Oct., 19581 The NICKEL BY A SOLVENT-EXTRACTION METHOD Determination of Direct Magnesium Photometry in 561 Solution BY R. 0. SCOTT AND A. M. URE (The Macaulay Institute for Soil Research, Craigiebuckler, Aberdeen, Scotland) The spectrochemical determination of magnesium in solution by the porous-cup - spark method has been greatly facilitated by the direct photo- meter described. The magnesium line at 2802 A is used with the strontium line at 4077 A as internal standard. Solutions containing from 0.3 to 24 p.p.m. of magnesium are analysed directly with a coefficient of variation of about &2-0 per cent. at a rate of forty determinations per hour. Inter- ferences are found to be negligible for most types of agricultural samples, such as extracts of soils and plant materials. IN agricultural laboratories, many thousands of samples have to be analysed each year for potassium, sodium, calcium and magnesium.Of these elements, the first three can be determined conveniently by flame photometry, especially when a multi-channel instrument is avai1able.l The determination of magnesium in this manner is unsatisfactory. In the first place, magnesium is not very sensitive in the flame and solutions generally have to be concentrated before adequate sensitivity is achieved. Secondly, the most sensitive flame line at 2851 A lies in an (OH)-band system, and, at the low concentrations required, some form of background correction is necessary. For the past few years, a spectrographic technique in which the porous-cup - spark method of excitation2 is used has been employed at the Macaulay Institute for determining magnesium in soil and plant extracts.The sensitivity is such that magnesium, in the range 0.3 to 24 p.p.m. in solution, can be determined without further concentration in the aceticSCOTT AND URE: THE DETERMINATION OF [Vol. 83 562 acid soil extract used for the flame-photometric determination of readily soluble potassium, sodium and calcium in soils. This spectrographic method proved to be so satisfactory that a direct-reading attachment for a Hilger small-quartz spectrograph has been constructed specifically for determining magnesium by the porous-cup method. DESCRIPTION OF APPARATUS DIRECT-READING ATTACHMENT- The plate-holder mounting of an E484 Hilger small-quartz spectrograph has been removed, and, in its place, the direct-reading attachment is fitted to the spectrograph casting.Two optical channels are provided, one for the magnesium line at 2 8 0 2 ~ and the other for a strontium internal-standard line at 4077 A. Fig. 1 shows schematic diagrams of the attachment and Fig. 2 shows a view of the interior. A steel backplate, A, slotted for the light beam, is bolted to the spectrograph casting. This plate bears guide rails, B, which carry two slides, C. These slides provide a coarse adjustment parallel to the focal plane and can be clamped by the guide rails. The magnesium slide is slotted to pass the magnesium line and carries a small 90" quartz prism, D, which reflects the magnesium line to an exit slit, as shown in Fig. 1 (c). The slides carry shelves, E, hung from brackets, F, by beryllium - copper springs, G.On these shelves are mounted the slit units and holders for the photomultiplier tubes. An independent movement of each shelf along the focal plane is provided by micrometer screws, H and J, which act against the appropriate double-spring assembly. A Scale inches 0 1 2 3 Magnesium shelf Strontium shelf (c) Fig. 1. Schematic diagram of the direct-reading attachment for a Hilger small- quartz spectrograph: A, steel backplate; B, guide rails; C, slides; D, quartz prism; E. shelves; F, shelf brackets; G, beryllium - copper springs; H and J, micrometer screws; K, exit slits; L, slit-focusing screws ; M, pivots ; N, photomultiplier tubes Fig. 1 (a). Elevation Fig. I@). Side view Fig. l(c). Plan The slits, K, Fig.1 (c), are supported on pivots, M, mounted on the shelves, E. The slits are fitted to brass discs that can be rotated to align them parallel to the entrance slit. The slits can be moved approximately 1 2 mm along the axis of the emergent light beamsSCOTT AND URE: THE DETERMINATION OF [Vol. 83 562 acid soil extract used for the flame-photometric determination of readily soluble potassium, sodium and calcium in soils. This spectrographic method proved to be so satisfactory that a direct-reading attachment for a Hilger small-quartz spectrograph has been constructed specifically for determining magnesium by the porous-cup method. DESCRIPTION OF APPARATUS DIRECT-READING ATTACHMENT- The plate-holder mounting of an E484 Hilger small-quartz spectrograph has been removed, and, in its place, the direct-reading attachment is fitted to the spectrograph casting. Two optical channels are provided, one for the magnesium line at 2 8 0 2 ~ and the other for a strontium internal-standard line at 4077 A.Fig. 1 shows schematic diagrams of the attachment and Fig. 2 shows a view of the interior. A steel backplate, A, slotted for the light beam, is bolted to the spectrograph casting. This plate bears guide rails, B, which carry two slides, C. These slides provide a coarse adjustment parallel to the focal plane and can be clamped by the guide rails. The magnesium slide is slotted to pass the magnesium line and carries a small 90" quartz prism, D, which reflects the magnesium line to an exit slit, as shown in Fig. 1 (c).The slides carry shelves, E, hung from brackets, F, by beryllium - copper springs, G. On these shelves are mounted the slit units and holders for the photomultiplier tubes. An independent movement of each shelf along the focal plane is provided by micrometer screws, H and J, which act against the appropriate double-spring assembly. A Scale inches 0 1 2 3 Magnesium shelf Strontium shelf (c) Fig. 1. Schematic diagram of the direct-reading attachment for a Hilger small- quartz spectrograph: A, steel backplate; B, guide rails; C, slides; D, quartz prism; E. shelves; F, shelf brackets; G, beryllium - copper springs; H and J, micrometer screws; K, exit slits; L, slit-focusing screws ; M, pivots ; N, photomultiplier tubes Fig. 1 (a). Elevation Fig. I@). Side view Fig. l(c).Plan The slits, K, Fig. 1 (c), are supported on pivots, M, mounted on the shelves, E. The slits are fitted to brass discs that can be rotated to align them parallel to the entrance slit. The slits can be moved approximately 1 2 mm along the axis of the emergent light beamsSCOTT AND URE: THE DETERMINATION OF [Vol. 83 562 acid soil extract used for the flame-photometric determination of readily soluble potassium, sodium and calcium in soils. This spectrographic method proved to be so satisfactory that a direct-reading attachment for a Hilger small-quartz spectrograph has been constructed specifically for determining magnesium by the porous-cup method. DESCRIPTION OF APPARATUS DIRECT-READING ATTACHMENT- The plate-holder mounting of an E484 Hilger small-quartz spectrograph has been removed, and, in its place, the direct-reading attachment is fitted to the spectrograph casting.Two optical channels are provided, one for the magnesium line at 2 8 0 2 ~ and the other for a strontium internal-standard line at 4077 A. Fig. 1 shows schematic diagrams of the attachment and Fig. 2 shows a view of the interior. A steel backplate, A, slotted for the light beam, is bolted to the spectrograph casting. This plate bears guide rails, B, which carry two slides, C. These slides provide a coarse adjustment parallel to the focal plane and can be clamped by the guide rails. The magnesium slide is slotted to pass the magnesium line and carries a small 90" quartz prism, D, which reflects the magnesium line to an exit slit, as shown in Fig.1 (c). The slides carry shelves, E, hung from brackets, F, by beryllium - copper springs, G. On these shelves are mounted the slit units and holders for the photomultiplier tubes. An independent movement of each shelf along the focal plane is provided by micrometer screws, H and J, which act against the appropriate double-spring assembly. A Scale inches 0 1 2 3 Magnesium shelf Strontium shelf (c) Fig. 1. Schematic diagram of the direct-reading attachment for a Hilger small- quartz spectrograph: A, steel backplate; B, guide rails; C, slides; D, quartz prism; E. shelves; F, shelf brackets; G, beryllium - copper springs; H and J, micrometer screws; K, exit slits; L, slit-focusing screws ; M, pivots ; N, photomultiplier tubes Fig. 1 (a). Elevation Fig. I@). Side view Fig.l(c). Plan The slits, K, Fig. 1 (c), are supported on pivots, M, mounted on the shelves, E. The slits are fitted to brass discs that can be rotated to align them parallel to the entrance slit. The slits can be moved approximately 1 2 mm along the axis of the emergent light beamsOct., 19581 MAGNESIUM IN SOLUTION BY DIRECT PHOTOMETRY 563 by means of screws, L, which act through pivots M, and by a similar amount across the axis of the beam. On account of the curvature of the spectral lines, the width of each exit slit was made greater than that of the entrance slit. A 3-mm x 0.02-mm entrance slit is used, with fixed-width exit slits of 10 mm x 0.05 mm for magnesium and 10 mm x 0.1 mm for strontium. A piece of suitably exposed and developed photographic plate, which serves as a neutral filter, is mounted behind the strontium exit slit to reduce the light intensity on the strontium photomultiplier tube. To facilitate the initial setting up, a piece of photographic film was held against the face of the slit and the image of the line was recorded on it.Without being moved, the film was then exposed to light through the slit from the'back. The film thus recorded the position 'of the spectral line image relative to the slit. A convenient accessory for this operation is an iron film-holder held against the face of the slit by a button magnet (Eclipse type A) placed against the back of the slit. A miniature lamp mounted inside the magnet serves to record the slit image. For the final more accurate setting, provision is made in the electronic circuitry for profiling either line by using the instantaneous voltage developed in a resistor by the photomultiplier current.After removal of the bases, each photomultiplier tube, N, with its resistance chain was fitted into a 40-mm diameter brass tube so as to project about 45 mm above the rim of the tube. These brass tubes were then filled with Di-jell 171 wax (obtained from Astor Boisselier and Lawrence Ltd.), and opaque plastic covers, with windows opposite the photocathodes, were fitted over the photomultiplier tubes. The RCA 1P28 photomultiplier tube used for magnesium had, at 950 volts, a dark current of 5 x lo-" ampere, compared with 8 x ampere before removal of its base. The brass tube carrying the photomultiplier and resistance chain is a sliding fit in a second brass tube, which is mounted on shelf E.The photomultiplier can therefore be rotated and moved vertically into its correct position before it is clamped. A cross-movement of the whole photomultiplier mount of about +2 mm is provided. The direct-reading attach- ment is light-proof, and has slides at the top and back. Parts inside the box are blackened, and the entire attachment is electrically earthed. INTEGRATING PHOTOMETRIC EQUIPMENT- The instrument has an integrating circuit in which the photomultiplier output current charges a condenser. The voltage developed in the condenser is a measure of the total light energy received and is measured by a null-reading electrometer - bridge circuit. The method used is basically that of Naish and Ram~den.~ A 1P28 photomultiplier tube, V,,, is used for the magnesium line and a 931A photomultiplier tube, VI2, for the internal-standard strontium line.V,, is operated at about 950 volts, but V,, has its voltage reduced to about 700 volts by resistor R16. Both voltages are supplied from a stabilised power pack, as shown in Fig. 3 (a). In addition, the 50-cycle mains voltage supply to the laboratory is controlled by a Ferranti voltage regulator, with a voltage stability of k0.5 per cent. The complete apparatus is shown in Fig. 4. When a determination is made, the integrating condenser for magnesium, C,, is con- nected in by S, and that for strontium, C, or C,, is selected by s,. With the measuring potentiometer, R,,, at zero and switches S, and S, at position 1, the integration is started by setting S, to position 2.The photomultiplier currents then charge their respective condensers until, a t the end of the exposure, S, is switched to position 3 and the two con- densers are isolated preparatory to measuring their voltages. To measure the voltage of the magnesium condenser, S, is switched to position 2 and the measuring potentiometer, R,,, is turned until an equal and opposite voltage is applied to the electrometer grid to balance the condenser voltage, as shown by a zero reading on galvanometer M,. To measure the voltage of the internal-standard (strontium) condenser, R,, is returned to zero, S, is switched to position 3 and the required angular rotation of R,, is read as before. The cycle of operations is completed and the instrument is re-set for further determinations by returning R,, to zero and S, and S, to position 1.R,, is a linear-law potentiometer, and thus the ratio of the magnesium and strontium angular rotations is proportional to the ratio of the magnesium and strontium condenser voltages and hence to the ratio of the relative intensities of the magnesium and strontium The circuit is shown in Fig. 3.564 SCOTT AND URE : THE DETERMINATION OF [Vol. 83 spectral lines. as the scale for Rl,. A 5-inch diameter 360" plastic protractor, illuminated from below, is used q x x x ' z 8 Either line can be profiled by using the instantaneous voltage developed by the photo- To profile the magnesium line, for example, multiplier current in a resistor (R, or R2,).S, is set at position 3, S, at position 1 or 2, S, at position 1 and S, at position 2. EXCITATION AND OPERATING CONDITIONS The porous-cup electrodes used for the determination of magnesium are made from 56mm diameter carbon rods, obtained from C. H. Champion & Co. Ltd. (less-pure grade). The cups are approximately 16 mm long and have a bore of 3.2 mm (4 inch), a drill with an included angle of 135" being used. The base thickness is 0.60 0.01 mm and no pre-heating or pre-sparking of the empty electrode is necessary. The counter-electrode has a sharpOct., 19581 MAGNESIUM I N SOLUTION BY DIRECT PHOTOMETRY 565 point of 70" included angle. The spark gap is 2mm and is not adjusted during the exposure. The source is an uncontrolled Hilger-type spark of 15,000 volts with 0.02-mH inductance, 0.001-pF capacitance and no added resistance.With these parameters, a relatively weak undamped spark is produced. Strontium has proved to be a suitable internal standard and is present in only small amounts in the materials being examined, so that a constant standard addition can be made, provided it is large enough. The magnesium to strontium ratio for a solution is determined by filling the porous cup with solution and sparking for 56 seconds. The direct reader is set to the integrating position, with the slit open, immediately before the spark is made, and sparking is stopped by a time switch after the discharge has continued for 56 seconds. During this period, about 0.11 ml of solution is consumed. All solutions to be analysed contain 600 p.p.m.of strontium and 2.0 t o 2.25 per cent. of acetic acid. The voltage ratio is calculated from the angular rotations of R,,, as described previously. The voltage of the magnesium con- denser, which can be measured by R,,, is limited by B, (see Fig. 3) to about 8 volts. If the solution tested gives too high a voltage with the 0.05-pF magnesium condenser, C,, the series combination of the nominal O.1-pF condenser, C,, and resistor R,, is switched in in parallel with C,, the charge of which is then distributed over a total capacity of about 0.15pF. The voltage is thus lowered to a readable value without the necessity for a second exposure. Resistor R,, is included to limit excessive current surges at this switching action. Should even higher magnesium contents have to be determined, provision could be made to in- corporate further condensers in parallel with C, and C,.The porous cup is placed 20 cm from the entrance slit with no condensing lens. With the nominal 0.05-pF condenser used for the magnesium line at 2802 A and the 1P28 photo- multiplier tube operated at about 950 volts, 6 p.p.m. of magnesium produce a condenser voltage of about 5.6 volts, 0.3 p.p.m. of magnesium about 0.72 volt and a blank solution (containing only strontium and acetic acid) about 0.24 volt. The dark current produces a condenser voltage of about 0.05 volt. A reduced voltage of about 700 volts to the 931A photomultiplier tube and a neutral filter behind the exit slit are required to lower the condenser voltage to a readable value.Other available strontium lines are not suitable, those at 4161 and 4215 A are in a band system, that at 4305 A is interfered with by calcium and that at 4607 A, which is almost as strong as the line at 4077 A, is an arc line. It is undesirable to reduce the added amount of strontium to less than 500 p.p.m. because a few parts per million may be present in the original solution. EFFECT OF VARIATION IN ELECTRODE PARAMETERS- Base thickness of the porous cup-Table I shows the effect of change in the base thickness of the porous cup on the relative intensities and intensity ratios of the magnesium and strontium lines. There is apparently a depression of the observed magnesium to strontium ratio with increase in base thickness, the relative intensity of strontium increasing more rapidly than that of magnesium.Such changes in base thickness as occur during preparation of the cups should introduce only very small errors. TABLE I Each determination was carried out in presence of 3 p.p.m. of magnesium A 0-1-pF condenser is used for the strontium line a t 4077 A. Each result is an average of seven replicate determinations. EFFECT OF CHANGE IN BASE THICKNESS OF THE POROUS CUP Base Relative Relative Intensity Error in thickness, intensity of intensity of ratio magnesium mm magnesium strontium (Mg to Sr) content, yo 0.40 125.0 148.8 0.839 + 5.7 0.60 129.3 161.1 0.802 0.0 0.80 131.9 168-8 0.782 - 2.4 Counter-electrode shape-Table I1 shows the effect of different shapes of counter-electrode. Each result is an average of six replicate determinations and the apparent magnesium contents were read from a standard curve prepared by using counter-electrodes with 70" points.It can be seen that, when rods with flat tops were used as counter-electrodes, both the relative566 SCOTT AND URE: THE DETERMINATION OF [vo~. 83 intensities and the intensity ratio increased with the diameter, the relative intensity of magnesium increasing more rapidly than that of strontium. When pointed rods were used, both the relative intensities and the intensity ratio increased with the included angle of the point, but much more slowly than with increase in rod diameter. Little change in the intensity ratio is found for points with included angles between 70" and 90". TABLE I1 EFFECT OF COUNTER-ELECTRODE SHAPE Each determination was carried out in presence of 1.2 p.p.m.of magnesium Rod Relative Relative Intensity Error in diameter, Included intensity of intensity of ratio magnesium mm angle magnesium strontium (Mg to Sr) content, % 2.0 - 43.9 113.2 0.388 +8 58.8 136.3 0.431 + 23 2.5 71.7 152.5 0.468 + 37 3.0 4.0 - 96.7 174.3 0.556 + 73 5.5 - 158.9 237.3 0.669 + 110 - 50" 43.1 129.2 0.334 - 8.4 - 60" 44.9 133.6 0.337 - 6.7 - 70' 46.7 130.3 0,358 0.0 - 80" 52.0 147.3 0.353 - 1.8 - goo 53.9 148.3 0.364 + 2.7 - 110" 62.7 156.1 0.395 + 14.1 - 125' 74.3 167.1 0.443 + 30.8 Flat-top#ed counter-elecirode- - - Pointed counier-electrode- Length of spark gap-The effect of change in the length of spark gap on the intensity ratio is shown in Table 111. Increase in the gap length produces only a slight increase in intensity ratio and apparent magnesium content, although the individual relative intensities increase considerably.In practice, an optical projection system is used to position the electrodes and the gap is set to within 10.05 mm of 2.0 mm. No error should result from such slight changes in the gap length. TABLE I11 EFFECT OF THE LENGTH OF SPARK GAP Each determination was carried out in presence of 1.2 p.p.m. of magnesium Relative Relative Intensity Error in magnesium mm magnesium strontium (Mg to Sr) content, yo 1.5 42.1 130.5 0.323 - 0.5 2.0 53.1 163.5 0.325 0.0 2.5 58.7 174.5 0.336 + 4.0 3.0 72.1 209.9 0.343 + 6.9 Gap length, intensity of intensity of ratio TABLE IV EFFECT OF VARIATION IN POSITION OF ELECTRODES ACROSS THE OPTICAL AXIS Each determination was carried out in presence of 1.8 p.p.m.of magnesium Distance moved Relative Relative Intensity Error in across optical axis,* intensity of intensity of ratio magnesium mm magnesium strontium (Mg to Sr) content, Yo + 2.8 55.1 124-0 0.444 + 7.7 +2*1 52.2 125.6 0,416 - 0.4 + 1.4 56.6 133.3 0.423 + 1.4 + 0.7 59.4 142.7 0.416 - 0.4 0 58.5 139.7 0.418 0.0 -0.7 57.6 128.9 0.446 + 8.4 - 1.4 59.2 130.8 0,453 + 10.5 -2.1 57.7 119.1 0,485 + 20.3 - 2.8 61.0 111.3 0,547 $40.4 * Distance measured from an arbitrary zero position.Oct., 19581 MAGNESIUM IN SOLUTION BY DIRECT PHOTOMETRY 567 Position of electrodes on optical axis-Instantaneous current readings, by means of the profiling circuits, were taken while the discharge was moved horizontally across the optical axis. A constant current was obtained for magnesium over about f 3 mm and for strontium over about f 1.5 mm.Similarly, on the vertical axis, constant currents were found for magnesium over f 8 mm and for strontium over f 5 mm. The vertical position of the electrodes does not appear to be critical. The effect of moving the spark across the axis is shown in Table IV, an arbitrary zero position being taken within the constant current region. It can be seen that, over a range of 2.1 mm (from +2.1 mm to the zero position) the magnesium to strontium ratio is constant. With the working position set at +1.0 mm, small changes in the horizontal position of the electrodes should not affect the results. Discharge conditions-Several sparking procedures that gave good reproducibility are compared in Table V.The results are from twenty replicate determinations by each pro- cedure. It can be seen that procedure C is better than A or B and is also the simplest. In procedures A, B and C the electrodes were taken consecutively from storage racks of two hundred cups and points, as would be done in practice. Procedure D is the same as C except that the electrodes were taken at random from a stock of six hundred cups and points. Procedure C is now used with an expected coefficient of variation of +146 per cent. for a single determination. TABLE V STATISTICAL ANALYSIS OF DETERMINATION OF 1.8 p.p.m. OF MAGNESIUM In all instances the exposure was 56 seconds and a counter-electrode with a 70" sharp point was used. Each result is the average of twenty replicate determinations, and was obtained from a standard curve prepared by using procedure A BY DIFFERENT PROCEDURES Time Mean Amount of pre- amount of Coefficient Maximum Maximum of solution sparking, magnesium Standard of negative positive Procedure used seconds found, deviation, variation, error, error, p.p.m.p.p.m. % % % Full cup 15 1.942 0.0549 & 2.85 4.8 7.5 A * 0.11 ml 15 1.793 0,0496 & 2.76 4.8 6.1 Full cup Nil 1.648 0.0306 & 1-86 3.5 2.9 Bf D§ Full cup Nil 1.660 0.0401 12.41 3.5 4.6 c: * Spark gap was re-set to 2.0 mm after pre-sparking with cup empty. f Spark gap was not re-set after pre-sparking with cup empty. $ Spark gap was set to 2.0 mm and not altered during exposure. J As for procedure C, but cups and counter-electrodes were selected a t random from stock of 600 of each.EFFECT OF OTHER VARIABLES- No change in the magnesium to strontium ratio was observed when the input voltage t o the photomultiplier power pack was varied from 240 to 210 volts. Similarly, no change was produced by altering the current from the standard battery, B, (see Fig. 3), in the measuring circuit from 458 to 470 mA. Change in room temperature from 14.5" to 24" C has not caused curve drift, and it would appear that the exit slits are sufficiently wide to take care of any normal temperature change in the surroundings. Long-term curve drift has occurred at intervals of about 3 months, when it has been necessary to re-set the exit slits by means of micrometer screws H and J (see Figs. 1 and 2).As the exit slits require to be re-set in opposite directions, this is probably the result of the beryllium - copper springs, G (see Figs. 1 and 2), being twisted by the side pressure of the micrometers. The design, in fact, should be modified, so that the micrometers act along the central axes of the springs. To carry out a routine check of exit-slit positions by profiling, freshly prepared hemispherical MG5 aluminium-alloy electrodes are used for magnesium, and, for strontium, a porous cup in which is placed a solution containing 1000 p.p.m. of the element is used. Short-term curve drift was observed in the early stages of setting up the instrument, when a sphero-cylindrical quartz lens was used to focus the spark on the entrance slit. A slow decrease in intensity ratio was caused by the gradual fogging of this lens, which then568 SCOTT AND URE: THE DETERMINATION OF [Vol.83 absorbed more of the ultra-violet (magnesium) light than of the visible (strontium) light. No condensing lens is now used between spark and entrance slit. EFFECT OF EXTRANEOUS ELEMENTS- The amounts of calcium, aluminium, potassium and phosphorus liable to be present in soils and plants and in the extracts used for determining magnesium are shown in Table VI. The effects of the presence of different amounts of these elements on the apparent magnesium content are shown in Table VII. Aluminium, potassium and phosphorus have practically no effect on the apparent magnesium content, most of the errors being within the experimental error of the method.In practice, variation of these elements in soil extracts and plant materials should not produce significant errors. Increase of the calcium present appears to enhance the magnesium content, possibly because the specially purified calcium carbonate used in the preparation of the solutions contained about 50 p.p.m. of magnesium. (The purest commercially available calcium carbonate contained considerably more magnesium.) Extracts of Scottish soils normally contain 50 to 70 p.p.m. of calcium and seldom more than 200 p.p.m. At this level, the effect of variation in the calcium content can be ignored. TABLE VI AMOUNTS OF VARIOUS ELEMENTS NORMALLY PRESENT IN SOIL AND PLANT MATERIAL Acetic acid extract of soil A Amount of element present per 100 g of soil, mg Element Calcium .. . . 80to400 Aluminium . . 80 to 240 Potassium . . 2.4 to 80 Phosphorus . . 0.4 to 140 Amount of element present in final solution, p.p.m. 20 to 100 20 to 60 0.6 to 20 0.1 to 35 Plant material Amount of Amount of element present element present in dry material, in final solution, 0.05 to 6.0 0.4 to 48 0.001 to 1.0 0.008 to 8.0 1.2 to 64 0.15 to 8.0 0.05 to 1.0 0.4 to 8.0 I A > % p.p.m. TABLE VII ERRORS I N APPARENT MAGNESIUM CONTENT IN PRESENCE OF EXTRANEOUS ELEMENTS Calcium Aluminium Potassium Phosphorus and potassium &- A > Amount Error in Amount Error in Amount Error in Amount of Amount of Error in of element magnesium of element magnesium of element magnesium phosphorus potassium magnesium present, content, present, content, present, content, present, p.p.m.% p.p.m. % p.p.m. % p.p.m. Each determination carried out in presence of 0.6 $.P.m. of magnesium- 0 0.0 0 0.0 0 0.0 0 100 0.0 6 + 1.0 10 + 1.0 10 500 + 5.3 12 - 1.0 50 f 1.0 50 1000 + 8.5 32 0.0 100 - 1.0 100 - - 64 -4.1 _- - - 0 0.0 0 0.0 0 0.0 0 100 +0*2 6 +0*2 10 - 0.6 10 500 + 1.8 12 -0.8 50 -0.8 50 1000 +2.0 32 -0.8 100 -0.8 100 APPLICATION TO AGRICULTURAL MATERIALS Each determination carried out in presence of 3.0 $.p.m. of magnesium- - - - - 64 0.0 - ACETIC ACID EXTRACTS OF SOILS- present, p.p.m. 0 13 63 126 - 0 13 63 126 - content, % 0.0 + 1.9 + 1.0 + 1.0 - 0.0 + 0.6 - 0.4 + 0.2 - Ten grams of soil are shaken for 2 hours with 400 ml of 2.5 per cent. acetic acid, and the suspension is filtered. Part of the extract is used directly for determining sodium, potassium and calcium by flame photometry.For magnesium, 5 ml of strontium chloride solution (containing 5-0 g of strontium per litre) are diluted to 50 ml with the acetic acid extract. The standard solutions are prepared by diluting 5 ml of strontium chloride solution to 50 mE with stock solutions containing from 0.3 to 24 p.p.m. of magnesium in 2.5 per cent. acetic acid. These solutions are sparked in porous cups as described under “Excitation and Operating Conditions,” and the magnesium to strontium intensity ratios are calculated.Oct., 19581 MAGNESIUM IN SOLUTION BY DIRECT PHOTOMETRY 569 Two standard curves are plotted of the magnesium to strontium intensity ratio for the standard solutions against magnesium concentration (in milligrams of magnesium per 100 g of soil).One curve is plotted from 0.3 to 6 p.p.m., the 0.05-pF condenser being used, and another from 6 to 24 p.p.m., the 0.05-pF condenser again being used for the integration, but with the O a l - p F condenser connected in parallel before measurement of the voltage ratio. Samples with magnesium contents above 24 p.p.m. are diluted with a solution containing 500 p.p.m. of strontium in 2.25 per cent. acetic acid. AMMONIUM ACETATE EXTRACTS OF SOILS- Twenty grams of soil are mixed with 100ml of a N solution of neutral ammonium acetate, the suspension is set aside overnight, filtered, and then leached with ammonium acetate solution to a volume of 1 litre. Five millilitres of strontium chloride solution and 1 ml of glacial acetic acid are diluted to 50 ml with the extract.Acetic acid is necessary to ensure percolation through the porous base of the cup. Standard solutions containing from 0.3 to 24 p.p.m. of magnesium and 500 p.p.m. of strontium are prepared in a base solution of N ammonium acetate and 2 per cent. acetic acid. Standard curves are plotted as before for the ranges 0.3 to 6 p.p.m. and 6 to 24 p.p.m. of magnesium. PLANT MATERIAL- Ten grams of dry plant material are ashed at 450" C overnight and the ash is twice evaporated to dryness with 2.5 ml of concentrated hydrochloric acid to precipitate silica. The residue is extracted with 25 ml of dilute hydrochloric acid (1 + 4 v/v), filtered, and diluted to 500 ml with distilled water. An aliquot of this solution, generally 2 ml, is placed by pipette in a 50-ml calibrated flask, 5 ml of strontium chloride solution are added, and the solution is diluted to the mark with 2.5 per cent.acetic acid. Standard curves are plotted as before, standard solutions in acetic acid incorporating potassium dihydrogen phosphate, potassium sulphate, calcium carbonate and sodium chloride in proportions corresponding to an average plant ash being used. Twenty samples of turnips were analysed in duplicate by the proposed method and by a colorimetric procedure with Titan yellow? The mean values for magnesium were, respec- tively, 0.0687 and 0.0685 per cent. (standard deviations j0-00168 and -t0.00384 per cent.), but the coefficient of variation was better by the proposed method (k2-4 per cent.) than by the colorimetric procedure (k5.6 per cent.).From other statistical comparisons, it would appear, as might be expected, that with the proposed method the coefficient of variation is constant at both high and low magnesium contents, although the standard deviation is constant for the colorimetric procedure. OTHER MATERIALS- Because of the high sensitivity of the method and the small effect of extraneous elements, magnesium can be determined in solutions derived from practically any material of agricul- tural interest. At the dilution required for determining magnesium in most limestones, for example, there are no more than 150 to 200 p.p.m. of calcium present in the final solution, an amount that, as shown previously, has little effect on the results. Similarly, with other materials, after the required dilution the effect of extraneous elements is generally negligible.Materials that have been analysed include woods, limestones, sands and minerals. The instrument has now been in use for over 2 years and has proved to be thoroughly reliable, three to four hundred magnesium determinations per week being carried out. Some forty determinations per hour are possible. The life of the batteries is at least 1 year under these conditions. APPENDIX LIST OF COMPONENTS USED IN THE CONSTRUCTION OF THE POWER PACK AND STABILISER FOR PHOTOMULTIPLIER TUBES AND THE INTEGRATING AND MEASURING CIRCUITS (Fig. 3) = 1-megohm &watt resistor (five 200,000-ohm 1-watt carbon resistors in series). = 100-ohm #-watt carbon resistor. = 200,000-ohm 5-watt resistor (five 1-megohm 1-watt carbon resistors in parallel).570 SCOTT AND URE [Vol.83 R,, R,, R,, R7 = 1.5-megohm f-watt carbon resistor. = 25,000-ohm 15-watt resistor (five 5000-ohm 3-watt wire-wound resistors in = 15,000-ohm 3-watt wire-wound resistor. = 80,000-ohm 12-watt resistor (four 20,000-ohm 3-watt wire-wound resistors in = 35,000-ohm 3-watt wire-wound variable resistor. = 10,000-ohm 5-watt wire-wound resistor. Rl, = 20,000-ohm 3-watt wire-wound variable resistor. R14? R13 = 100,000-ohm 3-watt wire-wound variable resistor. R16 = 400,000-ohm 2-watt resistor (four 100,000-ohm )-watt high-stability carbon = 20,000-ohm 10 per cent. 10-watt wire-wound linear-law potentiometer = 5000-ohm 3-watt wire-wound resistor. = 4000-ohm 3-watt wire-wound variable resistor.= 5000-ohm 3-watt wire-wound variable resistor. = 50-megohm f-watt high-stability carbon resistor. = 500-ohm 3-watt wire-wound variable resistor. = 1000-ohm f-watt carbon resistor. = 0.5-pF condenser, 3.5-kV working. = 0.05-pF TCC Plastapack polystyrene dielectric condenser, 350-volt working. = 0.1-pF TCC Plastapack polystyrene dielectric condenser, 350-volt working. = ET.38 valve. series). series). R, Rs RlO Rll Rl, resistors in series). (Painton CV25). Rl, Rl,, R,,, R,, Rl, RZa, R24 R26 R2O R,, Cl, c,, c, ‘4, ‘8 V. c,, c7 - - - - -. -. Vi, V,, V,, V,, V, = CV1070 (7475) valve. V, = EF36 valve. = 250-volt 15-watt indicator lamp. = CVllll (R11, V1907) valve. = &volt 0-3-ampere indicator lamp. = RCA 1P28 photomultiplier tube. = RCA 93lA photomultiplier tube. = 9-volt grid bias. = 4.5-volt section of 9-volt grid bias. = 6-volt section of 9-volt grid bias. = 15-volt Ever Ready Batrymax B121. = Voltage dividers (ten 100,000-ohm )-watt high-stability carbon resistors in = Low-frequency choke. = 0-2000-volt electrostatic voltmeter. = 0-500-pA moving-coil microammeter. = 450-ohm Cambridge spot galvanometer, 180 mm per pA. = Single-pole ON - OFF switch. = 2-section 3-postion shorting switch, with silicone-treated ceramic insulation and non-shorting moving contacts. = 3-section 3-position switch, with silicone-treated ceramic insulation and non- shorting moving contacts. = Double-pole ON - OFF switch. = Mains transformer : primary windings, 0-210-230-250 volts : secondary windings, (a) 0-1800-2000 volts, 20 mA, (b) 4 volts, 2 amperes, 2-kV working. = Mains transformer: primary windings, 0-210-230-250 volts; secondary windings, two of 6.3 volts, 1.5 amperes, 2-kV working. = Mains transformer: primary windings, 0-230-250 volts: secondary winding, 6.3 volts, 1.5 amperes. = Mains transformer : primary windings, 0-200-230-250 volts : secondary winding, 0-30 volts, 2 amperes, multi-tappings, 0-8-volt tapping used. v, v, Vl, Vll Vl, VlS = 28D7 (Sylvania) valve. B,, B, B, B, Dl, D, Ll Ml M, Ma Sl, s,, s7 ss B6 series). s,, s,, s, s, TI T* T, T4 We acknowledge the assistance of Mr. A. M. Fraser in the design and construction of the adaptor fitted to the small-quartz spectrograph. REFERENCES 1. 2. 3. 4. Mitchell, R. L., Spectrochim. Acta, 1950, 4, 62. Mitchell, R. L., and Scott, R. O., Appl. Spectroscopy, 1957, 11, 6. Naish, J. M., and Ramsden, W., Spectrochim. Acta, 1952, 5, 295. Hunter, J. G., Analyst, 1950, 75, 91. Received February 28th, 1958