ANALYTICAL CHEMISTRY.ALTHOUGH this Report deals with only four topics, the first three sectionsillustrate a large variety of current analytical trends.I . Zinc.No spectacular advances in the analytical chemistry of zinc have beenrecorded, but there has been fairly steady development, stimulated ho someextent by recognition of the biological importance of the element.(1) Separations and Qualitative Tests.-A procedure is given 1 fordetermination of zinc and other metals in foodstuffs, by precipitationwith hydrogen sulphide a t pH 8 after a wet ashing process. The pre-cipitate is dissolved, and iron, etc., precipitated with ammonia. Afteracidification, copper is separated with hydrogen sulphide and zinc determinedwith 8-hydroxyquinoline.McLellanpoints out that in 1 : 19 hydrochloric acid there is appreciable coprecipitationof zinc with Group 2 metals.Laws uses a formic acid buffer for separationof zinc from many other metals in light and in nickel alloys, but S. A. Colemanand G. B. L. Smith prefer to use citric acid as a buffer. E. A. Ostroumov 5says that post-precipitation of cobalt on zinc sulphide can be prevented bythe use of acraldehyde in the solution and that filter-paper pulp, unlesspreviously treated with buffer solution, materially increases the contamin-ation. C. Zollner stated that in absence of hydrochloric acid, cadmium isquantitatively precipitated by hydrogen sulphide from a solution containing15 ml. of concentrated sulphuric acid per 100 ml. ; the solution is boiled, andallowed to cool while the gas is passing.There are also several references 7**, 9to separation of cadmium from zinc by means of aluminium foil or powder,which precipitates cadmium from a somewhat acid hydrochloric acid solution.It would appear best to separate most of the cadmium in this way, andrecover the remainder from the filtrate by means of hydrogen sulphide.If to a solutioncontaining cadmium and zinc, thiourea and Reinecke's salt are added in thecold, cadmium is quantitatively precipitated asleaving all the zinc in the filtrate. Complex thiocyanates have been studied ;e.g., W. C. Vosburgh, G. Cooper, W. J. Clayton, and H. P. Pfann l1 state thatOther authors 2 * 3 write on the use of hydrogen sulphide.A novel method is put forward by C.Mahr and H. OhleeloCd( CS=N,H,),, [Cr(CNS),"H,) 212,J. H. Hamence, Analyst, 1937, 62, 18.E. Q. Laws, Analyst, 1941,66, 54.Ann. Chim. anal., 1937, 19, 145.7 J. J. Lurie and V. F . Neklyntina, Zavod. Lab., 1936, 5, 87.8 E. I. Nikitins, ibid., 1939,8, 1172.Q F. E, Townsend and G. N . Cade, Ind. Ewg. Ghem. Anal., 1940,12, 163.lo 2. anal. Chew., 1937, 109, 1.* G. McLellan, J . Assoc. Off. Agric. Chem., 1941, 24, 728.Ind. Eng. Chem. Anal., 1941, 13, 377.2. anal. Chem., 1938, 114, 8.l1 Ind. Eng. Chem. Anal., 1938,10, 393WILSON: ZlNU. 273for precipitation of ZnHg(CNS),, 100 ml. of solution may contain not morethan 2.5 g. of nitric or 5 g. of sulphuric acid. The precipitate can be weighedafter drying at 105', but W. Hoffman and G. B. Thackray l2 prefer to titrateit by R.Lang's procedure,l3 or to use potassium iodate. In analysis of brassor bronze they precipitate copper as Cu,(CNS), as usual, and to the filtrateadd (NH,),Hg(CNS), to precipitate the zinc salt. Similarly, a micro-methodfor zinc in ores l4 precipitates ZnHg(CNS), and titrates it with potassiumiodate. Iron may be masked with potassium fluoride,.citric acid, or (best)phosphoric acid, but large amounts of iron or aluminium prevent theprecipitation -It is stated l5 that p-naphthaquinoline in 0-1N-sulphuric acid and potass-ium thiocyanate give a very sensitive qualitative reaction for zinc. Char-acteristic crystals appear from a 2~-acid solution containing more than 1part of zinc per 200,000 parts. Less than 0.5% of cadmium does notinterfere, but numerous other metals do.Forty-five proposed qualitative reagents are tabulated by P.Wenger andR. Duckert.16 They recommend pyridine and potassium bromide in neutralsolution (nickel also reacts), potassium ferrocyanide in acid solution (cad-mium also reacts), K,Hg(CNS), and cobalt chloride (iron and manganeseinterfere). As " spot " tests they recommend metanil-yellow with potassiumferricyanide, and also dithizone. Benzoin in presence of alkali and mag-nesium ions is a highly specific reagent for zinc,17 giving a compound with avivid green fluorescence in U.V. light. Only bismuth, boron, and antimonyinterfere and the test will detect 1 pg. of zinc in 1 ml. of solution.(2) Use of Organic Reagents.-Anthranilic acid, first proposed by H.Funkand M. Ditt,ls was thoroughly investigated by R. J. Sherman, J. H. F. Smith,and A. M. Ward,lg who dealt with the solubility of the Zn, Cd, Cr, Ni, andCu salts : the solubility of the zinc salt is increased by sodium acetate, buteven in absence of salts, results are somewhat low. Bromination of theprecipitate with standard bromide-bromate solution to the tribromo-compound is preferred, but C. W. Anderson 2o weighs the precipit'ate, whichis formed in acetic acid solution, salts being absent. Preliminary separationof the zinc is essential, and P. Wenger 21 points out that the temperature ofprecipitation is unimportant.R. R. Ray and M. K.Bose 22 claim very good results by a micro-precipitation with sodium quin-aldinate from a hot solution faintly acid with acetic acid.The precipitate isdried at 125". A thorough investigation was made by R. J. Shennan,*3who showed that the reagent completely precipitates copper between pHQuinaldinic acid has also received attention.l2 Analyst, 1941, 66, 321.l* J. J. Lurie and L. A. Philippovs, Zavod. Lab,. 1939, 8, 1047.l 5 E. B. Sandell, D. B. Wishnick, and E. L. Wishnick, Microchim. Acta, 1938,3, 204.l6 Helv. Chim. Acta, 1942, 25, 400.17 C. E. White and M. H. Neustadt, Ind. Eng. Chem. Anal., 1943,15,599.2O In&. Eng. Chem. Anal., 1943, 15, 367.22 Mikrochem., 1935, 17, 11.l3 2. anal. Chem., 1929, 79, 161.18 2. anal. Chem., 1933, 91, 332. lo Amlyst, 1936, 61,. 395.21 Helv. Chim. Acta, 1942, 25, 1499.23 Analyst, 1939, 64, 14274 ANALYTICAL CHEMISTRY.2.5 and 6.9, cadmium from pH 3.9 to 7.2, and zinc a t pH 2-3-66, NOseparations are possible, and acetates have a marked effect on the solubility,but P. R.Riiy and T. C. Sarkar24 showed that if the copper is reducedto cuprous in the presence of thiourea, it is masked and does not interfere;A similar process can be used in presence of mercury.5-Nitroquinaldinic acid precipitates zinc from feebly acid solutions, andhas been made the basis of a colorimetric method.25 Most other metalsinterfere. Another reagent with which most common anions and cationsinterfere, and which would therefore appear to be of very limited use, istetraphenylarsonium chloride.26 Salicylaldoxime also precipitates zinc.L. P.Biefield and W. B. Ligett 27 give the pH at which it precipitates zinc,copper, and lead. T. G. Pearson28 says i t is not a suitable reagent, andJ. H. Flagg and N. H. Furman 29 describe conditions under which it couldbe used. A really useful (but not selective) reagent, however, is 8-hydroxy-quinoline, probably the only one of these reagents which is often used. In amost important paper,30 the authors examined a number of “ oxinates ” byX-ray diffraction, and showed that the dihydrated oxinates of zinc andmagnesium (as of copper and iron) are amorphous, a fact of very greatimportance in explaining the difficulty of many separations which wouldappear to be possible with this reagent. They also showed that fromsolutions containing both ions, the precipitates are solid solutions.Thisdid not confirm other work31 which explained the same phenomenon byadsorption. An earlier paper 32 on separations states that zinc can beseparated from manganese in two precipitations a t pH 5-6, from iron andbismuth in presence of tartrates at pH 8, from arsenic and antimony, but notfrom cobalt or nickel, by this reagent.( 3 ) Colorimetric and Micro-procedures.-Although diphenylthiocarb-azone (“ dithizone ”) is one of the least specific of reagents, it has been usefulin the determination of really small amounts of zinc. It was a t first appliedrather tentatively, but the investigations of many recent authors, in particularthe members of the American “ Association of Official Agricultural Chemists,”have rendered it most serviceable in an enormous number of ways.Thefirst qualitative use of the reagent appears to have been by H. Fischer,=and later with G . L e ~ p o l d i , ~ ~ he showed how, in a barely acid solution, mostmetals could be masked with thiosulphate, rendering the reagent practicallyspecific for zinc, potassium cyanide being added if cobalt or palladium waspresent. I n fhe first application of the reagent to foods,35 a chloroformMikrochem., 1939, 27, 64.3 5 W. L. Lott, Ind. Eng. Chem. Anal., 1938,10, 331.* 6 H. H. Willard and G . M. Smith, ibid., 1939, 11, 269.37 Ibid., 1942, 14, 359.3O R. C. Chirnside, C. F. Pritchard, and M. P. Rooksby, Analyst, 1941, 66, 339.3l H. V. Mayer and W. J. Remington, Ind. Eng. Chem. Anal., 1938,10,212.33 C.Cimerman and P. Wenger, Mikrochem., 1939, 27, 76.33 Ibid., 1930, 8, 319.36 N. D. Sylvester and E. B. Hughes, Analyst, 1936, 61, 734.26 2. anal. Chem., 1938, 112, 179.Ind. Eng. Chem. Anal., 1940, 12, 663.34 2. and. Chem., 1936, 107, 241WILSON: ZINC. 275solution was used to extract a solution of the ash buffered a t pH 4.5 withammonium acetate. Zinc (with bismuth and cadmium if present) wasextracted from the chloroform solution with diluted hydrochloric acid andfinally titrated either with potassium ferrocyanide or, after addition ofpotassium ferricyanide and iodide, with N /500-thiosulphate solution.E. A. Coakhill 36 improved upon P. L. Hibbard’s colorimetric process.37In the analysis of commercial lead, the bulk of the lead (from a 2-5 g.sample) is removed as sulphate, and most of what remains by hydrogensulphide in a solution acid enough to prevent precipitation of zinc sulphide.Thioglycollic acid prevents reaction of lead with dithizone, so a just ammoni-acal solution is extracted with a chloroform solution of both reagents until nofurther change of colour takes place.A “ blank ” solution is now similarlytreated, and standard zinc solution run in until the colours match. A veryimportant colorimetric application is described by F. H. Vogelenzang 38in the determination of zinc in blood, etc. All reagents contain zinc andshould be extracted by the dithizone reagent before use ; glass often containszinc and such glass must not be used. Zinc and copper are extractedtogether from a weakly ammoniacal solution, zinc stripped out with 2 ~ -hydrochloric acid, and finally collected in a carbon tetrachloride solution ofthe reagent; after filtration, the solution is diluted to known volume, andthe red colour measured in a Pulfrich photometer.It has a maximumextinction at 5250 A., and obeys Beer’s law.Several investigators used dithizone to isolate the zinc, and bhen deter-mined it in some other way,39 but generally, progress has been in purecolorimetry, although E. B. Sandell 40 proceeds by “ extraction titration ”a t pH 4.1, sodium thiosulphate masking other elements in soils, and P. L.Hibbard,41 having isolated the zinc compound, titrates it with bromine.H. J. Wichman 42 points out that sodium diethyldithiocarbamate masksalmost all elements except zinc, and that photometry at the proper wave-length enables zinc to be determined without removing excess of reagents.This important modification is being extended by R.A. Caughley, G. B.Holland, and W. S. R i t ~ h i e , ~ ~ and the last two authors44 emphasise thepresence of zinc in most reagents and in glass, and include a valuable tablegiving the reaction of many metals in 0*02~-ammonia or O.O2~-hydro-chloric acid to dithizone, “ carbamate,” and both together ; in 0 . 0 2 ~ -ammonia, this combination is specific. The analytical procedure is typicalof this kind of technique. The extract of the plant ash or the like, containingammonia and ammonium citrate, is treated with a carbon tetrachloridesolution of dithizone, until all reacting metals are extracted.The tetra-chloride layer is then shaken with O*02~-hydrochloric acid, which removes36 Analyst, 1938, 63, 800.38 Phurm. Weekblad, 1939, 76, 89.39 E.g., J. H. Peekman and J. E. Menshing, J . Assoc. Off. Agric. Chem., 1937,20, 627.40 Id. Eng. Chem. Anal., 1937, 9, 464.42 J . Assoc. Off. Agric. Chem., 1938,21, 197.O3 Ibid., p. 204.37 Ind. Eng. Chem. Anal., 1937, 9, 127.4 1 Ibid., 1938, 10, 615.44 Ibid., 1939, 23, 333276 ANALYTICAL CHEMISTRY.lead, zinc, cobalt, silver, and cadmium, leaving copper and mercury in thesolvent layer. The aqueous layer ie made 0 . 0 2 ~ with ammonia, bufferedwith ammonia citrate, and extracted with carbon tetraehloride-dithizoneafter addifion, of “ carbamate,” which brings zinc into the solvent layer.Unfortunately, in presence of “ carbamate ” extraction of zinc is not quitecomplete,45 and the partition ratio between the tetrachloride dithizone phaseand the aqueous phase is influenced by pH and concentration of the reagentsin the two phases.By keeping these factors constant, accurate results canbe obtained, preferably by finally measuring extinction a t 5400 A., a moreconvenient process than extracting excess dithizone with O.OIN-ammoniayleaving zinc dithizone in the carbon tetrachloride phase. For zino in soils(and also copper) G. D. Sherman and J. S. McHargue4‘j proceed somewhatsimilarly to Holland and Ritchie, but before the final extraction of zinc,they mask other metals (e.g., lead) with sodium fhiosulphate.Finally, di- p-naphthylthiocarbazone is proposed as a superior reagentwith many features similar to d i t h i z ~ n e .~ ~ Transmission diagrams are given,and the reagent quantitatively extracts zinc from solutions of pH >76,even when “ carbamate ” is present. Full instructions are given for theelimination of interferences. If less than 0~05 mg. is present, and no cadmium,photometry of the carbon tetrachloride aolution is recommended, but if morezinc or zinc and cadmium are present, the tetrachloride solution is ‘‘ stripped ”with O.S~-hydrochloric acid and polarographed from - 0.3 to 1.3 v., thezinc step taking place a t - 1.1 v. The addition of cadmium (0.6 mg.) as an“ infernal standard ” is a useful device, or after polarography, aknown amountof zinc is added, and from the heights of the zino steps in the two polarograms,the zinc content of the sample is calculated.(4) MisceUaneous Devehpme?ets.-Poiarography aa a micro-method maybe far simpler than colorimetric methods.J. F. Reed and R. W. Cummings 48describe the polarogrephic determination in soda and plant ash. None ofthe constituents of plant ash interferes when the pH of the solution is 4.6,and the solution is 0.025~ with respecf t o thiocyanate ions. On a l-g.sample, zinc may be determined with an error of 5 5 yo when present between0.5% and 5 parts per million. The determination of zinc polamgraphicallyin paints 49 and in brass plate 50 is also described, the latter determinationbeing made in 20 minutes.A publication by the British Aluminium Com-pany, Lfd., discusses analysis of aluminium (and magnesium) alloys by meansof the polarograph and the spectrograph.61R. T. O’Connor b2 describes a spectrographic method for traces of zincin fertdisers, beryllium being added to act as an internal standard, and the45 H. Cowling and E. J. Miller, Ind. Eng. Chem. Anal., 1941,13, 145.46 J . Assoc. 08. A&. Chem., 1942, 25, 510.47 J. Cholak, D. M. Hubbard, and R. E. Burkey, I n d . Eng. Chern. Anal., 1943, 15,48 Ibid., 1940,12, 489.60 W. P. Tyler and W. B. Brown, ibid., 194&18, 520.61 Publication Ml, 1943.754.49 B. M. Abraham and R. S. Huff?nan, dbid., p. 656.52 I n d . Eng. Chem. Anal., 194i, 13, 697WILSON: ZINC. 277Be line 2348 A.being used. Zinc can be determined between 0.0008~0 andabout 1%. E. J. Magaziner and N, A. Zventitzki 53 suggest striking the arcbetween a copper electrode and the sample, using for comparkon the linesCu 2824 and Zn 2756, but if lead is also present Cu 2883 is used.In deposition fromneutral citrate solution, very many metals interfere,= but excellent resultsare obtained by electrolysis at 2 amps. from a slightly alkaline solution onto a copper-plated platinum electrode, after a preliminary separation ; 55the method, for alloys, is very rapid. A. Cohen 56 recommends electrolysisfrom alkaline tartrate solution in analysis of aluminium alloys, and alsodiscusses the mercury thiocyanate precipitation. He emphasises that bythe usual procedure of dissolving the alloy in sodium hydroxide, some zincalways remains insoluble.One or two papers discuss the use of complexcyanides in alloy analysis. C. C. Casto and A. J. B ~ y l e , ~ ' in analysis ofmagnesium alloys, remove manganese and copper, add citrate and dilutesulphuric acid, and proceed by Lang's method.l3 If less than 0.5% of zincis present, a larger sample must be used. In the ferricyanide titration,p-ethoxychrysoidine 58 and o-anisidine 59 are recommended as indicators.In brass analysis, by heating to 800-850" in a vacuum mm.), zinc(and also lead) are distilled off, and the efficiency is greater the higher theproportion of zinc.60 Finally, three papers on commercial analysis for zincare noteworthy. Zinc oxalate is found to be insoluble in 70% acetic acid ifammonium chloride is absent ; 61 the precipitate is washed twice in a centri-fuge tube with 10% acetic acid, and the oxalate finally titrated with per-manganate (nickel, lead, and copper interfere). L.G. Miller, A. W. Boyle,and R. B. Neill 62 dissolve magnesium alloys in dilute hydrochloric acid,adding lead if necessary to prevent solution of copper, then add a littleferricyanide to mask any iron usually present, make the solution ~ / 1 inrespect to hydrochloric acid, add excess of ferrocyanide solution, and afterfiltration, back-titrate excess with ceric sulphate. The method is very rapidand accurate to 1-6%. Zinc in magnesium alloys is also dealt with byS. Weinberg and T. F. Boyd.63 The sample is dissolved in diluted sulphuricacid, a large excess of ammonium chloride and tartaric acid added, thenexcess of ammonia, and the solution electrolysed for 20 minutes at 2 amps.If significant quantities of Group 2 metals are present, they must be removedby means of hydrogen sulphide.Apart from this case, the determinationis complete in 25 minutes.There are a few references to electrodeposition.H. N. W.b3 Zuvod. Lab., 1940, 9, 992. ". R. Winchester and L. F. Yntemrt, Ind. E i q . Chem. Anal., 1937,8,254.5 5 G. H. Osbourne, AnuEyst, 1941, 66, 412.5 6 Helv. Chim. Acta, 1943, 26, 75.5 7 I n d . Eng. Chem. Anal., 1943, 15, 624.G 8 W. I?. Tyler, ibid., 1942, 14, 114.50 H. F. Frost, AimZyst, 19-13, 68, 51.s1 P. J. Ewhg and J. C. Lamkin, Ind. Eng. Chem. Anal., 1944,16, 194.W.P. Treadwell and G. Frey, Helv. C'him. Ada, 1044, 27, 42.62 Ibid., p. 256. Ibid., p. 460278 ANALYTICAL CHEMISTRY.11. Arsenic.The following report is divided into sections, each dealing with the applica-tion of a particular reaction or technique.The Gutzeit Method.-A thorough investigation discusses various errorsto which the process is subject. Since zinc varies in activity, if densemassive zinc be used, the temperature of evolution of arsine should behigher (40-60"); even in the presence of stannous chloride, all thearsine is not evolved from quinquevalent compounds, and reduction bysulphurous acid on a water-bath is recommended, followed by brief boiling,before evolution of arsine, but addition of a little potassium iodide (as acatalyst) and then stannous chloride, following the A.O.A.C.method, isequally effective.A very important paper describes the determination of minute amounts,as little as 0.1 pg., with a probable error of less than 5% and sensitive to0.01 pg. The author uses thin cotton threads impregnated with mercuricchloride as absorbents ; they are enclosed in capillary tubes, into which theyjust fit. All conditions must be rigidly standardised, including temperatureof evolution and of the absorption tubes. Quinquevalent arsenic is reducedby potassium bisulphite, stated to be preferable to the iodide, and a littleferrous salt as catalyst. The stains are developed by ammoniacal silvernitrate solution and measured with a Vernier caliper. This technique hasbeen further ~tudied.~L.Truffert * returns to the reaction of arsine on a silver salt in determiningarsenic in wines. Platinised zinc and sulphuric acid are used to generatehydrogen, and the gases are passed first through potassium hydroxidesolution. The paper, used in strips, is coated with silver citrate, and is stablein the dark; 1-30 pg. can be estimated.The Sub-committee of the Institute of Brewing considers two methods ofestimating the evolved arsine-those of Marsh and Gutzeit. In the latter,discs are preferred to strips, and mercuric bromide is about twice as sensitiveas the chloride, but preparation of the paper, which will not keep, needsgreater care. The lowest limits are stated as 0.001 mg. for the Marsh-Berzelius method, and 0.0004 mg.for the Gutzeit process. Despite theimpermanent nature of the stains, the latter is preferred. For hops, malt,beer, sugar, and finings, methods of preparing the solutions which do notrequire destruction of organic matter are described. Coal is ashed with equalweights of magnesia and potassium permanganate, and in all cases potassiumiodide and sodium sulphite are used to reduce AsV. Alternatively,G coal orcoke is ashed with a mixture of magnesia, sodium carbonate, and potassiumnitrate ; after solution of the residue AsV is reduced by sulphite and deter-mined by the mercuric bromide modification of Gutzeit's method. A con-W. A. Davis and J. G. Maltby, AmaZyst, 1936, 81, 96.A. E. How, Id. Eng. Chem. Anal., 1938,10, 226.E. Cahill and L.Walters, ibid., 1942,14, 90.Ahn. Falsif., 1938, 31, 73.D.S.I.R. Fuel Research Paper No. 44, Oct. 28th, 1940.J . Inst. Brewing, 1938, 44, 359WILSON : ARSENIC. 279tinuous apparatus is described 7 in which hydrogen is evolved from a longcadmium cathode, from an electrolyte containing sulphuric acid and hydroxyl-amine ; from time to time a sample is introduced below the cathode, flows uppast it to the platinum anode, and so to waste. The evolved hydrogen ispassed through. a heated glass tube. As most samples examined are practicallyfree from arsenic, they can very rapidly be dealt with in this apparatus :a contaminated sample is immediately recognised and the mirror can becompared with standards.Less than 0.1 mg. of selenium is stated to have no effect on arsinemethods. *The estimation of arsenic in foodstuffs, etc., contaminated with wargases 9s lo cannot be suitably summarised.G. Taylor and J. H. Hamence l1state that if zinc alloyed with 0.3% of copper is used, the whole of the arsenicis liberated without the use of sulphites in the presence of “ stannated”hydrochloric acid in the Gutzeit method.Reduction by Means of Hypop7tosphite.-This method continues to receiveattention. White metals are dissolved in hydrochloric acid in presence ofbromine, arsenic precipitated with hypophosphite, redissolved, reduced withsodium sulphite after addition of sulphuric acid, and finally titrated withbromate solution. Antimony does not interfere.12 As internal indicatorsfor the bromate titration, Bordeaux, naphthol blue-black, and brilliantPonceau-SR are re~ommended.1~ If reduction is carried out in the solutionat 90°, it is so rapid that there is no loss of ar~enic.1~ The precipitated metalis dissolved in N/5o-CeriC sulphate solution, excess being titrated with ~ / 2 0 0 -arsenious oxide solution, with osmic acid as catalyst. Antimony and tin donot interfere, and the method is valid for 0.1-2 mg.of arsenic. W. J.Agnew l5 proceeds similarly, but uses N/lOO-potassium dichromate to dissolvethe metal, and back-titrates excess with N/lOO-ferrous sulphate. The effectof selenium and tellurium is discussed; H. J. G. Challis l6 points out thatthese elements are also precipitated by hypophosphite but below 50” areprecipitated free from arsenic.After filtration, arsenic can be precipitatedby more hypophosphite on boiling, but B. S. Evans l7 says that this is onlytrue of “traces” of arsenic. He precipitates the three elements (say, inanalysis of copper) together by hypophosphite, redissolves them, andprecipitates selenium with potassium iodide and tellurium with sulphurdioxide, leaving arsenic in solution to be precipitated later with hypophosphiteas usual. The process is applied la to organic arsenicals, after wet oxidation8 Fuel Research Board : Report for year ending March 31st, 1939.Q H. A. Williams, Analyst, 1941, 66, 228.lo A. McM. Taylor and W. J. Stainsby, ibid., p. 233.l2 C. W. Anderson, I n d . Eng. C‘hem. Anal., 1937, 9, 569.13 G. F. Smith and R. L. May, ibid., 1941, 13.460.l4 J. M. Iiolthoff and E. Andrew, ibi&., 1940,12, 177.l5 Analyst, 1943, 68, 111.H. C. Lockwood, Analyst, 1939, 64, 657.l1 Ibid., 1942, 67, 12.l6 Ibid., 1941, 66, 58. l7 Ibid., 1942, 67, 346.H. A. Sloviter, W. M. McNabb, and E. C. Wagner, Ind. Eng. Chem. Anal., 1942,14, 516280 ANALYTICAL CHEMISTRY,with sulphuric and nitric acids, but is not applicable 1Q to cacodyl derivatives,which should be decomposed with potassium birjulphate and sulphuric acid ;after solution is complete a version of the hypophosphite process 2o is used.There are two mentions of a curious process in which the reduced metalis kept in a colloidal suspension. J. V. Harispe*l applies this to the semi-rnicro-analysis of urine : after destruction of organic matter, hydrochloricacid, hypophosphite, and a dilute solution of potassium silicate (protectivecolloid) are added, and after the mixture has been heated on a water-bath,the colour is compared with standards. J.Thuret 22 says that the arsenicslowly flocculates even in presence of stabilisers, and recommends a standardsolution containing borax and colophony, which has the same appearance andremains stable for one month.Distillation Methods.-For other than traces, distillation of arsenictrichloride and titration of the distillate continues to be largely used. Thedetermination in wood of arsenic added as preservative is described : 23after wet oxidation with sulphuric and nitric acids, the latter acid is removedby repeated evaporation, and the trichloride distilled as usual.The distillateis oxidised with nitric acid, excess of this removed, AsV reduced in acid solu-tion with potassium iodide, and finally titrated with iodine in presence ofexcess of sodium bicarbonate. For smaller amounts, a modification ofGutzeit's process is used on aliquots of the distillate. An important paper 24describes the quantitative separation of arsenic, antimony, and tin by frac-tional distillation of the aqueous solution of the chlorides. For arsenic inpyrite~,~5 the mineral is fused with sodium carbonate and peroxide, the massacidified with hydrochloric acid, and the trichloride distilled, hydrazine andpotassium bromide being used aB reducing agents ; the distillate is titratedwith bromate. In the analysis of soils which have been treated with leadarsenate, L.Koblitsky 26 oxidises interfering organic matter if necessary with30% hydrogen peroxide, and distils after reduction with hydrazine andpotassium bromide. H. N. Wilson2' determines total arsenic in glass byfusion with sodium hydroxide, acidification of the melt with hydrochloricacid, in such a way that silica is not precipitated, distillation with this acid,hydrazine and br.omide being the reducing agents, and titration of thearsenious chloride in the distillate (which is of the correct acidity) by ~ / 2 0 0 -potassium iodate. The method, applicable to 0.02 yo arsenious oxide andupwards, is rapid and accurate. V. Dimbleby 28 discusses the same process.Colorimetric Processes.-The most interesting development in arsenicdeterminations is the growth in popularity-of these processes, in all of which19 V.Levine and W. M. McNabb, ibid., 1943,15, 76.20 Ref. 18.2 1 J. P h m . Chim., 1939, 30, 58.22 Ann. Falsif., 1939, 32, 328.23 Commonwealth of Australia.24 J. A. Schemer, Bur. Stand. J. Res., 1938, 21, 95.25 T. A. Fedorkin, Zavod. Lab., 1940, 9, 1324.26 J . Assm. Off. Agric. Chem., 1939, 22, 680.27 Analyst, 1943, 68, 361.Division of Forest Products, Reprint No. 29, 1936.28 Glass Review, 1943, 19, 120WILSON : ARSENIC. 281arsenomolybdate is formed and reduced, the blue colour being a function ofthe arsenic concentration. The arsenic is first isolated from interferingsubstances by distillation as chloride or hydride.Organic matter in must orwine 29 is destroyed with nitric and sulphuric acids and hydrogen peroxide,arsenic distilled as chloride, and the distillate evaporated to dryness withnitric acid. The arsenic, now quinquevalent, is colorimetrically determinedby 0. Zinzadze’s reagent,30 preferably with a Zeiss photometer. Theextinction is proportional to the arsine content from 0.01 to 0.8 mg. J. A.Schemer 31 describes a superior form of Zinzadze’s reagent. D. M. Hubbard 32distils the chloride in a current of carbon dioxide, oxidises the distillate asabove, and obtains molybdenurn-blue by a molybdate-hydrazine sulphatereagent, reacting at 70-75’ for 30 minutes. The colour is stable for 24hours. The maximum absorption is in the near infra-red a t 8400 A., whichgives twice the optical density of that a t 7400 A .~ ~ An elegant, oxidation ofthe chloride is effected 34 by distilling it in a special apparatus, so that thevapours pass through a few ml. of potassium iodate solution, whereby itis oxidised, steam and hydrogen chloride passing on to be condensed andreturned to the flask.Two methods avoid distillation by extracting the acid arsenious solutionwith a carbon tetrachloride solution of sodium ethylxanthate. A. Kleinand F. A. Vorhes 36 evaporate the extract to dryness and oxidise the residuewith bromine water, then applying Zinzadze’s reagent. T. B. B. Crawfordand I. D. E. Storey36 extract inorganic arsenite from blood, urine, etc., bymeans of the xanthate (three extractions) and then proceed as in the foregoingor by means of G .A. Levvy’s method.37Several papers deal with the evolution, oxidation, and colorimetricdetermination of arsine, the arsenomolybdate being used. E. B. Sandell 38passes the arsine and hydrogen (from 20-30 mesh zinc) through mercuricchloride and potassium permanganate, evolution of arsine being completein 30 minutes; the solution is filtered and treated by Hubbard’s method.*2The procedure is suitable for 1-10 pg. of arsenic. M. B. Jacobs andJ. Nagler 39 oxidise the arsine by hypobromite, and R. Milton and W. B.Driffield 40 by iodine and sodium hydrogen carbonate solution. In all casescareful adherence to standard quantities of reagents is necessary, and alsoto exact control of acidity, as is always the case for quantitative formationand reduction of molybdenum complex ions.In controlled conditions, theblue colour obeys Beer’s law from 0.001 to 0.1 mg. of arsenious oxide per10 ml.Arsenic can be colorimetrically determined in lead without distillation29 J. Burkard and B. Wallhorst, 2. Unters. Lebensm., 1935, 70, 308.30 I d . Eng. Chem. Anal., 1935, 7, 227.32 Ind. Eng. Chem. Anal., 1941, 13, 915.33 J. A. Stultzaberger, ibid., 1943, 15, 408.34 A. L. Chaney and H. J. Magnuson, ibid., 1940,12, 691.36 J. Assoc. 08. Agric. Chern., 1939, 22, 121.37 Ibid., 1943, 37, 598.3B Ibid., p. 442.The process is very rapid.31 Bur. Stand. J . Res., 1938, 21, 96.313 Biochem. J., 1944, 38, 195.a6 Ind. Eng. Chem. Anal., 1942, 14, 82.40 Analyst, 1942, 67, 279282 ANALYTICAL CHEMISTRY.the latter being nearly all removed from the nitric acid solution by sulphuricacid, and the molybdenum-blue reagent applied to the filt~ate.~lC.C. Cassil 42 recommends passing arsine into mercuric chloride solution,titrating this by adding excess of an iodine solution containing enoughiodide to hold excess of mercury in solution, and back titrating with thio-sulphate.Vuriow Methods.-Polarography is said43 to be possible only in acidsolution; the shape of the polarogram is influenced by the supportingelectrolyte, and sometimes four waves are visible. K. Bambach 44 collectsarsine in mercuric chloride solution, heats this to convert arsenides intoarsenites, and at pH 6 precipitates mercury with hydroxylamine.The clearsolution is polarographed after addition of hydrochloric acid. In 1*5~-acidthe half-wave is at - 0.35 v., in O-S~-acid at - 0.5 V. The process is moreaccurate than that of Gutzeit and quicker than colorimetry. J. J. Lingane 45says that AsVis not reduced at the dropping-mercury cathode, but As111 isreduced in two steps; in N-hydrochloric acid the wave height a t -0.8 V. isproportional to concentration. An ill-defined wave at - 0.9 v. may corre-spond to reduction to arsine. Tartrate or alkali suppresses the wave, and0-1 yo of gelatin should be present.S. Torrance 46 states that if a solution of copper containing less than one-fifth of its weight of arsenic in hydrochloric acid is electrolysed, all the arsenicis deposited with the copper.If the deposit is redissolved and electrolysedin sulphuric solution, only copper is deposited, leaving all the arsenic insolution.Among noteworthy miscellaneous methods are the distillation of arseniouschloride from solutions containing tungsten,47 the determination of arsenatesand selenates in presence of one another,48 a microtechnique for determinationas magnesium ammonium a r ~ e n a t e , ~ ~ the mechanism of Bettendorf's test,50and a method for quantitatively separating arsenic from copper by co-precipitation with hydrated manganese oxide.51 In determining arsenic insulphur, S. J. Fainberg and G. A. Taratorin 52 dissolve the latter in boilingsulphite solution, and add copper sulphate and acetic acid, whereupon allthe arsenic is carried down as trisulphide by the copper sulphido; or thesulphur is dissolved in alkaline hydrogen peroxide, and the arsenic determinedcolorime trically .41 E.A. Coakill, Analyst, 1938, 63, 801.42 J . Assoc. Off. Agric. Chem., 1938, 21, 198; (with H. J. Wichmann) 1939, 22,436;1941, 24, 196.T. A. Krinkova, Zavod. Lab., 1940, 9, 950.44 I d . Eng. Chem. Anal., 1942, 14, 265.4 6 Analyst, 1938, 63, 104; 1939, 64, 263.4 7 T. Millner and F. Kiinos, 2. anal. Chem., 1936,107, 96.4g F. Hecht and M. von Mack, Mikrochem. Acb, 1937, 2, 218.Eo W. B. King and F. E. Brown, J. Amr. Chem. Xoc., 1939, 61, 968.0 1 C. L. Luke, I d . Eng. Chem. Anal., 1943, 15, 626.52 Zavod. Lab., 1940, 9, 1223.46 Ibid., 1943,15,583.J. Milbauer, ibid., 1937, 109, 171-SON : FLUORINE.283Finally, an important paper is devoted to analysis of large numbersof samples of foodstuffs or the like : after wet oxidation and removal ofnitric acid, a few mg. of cadmium sulphate are added to the diluted solutionand precipitated as sulphide, carrying down all the arsenic. The cadmiumlater serves as an internal standard. The sulphide is collected on a smallsuction filter dressed with powdered graphite, and after being dried is arcedin a low-tension D.C. arc between graphite electrodes using a Hilger mediumspectrograph to record the spectrum. Normally, visual comparison betweenthe spectrogram and that of a standard sample suffices to show whether ornot the specified limits for arsenic, etc., have been exceeded. The method isvery expeditious and suitable for mass production methods of working, butas only 1 g.of sample is used, scrupulous care must be taken over " blanks "and absolute cleanliness of working.H. N. W.111. Fhmrine.Probably the determination of fluorine has changed more than that of anyother element in recent years. H. H. Willard and 0. B. Winter's procedure,lin which the fluorine was distilled as hydrosilicofluoric acid from an aqueousacid solution in which quartz powder was suspended, hydrolysed again tofluoride in the distillate, and titrated with N/lOO-thoriurn nitrate solutionwith alizarin43 as an indicator, wag a remarkable advance in analyticaltechnique. Much subsequent work has simply been refinement and ampli-fication of this method.Ashing.-The possibility of losing fluorine during ashing or calciningoperations cannot be too greatly stressed.J. M. Sanchis2 adds sodiumhydroxide solution before ashing in platinum, and distilling at 135-140'with sulphuric acid ; chlorine interferes, especially in presence of manganese,and should be removed by means of sodium nitrite. Finally, thedegree of fading caused by various aliquots of the distillate is comparedagainst standards, using zirconyl alizarin-S lake as is now usual. In themicro-determination of fluorine in bl00d,3 loss of fluorine is reduced byfirst charring in a porcelain crucible, then transferring to a gold one ! Theash must not be sintered. is recom-mended (i.e., zirconyl chloride and purpurin solution). Wine can be ashedwithout loss of f l u ~ r i n e , ~ and one distillation at 135" with perchloric acidremoves all the fluorine.For determining fluorine in impregnated wood,B. Ikert adds chromium acetate and calcium acetate solutions beforeashing, as this is stated to lead to a better recovery. In analysis of woolmoth-proofed with fluorine compounds, F. F. Elsworth and J. Barritt 'moisten the wool with sodium carbonate solution, and ignite below aJ. M. Kolthoff's colorimetric reagent53 D. A. Harper and N. A. Strafford, J . SOC. Chem. Ind., 1942, 61, 74.Ind. Eng. Chem. Anal., 1933, 5, 7.H. Wulle, Z. physiol. Chem., 1939, 260, 169.Ind. Eng. Chem. Anal., 1934, 6 , 118.Chem.-Ztg., 1939, 63, 754.a Ibid., 1934, 6 , 134.5 €I. G. Rempel, ibid., 1939, 11, 378.7 Analyst, 1943, 68, 298284 ANALYTIOAL UHEMISTRY.red heat.For fluorine in coal (26-150 parts per million), H. E. Crossley 8burns it with sodium carbonate at a low temperature, or ignites i t in acalorimetric bomb. I n the first case he follows combustion by a fusion,separates silica as usual, and distils. In the second case interference bynitrates is overcome by reduction with a zinc-copper couple before distill-ation. The distillates are compared visually with standards, the authorcorrectly observing that the zirconium lake colours are not suitablc forabsolute colorimetry. Calcium or magnesium peroxide is suggested' 9 asan adjunct in the ashing of soils. In the presence of much silica, magnesiumperoxide leads to low recoveries, but in presence of much organic matter andlow silica content, it is preferred to calcium peroxide, and leads t o a lessviolent reaction.Magnesium acetate may be used as an ashing agent ;loit is free from fluorine and gives very concordant results after ashing a t570'.DistiZEation.-This has been very thoroughly studied. Recovery offluorine decreases with increasing volume in the distilling flask, and thepresence of non-volatile acids hinders the distillation; the effect of volumeis less marked with sulphuric acid, volatilisatio? is slowest with phosphoricacid, and recovery is more rapid at higher temperatures.11 D. S. Reynoldsand his co-workers l2 describe a steam-distillation in which the temperatureis regulated by blowing in steam rather than by dropping in water, and itsapplication to phosphate rock.The process is typical; 0.5 g. is distilledwith 15 ml. of 2 : 1 perchloric acid, the temperature being maintained a t125-150". Ifpyrites or organic matter is present, distillation is preceded by oxidation withpermanganate. I n the presence of colloidal silica or alumina, which obstin-ately retain fluorine, it is necessary 13 to make a preliminary distillation withsulphuric acid at 165".Titration.-The use of thorium salts for this purpose has been thoroughlystudied. The amount of indicator in Willard and Winter's titrationwith N/lOO-thorium nitrate must be kept constant,14 and changes in pHcause errors proportional to the change and to the fluorine content. Thebest pH is 2-5-3.0, and a back titration is preferred.By using a buffer ofhalf-neutralised N-chloroacetic acid, pH 3, R. J. Rowley and H. V. Churchill l5avoided the use of alcohol, and titrated with N/lO-Th(NO,),, extending therange considerably. R. A. Clifford l6 measured the colour during the titra-tion, using a photometer, and showed that for very small quantities there is9 W. H. MacIntyre and J. W. Hammond, J . Assoc. Off. Agric. Chem,, 1939,22, 231.10 W. E. Crutchfield, jr., Ind. Eng. Chem. Anal., 1942, 14, 57.11 D. Dahlc and H. J. Wichmann, J . Assoc. 08. Agric. Chem., 2936,19, 303; 1937l2 Jbid., 1936, 19, 156; I d . Eng. Chern. Anal., 1939, 11, 21.l 3 D. Dahle and H. J. Wichmann, J. Assoc. Off. Agric. Chm., 1936 .19, 320.14 D. DaNe, R. W. Bonar, and H.J. Wichmann, ibid., 1938,21,469.l6 Id. Eng. Chem. Anal., 1937, 9, 551.lA R. A. Clifford, J . Assoc. Ofl. Agric. Chem., 1940, 23, 303.Recovery is complete when 150 ml. have been distilled.J . Soc. Chem. Ind., 1944, 63, 280.20, 297WILSON : FLUORINE. 286no “end-point,” but a gradual change. He recommends titration to anintermediate colow, adding the same amount of 0.0004~-Th(NO,), solutionto a “ blank ” containing the same quantity of indicator, and back-titratingwith a standard (1 ml. = 0-01 mg. F) solution until the colours match.The difficulties of titrating very small amounts are reviewed. l7 Fluorineconcentrations of 2-50 pg. and 0.2-5 mg. per 10 ml. were studied a t pH 3 inpresence and in absence of chloroacetic acid, and the thorium nitrate was0.0175~.In alcoholic solution in the lower range, fair agreement wasobtained, but the buffered solution gave high results; in aqueous solutionsall results were too high, the buffered solutions again being the higher. I nthe 0.2-5 mg. range all the systems gave good results. Micro-quantitiesof fluorine in aqueous solution must thus be titrated by empirically standard-ised solutions, but for alcoholic solutions the stoicheiometric factor can beused.Colorimetric Methods.-Numerous methods exist for colorimetric esti-mation of small amounts of fluorine, apart from the “ titralion-colori-metry.” 16# 260 27 Except in waters, it is almost always necessary to isolatethe fluorine by distillation. The development of the zirconium-alizarinlake method is shown by three papers.In waters,1* the usual amounts ofmanganese, aluminium, iron, silicate, sodium chloride and sulphate have verysmall effect, but ferric iron and phosphate completely vitiate the procedure.The sample is made acid with sulphuric and hydrochloric acids, a solution ofzirconyl nitrate + alizarin-S added, and the whole heated to boiling andallowed to stand overnight. It is compared in Nessler cylinders with standardssimilarly treated. W. L. Lamar and C. G. Seegmiller l9 describe a procedurein which only sulphuric acid is added to the water, which is not boiled, butallowed to stand overnight with the reagent, before comparison with stahd-ards containing 0-02-0-24 mg. of fluorine. What is perhaps the bestmethod for fluorine in small amounts in waters, etc., is given by A.P. Blacket aL20 and by R. D. Scott.21 The reagent contains alizarin43 and zirconiumoxychloride, made 1 . 5 ~ with regard to both hydrochloric and sulphuric acid,thus reducing to a minimum interference from chlorides and sulphates. Thecolour is bleached by fluoride ions, and the best range of standards is 0.01-0.1 mg. of fluorine per 100 ml., but it can be extended to 0.18 mg. Iron andaluminium both interfere if present to more than 0.5 parts per million. Thereaction is complete in 2 hours, and the reagent keeps very well. N. A.Talvitie 22 suggests a thorium nitrate reagent buffered at 3.5, and containing0.008% of alizarin-S. The sample is neutralised with dilute nitric acidbefore applying the reagent.The method is simple and rapid, and woulddetect 0.1 part per million on a 100-ml. sample.Two papers utilise the bleaching of the iron complex with 7-iodo-17 J. W. Hammond and W. H. MacIntire, J . Assoc. Off. Agric. Chem., 1940,23, 398.l9 Ind. Eng. Chem, Anal., 1941, 13, 901.*O J . Amer. Water Works ASSOC., 1941, 33, 1965.22 Ind. Eng. Chem Anal., 1943, 15, 620.0. J. Walker and G. R. Finlay, Canadian J . Res., 1940,18, 151.21 l b i d . , 1941, 33, 2018286 ANALYTICAL CHEMISTRY.8-hydroxyquinoline-4-sulphonic acid (" ferron ") by fluoride ions. Forfluorine in rocks,23 the sample is fused with alkali, and silica is separated bya modified Berzelius-Rose method, finally with zinc oxide and ammoniumcarbonate. An aliquot of the filtrate is taken, and a standard preparedcontaining the same concentration of salts. Each is treated with 2 ml.ofthe ferric " ferron " reagent, and the standard titrated with a ~/50-sodiumfluoride solution until the colours match; a difference of 0.05 mg. of fluorineis readily perceptible, and the range is 0-1-1-5 mg. Alternati~ely,~~ thefluoDine may be isolated by a preliminary distillation at 165" with sulphuricacid, the distillate being neutralised, concentrated, and redistilled at 135"with perchloric acid. An aliquot containing < 5 mg. of fluorine is treatedwith the reagent, and its extinction measured with a photometer, using a redfilter. In the peroxytitanicsulphate method 25 aluminium ions largely counteract the bleaching effectof fluorine on the colour ; if the same amount of pertitanate solution is addedto each of two aliquots, and aluminium to one of them, provided the pH benot changed, the difference in colour will be proportional to the fluorinecontent, irrespective of the colour of the original solution.The twomethods are similar; in the English method on a I-g.sample, 1 part permillion can be determined with an accuracy of 0.15 part; in the American,the best amount of fluorine to have present is 30-70 pg. The originalsmust be consulted for details, but reagents must be prepared especially tobe free from fluorine, and a little (< 1.5 pg.) always comes from the glass.Thorium nitrate solution (0-025%) is added to the final distillate to a faintpink colour, the same volume added to the " blank," and then standardsodium fluoride solution (1 ml.= 0.01 mg. F) to the blank until the coloursmatch. Further applications are recorded in various papers.28Miscellaneous.-The lead chlorofluoride method is applied to insecti-c i d e ~ . ~ ~ For the determination of fluoride in organic compounds two pro-cedures are given. A new technique is described by P. J. Elving and w. B.LigetL30 The compound is heated in a closed tube, like a Carius tube, withmetallic potassium cut into small pieces, air having been displaced by addinga few ml. of ether, which is sucked off as vapour, removing also water.After exhaustion, the tube is sealed, and heated to 400". After cooling,excess of potassium is destroyed by ethanol, the residue dissolved, thesolution filtered, and fluorine determined, e.g., as lead chlorofluoride.M.L. Nichols and J. S. Olsen 31 commence by fusion with sodium peroxide,potassium carbonate, and sugar in a Parr bomb; after extraction andneutralisation, fluorine is titrated potentiometrically with cerous nitrate.23 J. J. Fahey, Ind. Eng. Chem. Anal., 1939, 11, 362.24 P. Urech, Helv. Chim. Acta, 1942, 25, 1115.25 D. Dahle, J . Assoc. Off. Agric. Chem., 1937, 20, 505.2 G Society of Public Analysts Sub-committee, Analyst, 1944, 69, 243.2 7 Anon., J . Assoc. Off. Agric. Chem., 1944, 27, 90.2 * P. A. Clifford, ibid., p. 246; Anon,, ibid., p. 98.30 Ind. Eng. Chem. A n d . 1942, 14, 449.The fluorine content is read from a graph.There are two important papers on fluorine in foods.26#27Anon., ibid., p.75.3 l Ibid., 1943, 15, 342HERON AND WILSON : ORGANIC MICROCHEMICAL ANALYSIS. 287The concentration of fluorine in air is determined automatically byaspirating i t through a solution containing ferric chloride, potassium thio-cyanate, and persulphate. Fluorine bleaches the red colour, and the changeis measured photoelectrically.32 Full details of apparatus and procedure forsampling anhydrous hydrogen fluoride are given.33.34 It is then dilutedby addition to ice in a special weighing vessel, after which the solution isanafysed for hydrofluoric, hydrosilicofluoric, sulphurous and sulphuric acids.Qualitative Tests.-J. Fischer 35 adds to the neutral fluoride solutioneosin, lanthanium acetate, and sodium acetate ; on boiling, cooling, andcentrifuging, a red precipitate indicates fluorine ; 2 pg.can be detected.E. R. Caley and J. M. Perrer 36 describe apparatus for a micro-etching test,and suggest the preparation of a series of standard etched microscope cover-glasses, the amounts dealt with ranging down to 0.05 mg. of calcium fluoride.IV. Organic Microchemical Analysis.H. N. W.In 1942 an excellent review of the recent literature was published,land the following is a brief note on progress since then. The determinationof specific compounds is not reported, but only analysis for elements orradicals, with notes on new apparatus.The American Chemical Society recommendations for apparatus for thedetermination of sulphur and the halogens (Pregl’s methods) have beenpublished,2 and G.H. Wyatt 3 reviews all kinds of micro-volumetric appara-tus. The error of a single weighing on a microbalance is 3 pg., most ofwhich is due to the placing of the rider, but results are stated to be describedas surprisingly the standard deviation being 3.4 pg., and the largestemor to be expected from a single weighing being 7 pg. isdevoted to the errors of a Kuhlmann balance ; the tracing of errors is described,and every possible error due to environment, construction, wear, design, etc.,is considered. Random errors amount to about 5 pg., and suggestions aremade to reduce this to 1 pg. Simple apparatus are described for micro-ammonia distillation 7 and for semimicro-alkoxyl determinations,* preferablyby F.Viebock and C. Brechner’s m e t h ~ d . ~ The latter apparatus can alsobe used for wet methods of halogen determination. A V-shaped micro-pyknometer made of capillary tubing, and with a capacity of 0.01-0.02a2 L. S. Tschemodanova, Zavod. Lab., 1939, 8, 1248.33 U.S.A. Manufacturing Chemists’ Association, Ind. r i g . Chem. Anal., 1944, 16,84 C. F. Swinehart and H. F. Flisik, ibid., p. 419.35 Z . anal. Chem., 1936, 104, 344.A long article483.a8 Microchim. Acta, 1937, 1, 160.L. T. Hallett, Ind. Eng. Chem. Anal., 1942, 14, 956.G. L. Royer, H. K. Alber, L. T. Hallett, and J. A. Kuck, ibid., 1943, 15, 230.Analyst, 1944, 69, 81.M. Corner and H. Hunter, ibid., 1941, 66, 149.Committee on Micro-balances, American Chem. SOC., Ind.Eng. Chem. Anal., 1943,A. H. Corwin, ibid., 1944, 16, 258. R. Markham, Bwchem. J . , 1942, 36, 790T. White, Analyst, 1943, 68, 366.15, 415.@ Ber., 1930, 63, 3207288 ANALYTICAL CHEMISTRY.ml., is described; lo it can be mounted on a board carrying 2 scales, andcalibrated with the liquid a t various levels, read on the scales, and a graphplotted to show volumes a t various levels in the arms. Results are quotedcorrect to 3 or 4 digits. Methoxyl and ethoxyl groups may be determinedin the same molecules ; l1 the two alkyl iodides are collected in a 10% solu-tion of frimethylamine in alcohol, three receivers in series being used. After50 minutes the receivers are washed out with alcohol and water, thesolution evaporated to dryness, and trimethylethylammonium iodide ex-tracted with a saturated solution of tetramethylammonium iodide inabsolute alcohol.The tetramethylammonium iodide thus remaining isoxidised to iodic acid, and determined as usual. Results quoted forsamples of 10-15 mg. are good. An improved semimicro-method foralkoxy-groups in cellulose ethers, etc., is given.12 A. A. Houghton la describesa modified apparatus for simple and rapid working of Viebock’s method ; theapparatus requires very little attention and is suitable for repetition work.An analysis is completed in less than an hour, and errors on 10-mg. samplesare insignificant.Hydroxyl groups are determined 1* by weighing 2-10 mg. into a melting-point tube sealed a t one end, followed by a weighed amount (20-25 mg.) ofacetic anhydride, and excess (not weighed) of pure pyridine.The tube issealed, centrifuged to mix the reagents, and allowed to stand for 24 hours.It is broken under water, and excess of acetic anhydride titrated with 0.04.~-sodium hydroxide.E. W. Peel, R. H. Clark, and E. C.Wagner l5 fuse 15-20 mg. of substance in a Parr micro-bomb, with sodiumperoxide ; they prefer gravimetrio procedures, and weigh liquids in gelatincapsules rather than capillary tubes. A. F. Colson16 describes a modifiedapparatus for volatile compounds, for which the Parr bomb is unsuitable.In it modified Pregl’s apparatus, the compound is volatilised in a stream of airor oxygen, the stream passing through a heated platinum cylinder to act asa catalyst. In the cylinder is a glass rod, which forces the gas stream throughthe small annulus between the rod and the cylinder, thus ensuring efficientcontact.The chlorine is absorbed in solid sodium carbonate and granularlimo, the mixture dissolved in dilute nitric acid, reduced with hydrazine,and the chloride finally weighed as silver salt. Excellent results are quoted.P. J. Hardwick l7 determines bromine in biological fluids by heating in asealed tube with sodium ethoxide solution, evaporating the product, andafter gentle calcination, dissolving the residue, neutralising it, and oxidisingit either (i) to bromate with hypochlorite or (ii) to bromine with chloramine-TA blank is run simultaneously.Several papers treat of halogens.lo A. A. Houghton, Analyst, 1944, 69, 346.l1 L.M. Cooke and H. Hibbert, Id. Eng. Chem. Anal., 1943,15, 24.l2 E. Y. Sarnsel and J. A. McHard, ibid., 1942,14, 750.l4 J. W. Peterson, K. W. Hedberg, and B. E. Christensen, Ind. Eng. Chem. Anal.,l5 Ibid., 1943, 15, 149. 16 Analyst, 1942, 87, 47.Analyst, 1944, 69, 363.1943, 15, 225.1 7 Ibid., p. 223HERON AND WILSON : ORGANIC MICROCHEMICAL ANALYSIS. 289in presence of fluorescein, the red colour (eosin) produced being a measureof bromine present.G. Ingram18 deals with the uses of mercuric oxycyanide in micro-volumetric analysis. In neutral solution, this reagent reacts with halogensalts to liberate an equivalent quantity of hydroxyl, to be titrated with ~/100-sulphuric acid. He describes applications to the determination of halogens,a special apparatus for the combustion being described.By using oxygensaturated with water, the formation of sulphur trioxide mist from sulphurcompounds is avoided. If both sulphur and halogens are present, total acidformed may be titrated as usual, then halogen by means of the oxycyanide.The reagent may also be applied to the determination of alkoxyl groups.W. B. Price and L. Woods l9 describe a technique for the analysis ofminute bubbles of gases, e.g., from the bubbles in glass. The bubble ismeasured under a slide in glycerol by means of a microscope, and aftertreatment with reagents, measured again. The smallest bubble can be 0-2mm. in diameter or 0.004 mm. in volume, and hydrogen, oxygen, hydrogensulphide, carbon monoxide and dioxide can be determined.A semimicro-apparatus is described 20 for determining Reichert, Polenske,and Kirschner values in fats.The formyl group may be determined 21 similarlyto acetyl, except that the formic acid produced may be either titrated withalkali or oxidised by bromine in N/loO-solution, thus making the determinationspecific. The semimicro-determination of esters, which are heated to SO" inclosed 25-ml. flasks with 2~-sodium hydroxide in 90% methanol, is discussedwith reference to steric hindrance.22R. Belcher and C. E. Spooner 23 deal with attempts to hasten the usualLiebig or Prcgl form of analysis by using rapid air rates, and to simplify theapparatus by use of empty tubes. Originally put forward as a macro-method for the ultimate analysis of coal, the method embodies novel features.The coal sample is gradually introduced into a tube heated to 1350", and nopacking is used except a silver spiral at the cool exit end to absorb halogensand sulphur. The oxygen rate is 300 ml./minute, and a complete combustionoccupies ten minutes. Carbon, hydrogen, sulphur, and chlorine can bedetermined simultaneously. Micro- and semimicro-procedures are alsogiven. Combustion is conducted in an empty silica tube, a roll of silvergauze serving to trap halogens and sulphur ; oxides of nitrogen are trappedby ~/5O-potassium permanganate or dichromate in sulphuric acid. Theoxygen rate is 50 ml./minute. Carbon, hydrogen, and sulphur can bedetermined simultaneously, the last being dissolved off the silver spiral assilver sulphate, after which the spiral is again weighed. It is claimed thesemethods are more rapid and simpler than the standard procedure. The lastAnalyet., 1944, 69, 265. 19 Ibid., p. 117.2o B. Dyer, G. Taylor, and J. H. Hamence, ibid., 194.1, 66, 355.21 J. F. Alicano, Ind. Eng. Chem. Anal., 1943, 15, 704.22 J . Mitchell, D. M. Smith, and F. S. Money, ibid., 1944,16, 410.23 Fuel, 1941, 20, 130; Ind. Chem., 1943, 19, 653; J . , 1913, 313; R. Belcher, J .Inst. Fuel, 1944, 17, 160.REP. VOL. XLI. 290 ANALYTICAL CHEMISTRY.paper was followed by an interesting discu~sion,~~ the opinion being thatgood results were obtained by the macro- and the semimicro-procedure,but that on the micro-scale, trouble was caused by formation of oxides ofnitrogen.G. Ingram 25 describes a new method for determining carbon and hydro-gen. In a silica combustion tube the packing is placed in boats. Thecatalyst is a special mixture of ceria, litharge, silver dichromate, and silveroxide, and also a roll of copper gauze filled with a mixture of ceric andvanadic oxides. The combustion is conducted a t 500°, and is complete in50 minutes. Silver vanadate on pumice as the main oxidation filling issuperior to copper oxide and lead chromate, or to platinum contacts.A. E. H.H. N. W.A. E. HERON.H. N. WILSON.24 J . Inst. Fuel, 1944, 18, Suppl. p. 51.26 J . Soc. Chm. Id., 1942, 61, 112; 1943, 62, 175