti,C,S, to be the predominant precipitate. Steel 2 contains a compound insoluble in hydrochloric acid but soluble in iodine and methanol. For steel 4, there is a difference in the amount of titanium found to be insoluble in iodine - methanol in comparison with bromine - methanol and hydrochloric acid.This difference indicates that some titanium compound partially or wholly dissolves in both bromine - methanol and hydrochloric acid. The difference probably represents a titanium sulphide such as hexagonal TiS or (Mn,Ti)S. Table I11 also shows that only steel 4 contains titanium compounds that are insoluble in fuming perchloric acid. This could be Fe,Ti or a similar compound. Therefore, this sample contains titanium as the oxide or titanate. TABLE I11 AMOUNT OF TITANIUM IN RESIDUE AFTER TREATMENT OF STEEL WITH VARIOUS ISOLATING REAGENTS Titanium isolated, % of steel I 7 h Steel No. In HCl In Br, - CH,OH In I, - CH,OH In HC10, 1 0.14 N.h.* 0.14 N.D.t 2 0.44 N.A.0.30 N.D. 3 0.026 0.026 0.025 N.D. 4 0.083 0.088 0.12 0.012 * N.A. = not analysed. N.D. = none detected (<0.01%).816 BAND1 AND KRAPF: IDENTIFICATION AND DETERMINATION OF Analyst, Val. 104 N b J 3 Resolution EGA response 610°C TiC,N, C (Nb,lVlo)C I Fe,C of EGA 600 500 400 300 Temperature/ "C Fig. 1. Resolution of EGA carbon dioxide response for 1.91 mg of residue isolated from steel sample 4. Fig. 1 shows the EGA response obtained for steel 4. It is typical of the responses used to determine TiC,N1/, Nb,C, and (Nb,,,Mo,,)C in HSLA steels and illustrates the use of the DuPont curve resolver to deconvolute the E.GA response into the peaks shown for the isolated residue. The resolution and identification of Nb,C, and (Nb,,,Mo,,,)C are con- sistent with past experience in the analysis, of precipitates in niobium - molybdenum- bearing steel^^,^ and also agree with the microchemical analysis results for the residue.X-ray diffraction was unable to differentiate between these two niobium carbide precipitates, but a diffraction pattern for cubic niobium carbide was obtained (Table 11). Similarly, an X-ray diffraction pattern was obtained for TIC, but DTA - EGA results showed that in some instances the compound was TiC,N,. The temperatures for the Fe,C, C, TIC and TiC,N, responses are also consistent with past experience.798 TIN and N, in TiC,N, were deter- mined from Fig. 2, and the results agree with ithose described in previous publication^.^^^^^^^^ I 400 500 - I I 600 700 800 Be0 Tern per atu re/ "C 1000 Fig.2. EGA nitrogen response for 4.78 mg of residue isolated by acid dissolution from steel sample 3. Table IV shows the results obtained for TiC, TiC,N,, TIN, Nb,C, and (Nb,,Mo,,,)C from However, Fig. 1 shows unidentified evolution of carbon dioxide Figs. 3 and 4 show DTA - A sulphur The difference in the responses figures such as Figs. 1 and 2. at 610 "C, which was observed for residues from all four steels. EGA responses for residues isolated in hydrochloric acid from steels 1 and 2. dioxide trap was used in obtaining these EGA responses.September, 1979 TITANIUM SULPHIDE AND.CARBOSULPHIDE COMPOUNDS IN STEEL 817 TABLE IV AMOUNTS OF TITANIUM AND NIOBIUM PRESENT AS TiC,Ny, TIC, TIN, Nb,C,, AND (Nbo.,Moo.,)C (70) Ti as Tic and Nb as Steel No. TiC,N, Ti as TIN Nb as Nb,C, (Nb,.,Mo,.,)C 1 0.004 N.D.* N.D.N.D. 2 0.09 N.D. N.D. N.D. 3 0.01 1 0.006 0.006 0.003 4 0.018 0.010 0.042 0.015 * N.D. = none detected (<0.001%). reflects the fact that there is 10 times more carbon in steel 2 than in steel 1. Therefore, the residue from steel 2 contains much more amorphous carbon, cementite and titanium carbide than the residue isolated from steel 1. Fig. 5 shows the resolution of the EGA responses and the response for the TIC identified by X-ray diffraction in steel 2. Temperature/"C Fig. 3. DTA - EGA resDonses for 2.04 ma of residue isolated TemDerature! "C Fig. 4. DTA - EGA response for 2.31 mg of residue isolated by acid dissolution from steel sample 2 (sulphur trap in the system).818 BAND1 AND KRAPF: IDENTIFICATION AND DETERMINATION OF AnaLyst, Vol 104 E 6 -.MC Fe,C 620 "C 600 500 400 Temperature/"C Fig. 5. Comparison of carbon dioxide evolved at 620 "C from (a) steel sample 1 ; and (b) steel sample 2. To determine whether any sulphur dioxide was simultaneously evolved with the carbon dioxide at 610-630 "C, the sulphur dioxide trap was removed from the evolved-gas train and the DTA - EGA responses were again recorded for the residues heated in oxygen. Figs. 6 and 7 show larger EGA responses at 610-630 "C than were observed in Figs. 3 and 4. The increase in magnitude is more noticeable for steel 1 than for steel 2. The EGA response and resolution for steel 1 with and without the sulphur dioxide trap are presented in Fig. 8. Not only is the response at 610-630 "C larger when no sulphur dioxide trap is used in the combustion of TiaC2S2, but also the resolution shows a second EGA response is present at 580 "C.Because of the composition shown in Table I, the X-ray diffraction results shown in Table I1 and the analytical results shown in Tables I11 and V, it is possible to state that the response at 580 "C is produced by the combustion of y-Ti2S. There is also an indication that the y-Ti2S may contain some carbon, because even with the sulphur dioxide trap inserted I I 100 200 300 400 500 600 700 800 Tempe rature/"C DTA - EGA response for 1.96 mg of residue isolated by acid dissolution from steel .,ample 1 (no sulphur trap in the system). Fig. 6.September, 1979 TITANIUM SULPHIDE AND CARBOSULPHIDE COMPOUNDS IN STEEL 819 i! DTA I I I I I I 1 I 100 200 300 400 500 600 700 800 Temperature/"C Fig.7. DTA - EGA response for 2.25 mg of residue isolated by acid dissolution from steel sample 2 (no sulphur trap in the system). in the gas train, a small carbon dioxide response was observed at 580 "C. This could be due to inefficiency of the trap as it had a small absorption capacity and had to be renewed every day. Although no thermal conductivity response for Ti,S could be observed visually in the EGA trace for steel 2 (Fig. 9), when the response was resolved into its component parts a small EGA peak at 580 "C was needed in order to obtain a good fit. It therefore appears that steel 2 contains a small amount of y-Ti,S and a larger amount of Ti,C,S,. In practice we do not use the EGA apparatus to determine sulphur dioxide, because it is known that some of the sulphur dioxide will condense in the train and eventually either react with the metal tubing to form sulphur trioxide or form this gas by another mechanism.Further, it is known that thermistors operating- with a bridge current of 6 mA deteriorate in the presence of sulphur dioxide. obtained by conventional gravimetric or combustion procedures. Therefor6 quantitative sulphur dioxide values were 1 / '\ 58O"CSO, inrTi,S Fe,C Residual CO, or SO, in yTi,S CO, in C ---- _ _ ., CO, in Fe,C -- ___ - ~- - Temperature/"C Fig. 8. Effect of SO, on thermal conductivity of gas evolved from residue of steel sample 1. Resolution of EG-4 responses for 2 mg of residue isolated by acid dissolution from steel sample 1 : (a) without SO, trap; and (b) with SO, trap.820 BAND1 AND KRAPF : IDENTIFICATION AND DETERMINATION OF Analyst, Vol.104 600 500 400 Temperature/"(= Fig. 9. EGA response and resolution for 2.25 mg of residue isolated by acid dissolution from steel sample 2 (no sulphur trap in the system). Tables V and VI summarise the distribution of sulphur and titanium precipitates found in the steels. For steel 1, Table V shows that the quantitative results obtained for Ti,C,S, from the EGA response and sulphur calculated to be y-Ti,S from the classical chemical determination agree with the total sulphur content in the steel, the X-ray diffraction data and the calculation of the y-Ti,S concentration from results in Table 111. TABLE V COMPARISON OF Ti,C,S, CALCULATED FROM DTA - EGA RESULTS WITH SULPHUR FOUND AS TITANIUM - SULPHUR COMPOUNDS BY CHEMICAL METHODS (70) S found as y-Ti,S DTA - EGA C as S equivalent and Ti,C,S, by Amount of S as Steel No.Ti,C.,S, DTA - EGA. C chemical methods y-Ti,S 1 0.005 0.014 0.051 0.037 2 0.023 0.060 0.059 0.000 3 0.0005 0.001 3 0.003 0.002 4 0.003 9 0.010 0.018 0.008 For steel 2 the calculation of the sulphur equivalent of Ti,C,S, in Table V indicates the absence of y-Ti,S, although it was indicated as present in the EGA resolutions. A probable explanation is that some of the sulphur dioxide passed through the trap, causing the EGA TABLE VI DISTRIBUTION OF TITANIUM IN SAMPLES (yo) Total Ti Total Ti as Tic, TIN, found in Steel As Tic As As TiO,,Ti,CIS, residue, No. TiC,N,, As TIN As TiO, Ti,C,S, y-Ti,S and y-Ti,S I, - CH,OH Difference Remarks 1 0.004 <0.001 <0.005 0.042 0.10 0.15 0.14 - 0.01 All titanium precipitated as Tic, Ti,C,S, and y-Ti,S 2 0.09 <0.001 t0.005 0.18 ? 0.27 0.30 +0.03 HCl difference 0.17.Some titanium precipitated as Fe,Ti, M,C; y-Ti,S also present 3 0.011 0.006 <0.005 0.004 0.005 0.026 0.026 - 0.000 Results agree 4 0.018 0.010 0.012 0.030 0.018 0.088 0.12 +0.032 Some titanium present (0.088 total TI in Br, - CH,OH) probably as TiS or (MnTi)SSeptember, 1979 TITANIUM SULPHIDE AND CARBOSULPHIDE COMPOUNDS IN STEEL 821 result for carbon dioxide for Ti4C2S2 to be high and therefore the sulphur equivalent, which is based on the premise that only carbon dioxide was being detected, is also high. When this high sulphur result is subtracted from the total sulphur present as Ti,C,S,, no sulphur as y-Ti,S is detected.The isolation of 0.17% more titanium in hydrochloric acid (Table 111) than is shown in column 7 of Table VI indicates the presence of titanium in other precipitates such as (Fe,Ti),C and does not necessarily indicate an error in the analysis of the titanium - sulphur precipitates. The chemical value shown in Table V for the total amount of sulphur associated with titanium is greater than the amount found to be present as Ti,C,S, in steel 3. This indicates the possibility of a trace amount of y-Ti,S being present. However, because the sulphur bound as Ti,C2S, was calculated on the basis of detecting 0.0005% of carbon, it is impossible to say with certainty that a titanium - sulphur compound other than Ti,C,S, is present in this steel.The results shown in Table VI for the residues isolated from steel 4 definitely confirm that the steel contains a third unidentified titanium - sulphur compound, as indicated by the results in Table 111. Additional DTA - EGA runs confirmed the carbon dioxide value for Ti,C,S2, and the determinations of total titanium isolated in bromine - methanol agree with acid isolations and DTA - EGA results. Therefore, the larger amount of titanium precipitate isolated in iodine - methanol indicates the presence of TiS or (Mn,Ti)S. Conclusions In conclusion, two Fe - Ti - C - S alloys and two experimental HSLA steels containing titanium additions for the purpose of forming titanium - sulphur compounds were examined by X-ray diffraction, DTA - EGA and classical chemical methods.The results showed that every sample contained Ti4C,S2 and that one also contained y-Ti,S. A second unidentified titanium sulphide was found in the HSLA steels. I t was demonstrated that DTA - EGA is the best method for identifying Ti,C,S,. It is also the best method for differentiating between Ti,C,S, and y-Ti,S because their thermal responses are different, even though the X-ray diffraction patterns are similar. Finally, DTA - EGA has been used to differentiate Nb4C, and (Nbo,,Moo,,)C from each other and from cubic NbC and also has been used to differentiate between TIC and TiC,N,. The authors acknowledge the assistance of P. A. Stoll and R. P. Bacco of the Research Laboratory who provided the X-ray diffraction results. 1. 2. 3. 4. 6. 6. 5. 8. 9. 10. 11. 12. 13. References Bandi, W. R., Lutz, J . L., and Buyok, E. G., in deBarbadillo, J. J., and Snape, E., Editors, “Sulfide Kiessling, R., and Lange, N., “Nonmetallic Inclusions in Steel,” Part 11, Iron and Steel Institute, Karp, H. S., Bandi, W. R., and Melnick, L. M., Talanta, 1966, 13, 1679. Bandi, W. R., Straub, W. A., Buyok, E. G., and Melnick, L. M., Analyt. Chem., 1966, 38, 1336. Bandi, W. R., Science, N . Y . , 1977, 196, 136. Lloyd, M. H., and Shanahan, C. E. A., J . Thermal Analysis, 1977, 12, 321. Bandi, W. R., Lutz, J . L., and Melnick, L. M., J . Iron Steel Inst., 1969, 207, 348. Krapf, G., Lutz, J . L., Melnick, L. M., and Bandi, W. R., Thermochim. Acta, 1972, 4, 257. Bandi, W. R., and Krapf, G., Analyt. Chem., 1977, 49, 649. “1978 Annual Book of ASTM Standards, Part 12, Chemical Analysis of Metals; Sampling and Analysis of Metal-bearing Ores,” American Society for Testing and Materials, Philadelphia, Pa., 1978, p. 25. Lundell, G. E. F., Hoffman, J. I., and Bright, H. A., “Chemical Analysis of Iron and Steel,’’ John Wiley, New York, 1931, p. 612. “1978 Annual Book of ASTM Standards, Part 12, Chemical Analysis of Metals; Sampling and Analysis of Mctal-bearing Ores,” American Society for Testing and Materials, Philadelphia, Pa., 1978, p. 527. Randi, W. R., Buyok, E. G., Krapf, G., and Melnick, L. M., in Schwenker, R. F., Jr., and Garn, P. D., Edztors, “Thermal Analysis,” Volume 2, Academic Press, New York, 1969, p. 1363. Received October 4th, 1978 Accepted February 27th, 1978 Inclusions in Steel,” American. Society for Metals, Metals Park, Ohio, 1975, p. 178. London, Publication 100, 1966, p. 138.