U. S. Department of Commerce Research Paper RP1871 National Bureau of Standards Volume 40, Aprill 1948 Part of the Journal of Research of the National Bureau of Standards

Determination of Sulfur in Bone Char 1 By Victor R. Deitz, Helen R. Higginson,2 and Cola Parker

In sugar technology t he established a na l.dieal procedures we re found to give erroneous results for the "sulfide" sulfur a nd " total" sulfur in bone char. The error was fou nd to be caused by the presence of organ ic sulfur in t he carbo naceous co mpo ne nt. Solution of t he total sulfur in bone char was obtain ed by three independent methods : (1) After oxida_ tion wi th nitr ic a nd perchlori c acids, (2) after a ir oxidation in an intimate mixtme ,,"ilh sodium carbonate a nd magnesium oxide, (3) a fter ox ida t io n with nitric acid in a modified Ca rius tube. The precipi tate of barium sulfate ,,·a_ found to be co nta minated wi th calcium a nd phosphat e from the bone char. It was di sco,·e red in t he present WOrk Lhat barium sul­ fate \\·as eas il y and co mpletely soluble in hot 70-perce nt perchloric acid, thus making po,,,ible a purificatio n b~ ' rep recipitaiion. Quantitative precipitat ion was verified by Ll sing known mixtures of calciu m chloride, phosphori c acid , and sulfuric acid. The distribution of the total sulfur between th e carbonaceous residue a nd t he filtrate resulting from the conven­ tional acid t reatment of bone char was determined after digest io n in t hree mi xtures: ni t ri c and hydrochlo ri c acids, potassium chlo rate and h.v drochlo ri c acid, and hydrochloric acid alone. It was shown that t he s ulfur co ntain ed in the ga se~ evolved wi th hydrochlo ri c acid (determined b~' combustion wit h a ir follo wed b.,· gra,·imetri c barium sulfate) was greater in most samples than t hat determiner! as s u l fid e - ~u l fur (precipitated as cad mium ulfide and determined iodometricaJl y). The di tribution of the total sulfur between filtrate, residue, and evolved gases was influenced by delib rate vari ations in t he aeidleachi ng proees)';.

I. Introduction then made in all aliquot part, of the filtrate. The re ult was calculated as percentage "calcium In sugar technology the established methods of sulfate" of the original sample of char. In order determining sulfur in bone char have been con­ to determine "calcium sulfid e", a sample of the cerned with the calcium sulfate and calcium sulfide. original char was first digested with either nitric The general procedure outlined by Spencer and acid or a mixture of potassium chlorate and hydro­ M eade [IP is substantially the same as that of the chloric acid. The subsequent procedure was the original methods outlined in 1881 by Tucker [2]. sam e as that given for the calcium sulfate. Sub­ Discordant results obtained by these methods tracting from the result of this determination, the made it desirable to make a critical study of the barium sulfate already found in the calcium sulfate procedures for determining sulfur in bone char. detennination gave the barium sulfate equivalent According to the old procedure, a sample of to the calcium sulfide. One obj ect of this paper bone char was heated with hydrochloric acid, is to show that these methods may give erro­ cooled, and diluted to a known volume. The neous results for the sulfide sulfur and the total residual carbonaceous matter was removed by sulfur in bone char. It will also be shown that filtration . A barium sulfate determination was thi s is due to a significant quantity of organic

, R" search Associate at the National Bureau of Standards, re presenting the 1 T his investigation has resulted rrom fl joint research project undertaken Gnitcd States Cane Sugar Refin ers and Bone C har Manufacturers and the by t he United Stales Cane Sugar Refin ers and Bone C har Manufacturers, greater part of the refi ning industry of the British Dominion. a greater part of the refi ning industry of t he British Commonwea lth , and 3 Figures in brackets indicate the li terature references at t he end of this t he National Bureau of Standard s. paper.

Determination of Sulfur in Bone Char 263 sulfur contained in the carbonaceous component ions contained in a solution of 5 g of a bono char. of bone char. In these investigations, the precipitation of An accurate determination of the sulfur in bone barium sulfate was conducted at an acidity char is important in a study of its revivification. between pH 1 and 2. It was found possible to In kilns using direct firing, the bone char is in con·· obtain an easily filterable precipitate and to avoid tact with the products of combustion of the fuel, excessive loss of barium sulfate by solu tion. This which may have several possible effects on' the value for acidity is in agreement with the value of char. 0.05 N hydrochloric acid recommended by Kolt­ The total sulfur content of service bone char hoff and Sandell [4]. Fortunately, the solutions may be of the order of magnitude of 1 percent, prepared from the bone char samples were readily when expressed as S04. The gross heterogeneity filterable at pH 1, although they were almost non­ of char and the difficulty of obtaining uniform filterable at pH 2. A glass electrode and pH samples are factors that limit the reproducibility meter, calibrated with a buffer of pH 3.0, were of the analytical results. For aliquot parts of a used in determining the pH of all solutions. given well-mixed powdered sample, it is possible Data for the weights of barium sulfate precipi­ to realize a precision adequate for this work.4 tated from known amounts of sulfuric acid in mixtures of calcium chloride and phosphoric acid II. Precipitation of BaS04 in the Presence are given in table 1. The precipitation was made

of Ca and PO 4 Ions by the slow addition of 50 ml of a 5-percent solu­ tion of barium chloride. The first four determi­ Many investigations with contaminatcd barium nations were with mixtures of the sulfuric acid in sulfate precipitates in analytical chemistry have calcium chloride only, and the rcsults indicate a shown the difficulties attending the exact deter­ mination of sulfur as barium sulfate. An inter­ T ABLE 1. Weights of barium sulfate derived from a single precipitation in a mix/w'e of sulfuric acid, calcium chloride, esting discussion of the possible compensation of and phosphoric acid, I1djus ted wtth hydrochloric acid to considcrable plus and minus errors has been given pH 1 by Hillebrand and Lundell [3]. Kolthoff and Sandell [4] have also reviewed the evidence for the W eight of p recipitate Oonditions of p recipitation contamination of barium sulfate precipitates by FOllnd I Theory IDifference co precipitation under various conditions. A great deal of difficulty was experienced in pre­ OALOIUM OHLORIDE AND SULFURIO ACID cipitating the sulfate quantitatively from known y g Percent solutions containing the sulfate, calcium, and 4 hr stirring at 30° 0 _____ :______0.3273 0.3239 +1.05 phosphate ions. For bone char the problem is Do ______. 3228 .3239 -0.33 Do______. 3263 .3239 +.74 that of determining about 1 percent of total Do ______.3278 .3239 +1.2() sulfur, expressed as S04 in the presence of about 40 percent of calcium ion and 50 percent of phosphate PHOSPHORIO ACID AND SULFURIO AOID ion. The extensive literature on the barium sulfate determination has unfortunately not dealt 4 hI' stirring at 30° 0 ______0.3346 0.3239 +3.30 Do ______.______. 3440 .3239 +6.21 with this situation. The precipitation of barium sulfate from known CALOI UM OHLORIDE, PHOSPHORIC AND SULFURIC AOIDS solutions of calcium chloride, phosphoric acid, and sulfuric acid was first investigated in some detail. 4 hI' stirring at 50° 0, stood overnigbt______0.3044 0.2790 +9. 10 The weights of calcium chloride and phosphoric Do______.3081 .2790 +10.43 6 hI' stirring at 90° 0, stood overnight______. 3375 . 3244 +4.0 acid in each solution were maintained constant, Do______.3387 .3244 +4.4 and the values were chosen so as to give 1.75 g of 4 hr stirring at 50° 0, stood overnight______' .2961 . 2790 + 6. 13 Do______'.3009 . 2790 +7.85 calcium ion and 2.75 g of phosphate ion in each Do______'.2979 .2790 +6.77 test solution. These weights correspond approxi­ Do ______"____ '.3010 . 2790 +7.88 Do______'.3056 .2790 +9.53 mately to the weight of calcium and phosphate

'A coefficient of variability of 1.65% as a m easure of t he precision was cal· • Saturated barium h ydroxide used instead of 5-percen t solution of barium culated from t he data o[ table 5. chloride.

264 Journal of Research ------

percentage difference of about 1 percent. With TABLE 2. W eights of bW'ium sulfate after do uble sulfuric acid in solu tions of phosphoric acid, the precipitation indicated contamination was greater. With mix­ W eight or precipitate tures of calcium chlroride, phosphoric ·acid, and pH sulfuric acid t he positive error has been found to Found Theoretical Difference be as much as 10 percent,. Qualitative examina­ ------(j U Percent tion showed that the precipitates from tho phos­ 2. 0 0. 1336 0. 1298 + 2. 93 phate mixtures wore contaminated with phosphate 2. 0 . 1306 . 1298 + 0.62 1.0 . 1283 . 1298 - 1.16 LOll. 1.0 . 1348 . 1298 +3.85 In the co urse of the experiments to be described 1.0 . 2907 . 2790 + 4. 19 on the total oxidation of the carbonaceous matter 1.0 . 291 2 .2790 + 4.37 in bone char, it was discovered that the barium 1.0 . 2923 .2790 + 4. 77 1.0 . 2897 .2790 +3.84 sulfate precipitate was freely soluble in hot per­ 2.0 .3286 .3244 + 1.29 chloric acid. 5 6 It was thus possible to purify the 2. 0 . 3328 . 3 2~4 + 2. 59 precipitate by a reprecipitation from perchloric 2.0 . 3360 . 3244 +3. 58 acid. The only other solvent for barium sulfate 2. 0 .3333 . 3244 + 2.74 1.0 .3384 .3244 + 4. 31 to which reference has been found is hot concen­ 1.0 .3376 .3244 + 4. 06 trated suUmic acid. Its use here is obviously 1.0 .3352 . 3244 +3. 33 no t feasible. 2. 0 . G478 . 64 88 - 0. 15 The procedure for reprocipitating the barium 2. 0 .6506 . 64 88 +. 28 sulfate was as follows: The first precipitate was 2. 0 . 9873 . 9732 + 1.44 2. 0 1. 0000 .9732 + 2. 75 obtained from a known volume of th e standard solution of sulfuric acid in solutions of calcium chlorid e and phosphoric acid by the slow addi­ crucible. It was ignited at 600 ° C for at least tion in drops of 50 ml of a 5-percent solution of 2 hours. barium chloride to the stirred mixture at 50 ° C. The pH was adjusted with ammonium hydroxide The results for tho dOll ble precipitation arc with the aid of a glass electrode to either pH 1 or shown in table 2. The quantity of standard sul­ 2, as designated in table 2. The precipitate was furic acid was varied over a wide range. An separated by filtration through analytical-grade analysis of the results in table 2 was made, dis­ filter paper and washed with 200 ml of distilled regarding the two values 1'01' pH. A measure of water. Both precipitate and paper were digested precision and also of accuracy was obtained and with concentrated nitric acid until the paper was found to be very satisfactory. A very slight decomposed. P erchloric acid was then added, and contamination of the barium sulfate precipitate the mixture boiled in a covered beaker until a was found equal to about 1 percent.7 clear colorless solution was obtained. Distilled The better agreement realized on double pre­ water was added slowly to the warm solution until cipitation could be attributed to either a purging the volume was approximately 400 ml ; 5 ml of a of the barium sulfate of the contaminants or to 5-percent solution of barium chloride was added, the closer compensation of the minus errors (for and the stirred mixture wa adjusted to pH 1 example, the possible loss of barium sulfate ill with ammonium hydroxide. The precipitate was the acid solution because of the additional analyti- allowed to stand overnight at room temperature 7 The authors are indebted to J. Mandel and ,V. V . LoebensLein for the and then filtered through a porcelain filtering following analysis of the data in table 2. 'rhey have plotted the weigh t of the precipitate found against the corresponding weigbt as calculated from ,j'rhroughout this paper, whenever acids arc mentioned, pcrchloric acid Lhe vo lume and tbe normality of tbe standard sulfuric acid. 'rhe plot of 19 will mean the co ncentrated acid or commercc 70 Lo 72 pcrce nL ; hydrochloric points appeared to be sufficiently lincar to warrant a fit by the method of acid , the concentrated acid of speci fi c grav ity 1.1 8; nitric ac id, the CO ncen­ least squares. The intercept of the resulting line was not significantly differ­ trated acid of specific gravity 1.42. DiluLcd acids will be designatcd as ent fro m zero as indicated by a "Student's" f, va]ue or 1.64. This is smaller follows: For cxample, d iluted bydrochlorie acid (1+ 20) will mean 1 volume thau the 2.1l value, corresponding to the 5·percent significance level for 17 of the concentrated acid or specific gro.v ity U 8 diluted wi Lh 20 volumes or degrees ofrreedom (R. A. Fisher, StatisLical Methodsror Research Workers). water. Ammonium hydroxide will m ean thc co ncentrated so luLion of The slope of t be fitted line was round to be only sligbtly diil'erent from I , specific gravity 0.90. since the t value in this case was 2.14 compared to 2.11 . This might indicate 6 Because of tbe possible explosive nat ure or pcrehlorie acid mixtures, care a con Lamination in the second precipitation of the order or 1 perce nt of the mus t be taken that cxcess niLric acid be always presenL before adding PCI" theoretical value. The standard deviation from regression is 0.0062 when chloric acid. expressed in grams of precipitate.

Deter:mination of Sulfur in Bone Char 265 cal step) with t he plus errors (the gain due to the diluted mixture IritIt ammonium hydroxide at contamination). In order to test the latter pH 1 by the addition of sufficient barium chloride possibility, an additional precipitation of the (see table 4). The addition of saturated barium barium sulfate " 'as made. The procedure for the hydroxide solution to t he diluted mixture (instead second reprecipitation was exactly the same as of ammonium hydroxide and barium chloride) foJ' the first. The results are shown in table 3 ,vas tried in order to raise the pH and to depres!" where a comparison is made with those Jor a the solubility hy the common ion eff ect. Un single and for a double precipitation. The results fortunately, a considerable amount of barium of a spectro-chemical analysis of the second and hydroxide was l'eq uired to decrease the acidity third precipitates did not show any appreciabll' to pH 1 and the corresponding amount of barium difference in the presence of calcium or any other ion introducE'd (50 times that contained in 5 ml possible contaminants. This would indicate that of a 5-percent solution of barium chloride) wa, the observed smaller plus errol' obtained in the found to yield a large positive error (see table 4). third precipitation might be attributed to a <'loser It would be of interest to ascertain how much compensation of the plus and minus errors. barium hydroxide should be added to the diluted mixture in order to compensate for the negative TABLE 3. T )'i ple precipitation oJ the ba1"i1l1IL sulfate Jr om errol' caused by the solubility. The addition of known sol1ttions oJ s1i1.fw"ic acid conta1'ning calcill II! the barium ion would depress the solubility by the chloride and phosphon-c acid common ion effect alone. Another mode of

" -eight of prpc ipitatf' attack would be to raise tile pH by addition of the Me·thod base of a large organic group and to furnish the Found 'rheol'etical "Difference barium ion by the bm'j urn salt of an organic acid ------·---11----

y (/ Percent, (for example, barium salicylate). The use of' T hree pl'Pcipitations ______0.2845 O. 2790 + 1.97 large ions that do no t copl'ccipitate has recently .2798 . 2790 + 0.29 ~ ~ :::::::::::::::::::::::::::::::I . 2837 . 2790 + 1.68 been used for the purpose of preparing a crystallo . graphically pure barium sulfate [8]. The preci­ sion so far indicated by the work described appears The best procedure for the repreclpltation of to be sufficient for the present purposes. barium sulfate from its solution in hot perchloric

acid has yet to be cstabli~hed. It is, of course, T A B LE 4. Effect of some v a l'i at ·i on.~ in the conditions em· most desirable to a void the addition of any reagent ployed to TepJ'eci pilate bW"iwn s1dfale .from soluhons in that will J'econtltminate the harium sulfate. A hot concentmted pe/'chloric acid few determinations \\l'I'(' madc in which the pre­ [All first precipitates made in same malmer] cipitate was dissolved in 15 ml of perchloric acid and reprecipitated by dilution with 600 ml of Weigh t of precip itate distilled water only (pH of the resulting mixture, M ethod pH Found TheOryi Differ· 0.3). The re'i ults obtained by this proccdure are . ence low, as may be seen in table 4. Since barium ------g g Percent sulfate was known to be soluble in strong acids, a Dilution w ith watcr only ______0.3 0. 2732 O. 2790 - 2. 08 few determinations wcre made in which the di­ Do _____ ~ ___ ~ ~ __ ~ _~ ______~. _____ ~ ___ ~ __ . 3.2688 _2790 - 3.66 luted mixture from the perchloric acid solu tion Dilution +NH , OH ~ _____ ~ __ ~ _~ ____ ~ ~~ _~ _~_~ l.0 .0904 1298 - 30.4 was adjusted to pH 1 with ammonium hydroxide. Do ~_~_~~ ~ _~_ ~ ___ ~~ ~_ ~ ___ ~~ __ ~ ______~_~ 1.0 . 1017 . 1298 -21. G

The very low results so obtained indicate the Dilutioll+NB,OH+l ml BaC" _____ ~.~ ____ LO . 3069 .3244 - 5.39 relatively strong effect of ammonium salts. This Do ____ ~ _~ ____ ~ ~ ______. ______1.0 . 3076 .3244 - 5. 18 lllay have affccted the solubility of the barium Dilution +NR ,OH+ lO ml BaCJ, ______l.0 . 3384 . 3244 +4.31 sulfate or the ammonium ion may have replaced D o ______. _. ______. ______._ 1.0 .3376 .3244 + 4.06 Do ______. ______. ____ some barium in the prec ipitate. Thc res ults 1.0 . 3352 .3244 +3.33 were also low when only 1 ml of a 5-percent solu­ D ilut ion+saturated Ba(OE), (400 ml) _____ 1.0 .2990 . 2790 + 7. 17 D o_... __ . __ . ______. __ . _____ 1.0 . 2994 .2790 +7.31 tion of barium chloride was added to the mixture Do ______. ____ ... _____ . _____ . ______1.0 . 2980 .2790 +6.81 containing ammonium hydroxide at pH 1. Ho,,'­ Do ______.. ______•. _.. ... ___ 1. 0 . 2925 . 2790 + 4.84 ever, it was possible to obtain a positive error from

266 Journal of Research III. Total Sulfur Content of Bone Char 1.2. The sample '\"as \\"etted wi t h abou t 10 to 15 ml of distilled water, and 25 ml of nitric acid was add cd " This In order to obtain the complete solution of a mixture was boiled gently for 1Yz hours in a co '"crcd bcaker. sample of bone char, it is nocessary to decompose Fifteen m illiliters of perchloric acid (see f oot ll otc~ 5 a nd 6) the insoluble carbonfLceoll s component under con­ was added and the beaker heated until all fumes of 11 itric acid had disappeared and the carbonaceous residue \ms ditions that insure no loss of the sulfur. Some of completely oxidized . During t his process the mixture the total su lfu l' is contained as sulfate-sulfur, some changed from t he original black to orange, then to bril­ as sulfide-sulfur, and some in combination with liant yellow , and finally to a clear solution \\"ith a slightly the carbonaceous component. Three different pinkish tinge. methods have been employed in this inves tigation 1.3. The clear solution was cooled and diluted with about 100 III 1 of distilled water. The small residue to prepare a complete solu tion of the sulfm free of (insoluble sili ca) was removed by filtration through a organic matter: (1) complete oxidation with nitric Whatman No. 40 fil ter paper or its eq uivalent. The and perchloric acids, (2) the Eschka method, i. e., residue was ,,"ashed thoroughly with 150 t o 200 ml of air oxidfLtion in a muffle furnace of the int.imfLte distilled water, and all fil t rates were collected in a 1,500- mLxture with sodium carbonate and magnesium 1111 beaker. 1.4. Water was added until the vol u me of filtrate oxide, (3) nitric fLcid treatment in a modified was 750 to 800 m1. The pH of t his solution waf' adjusted Carius bomb. to 1.0 by adding a mmoni um hydroxide drop,,"ise during mechanical stirring (add hydrochloric acid if the pH is 1. Complete O xidation With Nitric and Perchloric greater t han 1.0 at first). A glass elect.rode and pH Acids meter were used in determining t he pH. 1.5. The solution was heated to a temperature of The destruction of orgfLnic matter by heating 50° C, and 50 ml of a 5-percent solution of barium chloride with concentrated nitric and perchloric acids has was added drop \Y ise, the solution being stirred mechani­ been incorporated in a number of analytical cally during the entirc process. The temperature was maintained at 50° C for 4 hours. schemes. A 72.4-percent perchloric acid solution 1.6. The solu t ion ,,"as then filtered through a \rhatman in water boils at 203° C (1 atm). This is approxi­ No. 42 fi lter paper or its eq uivalent and t he precipitate mately 100 deg lower than the temperature washed with di sti lled water untii on ly a faint test for reached in a boiling sulfuric acid solution at atmos­ chloride was obtai ned wi th silver ni t rate. pheric pressure. Tn these experiments, as well as 1.7. The fil ter paper conta.ining t he barium sulfate precipitate was transferred to an 800-ml beaker, and as many others, it has been shown that the oxida­ 25 ml of nitric acid was added. It was hcated gently tion with a mixture of nitirc and perchloric acids u ntil the paper had decomposed ; 15 ml of perchloric acid can be performed without fear of dangerous explo- was added, and t he solution was boiled until the ni tric r sions. The only precaution needed is first to heat acid had been eliminated and the residue had dissol ved. the mixture at about 100°C until subsidence The covered beaker at this point became filled wi th whi te f umes, and the con tant boiling mixture of perchloric indicates the destruction of the more easily decom­ acid and ,,"ater condensed on the \\·all of t he beaker. posable portion of the organic matter. This state of reflu x was continued for at least 15 minutes. The detailed procedure for the analysis of a bone The solution was perfectly clear and colorl ess or sli ghtly char for total sulfur by this method follows. A yellow. screen test of samples for analysis indicated that 1.8. The solution wa cooled to room temperature and 400 ml of disti ll ed water was added; 5 ml of a 5-percent 90 percent passed tlu-ough a U. S. Standard Sieve barium chloride solution was added, and the pH readjusted 80 and that 60 percent passed through a U. S. to 1.0. It was a ll o\\"ed to stand overnight. The mi xture Standard Sieve BO. The procedure is divided was filtered through a porcelain filtering crucible, \\"ashed into the numbered paragraphs for convenience in with 100 ml of distilled water, and ignited for about cross reference when the other methods are de­ 2 hours at 600° C, t hen cooled in a desiccator and weighed. The precipitate was weighed as barium sulfate, and t he scribed. The following procedure is designated as result was expressed as percentage of S04. method 1. 2. Eschka Method Applied to Bone Char 1.1. Five grams of th e powdered bone char was placed in a weighing bottle and " "as dried bet\yeen 105° to noo c The Eschka method is commonly used in the in an air oven for a t least 4 hours. The bottle containing determinat.ion of sulfur in "coal. An intimate the sample was tightly stoppered upon removal from the oven and cooled in a desiccator. The stoppered bottle mixture of the sample with s0dium carbonate was weighed, the contents then t ransferred to an 800-m) and magnesium oxide is heated to convo-t all the beaker, and the bottle 'yeigh ed again. sulfur in to water-soluble sulfate. The method

Determination of Sulfur in Bone Char 267 can be readily applied to the determination of The following analytical procedure is designated sulfur in bone char. There is little danger of as method 3: losing oxides of sulfur, since the quantity of sulfur 3.1. A 5-g sample of powdered bone char was taken. expressed as S04 contained in the organic matter However, the sample in t his case was transferred to a to be oxidized is usually less than 1 percent of Carius t ube [6] . The tube was cooled in liquid air, and 5 the weight of bone char. ml of fuming nitric acid was added. This was done to A modification of the Eschlm method described freeze the acid in order to delay any reaction. The tube was sealed while it was still immersed in t he cooling by Treadwell and Hall [5] was used. The follow­ mixture. The tube was placed in t he steel shell and sur­ ing procedure is designated as method 2. rounded with 50 g of dry ice and t he top screwed on 2.1. A 5-g sample of powdered bone char ,vas taken. securely. The steel shell was tested for leaks by immers­ However, t he sample in t his case was transferred t o a ing it in water to note the escape of gas bubbles. It was 200-ml evaporating dish instead of an 800-ml beaker. then placed in a muffle furnace at 215 0 C, slightly inclined 2.2. The sample was mixed thoroughly in the dish so that the seal of t he glass tube was elevated. The tube with 10 g of magnesi um oxide and 5 g of sodium carbonate. was left overnight in t he furnace. A second mixture containing 6 g of magnesium oxide 3.2. The steel shell was cooled to room temperature and 3 g of sodium carbonate was sprinkled over t he top (about 2 hours) and t he carbon dioxide cautiously released of the mixture in t he dish. The mixture with t he sample by unscrewing t he cap. The Carius t ube was taken out should be completely covered. The dish was placed in and the contents fro zen b y placin g it in li quid air.B It an electric muffle furnace at room temperature, heated was opened by heating with a tiny gas flame from a blow torch a t the very end of t he seal. The material inside the slowly to 600 0 C, and left overnigh t. 2.3. The sample ,vas removed from t he furnace and al­ Carius tube usually had a deep blue color t hat di sappeared lowed to cool. The mixture was transferred to a 1 500-ml soon after the tube was opened. The residue in the t ube beaker with distilled water, stirred mechanically, ~nd 100 was t ransferred to an 800-ml beaker with a stream of ml of saturated bromine water added. The mixture was hydrochloric acid. The analysis was continued exactly as heated to boiling on a hot plate while co ntinuously stir­ in paragraphs 1.3 through 1.8. ring; 50 to 80 ml of diluted hydrochloric acid (1 + 1) was added unt il the solut ion was acid to litmus, and t he hot 4. Percentage of Total Sulfur mixture was stirred for about an hour. Co nce ntrated The method of determination for a given service hydrochloric acid was then added dropwise until the solids bone char (Y - 1) was further tested by makin g a dissolved and t he solution became clear yell ow wi th only a small insoluble residue of silica remaining. The amount of number of determinations with the addition of hydrochloric acid needed to make the solu tion clear varied known quantities of standard sulfuric acid to the from 100 to 200 ml. After t he final addi tion of h ydro­ bone char. The determinations were made ac­ chloric acid, the solution was kept on the hot plate an cording to method I in which the total sulfur was additional hour while continuously stirring. oxidized to sulfate-sulfur with a mixture of nitric 2.4. The solution was allowed to cool and the silica resi­ due filtered according to the procedure used in paragraph and percl1loric acids. The results are summarized l.3. The analysis of t he fil trate was performed exactly in table 5. The amoun t of standard sulfuric acid as in paragraphs 1.4 through 1.8 of t he perchloric-acid was varied from 5 to 25 ml (0.1112N). With the method (method 1). smallest addition of sulfuric acid, approximately 3. Complete Oxidation in a Modified Carius Tube 20 percent of the barium sulfate came from the acid and 80 percent from the char. With the A Carius combustion tube as modified by Gor­ largest addition of sulfuric acid, approximately 60 don [6] and the technique described by Gordon, percent came from the acid and 40 percent from Schlecht, and Wichers [7] for the use of sealed the char. The average value of the total sulfur tubes in preparing acid solutions of samples for for the eight determinations was 1.91 percent analysis were employed in these al).alyses. The expressed as S04' There is no apparent correla­ sealed tube was placed in a protective shell of tion of the individual r esults with the total amount cold-rolled steel, made by drilling a I-in. hole of barium sulfate involved in the determination. into 1 4-in. length of n~-in . hexagonal stock and In view of the consideration already given to the surrounded with solid carbon dioxide to provide a reprecipitation of the barium sulfate (derived from compensating pressure that served to reduce the known solutions of sulfuric acid , calcium chloride probability of bursting the tube. It was found , and phosphoric acid) after solution in perchloric possible to oxidize the carbonaceous matter in bone char completely with fuming nitric acid and 8 As a safety precaution, it is recommended that the steel shell be cooJed in dry ice and the transfer of the Carious tube to the li quid air be made behind to reuse the tubes for several analyses. a window of safety glass.

268 Journal of Research

~-- TABLE 5. Determinations by the nitric-perchloric acid T ABLE 6. Single, double, or triple precipitation of the method of the total S11/.fW· in samples of a service bone char barium sulfate in the determination oj total sulfur in bone (Y- l ) with additions of standard sulfw'i ~ acid char (ex pressed as S04)

lAll determinations wore made with a service bone char, Y- l] Weights of barium sulfate Total mlfur Volume of standard 'Voight in cbar H,SO, (O.11l2N) From the From the of char ex pressed Single Double Triple Total H,SO,- char as SO, precipitation [ precipitation precipitation -----_.. _------I ml 0 0 9 9 Percent COMPLETE OXIDATION 0 1' THE SAMPLE WITH N ITRIC AND PERCllLORIC ACIDS 5._ ... _._. ___ ... _.. . _.. 0. 2901 0. 0649 0.2252 4.8465 l. 91 10. __ . ______. ____ .. _ .3645 . 1298 .2347 4. 9098 1. 96 20 ... ______._. _. _. _.. . 4883 .2596 . 2287 4.8937 1. 92 P ercent Percent Percent • 2. 01 I. 89 1. 89 25 ___ . ______. __ . _____ .5545 .3244 .2301 4.8797 1. 94 • 1.97 1.90 1.90 5 ______. _. ______. 2920 . 0649 .2271 4.9799 1. 88 1. 97 l. 91 1.88 10 ______. 4447 . 1298 .3149 6.9572 1. 86 1. 96 1.92 ----

20 ______. ______. 5169 . 2596 . 2573 5.5377 I. 91 ESCHKA'S METHOD APPLI ED TO nONE CHAR 25 ______.5497 .3244 .2253 4.8260 1. 92

Avg_.______I. 91 1. 98 1. 92 - --- l.96 1. 92 -- --

- Weights calcul ated from t he uormality and Quantity of the stand ard Avg_. l. 97 1. 91 1. 89 sulfuric acid in co lumn "1 . I - Standard sulfuric acid added, and the co rres ponding amount of barium acid, it is important to study the effect of multiple sulfate subtracted from the q uantity detenuined experimentall y. precipitation employing bone char alone. Accord­ ingly, several determinations were made of the T AB LE 7. Total sulfUJ' (expresse d as S04) in six sel'vice bone chaTS by three methods total sulfate after a nitric-perchloric acid oxidation of the char sample in which the barium sulfate was [Values in itali c type arc average values] reprecipitated twice. The results are contained in Av(~ l' age method 1 metbod 2 method 3 of metb­ table 6, in which a comparison is made ''lith the Cbar ods 1, 2, results from a single and doubJe precipitation. a ncl 3 ------_.------Included, also, are some determinations in which Percent Percent Percent Percent Y - 1______. ______. ______the initial oxidation of the bone char was effected 1.91 1.92 1.96 1. 93 36. ______1.98 J. 99 2. 04 2.00 by the Eschka method already described. It is P- l . ______. __ . __ . ___ . ___ 1. 01 0.95 1. ()() 0.99 apparent from table 6, and can be shown by P - 2 ______0.80 .81 0. 85 . 82 R - 22 ___ . ___ . ______. ____ . __ . 96 .95 1. ()() .97 statistical consideriitions, that a signifi cant dif­ R - 23 ______. 97 . 94 1.04 .98 ference exists between the average values for single and double precipitations, namely, l.97 and l.91 percent, respectively. The average value for rather than to loss of sulfur during the oxidation the determinations employing triple precipitation step. is l.89 percent, which can be shown not to be Further discussion of these results will be post­ significantly smaller than l.91 percent. It appears, poned until presentation of the analyses according therefore, that while a single reprecipitation of the to methods 4, 5, and 6. barium sulfate precipitate is justified, there is not mueh to be gained by another similar operation. IV. Sulfur in fhe Residue and in the Table 7 contains the results for the determina­ Filtrate From Acid Leaching tion of total sulfur (expressed as 804) in six dif­ When bone char is digested with nitric acid, it ferent service bone chars. The sulfur in each char is well known that the carbonaceous residue takes was completely oxidized to sulfate-sulfur by the on a different appearance than when hydrochloric three m ethods already outlined. The results by acid alone is used. The treatment with nitric acid initial oxidation with nitric acid in a Carius tube leaves a definite brownish residue that has been (method 3) appear to be higher than the results attributed to organic oxidation reactions. It is by methods 1 and 2. This may be due to un­ also known that the inorganic sulfides or sulfites known factors inherent in the methods of analysis are completely oxidized to sulfates. It is possible

Determination of Sulfur in Bone Char 269 77S365- 48--2 also to oxidize sulfides and sulfites to sulfates with 4.6. The percentage total sulfate in t he sample ,ras a dilute mixture of potassium chlorate and hydro­ calculated from t he sum of t he ,,,eights of barium sulfate chloric !l cid . This does not appear to alter the obtained in 4.4 and 4.5. The volume of t he r esidue was estimRted from t he weight of the residue and a vallie of carbona,ceous matter to the extent that the nitric 1.8 for the density. The weight of t he residue was deter­ acid treatment does. Dilute hydrochloric acid mined by t he acid \\'ashing of independent samples. alone niay be expected to oxidize the carbona­ Subt racting t he small volume of the residue from 250 ml ceous matter to the least extent. The extent to yielded the total volume of the fil t rate. The total sulfur which the above three acid mixtures oxidize the in t he filtrate, expressed as S0 4, was t he value obtained for 200 ml (paragraph 4.4) multiplied by t he rat io of t he total carbonaceous residue in bone char should deter­ volume of fi ltrate to 200 ml , approximately 1.25. The mine the amount of sulfur that is extracted from total sulfur expressed as S04 in t he residue was calculated the residue, providing, of course, there is appreci­ by subtracting t he value for the fi ltrate from t he value for able sulfur in the residue. Thus, the sulfate­ the total sulfate. sulfur contained in the filtrate need not be derived 2. Digestion With Potassium Chlorate and Hydro­ solely from the inorganic salts in bone char. chloric Acid Three ind ependent analytical procedures will now be described in which the sulfur may be The following procedme is designated as method determined in both the r esidue and the fil trate 5 for the determination of sulfur in bone char: obtained in acid leachings of bone cha,]'. 5.1. A 10-g sample of fin ely powdered bone char ,,-as taken. 1. Digestion With Nitric Acid and 5.2. The sample contai ned in the 800- 1111 beaker ",as Hydrochloric Acid wetted ,,-it h 25 ml of dbtilled ,,-ater and 50 1111 of a solu­ tion of potassium chl oratp (1 0 g of potassium chlorate in The follo wing procedure will be designated as a li ter of water) \ras slo\\-Iy added , The solution " -as method 4 for the determination of sulfur in bone heated to about 80° to 90° C, and 100 ml of di luted hydro­ char: chloric acid (I + 1) was added nropwise from a buret, whil e stirring mechanica ll ,\', The acid mU iit not be added 4.1. A 10-g sample of fin ely powdered bone char ""as too rapidly lest sulfur e~cape as hydrogen sulfid e. After taken . t he addition of all t he acid, the mix tl:rc \I-as boiled gently 4.2. The sample contained in t he 800- 1111 beaker was in the covered beaker for 1 Y:! hours. wetted ,,-it h 25 ml of distil led water ; 25 ml of concentrated 5.3. The analysis ,, - a~ concluded according to para­ nitric acirl was added dropwise a nd t he mixtllre allowed to graphs 4,3 t hrough 4.6, evaporate slowly to dryness on t ile hot plate. It. was necessary to r emove t he mixture from t he hot plate imme­ 3. Digestion With Hydrochloric Acid diately after the evaporation was completed so t hat no carbonaceous matter would be lost by ox idation. Three The fo llowing procedure is designated as method a,dditional e\'aporations to dryness ,yere then made using 6 for the determination of sulfur in bone char: 15 ml of diluted hydrochloric Rcid (1 + 1) each time. This 6.1. A 10-g ~a mpl e of finely powdered bone char \\'as was done in order to expel t he nitric acid completely. taken. After t he t,hird evaporation , 40 ml of diluted hydrochloric 6.2. '1'lw sample contai ned in t he 800-1111 beaker ,,,as acid (l + 1) was adden and the mixt llr(, boiled gently for wetted ",it h 25 ml of di~till(>d \I'ater and 100 ml of diluted about 1 hour. hydrochloric acid (1+ I) slo,,-Iy adden , The solu tion 4.3. The mixturp was cooled and trall5ferren to a vol1l ­ ",as heated in a covered beaker for about l )of hours at 80° metric fla~k, where it was ma de up to exactly 250 ml. to 90° C. The mixture \I-as filtered t hrough a Ko. 40 W hat man fil ter 6.3. The a n u l~- s i s ,,,u, concl1lded according to para­ paper or its equivale nt and exactl y 200 ml collpcted in a graphs 4,3 t hrough 4_6. H owever. wh en iieparating the second yolu metric flask. The remaining solution, the residue from the 200 ml of fil t r3te, Ko. 42 " -hat man fil ter residue and filt er paper, a nd t he funnei ,ya~ hing s " 'ere paper, or its eq ui valent was lI sed. t.ransferred to an 800-ml beaker. 4.4, The fil trate contained in the 200-ml volumetric 4. Distribution of the Sulfur Between Residue a nd flask \I'as transferred to a 1,500-1111 beaker a nd t hi s part of Filtrate the analysis concluded according to paragraphs 1.4 t hrough 1.8. The results of a number of analyses by meth­ 4.5. T \renty-five mi llil iters of nitric acid was added ods 4, 5, and 6 haye sho"'n good checks for the to the beaker contai ning t he r esi due, t he paper, and re­ sum of the sulfur contained in the residue and maining fi ltrate and heated until t he filter paper had dis­ integraten; 15 1111 of perchlori c acid was added, and the that contained in the fil trate. T he data for the six procedure outlined in paragraphs 1.2 t hrough 1.8 was then service bone chars already mentioned are giyen follo\l·ed. in table 8. The values for total sulfur contained

270 Journal of Research TABL~~ 8. Distribution of the total sulfur (ex pnssed as SO,) between th e rei idues and filtrates derived from the acid­ leaching of sel'vice bone chars

36 ______2.00 1.77 . 23 2.00 1.53 . 48 2.01 1.41 .35 . 13 1. 89

P- I __ - __ - --- _____ - ______- __ - __ _ 0. 99 0.94 .04 0.98 0.98 .05 1. 03 0.78 . 16 .07 I. 01

P-2 ______. Cm . 15 .81 .52 . 30 . 05 0.87 .82 r .72 . 12 0.84 1 . 73 . 09 .82 } . 54 . 25 .05 .84 R- 22 ______.88 . 10 .98 .97 . 76 . 23 . 99 . 48 . 40 . 08 .96 { . 86 . 06 . 92 }

R -23 ______.57 .29 .10 . 96 .98 .89 . 08 .97 .83 . 11 . 94 I .54 .33 . 10 .97

s Drl.f' J'llIlnl'Cl hy co rnhllsl ion or ('\,01\"(..'(1 l!:1SCS in ai r. in column 2 a rr thosr givrn in tabl e 7, and thcsc V. Analysis of Gases Evolved With Hel arc inserted for compari so n. The use of ni t ri c The prese nce of hydroge n sulfide is readily ap­ and hydl'Ochlori c acids (mrthod 4) res ul ted in the j)firrnt in the gases evolvcd from bone char by the least value for the sulfur in the rcsidu e. It is of f1.ction of hydrochloric acid . Tile hydrogen sul­ interest to note that the sulfur con tai ncd cithcr hell' has bcrn a,ttributcd to thc decomposition of in t~e res idue or in the filtrate was found to vary in organic sulfides in thc char. Some may possibly conslde]'ably for individual detrl'minations. ,\Vhen r esult from an acid hydrolysis of an organic-sulfur ~h e sulfur was lowe)' in thc residue, it wa higher componcllt in the cal'bona crous r csidu e. If the JLl thc fU trate; however, as mrntioneci apovc, thc latter possibility exists, there ma.y br other volatile sum of that in the rcsidu e a nd that in the filtrate sulfu ['-co n taining gases in tho c evolved with was CO D tant_ hydrochloric acid. Without exception the sulfur contain ed in the With this possibility in mind, th e gases derived residue was greatest when the char was extracted from bone char by evolution with hydrochloric with hydrochloric acid (method 6). There is acid wc re analyzed by two methods. F irst, the evidence whi ch points to some oxidation of the ltyd rogen sulfide was determined alonc; second, sulfur contained in the carbonaceous matter cven the e." 91ved gases were burned in ail' and a gravi­ when this acid alone is uscd. Samples of the metrJC determination fo r barium sulfate made in same char (P--2) were digested in dilu ted hydro­ the products of the reaction. chloric acid (1 + 1) for different periods of ·time.

Three samples after dio'b estion for 1 !1 ~ 110m's.. bO 'ave 1. Iodometric Determination of the Hydrogen residues in which the resul t for sulfur (as S0 4) Sulfide \\'ere 0.307, 0.262, and 0.246; two valu es after The sulfid e sulfm was scparatrd from the char 8 llOUl'S were 0.234, and 0.222; one value after 50 by cyolution with hydrochloric acid in a closed h01l1's was 0.226. These results indicate that for apparatus and by washing the gases in ammoniacal contact periods up to approximately 8 hours the cadmium sulfate soIu tion.9 A schematic appa­ hot hydrochloric acid is gradually' lcaching the ra tus is illustrated in figu re 1. The cadmium sulfur from th e carbonaceo lls residue. B eca ~ se of sulfide wa s determined without removal from the the loss of volatile sulfm compounds (principaHy gas wa sh bottle by it subsequent t itration with hydrogen sulfidr) during tIle hydrochloric-acid stancLanl iodine in the acidified solution. leaching, the slim of the sulfur co ntained in th e residue aDcL in the filtrate does no t equal the total 9 T~ C stock solution of ammoniaca l ca dmium sulfate contained 168 g of cadl:ll~lm sulfate. 960 ml of concentrated ammonium hydroxide, Bod 1,046 ml sulfur. of d IstIlled water.

Determination of Sulfur in Bone Char 271 nation will be given after the presen ta tion of the second method fo r determining sulfur in the evolved gases. A number of measurements of the hydrogen sulfide liberated from steel (N a­ tional Bureau of Standards Samples 9d and 8g) were made. The average of 10 determinations fo r samples of 9d was 0.036 percent of sulfur as compared with the standard value of 0.036 o percent of sulfur. The average of five determina­ tions of 8 g was 0.024 percent of sulfur as com­ pared with the standard value of 0.02 6 percent of sulfur. The results of table 9 indicate a possible oxi­ dation of the sulfur gases with the carrier gas used to sweep out the apparatus. It appears

FIGURE 1. Apparatus for the evolution of hydrogen sulfide. that the resul ts are higher when the carrier gas A, P ressure regulator for carrier gas; B, e volution unit w ith E rl enmeyer contains oxygen. However, the differences may flask cquipped witb in tercbangeable ground joint; C and D , gas-wasb bottles witb fritted disks cbarged with solu tion of ammoniacal cadmium sulfate. also be due to variable action of the hydrochloric acid on the carbonaceous component of th e bone Two gas-wash bottles with fritted disks were char. used in series and charged with a dilute solution of ammoniacal cadmium sulfate made by di­ TABLE 9. Determination of sulfur (expressed as SO.) in luting 1 ml of ' the stock solution to 100 ml of the gases evolved from bone char by treatment with hydro­ water. A 5-g sample of bone char was intro­ chloric acid duced into the Erlenmeyer flask, the inter­ changeable joint wetted, and the apparatus Gravimetric determina­ assembled . A test for possible leaks was made Iodometric determ inatiou of the tion aftcr Sam ple h yd rogen sulfide displaced witb combustion by application of air pressure. Fifty milliliters carrier gas noted in air in a q aartz far­ of diluted hydrochloric acid (1 + 1) was slowly nace added, and, as soon as the reaction slowed down, the system was swept wi th a flow of compressed Percent Percent 0. 135 AiL______0. 176 air. The fl ask was heated gently for 20 minutes. . 1l8 Commercial N2 ______. 151 The first gas-wash bottle was cooled with ice .105 _____ do ______. 160 y - 1. ______. 108 _____ do ______. 217 water. If the first gas-wash bottle alone con­ . 117 ____ _do ______. ](33 tain ed a yellow precipitate of cadmium sulfide, -. 148 N 2 and 02----______{ . 174 then only this mixture was titrated wi th iodine. . 151 Before titration the mixture was acidified with . 102 25 ml of hydrochloric acid. The heat evolved . 103 was usually sufficient to bring the solution to .131 room temperature. The titration with standard '" - j -. 126 iodine solution was made after adding 5 ml of a . 028 P - I ------{ . 040 ~~r~~~l:~i;lN ;------~~~ } . 072 freshly made starch solution. As the cnd point -. 035 N 2 and 02-- ______was approached the solution appeared yellowish . 031 Commercial N2 ______{ P- L ______,, ___ { . 046 red, orange, purplish red, and finally a deep -.036 N2 and 02--______permanent blue. A blank was run to determine .046 CommercialN2 ______} R - 22 . ______{ . 077 the amount of iodine necessary to color a similar -. 071 N 2 and 02 ______solution containing no sulfide. . 047 A number of results for six service bone chars R- 23 . ______{ .062 ~~r~~l~~~~ i~_l_~~-_-_~~::: } . 095 are given as part of table 9. Consid erable fluctu­ -.057 N 2 and 02 ______ations were observed in repeated runs with ali­ quot parts of the same sample. A possible expla- - Independent values by R. M . Thompson.

272 Jou rnal of Research 2 . Gravimetric Determination of the Evolved nations, is the variable action of Lhe hydl'ochloric Sulfur by Combustion in Air acid on the carbonaceous residu e of the char. It would be very difficult to control the exLenL of the In this determination the gaseous products of acid hydrolysis of the organic-sulflll' component evolution were dirrcted by a stream of air to pass in the residue. It is important, therefore, in the through a quartz tube, 1-cm inside diameter and analysis of a bone char by method 6, to determine 10 cm long, in which pieces of solid quartz rod first the sulfur in the evolved gases, and to usc were packed s.nugly. The quartz tube covered the mixture remaining in the flask to det rmine with wire gauze was heated to about 800 0 C with the sulfur in the residue and the filtrate. Un­ two NIeker gas burners for at least 15 minutes fortunatelv the values in table 8 were obtained before the reaction W ~LS allowed to take place in by 6 with two samples of char, one for the evolution flask. While the quartll tube was m etho ~l ' the evolu tion experiment and the other for the heating, the system including two gas-wash bot­ distribu tion of sulfur between residue and filtrate. tles containing a 5-percent solution of sodium Three service chars, however, were analyzed by hydroxide was swept with air. method 6 with the precaution of using but one An attempt was made to get a very slow sample for analysis. The results for these chars evolution of gase;::; by the following procedure fo r are given in table 10. A satisfactory agreement tbe addition of the hyclrocbloric acid: (a) The was obtained between the total sulfate deduced sample was wetted with a few milliliters of dis­ from method 1 (column 2) and the total , ulfate ti lled water, (b ) ten 1-ml portions of dilu ted hy­ derived as the sum of that in the residue, in drochloric acid (1 + 20) were added during a the fil trate, and in the gases evolved with hydro­ period of about 20 minutes, (c) ten 2 ml portions chloric acid. Independent data for these three of diluted hydrochloric acid (1 + 20) were added samples arc available for the ulfate content of in about 20 minntes, (d) 20 ml of concentrated the filtrate after lwdrochlorie acid treatment. hvdrochloric acid was added in one portion. These were through the courtesy of F'inally, the flow of air was increased, and the flll'nish~d H . 1. Knowles , chief chemist, Atlantic Sugar contents of the fl ash was boiled gently for 40 Refineries, L td. , Saint John, N. B., Canada. minutes. During this process the quartz tub e was continually heated with two Meker burners. TABLE 10. Determination oj sullvr (expressed as S04) At the conclusion of the experiment the quartz in service bone chars t ube was allowed to cool. All the connecting Pyl'rx tubing was washed with 200 ml of distilled Total slll­ Diges tion of Saml)le with n OI Obar f~~ I ~;.r~- I---.---.---.---.-F -I. I.-ra-.e- water into the first gas-wash bottle. The con­ acid ' 1 Evolvcd Total "" tents of both gas-w ash bottles were transferred m et hod) Filtrate Res le ue gases onl y ' to a 2-liter beaker; 100 ml of sa.turated bromine Percent Percent Percent Percent Percent Percent A ______ll ______1.07 0. 70 0.32 0. 0'19 1. 07 0.60 water was added to oxidize any sulfites to sul­ 1. 48 1. 00 . 45 . 05 1 1. 50 .99 0 ______fates and eoncen trated hydroch loric acid added 1.53 1. 13 .34 . 056 1.53 1.10

until the solution turned a definite orange. The • Sam pI es and the results in this colu mn were furn ished through tbe solution was boiled until colorless, cooled, adjusted co urtesy of JI. r. Knowles, chief chemist, Atlantic Sugar R efineries, Ltd. to pH ] , and a sjngle precipitation of barium sulfate made. VI. Discussion of Results The results for the six service chars are con­ The res ults of analysis of the several bone chars tained in the last column of table 9. Although for sulfur show clearly the errors in the existing definite flu ctuations were observed in repeated methods used by sugar refining laboratories. The determinations with aliquot parts of the same sam­ major error is in the value for so-calle? cal.ciu.m ple, the sulfate values so determined arc grrater sulfid e, which is based on a faUaey contawedw ItS than those cOlTesp oncling to simple evolution of defi ni tion. The sulfur reported as calcium sulfide hydroge n sulfide. In one case, sample 36, the is in reality that contained in the sulfides liberated two values are the same. by evolution with hydrochloric acid and a part of A reasonable explanation of the flu ctuations, the combined oro'anic sulfur in the carbonaccous apart from inherent inaccuracies in t.he determi- residue. Othpr ~ lTors arc due to the variable

Determination of Sulfu r in Bone Char 273 amount of organic sulfur removed from the car­ a maximum of 1 to 2 percent in service char. bonaceous residue that depends on the oxidation This would indicate that an important source of media employed in the analysis. the slllfur during char operations in sugar refining It is recommended that for practical purposes is the adsorbed impurities removed from the sugar the sulfur contained in bone char be defined as liquors. follows: Some of the adsorbed impurities contain pro­ Total 8u? fur. - This value may be determined tein matter composed in part of sulfur-bearing from either methods 1, 2, or 3. Method 1 is pre­ amino acids. The sulfur could be retained by ferred for its convenience. the carbonaceous residue formed in the reactions Inorganic 8uUate-81t(fur.- The nearest vaJue to that occur in the char kiln. It is also possible this quantity is the sulfate found in the filtrate of that some sulfate-sulfur may be introduced if a dilute hydrochloric acid treatment of the bone phosphate defecation was used in the refining char, method 6. operations. Even good technical phosphoric acid Suljide-8ulfur.-This value may be obtained may contain approximately 0.3 to 0.5 percent of from a determination of the hydrogen sulfide sulfate. If manufactured by the wet method evolved by treatment of the char with hydrochloric (sulfuric acid and phosphate l'oek), it may contain acid, see section V, 1. more sulfate than if manufactured by the furnace Combined organic 8uljur.- The sulfllr in the car­ method (oxidation of phosphorus). The latter bonaceous residue may be determined by either product may be preferred for sugar refining. The of two alternative procedures: the residue obtained two mentioned sources of sulfur are independent after digestion with dilute hydrochloric acid may of the type of kiln used to revivify the char. be so analyzed (method 6) ; or, the total 8ulfm may be obtained (method 1) and the sum of the irwl­ VII. References gamic 8ulfate-su~fur and the 8uljide-8ulfur Sll b­ tracted. [1] G.]'.. SlJCnCCr and G. P . Meade, Cane s ugar handbook, Any proposed definition of the sulfur content p. 348, 8th ed. (John Wiley & Som, Inc., New York, N. Y., 1945). of bone char must take into account the rather 12] John H enry Tucker, A manual of sugar analysis, p. complex chemical composition. R esearch pur­ 320- 322, 1st ed. (D. Van Kostrand Co., Inc., New poses, of course, require the best available methods York, N. Y., 1881 ). of analysis, and the data obtained thereby may 13] W. F. Hillebrand and G. E. F. Lundell, Applied in­ organic analysis, p. 575- 579 (John Vi;iley & Sons., not fit into the three simple categories defin ed Inc., New York, N. Y., 1929). ( above. For practical considerations of sugar 14] I. M . Kolthoff and E. B. Sandell, T extbook of quanti­ -, technology work, however, it may suffice to know tative inorganic analysis, revised ed., chap. 20, the sulfur content of bone char with a precision D etermination of sulfur as barium sulfate, p. 329- and accuracy sufficient to warrant the proposed 344 (The Macmillian Co., New York, N. Y., 1943). definitions. [5] F. P . Treadwell and W. T . Hall, Analytical chemistry 2,322, 6th ed. (John Wiley & Sons, Inc., New York, Interesting changes take place in the sulfur N. y., 1924). content of bone char during cane-sugar refining 16] C. L. Gordon, J . Research NBS 30, 107 (1943) RP1521. operations, and the contrast in the percentage of [7] C. L. Gordon, W. G. Schlecht, and E. Wichers, J . sulfur of service bone char with that of new bone Research NBS 33, 457 (1 944) RP1622. char should be briefly mentioned. The order of 18] G. Walton and G. H . Walden, Jr., J. Am. Chern. Soc. G8 , 1742 (1946). magnitude of the total sulfur expressed as S04 in new bone char is 0.15 percent as compared ,,~ith vV ASHINGTON , November 19, 1947.

2'?4 Journal of Research