Document JNDJdyZYJo949KLQ83OOevMD2

LTD 45-57 December 14, 1945 DBTBS.iCEmyiQH OF LEAD BY DI5H1SQM A Critical Comparison of the Methods Described In LTD 42=5 and DSM 44-188 PURPOSE (1) To evaluate certain factors in the "mixed-color" method for lead given in LTD 42-5 (2) To compare this method with the "one-color55 method described in DSM 44-188. (3) To check the sulfite reduction method, given in DSM 45=39, for the removal of Iodine from the solutions obtained in the determination of TSL in air. The two methods have been compared with respect to the effects of concentration and pH of the lead solution, concentration and age of the reagent, presence of phosphate, and stability of the final solution. The methods are of compsrable sensitivity and, with proper control of experimental conditions, appear to bs equal in accuracy. However, it is believed that further, special experimental tests are needed to demonstrate unequivocally the quantitative recovery of lead from very dilute solutions in the one- color method. Contrary to the results given in DSM 44=188 It was found that the solutions obtained by the on-color method were relatively unstable on standing, so that careful timing.of the analysis was required. This point should receive further investigation before the method can be recommended for general use. At present, for laboratory use, the preferred method is that of LTD 42=5, modified to avoid a small interference.from phosphate in the second extraction. The sulfite reduction of iodine solutions is satisfactory, if the organic lead iodides are first decomposed by heating. The resulting solution can be analyzed for lead by the mixed-color method. K ? 00G5S82 1NTR0DPCT10H . There are two major procedures for the use of dithlzone (dlphenylthiocarbazone) for the determination of lead. These have been designated as the "one-color" and the nmi:xed-color" methods. . Both involve the extraction of the lead ion from an alkaline solution (pH 8.5 to 11} with an excess of dithlzone in chloroform, and obser~ vation of the colored lead dlthlzonate in the chloroform solution. S The procedure used in this Laboratory (LTD 42-5) for the determination of small amounts of lead by dithlzone, is based on the well known and generally accepted mixed-color method. Briefly, the lead in nitric acid solution is treated with citric acid, potassium cyanide, and ammonium hydroxide to bring the solution to pH 9,,5o The lead is then extracted with a chloroform solution of dithlzone. A portion of the excess dithlzone dissolves in the ammon iacal solution, but sufficient remains in the chloroform to give it a green to violet color. However, with a spectrophotometer, the red color of the lead dlthlzonate in the chloroform solution 1b readily determined. . Recently the Baton Rouge Development Section Laboratory (LSM 44-188) proposed a modification of the original one-color method. Instead of extracting the lead at a pH of about 9 and then removing 'he excess dithlzone by shaking with dilute ammonia, as in other one-color methods, the Baton Rouge method carries out the extraction at a pH of 1.2. At this pH, the excess dithlzone is almost completely extracted b the aqueous solution, leaving only the red lead dlthlzonate in the chloroform. It was known that lead dlthizonste is also extracted in part from the chloroform by aqueous solutions at pH .11 to 12, but in the present method, the large excess of free dithlzone in the solution apparently jjrevents this. Either dithlzone procedure is far from ideal. Several dependent equilibria are involved; '' < ' 1. Reversible reaction of lead with dithlzone. 2. Distribution of lead ditfiizonate between water and chloroform, . ' ----- -- , ; 5. Distribution of dithlzone between water and chloroform. ; 1 4. Shifting of these equilibria with variations in pH. *1 Systems involving such equilibria must be carefully standardized before they can oe used for quantitative analysis. Likewise, one - must recognize that in the use of such systems one cannot predict the effect of any variable but must specifically verify every conclusion .Vj* 1 by experiment. |  .. .#- 3* / When Clifford and Vlchraanm-*-' studied the one-color method of Winters, et. al,(2) , they stated? "Analysts using one-color methods find themselves in a dilemma. If they do not remove all the unbound dithizone, their results are high, and if they take meticulous care to remove it, they lose lead in the process .... The greatest accuracy of one-color methods is about 0.001 rag." For this reason, and because of the wide availability of 8pectrophotoraatric equipment, the mixed-color method has become greatly recommended. In this method, the presence of some excess reagent is not objectionable. The preparation of the blank for lead in the reagents, tinder the same conditions as the samples, provides an accurate correction of any errors thus introduced. Variable slit-width instruments such as the Beckman quartz spectro photometer or the Cenco Spectrophotelometer permit such compensation without decreasing the range of the method. Fortunately, the load dithizonfxte exhibits its maximum absorption at the wave length where the dithizone itself shows minimum absorption. Thus, for photometric measurement, the presence of the green color of the dithizone reagent, even in excess, is of no disadvantage. For visual comparison the change in hue seems to be even more easily distinguishable than a change in color intensity. With an appropriate choice of reagent concentration Clifford and Wichmann(l) could distinguish visually ten steps between 0 and 1 microgram of lead. Nevertheless, the one-color method proposed in DSM 44-188 offers a new. approach, to the problem, which may have advantages for certain uses. It seemed, therefore, desirable to make an independent check of this method. EXPERIMENTAL Effect of Lead Concentration. In the presence of an excess of reagent, both methods obey Beer's Law, as shown in Fig. 1. These standard aeries were run under the conditions described In DSM 44-188 and LTD 42-5, The procedure from'LTD 42-5 was changed slightly to eliminate phosphate Interference, which is discussed in detail in another part of this report. The data indicate the methods are comparable with respect to sensitivity. Using 1-cm. cells, from 0.004 to 0.082 mg. Pb can b determined using the maximum useful .range of the spectrophotometer. This range Is from 10$ to 90$ trsnsraittaney, as shown in Table I. Che decreased accuracy at high and low transmittancies results from the logarithmic relationship between color Intensity and concentration of colored substance. It S will be noted that In a careful study of the effect of a variable in / a colored system It 3 desirable to limit transmission readings be tween 20$ and 60$. K & 00C5S34 . TABLE I - Variation in Errors over Transa!ttanov Range Transmittancy 95 90 80 70 60 50 40 30 20 10 5 Concentration m&.Pb Absolute Error ,, for 1$ Transmittaney .._-.mg.Pb.. . ..... $ Error 0.002 .004 .008 .0 125 .013 .0245 *0525 .043 .058 .082 .117 j 0.00015 .00025 .00040 .00055 .00055 .00070 .00090 .0 0 110 .00175 .00400 .00800 7.50$ 6..2556 5 .00$ 4.30$ 3 .00# 2.30# 2.76# 2.50$ 3o0q 4.90$ 6 .80$ ' ,| '' "''b;4 -'.'Jr-' . '-.'.bf' . :v; . '.rv Because of the ease or extraction of the lead salt and its extremely favorable distribution between chloroform and water, both methods are extremely flexible, with respect to sensitivity. The applicable range of lead concentration can be extended considerably in both directions by varying the volume and concentration of the reagent solution. For example, Stengeri2) has applied the method v to magnesium-base alloys, using such a combination of these variables that the range for 1 cm, cells is 0,002 to 0.014 mg. Pb. However, such changes in procedure should be completely verified experimentally Effect of pH. In either dithizone method for lead, the pH is the critical variable In fact the only essential difference between the mixed-color method and the one-color method is in the pH used. In both methods the ease of control is of prime importance and the methods are comparable in this respect. Both systems are highly buffered by ammonium and cyanide ions and only a rough measurement of the various solutions used is required to furnish the desired pH. Chloroform solutions, containing approximately 0.025 mg. and 0*075"fflgKld:.wapeetiVely'*;jwe;prepared-to determine the solu bility of the lead dithizonate in the aqueous phase when extracted in accordance with the method described in DSM 44-188. Instead of determining the lead dithizonate photometrically after the first extraction as required in the method, as many as three repeat extreetions with fresh aqueous solution, at pH 11.4, were performed on individual samples. The results of these tests given in Table II show that the lead dithizonate distributes between the chloroform and the aqueous phases, when no excess dithizone la present, '' ' . 5 TABLE II - Solubility of Lead Dithizonato at S H . U . A Additional Extractions T.*510 ... mg, Pb _ Pb WR. Loss 0 i 2 JL0 i 2 3 - 52,5 59 67 b. 16,0 20,5 27,0 0,0255 ,0210 ,0160 ,,0105 ,030 ,,065 . ,,0565 *047 * Transmittancy m > a 0.0045 .0095 fiiSSL .0 15 ,,0235 o033 * These conditions are not applicable to the method described in PSM 44l88* The first extracting solution contained ' everything in the proper concentration* but* dulling the additional extractions* there was only a slight excess of dithisone. In an effort to determine whether any loss occurred in the presence of excess dithizone at this pH* another series was prepared and extractions were carried out with 50 ml portions of the water layer obtained from running a series of blanks. In this way 8 solution was prepared for additional extractions which corresponded to the usual extracting solution, The results of these tests are given in Table X U , TABLE XIX - Lead Loss of pH 11,4 in Presence of Excess Dlthizone Additional Extractions .? 51.0_____ Pb vim-, Loss 0 53 0,0225 1 50 ,,0250 0,0025 2 52 ,,023KT 0 ,,0005 A repetition of this series of tests gave the same results* including the same alight increase.in apparent lead for the sample which received only one additional extraction, Recognizing that the solubility of lead dithi2onate is small under these conditions and that there is sufficient lead in the water and reagents to cause an appreciable blank* further experiments seemed useless for our purpose. However* it is suggested that this solubility be. determined by some independent method on solutions which have been carefully freed of lead. We would likewise suggest that any such experiments should include more than one concentration of lead* specially one . in the concentration range requiring the use of 5-cm, cells. H0 . C0. G58S6 6 ^ Stability of Lead Plthlsonate. la an attempt to compare the two methods for low-concentration of lead, using 5-cm. cells, a standard series of solution containing between 0 and 0.020 mg. Pb was prepared by each method. Erratic results which did not fall about a straight line were obtained by the BSM 44-188 method. After standing a short time, one of the very weak solutions changed from pink to green. It was suspected that under ordinary laboratory conditions these solu tions were sufficiently unstable to require a careful control of time. To determine the effect of time, standard lead series for 5-cm. cells and 1-cra. cells were prepared and tlmei carefully. A straight line was obtained In each case. Readings were taken on some of the solutions after they stood In the dark, at room temperature, and in the transmission cells. The data for 5-cm. cells are given In Table IV and for X-cm. cells In Table V. . - M s M l l t y , studio3 Using 5-cm. Celia D5M 44-188 Method jniiu _1Q mln,, 1Q5 min. 2219. SSLfPb T51Q mg.Pb 1510 m<?,Pb 9 2 o5 78.0 57 oO 40.5 29*9 0.001 95 oC025 85 .005 65.5 .0075 44.0 02.0 33.1 0.0007 -- .0015 87.0 .0040 .0068 56.2 .0095 44.0 -- 0.0012 .0048 .0069 3512 92 77.8 5 5 08 42.9 33ol LTD 4g-5 Method . Wo min. 3512 S&Si Sm qg,?'b 0.001 92.2 .0025 77.5 .005 56 .0075 -- .010 ---- 0.0009 -- .0024 7706 .0051 -- -- 43*2 -- 33.1 -- 0.0024 -- ,,0076 o0i0 * Used to prepare Standard Curve 7ABLE.V - Stability Studies Using 1-cm, Cells _5_minc 80.2 0.010 61.5 .020 35.7^- -P.040 27.5 -v.050 20.8 'r.66o -T5.1.Q ~__ IgO-gin.. T510 83 63.5 37.8 29.6 22.2 0.0065 - -- .0160 0345^38.0 0435V.7--- .0535 -- 0.0345 Tmr. _ _It: Y-fJJ noted that the system prepared according to LTD 42-5 la stable and that of BSM 44-188 Is unstable, Thialnstobillty decreases with age and decreases with an Increase In the concentration H e 00C5SS7 7 Effect of Reagent Concentration. In order to determine the effect of reagent concentration, a standard curve was prepared each time a new 30mgo/l. solution of fiithizone was prepared and these curves were compared with a standard curve for 20mg./l. solution The results of these tests are shown in Fig. 2. It will be noted that there is no significant variation for the JO-mg./l. solutions which were prepared from different batches of chloroform and from dithizone obtained from different sources. At lower con centrations of lead no error is evident due to a decrease of onethird in the concentration of reagent. There is a slight deviation at high concentration, near the useable limit of the method. This is in the expected direction. Effect of Age of Reagent. Dithizone is oxidized to diphenyl thiooarbodiasoneo This is a yellow-brown substance, soluble in chloroform and insoluble In dilute acid or alkali. It shows signi ficant absorption at 510 raji. For this reason the blank will change with the age of the reagent and must b determined daily. The necessity for such a determination is important in both color end mixed-color methods. A 50-rag./Idithi zone solution was allowed to stand at room temperature for several days. A portion of standard lead was analysed by the mixed-color method described In LTD 42-5 using this solution and by the method described in BSM 44-188 using another ,, portion of this solution diluted to 20 mg./I. with chloroform. 7 The j?esiilts of these analyses compared with a similar series made with fresh reagent ar given in Table VI from which it will b noted that low results ar likely to b obtained by either method if a decomposed reagent is used. TABLE VI - Effect of Age of Reagent . Method mg. ?b T 510 Old Reagent "standard Curve Hew Reagent LTD 42-5 DSM 44-188 0.025 .025 58.0 56.2 ,-54.0 . ' 49.5 53*5 49.1 Effect of Phosphate. It is an accepted fact that, in the absence of ammonium citrate, phosphate will interfere with the deter mination of- lead by the mixed color dithizon method. In our earlier studies, the presence of phosphate impurity in the potassium cyanide was not recognized. In Fig. 2, curves for several different dithizone relations are shown using the seme stock of potassium cyanide. These were prepared following the directions given in LTD 42-5. In the original procedure, the lead was separated from possible interfering substances in a first series of extractions. Then the lead dithi2onate K E 00G5S88 was decomposed by dilute nitric acid and tn:ts solution adjusted to pH 9.5, without th addition of citric acid sine the solution was practically free of ell possible interfering subiitancea. Potassium cyanide was, however, added to this solution. Because of the possible interference of the email amount of phosphate present in the cyanide, and since citrate prevented phosphate Interference in the first extraction, its addition to this solution was tried. We found that the addition of 5 ml of 5$ ammonium citrate solution would eliminate the interference of up to 250 mg. of phosphate. It is recommended therefore, that the method of LTD 42-5 should be changed to include the addition of 5 ml. of a 5% solution of ammonium citrate.to the nitric acid solution of the sample before the addition of solution B (described In LTD 42-31$ or 12.5 g. ammonium citrate may be added to 1 liter of solution B. Application ,,to .Determination of-T3L in A ir. DS& 45-59 recommended the elimination of the long heat dehydration procedure, and the substitution of sulfite reduction. Th final lead deter mination could be made in two ways 1. The Iodine solution Is heated to 100s Q* and this temperature Is maintained for one minute. After cooling, it Is added, to a sulfite solution which reduces the iodine. Lead Is determined by the procedure recommended in DSM 44-188. 2, The unheated iodine solution Is added to sulfite solution containing both cyanide and ammonia. This solution, with a pH of 11.4, Is extracted with a chloroform solution of dithlgone. A compound with n absorption maximum at 4-90 tspl remains In the chloroform*. A measurement of this color -permits a determination of the lead. This color is attributed to the dithlsone complex of an orgeno-lead Ion. It is believed that the lead complex la other than the trialkylload iodide, since synthetic triethyHoad Iodide Is Insoluble In the usual solvents (water, ammonia}, Is liquid above l8'=> C.* and does not react ilka the other trialfcyllead .halides. The application of the mixed-color method to the determination of TEL In iodine solutions was studied. It was found that the regular mixed color procedure could be applied if first the organo-lead " cosjplexes were decomposed by heating in the presence of an excess of iodine and if the excess iodine was removed with sulfite. However, the mixed color method was unsatisfactory for direct reaction of the c-rgano-lead Ions. The expected reaction took place liufcat least 50$ of the methods sensitivity was lost due to the coincidence of the absorption maxima of the dithlaone reagent and the organa-lead d l t M - aonate. K 00C5S89 CONCLUSIONS ^ The mixed-color method for determining small quantities of lead, presented in LTD 42-5* la both sensitive and accurate* It requires the use of three different solutions for the three different cell sizes in order to include very low concentrations of lead. How ever, the reagent concentration is not too critical end these solu tions may be prepared by dilution of the stronger solutions without exact measurement* Because of the interference of the small quantities of phosphate present in the potassium cyanide, it is recommended that ammonium citrate be included in the solution for the second extraction* Otherwise, a new standard curve must be prepared for aeh new stock of potassium cyanide* The one-color method recommended by the Baton Rouge Devel opment Section laboratory compares favorably with the above method in all respects except stability* If the time of color development is not controlled, erratic results are likely to follow* This effect is more noticeable at low concentrations. In the presence of thallium end tin th one-colop methods are more desirable than the mixed-color method* In our opinion, the B3M 44-188 method is an improvement over existing one-color methods. This is attributable to the presence of over 10G^> excess dithizone in all alkaline solutions used to remove the excess reagent from the chloroform layer. . REFERENCES 1* Clifford, P. A. and WIchmann, H. H., J. Assoc. Off. Agr* Chem. 19, 130 (1936) 2. Sfconger, V* A., Proceedings A3TM 44, 754 (1944) 3. Winter, 0. B. et al., Ind Eng. Chem. Anal. Ed. 7265(1935) . DISTRIBUTION Reference?. 1053-XP-21 109S-X?-1,2,3 0. Edgar R . Kehoe . 0 E. Kurt J Ho Schaefer Work by? Report bys B. Hockey M. Griffing C. Gambrill 111 0. ,M. Gambrill 0005330 Approved bys /<" O005392 H f w r r c L A S S t t H c o . , N . y N O . 3 G* 5 1 : ' A v } - * Il ' :,*.> i ;v}< * L - i m i f. I . A.