1466 Artalyst, December, 1983, Vol. 108, pp. 1466-1170 Determination of Manganese in Precipitated Calcium Carbonate Samples Using Candol uminescence Spectrophotometry: Some Problems Associated with Preparation of Standards (The late) Ronald Belcher, E. Roy Clark, Malcolm H. Lloyd" and H el ena P uza nows h a -Tarasi ew i cz Department of Chemistry, University of Aston in Birmingham, Gosta Green, Birmingham, B4 VET The intense yellow candoluminescence given by manganese in a calcium oxide - calcium sulphate matrix when placed in a hydrogen - nitrogen - air flame has been used to investigate the determination of manganese (in the range 0.2-0.6 p.p.m.) in commercial samples of precipitated calcium carbonate. The results obtained by the candoluminescence emission technique have been compared with those obtained by atomic-absorption spectrophotometry and a kinetic spectrophotometric method.Keywords Manganese determination ; calcium carbonate ; candoluminescence spectrophotometry ; standard firefiaration Since the publication of the review on candoluminescence spectrophotometry by Belcher et al.1 in 1978 in which the history and practical applications of the technique were discussed very few other papers have been published. One by Sokolov2 deals with semiconductor applications and another by Dhaher and Kassir3 describes a new technique in which the matrix material of calcium oxide and calcium sulphate is fabricated into the form of a rod, which is then intro- duced into the hydrogen - nitrogen flame. This modification was used for the quantitative determination of manganese, bismuth and antimony in solutions.The use of a calcium oxide - calcium sulphate matrix in candoluminescence emissions has received much a t t e n t i ~ n ~ - ~ and the main parameters that determine reproducibility have been thoroughly investigated. The control of variables such as flame conditions, gas flow-rates and positioning of the matrix is relatively easy but the reproducibility of the matrix preparation is more difficult. Nichols et a1.' investigated manganese candoluminescence in calcium com- pounds but Belcher et aL5 were the first to develop a quantitative method for the determination of manganese over the range 0.1-3.0 ng in a procedure involving the application of 1 pl of the solution containing the activator to the centre of the matrix surface when a reproducibility of between 5 and 10% was obtained. In the surface application technique first described by Belcher et aL4 for the determination of bismuth, the moisture content and porosity of the matrix are critical because candoluminescence depends upon the concentration of the activator at the surface of the matrix exposed to the flame.If the absorption of a given volume of activator solution is not constant for a set of matrices then candoluminescence emission intensities will vary and this will result in poor reproducibility. Many of the applications of candoluminescence published so far have involved this surface technique and are thus subject to these problems, but one application that could avoid this step in the analysis is the determination of manganese in calcium carbonate, a compound that can be directly converted to calcium oxide.The oxide could then be converted into the calcium oxide - calcium sulphate matrix offering a simple procedure using no solutions and only one reagent, i.e., calcium sulphate, thus minimising contamination. The only problem is the production of suitable standards in which the manganese is uniformly distributed. Two approaches are possible. Firstly, standards could be produced by precipitation methods, or secondly, calibration could be made by another method. Both of these possibilities were investigated and are reported in this paper. The preparation of standards was investigated using two procedures and the results obtained are compared with those obtained by atomic- absorption spectrophotometry and a kinetic spectrophotometric method.8 * Present address : Ferro (Great Britain) Ltd., Wornbourne, Wolverhampton.BELCHER, CLARK, LLOYD AND PUZANOWSHA-TARASIEWICZ 1467 Experimental and Results Apparatus The candoluminescence emission intensity of manganese was measured at 580 nm using a Pye Unicam SP 900 flame spectrophotometer (slit width 0.15 mm) fitted with a Meker-type burner and a Pye-Unicam AR 55 chart recorder. The matrix was inlaid into the hexagonal aperture (3 mm deep, hexagonal sides 2 mm) in the head of an Allen screw and the matrix holder was positioned in the flame in a similar way to that reported previou~ly.~,5 A metal cover placed over the matrix holder ensured that no extraneous light reached the spectro- photometer. Gas flows of 2.0,2.0 and 7.25 1 min-l for hydrogen, air and nitrogen, respectively, were controlled by needle valves and monitored using RS Series “Meterate” flow tubes RS2/C, RSS/R and RSS/C (Glass Precision Engineering Ltd.). The burner assembly was as described previously and the horizontal position of the matrix near the edge of the flame was adjttsted until maximum candoluminescence intensity was obtained.This position was kept constant throughout, as was the height of the matrix surface above the burner (1.6 cm). Reagents Calcium oxide used in procedure 1 was prepared from analytical-reagent grade calcium carbonate heated at 800 “C in a muffle furnace with weighing checks to ensure that conversion was complete.Commercial samples of calcium carbonate containing various amounts of manganese were supplied by John and E. Sturge Ltd., Birmingham. Preparation of Stock Solutions De-ionised water was used for the preparation of all solutions. A solution of 57.383 g of calcium nitrate was prepared by dissolving 82.581 g of analytical-reagent grade Ca(N0,),.4H20 in water and diluting to 100 ml. A solution of 5.556 g of sodium carbonate was prepared by dissolving 15.000 g of analytical-reagent grade Na,CO,. 10H,O in water and diluting to 100 ml. The manganese calibration solutions were prepared by appropriate dilution with de-ionised water of various volumes of solution taken from 100 ml of an atomic-absorption standard solution containing 996 pg ml-1 of manganese in 2% nitric acid (Aldrich Chemical Co.Inc.). A solution of 0.1 p.p.m. manganese, prepared from a 10 p.p.m. manganese solution, was used to prepare the manganese-doped matrices in the coprecipitation procedures. Coprecipitation Procedures Procedwe 1 The first procedure used was an adaptation of the “matrix preparation method” described by Belcher et in which the matrix was prepared from a slurry of calcium oxide - plaster of Paris. The modification involved the addition of manganese solution to the de-ionised water used to prepare the slurry so that in the alkaline conditions manganese was precipitated as hydroxide into the calcium hydroxide - plaster of Paris suspension. (a) A 1.5-g mass of the calcium oxide mixed with 0.187 5 g of plaster of Paris was suspended in 11.25 ml of de-ionised water and the mixture was shaken for about 10 min.The resulting suspension after filtration by vacuum until sufficiently dry was inlaid into a set of weighed Allen screws in the manner previously de~cribed.~ After drying in an oven at 110 “C for 13 min the matrices were stored in a desiccator until ready for use. Before each matrix was introduced into the hydrogen - nitrogen - air flame, an excess of matrix material was carefully scraped off with a clean razor blade so as to leave a clean surface, which was level with the top of the screw. The Allen screw plus matrix was weighed and positioned in the flame using the apparatus described above. (b) The procedure given under (a) was then repeated using 11.25-ml amounts of de-ionised water containing amounts of 0.1 p.p.m.manganese solution such that the resulting sets of matrices based on calcium oxide were equivalent to calcium carbonate samples with manganese contents of 0.2,0.3 and 0.4 p.p.m. An average candoluminescence emission intensity for each set of matrices was calculated (Table I). Procedlure (2) In the second method, manganese carbonate was coprecipitated with calcium carbonate from a manganese-doped calcium nitrate solution using sodium carbonate solution. The filtered The plaster of Paris was of technical grade. The candoluminescence emission was then measured.1468 Manganese concentration, p.p.m. 0 0.2 0.3 0.4 0.6 BELCHER et a,?. : DETERMINATION OF Mn IN CaCO, AyzaZyst, VoZ. 108 TABLE I CANDOLUMINESCENCE EMISSION INTENSITIES n = 8 in all instances.Procedure 1 Procedure 2 A ~r -7 r--------A-------- Intensity* Intensity* Mean of corrected for Mean of corrected for 0.0408 11 - 0.0442 24 - 0.0434 33 22 ( 4 4 ) - 0.044 1 39 28 ( 4 6 ) 0.0468 50 39 ( 5 7 ) matrix mass/g Intensity" blank matrix mass/g Intensity* blank - - 0.042 5 42 18 ( A l l ) 0.042 2 60 36 (It211 - - - - - - * One chart division = 0.25 cm. 95% Confidence limits (chart divisions) are given in parentheses. precipitate was then heated to produce manganese-doped calcium oxide, which was subse- quently used to prepare the matrix with plaster of Paris. (c) A 3.5-g mass of calcium carbonate was obtained by adding 66.80 nil of the sodium car- bonate solution to 10.62 ml of the calcium nitrate solution, filtering off 21.0 ml of water, drying in an air oven at 110 "C for 1 h and then conversion into calcium oxide by heating in a muffle furnace at 800 "C for 4 h.A set of matrices was prepared, and the corresponding candoluminescence emission intensities were measured in the same manner as described under Procedure 1 (a) above. ( d ) Samples of calcium carbonate doped with 0.3 and 0.6 p.p.m. of manganese were similarly prepared by using the required proportions of de-ionised water and 0.1 p.p.m. manganese solution in total volumes of 21 ml. These carbonate samples were used to prepare samples of calcium oxide, which in turn were used to prepare two sets of matrices whose candoluminescence emission intensities were measured in the same manner as described under Procedure 1 (a). The results, together with those for samples containing no manganese [Z(c)], are given in Table I.Candoluminescence Emission Intensities of Commercial Samples of Calcium Carbonate Samples were converted into calcium oxide by heating in a muffle furnace under the same conditions as described under Procedure l(a). The matrices were then prepared using the same amounts of calcium oxide and plaster of Paris and the same procedure as previously described. The candoluminescence emission intensities of four samples containing 0.2, 0.3, 0.4 and 0.6 p.p.m. manganese are given in Table 11. Determination of Manganese by Atomic-absorption Spectrophotometry and a Kinetic Spectrophotometric Method A Perkin-Elmer PRS-10,460 atomic-absorption spectrophotometer was used for the deter- mination of manganese using the method of additions.A Pye Unicam SP8-100 spectrophoto- meter was used for the measurement of absorbance in the kinetic spectropnotometric method. Discussion The results for the determination of manganese in the commercial samples of calcium carbonate using atomic-absorption spectrophotometry and the kinetic spectrophotometric method are given in Table 11. There is good agreement between the two methods and both may be used as the basis for the standardisation of the candoluminescence method. Graphs of average candoluminescence emission intensities, using coprecipitation procedures 1 and 2 (Table I) and those for the commercial samples of calcium carbonate (Table II), are shown in Fig. 1, in which intensities are plotted against the manganese contents of calcium carbonate and the manganese contents of the doping co-precipitation procedures.Corrections are made for the candoluminescence emission intensity of "blank" determinations obtained by extra- polation back to zero manganese content. This emission arises from the manganese content of the plaster of Paris. Purer grades of plaster of Paris or freshly precipitated calcium sulphateDecember, 1983 SAMPLES USING CANDOLUMINESCENCE SPECTROPHOTOMETRY 1469 TABLE I1 CANDOLUMINESCENCE EMISSION INTENSITIES AND MANGANESE CONTENTS OF COMMERCIAL SAMPLES OF CALCIUM CARBONATE BY ATOMIC-ABSORPTION SPECTROPHOTOMETRY AND KINETIC SPECTROPHOTOMETRY For caiidoluinincscence intensity measurements n = 13 ; for both atomic-absorption and kinetic spectrophotometry n = 5. Sample B t Sample C$ Sample Df --7 ~ ---- h T,---------A--------7 r------h----- Sample A* h--- __-__ Matrix mass Intensity? Matrix mass Intensity7 Matrix mass Intensity? Matrix mass Intensity7 Meanlg 0.0415 13 0.0422 15 0.0427 21 0.0443 26 Relative standard deviation, ?& .. . . 3.0 23.1 5.2 13.3 4.0 19.0 4.5 15.4 Mean, after correction for blank - 9 - 11 - 17 - 22 95% confidince li&, . . *2 chart divisions . . . . - * Manganese content: 0.205 p.p.m. f 1.1% (r.s.d.) by AAS; 0.224 p.p.m. f 1.6% by kinetic spectrophotometry. t Manganese content: 0.301 p.p.m. f 1.1% (r.s.d.) by AAS; 0.305 p.p.m. & 1.8% ( r d . ) by kinetic spectrophotometry. 3.3% (r.s.d.) by kinetic spectrophotometry. 5 Manganese content: 0.608 p.p.m. f 2.4% (r.s.d.) by AAS; 0.609 p.p.m. 6 3.2% (r.s.d.) by kinetic spectrophotometry. 7 Candoluminescence intensity, 1 chart division = 0.25 cm.Standard deviationig : : 0.0016 3 0.0022 2 0.0017 4 0.0020 4 - *2 - *I - f 2 Manganese content: 0.405 p.p.m. f 0.8% (r.s,d.) by AAS; 0.416 p.p.m. dihydrate (CaS0,.2H20) subsequently heated to 120-130 “C to give CaSO4.4H,O should produce lower “blank” emissions. The peak heights for emissions obtained for coprecipitation procedures 1 and 2 are both higher than expected. This can possibly be attributed to a non-uniform distribution of manganese throughout the particles of the matrices, the manganese being more concentrated near or on the surface. 0.2 0.4 0.6 Manganese concentration, p.p.rn. Fig. 1. Candoluminescence emission intensities obtained using A, procedure 1; B, procedure 2; and C, commercial calcium carbonate.Procedure 1 (Fig. 1A) gives the highest candoluminescence emission and this could possibly arise from an “after precipitation” of manganese hydroxide, which coats the surface of the calcium hydroxide particles. In procedure 2 (Fig. 1B) higher emissions than expected are found. Here, it seems likely that manganese carbonate is more concentrated on the surface of the calcium carbonate than in the centre of the particles. Differences in particle sizes of the precipitates obtained in procedures 1 and 2 could account for the differences in emission intensities. Our results indicate that the preparation of standards by either of these coprecipitation1470 BELCHER, CLARK, LLOYD AND PUZANOWSHA-TARASIEWICZ procedures cannot be used and that the standardisation by alternative methods is thus justi- fied.The kinetic spectrophotometric method is time consuming but satisfactory for these low levels of manganese. The atomic-absorption spectrophotornetric method also involves a dis- solution step but the candoluminescence technique uses only one other reagent, calcium sulphate. The method is relatively simple once a calibration graph has been obtained and it is rapid in the hands of a skilled operator. The larger standard deviation by the candolumines- cence method compared with the other methods is due to the slight variations in gas flows. This problem could possibly be overcome by modification of the sample holder head so as to permit a change-over of samples without cutting off gas flows. One of us (H.P.-T.) thanks Dr. D. J. Harrison, Science Officer, and the British Council in Warsaw for the award of a visiting Research Fellowship. We also thank John and E. Sturge Ltd. of Birmingham for supplying samples of precipitated calcium carbonate and for their interest in this work. Refer en ce s 1. 2. 3. 4. 5. 6. 7. 8. Belcher, R., Nasser, T. -4. K., Shahidullah, M., and Townshend, A., Int. Lab., 1978, Jan./Feb., 45. Sokolov, V. A., Zh. Fiz. Klzinz., 1978, 52, 3090. Dhaher, S. M., and Kassir, 2. M., Ana2. Chew, 1980, 52, 459. Belcher, R., Ranjitkar, K. P., and Townshend, -4., Analyst, 1975, 100, 415. Belcher, R., Karpel, S., and Townshend, A., Talanta, 1976, 23, 631. Belcher, R., Ranjitkar, K. P., and Townshend, A., Analyst, 1976, 101, 666. Nichols, E. L., Howes, H. L., and Wilber, D. T., Carnegie Inst., Washington, Publ., 1928, No. 348. Dolmanova, J. F., Zolotova, G. A., and Ratina, M. A., Zh. Anal. Khim., 1978, 33, 1356. Received May 26th, 1983 Accepted July 4th, 1983