156 Analyst, March, 1967, Vol. 92, $9. 156-161 The Importance of Fuel Gas Composition in the Atomic-absorption Spectrophotometric Determination of Magnesium BY T. R. ANDREW AND P. N. R. NICHOLS (Central Materials Laboratory, The Mullard Radio Valve Company, New Road, Mitcham Junction, Survey) The specifications for commonly used “low temperature” fuel gases permit wide variations in composition. The effect, on the determination of magnesium in nickel, of using two samples of “propane” of differing com- positions is considerable, and the use of acetylene is advocated. The results emphasise the need for a complete specification of flame conditions to be given when reporting sensitivities and interference susceptibilities in atomic- absorption procedures when using flame atomisation.FLAMES used for atomic-absorption spectrophotometry can be loosely classified as “cool” (<2000” C) or “hot,” Examples of cool flames are air - town gas, air - butane and air - propane, while the most commonly used hot flame is air - acetylene, although nitrous oxide - acetylene is coming into use. Most of the metallic elements can be determined with more than one flame, but often there is a considerable difference in sensitivity between flames1 The difference in sensitivity of detection of elements in these flames and the extent of interference by other elements are usually attributed to the different temperatures reached in the reaction zone of the flame. In an earlier publication2 we reported the determination of low concentrations of mag- nesium in nickel by atomic-absorption spectrophotometry.In this work we used an air - town gas flame and found no interference from 500 p.p.m. of nickel in the determination of up to 0.50 p.p.m. of magnesium. After several months’ satisfactory application of this procedure, occasional runs showed clear evidence of slight interference by nickel. When all other variables had been studied and found to be adequately controlled, suspicion fell on the town gas supply. Although town gas is supplied at a constant calorific value, its composition varies widely from place to place and also from day to day at the same place, depending on the local gas producers’ requirements. Two typical analyses, not from the same supply, are shown in Table I. TABLE I COMPOSITION OF TWO SAMPLES OF TOWN GAS Sample A, per cent.Carbon monoxide . . .. . . 2.7 Hydrogen . . . . ,. . . 50 Nitrogen . . . . .. .. 8.0 Carbon dioxide . . .. .. 12.3 Oxygen . . + . .. .. 0-5 Methane and ethane .. . . 23.6 Other hydrocarbons . . .. .. 3.0 Sample B, per cent. 9.1 44.9 18.2 2.7 0.4 19.3 6.4 This variation could well explain the occasional interference of nickel encountered in the determination of magnesium and, for other determinations, could lead to variations in sensitivity and amount of interference encountered. We decided, therefore, to reject town gas as a standard fuel and turned our attention to other fuels. Air - acetylene was the most commonly used hot flame and air - butane or air - propane seemed a reasonable choice as a cool flame. Before proceeding to examine these, however, it seemed prudent to investigate the stated purity and constancy of composition of these three commonly used fuels.Acetylene is readily available at a purity of higher than 98 per cent. with the sole proviso that the last 5 to 10 per cent. of the cylinder contents may be contaminated with acetone.ANDREW AND NICHOLS 157 Butane as supplied commercially conforms to existing specifications, compiled without consideration of possible analytical applications, which require only that it consists principally of butane or butene, or both. The composition is dependent on the refinery from which it is obtained. Typical compositions are shown in Table 11. TABLE I1 COMPOSITION OF COMMERCIAL BUTANE Refinery Component, mol. per cent. A B C D E Ethane and ethene .. . . 0.1 Nil Nil 0.2 Nil Propene . . . . . . 1.3 Nil Nil 3.4 3.1 Propane . . . . . . 9.2 11.5 0.6 8.4 6.2 Butenes .. * . . . 12.3 Nil X il 31.9 46.5 19.8 3.9 15-6 36.2 68.7 95.5 30.5 8.0 Iso-butane . . . . * ' } 76.9 Butane . . . . . . Propane is available at higher than 99 per cent. purity, but the cost of the pure material Commercial propane is again a by-product of oil Typical values is such as to preclude its use as a fuel. refineries and its composition is also a function of the particular refinery. for the composition of propane from various refineries are given in Table 111. TABLE I11 COMPOSITION OF COMMERCIAL PROPANE Refinery Component, mol. per cent. A Propane . . . . . . 91.1 Propylene . . . . . . Nil Methane . . .. . . Nil Ethane . . " "} 2.3 Ethylene .. . . . . Butane . . . . . . 6.6 Butylene . . .. . . Nil B 72.5 24.5 Nil 2-4 Nil 0.3 0.1 C 89.1 Nil Nil 4.0 Nil } 6.9 D 57.4 39.5 Nil 0.7 0.1 2.3 Nil 7 E 95.2 Nil Nil 1-8 Nil 3.0 Nil Consideration of Tables I1 and I11 shows that propane is available from three out of five sources at about 90 per cent. purity, whereas only one source produced butane of a comparable purity. Accordingly, propane was considered to be a more appropriate subject for study than butane. The information given in Tables I1 and 111 was supplied by Shell Oil Company from whom we obtained small amounts of propane from refineries D and E, which were selected as providing the two extremes of composition given in Table 111. Their effects on the determination of magnesium in nickel were compared.EXPERIMENTAL To study the behaviour of magnesium in both air - acetylene and air - propane flames we constructed a burner having two separate interchangeable burner heads. This burner could then use the same atomiser conditions, and stable flames could be produced merely by altering the fuel-inlet jet. The basic design of the burner is shown in Fig. 1. The acetylene burner head was supplied by Messrs. Hilger and Watts; it has a slot 0.5 mm wide and 120 mm long, and it is a sliding fit on the main burner tube. The propane burner-head top-plate was made from 1-5-mm stainless steel carrying five rows of 1-32-mm holes on 3-mm centres. This top-plate was attached to the brass casting, as supplied with the early Hilger and Watts town gas burner, the stem of which had to be reamed out to be a sliding fit on the new burner tube.The sample spray and fuel enter the main burner tube through a single side-tube just above the base. To avoid any possible encrustation of the fuel-inlet jet with evaporated sample solution it is set back at an angle of 45". The base of the burner tube has a removable cap to facilitate cleaning.158 ANDREW AND NICHOLS: FUEL GAS COMPOSITION IN THE [AflakSt, VOl. 92 The pressure of air for maximum efficiency of the E.E.L. atomiser used in this work was 30 lb per sq. inch (2 x lo5 N/m2) and this therefore represented the amount of air to be admitted to the burner. The air-to-fuel ratio for stable flames with the two fuels was expected to be different, so provision was made for the use of fuel-inlet jets of different sizes.This was done by making the jet compartment removable (see Fig. 1). A series of jets, designed for the Amal Maximus burner, was turned down so that the jets readily slid into the jet compartment tube, the bases being slotted to permit insertion of a screwdriver for fixing. Fig. 1. Burner assembly: (a) interchangeable burner head ; (b) interchangeable fuel-inlet jet; (c) main burner The main burner assembly was made of brass, for ease of machining and ready avail- ability. Stainless steel could have been used, but at a considerably higher fabrication cost, and was hardly justified as no contamination by brass burners has so far been detected by us. Tile remainder of the apparatus used in this work was as previously reported,2 with the exception of the scale expansion unit which has been modified to permit a 10-fold increase in sensitivity.The circuit diagram of the scale expansion unit is shown in Fig. 2. recorder heater circuit ““4 TI From FA 17 qB \ TO recoi Fig. 2. Scale expansion unit -der Experiments were carried out with the two burner heads to determine optimum conditions for apparently stable flame formation at air flow-rates of 7 litres per minute through the E.E.L. atomiser. These conditions are shown in Table IV. TABLE IV CONDITIONS FOR STABLE AIR - PROPANE AND AIR - ACETYLENE FLAMES Inlet-jet diameter, Fuel gas flow-rate, Fuel gas pressure, Fuel gas mm ml per minute inches (water gauge) Propane .. .. 0.55 300 4 Acetylene . . .. 0.76 700 4March, 19671 ATOMIC-ABSORPTION SPECTROPHOTOMETRIC DETERMINATION OF MAGNESIUM 159 Study of the air - propane flames showed that the sensitivity of the two flames towards magnesium when using propane D and E, under otherwise identical conditions, differed widely as also did the effect of nickel.To obtain the maximum information from this study, the variation of absorption with the position of the optical path in the flame was examined with a wide range of air-to-fuel ratios. The results are presented in Figs. 3 and 4. DISCUSSION Consideration of the several graphs in Figs. 3 and 4 show that, except under a limited set of conditions, the presence of nickel has a marked effect on the profile of magnesium absorption in the flame and also that the composition of the propane fuel is of importance. It is worthy of note that the degree to which the presence of nickel affects the magnesium absorption and the variation of absorption through the flame is very different for the two compositions of propane, and conditions which might be established for any one supply of propa well need re-investigation when a fresh supply was obtained.0 I I I I I 0 1 2 3 4 5 Propane pressure, inches (water gauge) 90 c --+ c -- --+ - - -_ ‘t I I I I I 1 2 3 4 5 Propane pressure, inches (water gauge) 01 I I I _ I 1 0 1 2 3 4 5 Propane pressure, inches (water gauge) 9 0 c Propane pressure, inches (water gauge) e could Fig. 3. Variation of instrumentre sponse with propane pressure for magncsium and for magncsium plus nickel a t different settings of burner position below optic axis: (a) 0.40; ( b ) 0.55; (c) 0-60; ( d ) 0.65 inches.Propane from refinery E represented by -; from refinery D by - - - - -; curves (0) obtained with 0.13 p.p.m. of magnesium; curves (+) with 0.13 p.p.m. of magnesium plus 250 p.p.m. of nickel While it is possible to use an air - propane flame for the determination of magnesium in nickel, the need for close control of conditions and for the regular provision of standards renders the procedure less than satisfactory. When air - acetylene was used with the slot-type burner head, the results given in Fig. 5 were obtained. I t is clear that the air - acetylene flame is less sensitive to operating conditions than air - propane and that the use of air - acetylene is to be preferred for the determination of magnesium in nickel as it gives enhanced sensitivity, and the presence of nickel is without influence on the determination.160 ANDREW AND NICHOLS: FUEL GAS COMPOSITION 1N THE [Ana&St, VOl.92 9 0 0 90 t 0 8 0.40 0.45 0.50 0.55 0.60 0.65 Burner position below optic axis, inches t 0 4 040 0.45 0.50 0.55 0.60 0.65 Burner position below optic axis, inches oI 0.40 0.45 0.50 0.55 0.60 0.65 Burner position below optic axis, inches 01 I I I I I 040 0.45 050 055 0.60 0.6 Burner position below optic axis, inches Fig. 4. Variationzof instrument response with burner position for magnesium and for magnesium plus nickel a t different propane pressures measured in inches (water gauge) : (a) 1 ; (b) 2; (c) 3; ( d ) 4. Propane from refinery E represented by -; from refinery D by - - - - -; curves (0) obtained with 0.13 p.p.m.of magnesium; curves (+) with 0.13 p.p.m. of magnesium plus 250 p.p.m. of nickel This work has been confined to the study of one element, magnesium, but it seems to us probable that some of the differences of opinion expressed in the atomic-absorption spectrophotometric literature may well be a result of differences in composition of nominally identical fuels. It is also apparent that when using propane fuel it should not be assumed that the sensitivity and degree of interference obtained with one cylinder of propane will necessarily be the same with another cylinder. s- I (D o’ 0.b5 o!Io O!l5 0!20 0!2S OkO Ol35 0’40 0:s V Burner position below optic axis, inches Fig. 5. Variation of instrument response with burner position for magnesium and magnesium plus nickel a t different acetylene pressures measured in inches (water gauge): curve A, 1 ; curve B, 2; curve C, 3; curve D, 4; curve E, 5.The curves are the same for 0.13 p.p.m. of magnesium and 0.13 p.p.m. of magnesium plus 250 p.p.m. of nickelMarch, 19671 ATOMIC-ABSORPTION SPECTROPHOTOMETRlC DETERMINATION OF MAGNESIUM 161 We are grateful to Mr. C. H. R. Gentry, Head of the Central Materials Laboratory, and the Directors of the Mullard Radio Valve Company for permission to publish this work, and to the Shell Oil Company for provision of information on the composition of propane and butane. REFERENCES 1. 2. Gatehouse, B. M., and Willis, J. B., Spectrochirn Ada, 1961, 17, 710. Andrew, T. R., and Nichols, P. N. R., Analyst, 1962, 87, 25. Received August 9th, 1966 Appendix COMPONENTS LIST FOR THE SCALE EXPANSION UNIT (FIG. 2) = 500-ohm, l-watt resistor = 75,000-ohm, 0-5-watt resistor = 2000-ohm, %watt variable resistor with a tolerance of 5 per cent. = 100-ohm, 0.25-watt resistor with a tolerance of 5 per cent. = 200-ohm, 0-25-watt resistor with a tolerance of 5 per cent. = 300-ohm, 0.25-watt resistor with a tolerance of 5 per cent. = 510-ohm, 0.25-watt resistor with a tolerance of 5 per cent. = 680-0hm, 0.25-watt resistor with a tolerance of 5 per cent. = 900-ohm, 0-25-watt resistor with a tolerance of 5 per cent. = 4-pF capacitors, 50-volt working = Double-pole, double-throw switch = Single-pole, 6-way switch = Mullard OAZ203 = Transformer, ratio of input t o output, 1 : 2.5 V,, V, = Mullard OA5