Analyst, October, 1976, Vol. 101, pp. 803-80'1 803 Development of Fluxes for the Analysis of Ceramic Materials by X-ray FI uorescence Spectrometry H. Bennett and G. J. Oliver The British Ceramic Research Association, Queens Road, Penkhull, Stoke-on-Trent, ST4 7LQ From experience of the use of both lithium metaborate and tetraborate as fluxes for a wide variety of minerals and refractories, it is postulated that when the sample contains a preponderance of acidic oxides (e.g., silica), lithium metaborate has considerable advantages. The melts from meta- borate are much more fluid and decomposition can be carried out by heating over gas burners. For samples that contain high contents of basic oxides (e.g., carbonate rocks), the more acidic tetraborate is recommended. A 1 + 4 mixture of metaborate and tetraborate forms a universal flux for samples containing silica and/or alumina as major contents.The most suitable flux can be predicted from the acidity or basicity of the sample. It has been accepted for some time that if accurate analyses of minerals and refractories are required using an X-ray spectrometer, it is normally necessary to overcome various minera- logical and particle size effects.1-9 It is generally recognised that a satisfactory procedure is to fuse the sample in one of a variety of fluxes and cast the resultant melt into a glassy bead or button for presentation to the X-ray spectrometer. Complete decomposition of the sample and production of clear glass beads are obviously required for the best results. The process has developed from casting the bead into a graphite mould, then polishing the lower surface,1°-13 through casting into a dish made of a platinum - gold alloy (95 + 5) and using the lower s ~ r f a c e ~ o ~ ~ * , ~ ~ to using the slightly curved top surface.10J6 The platinum - gold alloy is not wetted by the melt and beads are released during the cooling process.The use of the top surface obviates the need for polishing either the lower surface of the bead or the surface of the metal casting dish to maintain a mirror finish, the results being equally satis- factory. Most workers have used lithium tetraborate as the f l ~ ~ ~ 9 ~ ~ 3 ~ ~ 9 ~ ~ - ~ ~ ; it decomposes a wide range of materials at suitable flux - sample dilutions and is composed of elements that yield a light matrix and whose determinations are not normally required.Lithium metaborate has also been used as a chiefly for atomic-absorption spectrophotometry, but it has received little attention in the field of X-ray spectrometry. Lithium tetraborate is not universally successful as a flux and the object of the present work was to develop fluxes enabling a wide range of minerals, oxides and ceramics to be decomposed and prepared for presentation to an X-ray spectrometer as a cast bead. Decomposition of Materials in the Silica - Alumina Range Most analyses conducted in the ceramic industry and in the field of geochemistry are on materials whose major contents are silica, alumina or both. For this reason, the problems associated with these types of material were investigated first. The materials range from almost pure silica, sand, flint and quartz through, for example, feldspar, clay, kyanite, sillimanite and bauxite, to almost pure alumina.Although it is possible to develop individual fluxes covering specific parts of this range, it is obviously desirable to have a single universal flux. As a result of co-operative work carried out by a Working Group of the British Ceramic Research Association on X-ray fluorescence analysis,1° a dilution of 1 part of sample to 5 parts of flux (lithium tetraborate) had been found satisfactory for this purpose. There are disadvantages to this procedure, however. Melts of lithium tetraborate are extremely viscous and as a result high temperatures (1 200 "C) are needed to produce an adequately liquid melt to permit swirling, pouring and casting.Even so, sufficient cooling occurs to prevent the complete transfer of the melt to the casting dish. Different operators will leave differing amounts of melt in the fusion vessel, resulting in different thicknesses of bead and therefore different curvatures of the top surfaces, thus affecting the results. The high804 BENNETT AND OLIVER: DEVELOPMENT OF FLUXES FOR ANALYSIS Analyst, VoZ. 101 temperature and the need for periodic swirling during fusion make this operation uncomfort- able for the analyst. A further disadvantage of lithium tetraborate is the slowness of fusion of high-silica materials. Whereas most aluminous and aluminosilicate materials will be decomposed after 20 min at 1 200 "C, decomposition may take 1 h for a high-silica material.Finally, there is increasing danger with higher temperatures of reduction of some of the oxides, e.g., those of cobalt and lead, and the metals alloying with the platinum. Lithium carbonate has been used in conjunction with tetraborate but this mixture has a number of disadvantages. Firstly, if a localised concentration of the carbonate occurs next to the dish, on decomposition to the oxide by heating damage to the alloy may result. Secondly, the release of carbon dioxide may result in spurting, with some of the sample adhering to the lid, and finally this same loss of carbon dioxide will cause a reduction in amount of the melt, which, unless consistent, will lead to variable results. The current more ready availability of lithium metaborate offers a way of avoiding these problems.Lithium metaborate has a much lower melting-point (849 "C) than lithium tetraborate (917 "C)24 and yields fluid melts even when heated over gas burners. This low melting-point has the advantage that swirling could take place not in a 1200 "C furnace but over a gas burner. The melt, after heating to 1 200 "C, was so fluid that almost all of the liquid could be transferred to the casting dish. Experiments with lithium metaborate as a flux showed that all samples ranging from lOOyo silica to 80-S5~0 alumina were decomposed after a period over the gas burner followed by 5 min at 1 200 "C, but that as the alumina content increased above this range there was a greater tendency towards devitrification on cooling.Microscopic examinations of beads from high-alumina materials fused in lithium metaborate showed tiny particles of unfused material even after treatment at 1 200 "C and these particles were acting as loci for devitrification. The situation therefore appeared to exist whereby lithium tetraborate alone was effective from lOOyo alumina to 95% silica and lithium metaborate from lOOyo silica to 80-85y0 alumina. It seemed possible that some admixture of the two might cover the range com- pletely and, bearing in mind the practical advantages in the use of lithium metaborate, it was desirable to use as much of this salt as possible. Examination of the phase diagram of the lithium oxide - boric oxide system2* showed a eutectic at 832 "C for a mixture of 73% boric oxide and 27% lithium oxide.One part of lithium tetraborate mixed with four parts of lithium metaborate yields 72.5% boric oxide and 27.5% lithium oxide so that this mixture was selected for fusion and casting trials. Full decomposition and clear beads were achieved over the whole range from pure silica through various materials with increasing alumina content to pure alumina. The method finally adopted for this range of materials was as follows: 1.5 g of the ignited sample is mixed with an amount. of the flux equivalent to 7.5 g after ignition at 700 "C in a 45-50-mm platinum - gold alloy dish.25 In order to ensure constant dilution it is imperative that the sample be ignited before weighing and the mass of flux taken should allow for the loss on ignition.The dish is transferred to a gas burner and fused, maintaining oxidising conditions,26 swirling being carried out at intervals. Fig. 1 shows a four-unit burner head in which the base rotates, altering the angle of the support progressively, thereby providing a swirling action. Decomposition is complete in 10-15 min, but experience has shown that for pure alumina samples it is desirable to complete the fusion by heating for 5 min at 1200 "C. As heating to this temperature can result in some loss of flux by volatilisation, all samples are subjected to the same treatment. After 3 min of this period the casting dish on its support (e.g., a zircon block) is placed in the 1 200 "C furnace. The dish and support are then rapidly transferred to a flat, horizontal surface and the melt is poured into it.Once the bead has solidified, the casting dish is placed over an air jet (Fig. 2) in order to assist in cooling and releasing the bead. Once the bead has released itself, it is allowed to stand for about 15min on an asbestos surface and is then cool enough for use. Fig. 3 shows the casting dish and resultant bead. Empirical Theory of Decomposition by Lithium Fluxes Although the whole range of silica - alumina materials were satisfactorily treated by using a 1 + 4 lithium tetraborate - metaborate flux, the total range of samples to be analysed was considerably wider. Reasons were sought for the successes and failures of the two types ofFig. 1. Four-unit rotary burner, fitted with multiple burners. The burners used here are Nordsea multiple Bunsen burners, Type No.122, but 14-in Amal burners are used on a similar device and are just as effective. Fig. 2. Simple air jet, Fig. 3. Cast bead with platinum - gold alloy mould. [to face p . 804October, 1976 OF CERAMIC MATERIALS BY X-RAY FLUORESCENCE SPECTROMETRY 805 borate. In the terminology of ceramics, silica is regarded as an acidic oxide, alkali and alkaline earth oxides are clearly basic and alumina is amphoteric, but with a tendency to be more basic than acidic. Lithium metaborate may be considered to be Li20.B20,, which in the presence of an acidic oxide can react as lithium oxide, with the surplus boric oxide effectively producing lithium tetraborate (Li,0.2B,03). Lithium tetraborate, on the other hand, may be regarded primarily as a source of boric oxide, which can react with basic oxides to form metaborates together with the equivalent of lithium metaborate.An oxide such as alumina wiU probably react more readily with boric oxide than with lithium oxide, for although it is amphoteric it reacts more readily as a basic oxide than as an acidic oxide. These suggestions fit in with the established facts that high-silica materials react more readily with metaborate and alumina materials with tetraborate. It is also significant that of the three fluxes lithium metaborate, tetraborate and the mixed flux (1 + 4), only the tetraborate will cast to produce a glassy bead in the absence of a sample. This result suggests that normally the proportion of tetraborate in the bead must exceed 1:4, except where other elements present assist the formation of a glass.Thus, it is to be expected that the presence in samples of high contents of basic oxides, e.g., alkali and alkaline earth oxides, will not produce castable beads except possibly in tetraborate. In fact, as will be shown, the ratio of tetraborate to sample has to be increased above 5: 1 with basic samples such as magnesite. Development of Fluxes for Other Materials The readily fusible 1 + 4 lithium tetraborate - metaborate flux already in use for the silica - alumina range of materials was tested on a wide range of samples in order to determine its limitations. Aluminosilicates of a more basic nature, feldspars, steatite and asbestos cements containing up to 60% of calcium oxide are all decomposed and cast into good beads.Similar treatment has been successful with samples containing up to 35% of sodium oxide, 25% of iron(II1) oxide and 55% of titanium(1V) oxide. Zircon materials also produce beads with no difficulty, as does a 1 + 1 mixture of ahminosilicate and bone ash in the form of bone china. Bone ash itself, which is similar to hydroxyapatite in composition, does not produce beads, nor do more basic minerals such as magnesite, dolomite and calcite. Bone ash fuses and casts readily in lithium tetraborate at a 3: 1 ratio of flux to sample, but materials more basic than bone ash do not. When the sample is too basic for the flux, it is frequently possible to decompose the material and produce a clear melt, only to find that the bead devitrifies on cooling.The presence of about 40% of phosphorus(V) oxide in the bone ash clearly helps in balancing out the effect of the high calcium oxide content (56%). Similarly, zircon (65% of zirconia, 35% of silica) fuses and casts in the 1 + 4 flux but zirconia will not cast even in lithium tetraborate a t the 5: 1 ratio. Suitable fluxes for a range of materials are shown in Table I, demonstrating clearly the validity of the hypothesis concerning acidity and basicity. Many analysts use a heavy-element absorber such as lanthanum to assist in overcoming inter-element effects. This procedure has the effect of increasing the basicity of the flux and would need to be allowed for in assessing the nature of the flux. Chrome-bearing materials such as chrome ores and magnesite - chrome refractories pose a special problem.The whole range of these materials as used in ceramics produces a wide spread of composition with variation of content of four elements. Chromium (as Cr,O,) may vary from 5 to 50%, magnesium oxide from 15 to SOY0, iron(II1) oxide from 5 to 25% and alumina from 2 to 30%. Iron(II1) oxide and magnesia will require an acidic flux, and an acidic flux is preferred for alumina, but chromium(II1) oxide is usually more readily decomposed by an air oxidation in a basic flux. A series of experiments showed that the best results were achieved by moderating the acidity of the tetraborate with a slightly smaller amount of metaborate (1 part of sample + 10 parts of lithium metaborate + 12i parts of lithium tetraborate). An adequate amount of melt to form a bead is achieved by the use of 0.4 g of sample.Other analysts taking part in the activities of the British Ceramic Research Association Working Group on X-ray fluorescence reduced the acidity of lithium tetraborate by adding either l a n t h a n ~ m ~ ~ ~ ~ ~ or strontium and lanthanum.10,30806 BENNETT AND OLIVER: DEVELOPMENT OF FLUXES FOR ANALYSIS Analyst, VoZ. 101 TABLE I SUITABLE FLUXES FOR VARIOUS TYPES OF MATERIAL Parts of flux to 1 part of sample (ignited) 7 A -l Material Lithium metaborate Lithium tetraborate Silica - alumina range . . Steatite . . .. .. Bone china .. . . Bone ash (apatite) . . Zircon . . .. . . Zirconia . . .. . . Titania . . . . . . Limes tone .. . . Dolomite . . . . .. Witherite . . .. . . Barium titanate . .. . Borax frit . . . . Glaze, <20% PbO . . Glaze, 20-60% PbO . . Lead bisilicate . . .. Enamel . . .. .. Chrome-bearing refractory Silicon carbide . . .. Boron carbide . . .. Magnesite*’ . . . . Iron(II1) oxide . . .. .. .. . . .. . . . . . . . . .. .. . . . . . . .. . . . . . . . . .. . . .. 4 4 4 0 4 0 0 0 0 0 0 0 4 4 0 0 0 0 10 6 3.52 1 1 1 3* 1 12 12 5* 5* 10 12 12 1 1 12 12 12 12 12.5 1.5 0.88 * These ratios of flux to sample are the minimum to produce clear cast beads. For purposes of X-ray fluorescence, greater dilutions may be advisable. It would appear, therefore, that the selection of a suitable flux for minerals or refractories can be assisted materially by consideration of its chemical composition. When the sample contains a preponderance of acidic oxides such as silica, lithium metaborate is to be preferred, but when the sample contains a high content of basic oxides the more acidic lithium tetraborate is more suitable.The authors thank the Director of Research, Mr. A. Dinsdale, OBE, for permission to publish this paper. They also thank Mr. M. Holines for help in producing this paper, Mr. W. Baker and Mr. A. Willis for producing the fusion swirler apparatus, and Mr. D. Basnett for the supports for the casting dishes. 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. References Ambrose, A. D., Br. Ceram. Res. Ass. Spec. Publ., 1971, No. 72, p. 7. British Iron and Steel Research Association Report, MG/D/671/70. Carr-Brion, K. G., Paper presented a t Fifth Conference on X-ray Spectrochemical Methods, Swansea, Jenkins, R., and De Vries, J.L., “Practical X-ray Spectrometry,” Philips, Eindhoven, Macmillan, Anderman, G., and Allen, J. O., in Mueler, W. H., Editor, “Ninth Denver Conference, Advances in Adler, I., and Axelrod, J. M., Sfiectrochim. Acta, 1955, 7, 91. Claise, F., Norelco Reptr, 1957, 3, 3. Blanquet, P., “L’ Analyse par Spectrographie et Diffraction de Rayons X (Madrid) ,” Philips, Eind- Claise, F., and Sampson, C., Province de Quebec Services des Laboratoires, Report 67, 1962. Ambrose, A. D., Rr. Ceram. Res. Ass. Spec. Publ., 1974, No. 86, p. 25. Ambrose, A. D., “The X-ray Fluorescence Analysis of Aluminosilicate Refractories,” Corporate Ambrose, A. D., Rutherford, R.. and Muir, S., Mt?taZZurgia, 1970, 82, 119. Ashley, D. G., BY. Ceram. Res.Ass. Spec. PubZ., 1!371, No. 72, p. 17. Herrman, H., Bull. SOC. Fr. Cham., 1974, 13, 103. Kroeft, U., Keram. Z., 1974, 26, 434. Bennett, H., Oliver, G. J., and Holmes, M., BY. Cevam. Res. Ass. Res. Pap., 1974, No. 665; Trans. BY. Ceram. Soc., 1977, in the press. Orrel, E. W., BY. Ceram. Res. Ass. Spec. Publ., 1965, No. 50, p. 70. Hickson, K., Br. Ceram. Res. Ass. Spec. Publ., 1963, No. 43, p. 37. September, 1966. London, 1967. X-ray Analysis,” Volume IV, Plenum, New York, 1960, p. 414. hoven, 1962. Laboratories, British Steel Corporation Report MG/CC/565/72.October, 1976 OF CERAMIC MATERIALS BY X-RAY FLUORESCENCE SPECTROMETRY 807 19. 20. 21. 22. 23. 24. 26. 26. 27. 28. 29. 30. Spectrogr. News Lett., 1954, 7 , 1. Sastry, B. S. R., and Hummel, E. A., J . Am. Ceram. Soc., 1959, 42, 216. Ingamells, C. O., Analyt. Chem., 1966, 38, 1228. Boar, P. L., and Ingram, L. K., Analyst, 1970, 95, 124. Van Loon, J. C., and Parissis, C. M., Analyst, 1969, 94, 1057. Levin, E. N., Robbins, C. R., and McMurdie, H. F., “Phase Diagrams for Ceramicists,” American Selman, G. L., Spender, M. R., and Darling, A. S., Platin. Metals Rev., 1966, 9, 130. Bennett, H., and Oliver, G. J., Analyst, 1971, 96, 427. Ambrose, A. D., BY. Ceram. Res. Ass. Spec. Publ., 1974, No. 88. Newell, D., Steetley Organization Research Department Workshop, personal communication. Tunney, A., British Steel Corporation Research Centre, Teesside, personal communication. Salt, P. D., Central Laboratories, English Clays, Lovering, Pochin and Co. Ltd., St. Austell, personal Received May llth, 1976 Accepted June 2nd, 1976 Ceramic Society, Columbus, Ohio, 1964. communication.