THE ANALYST. 185 BRITISH PHARMACEUTICAL CONFERENCE. The following are abstracts oj‘ sonte of the Papers q*ead at the Meeting at Shefield in August last :- NOTES ON PETROLEUM SPIRIT OR “BENZOLINE,’’ BY ALFRED H. ALLEN, F.C.S, ALTHOUGH it is well known to professional chemists that petroleum spirit is composed of hydrocarbons quite distinct from those coiistitutiiig coal-tar usphtha, among the general public, and to il certain extent among people possessed of some knowledge of186 THE ANALYST.chemistry, great coihsioii has arisen as to tlie nature of tlie liquids hiowii iii commerce its (‘ benzine,” ‘‘ benzene,” beiizol,” niid 6 L beiizoline.” Of these, tho liyclrocarboii h i d or ~ C I L : P I ~ C , C,H,, is tlic cliief niid clinractcristic coiistituent of coal-tar naplitliit, vhile it is present in very iiisigiiificaiit amouiit in petrolcum spirit or mineral naplitlin.‘The terms bewine ancl b e x o l i n e liave no scientific application, ancl are merely commercial names for petroleum spirit. It is owing, in a great measure, to tlie similarity of these names to those of tlie cliief constituent of coal-tar naphtha that coiifmion lias occurrccl, h i t it has been made far worse by tlie acciclental or intentional substitution of one liquid for the other, until it is clifficalt to obtain the coal-ttir prodnct retail, even when it is asked for by its proper name.This would be-of but little consequence if the two liquids were of exactly similar naturc, but ili certain cases they present decided differences of beliavionr, although in general characters tliere are very close resemblances.The following tabular s t 5 t em en t of tlie character is t ic cliff ereii c c s bet me cii petroleum spirit and coal-tar naplitha has been compiled froin various sourcos ,znd iiicludes a few original tests. All the characters given have been carefully verified by acturtl experiment on representative samples of commercial petroleum spirit and coal-tar benzol.Petroletm Spirit, ‘‘ bcnzoliire,” or ’‘ beniiiic.” 1. Consists of hptnne, C7111 $, aiiJ its honio- 2. Heptane ccntains 84.0 per cent. of carbon. 8. Burns with a somewhat smoky flame. 4. Commences to boil a t 619 to GOD C. 6. Specific grnvity about a 6 9 to *72. G . Smells of petroleum. 7. Dissolvcs iodine, forming a solutioii of a lognes.rsqberry rcd colour. 8. Does not sensibly dissolve pitch, auci is s-arcely coloui.ed by it, even on proloiigecl contact. 9. Wlien slii&en in the cold with oce-third of its volume of fused crystals of absolutc carbolic i-xid, tlie latter reniniiis undissolved. 10. Requires two volumes of absolute alcohol, or four or five volumes o€ inetliylated spirit of -838 specific gl’avity for coinplete solution a t the ordinary temperature.Cod- Tll)’ NclphtlLn, 0)’ ‘( lMl1201.” I. Consists of benzene, C,H,, and its honio- 3. Benzene contains 02.3 per cent. of carbon. 3. Curils with a very smoky flame. 4. Commences to boil at, about 80° C. 5. Specific grnvity about ~88. 6. Smells of coal-tar. 7. Dissoires iodine, forming a liquid having the colour of a solution of potassium perinan- ganate.8. licaclily dissolves pitch, forming a deep browi soln tioii. logucs. 9. Miscible with nbsolnte carbolic acid in all yroprtions. 10. Miscible with absolute alcohol in all pro- portions. Forins tl homogeneous liquid with an equal measure of methylated spirit of -825 specific gravity. Althongli the foregoing tests are abundantly sufficient for the distiuzti0:l of petroleum spirit aiici bmzol, wlien applied to mixtures of the two products they aye of but little value even as qualitative indications, and in that case the density is the only one of thc above characters which is capable of giving even an approximatiou to the quantities in whicli the coiistitueiit liquids are mixed.Tlic actioii of nitric acid 011 coal-tar naplitlin is well known to result h the formatioil of iiitrobenzerie niid its liomologues, and litts been employed by Schorlemmer for detecting traces of benzenc, kc., in petroleum.On the other hand, tlie action of nitric acid on the liydrocsrbons of tlie paraffin series, which constitute practically the whole of petroleum spirit, is almost nil in tlic cold, even if fiming acid be used, aiid is veryTHE ANALYST.187 limited iii csteiit if liot acid be employed, p:.oviC!ccl that tlie very strongest be avoitled. I found by experiment tlist the action of nitric acid on petroleum spirit was, uiider certain conditions, even more limited than I had stpposed, and eventually I found that by employing the acid in a particular manner, it was not only possible but easy to effect a tolerably perfect quantitative separation of coal-tar naphtha aiid yetrolenm spirit.The following was the moiIe of treatment eventually employed-, and, by adliering to it, it is possible to detect and approximately estimate the proportion of petrolemu spirit existing in a mixture of it with benzol, in ten or fifteen minutes : A known measure of the sample (from 4 to G c.c.) vas treated with four times its measure of yellow nitric acid of 1.45 specific gmvitj-.Tlic mixture i ~ n s mads in n flask, to which a coiideiising arrangement was attached. Slight licat was npplid externally by ixeans of a flame if the spontaneous action mas not sufficiently vigorous. After about five miniites tlic contents of the flasli were cooled, and then poured into a iiarrow graduated tube.Any oily layer was measured and removed with a pipette, and the remaining liquid poured into a large excess of water. When ordinary petroleurn spirit is thus treated the nitric acid becomes coloured more or less brown. Very little heat is evolved, but on applying moderate liest externally tlie production of red fumes proves the occurrence of a certain amount of action.The effect, however, is not so violent as I had anticipated from a perusal of Schorlemmer’s description of the reaction, and, if the experiment be carefully made, the petroleum spirit employed forms a layer on tlie nitric acid, and 011 transferring tho liquid to a graduated tube is found to occupy the original volume of tlic sample usr~l. This is true whether petroleum spirit alone be operated on or wlietlier it be previously mixed with various proportions of coal-tar naphtha.If the proportion of the latter be large, the quantity of nitrobenzene forrne.1 is larger than can be retained in permaiieiit solution in the nitric acid. This fact causes no iiiconveiiience, for the nitrobenzene forms a separate layer bclow the petroleum spirit, and in preseiice of nitric acid is not miscible with it.Nitrobenzene and petroleum spirit are perfectly miscible alone, but on shaking the mixture with strong nitric acid tlie nitrobenzene is dissolved out. It will be seen, therefore, tliat the react’ion with nitric acid may be conveiiiently employed for the determinatioii of petroleum spirit in admixture with benzene, Very fair approximate results are obtainable. If t h e layer of petroleurn spirit be removed mitli a pipette and shaken with water to remove dissolved nitrous fumes, it is obttiined in a suitable condition for further exaixination.If, after removing the layer of unacteci-on petroleum spirit, the nitric acid solution be poured into water, a very sensible turbidity is usually produced, even with petroleum spirit free from coal-tnr prodncts, aiid on filtc~iiig of7 the precipitate, or allowing it to settle, and decanting the liquid, distinct evidence of tlie formation of nitrobenzene is obtainable by the aniline test.When the more volatile portion of petroleum spirit is thus treated, tlie nitric acid is scarcely coloured at all, and hardly a trace of milkiness is produced when tlie acid is poured into wfxr.Tlie brown colour and turbidity 011 dilution increase with the boiling point of the sample of naphtha, and are strongest with kerosine oil ; but in a11 cases in which petroleurn prodncts arc treated with nitric acid, the quantity of precipitate 011 ililution is very insignificnnt. A1 ttli~i,o1i nitrobenzene is188 T€IE ANALYST. recognizable among the products of the sctioii of nitric acid on petroleum spirit, as was shown long since by Schorlemmer, I do not think tlie turbidity prodncecl on dilution is clue solely to its formatioil.It' is probably rather due to tlie production of various nitro-substitutioii products, as it is wll known tliat the higher numbers of the paraffin series are far inore readily acted on by nitric acid tlian their lower liomologues.I lime attempted to determine tlie proportion of benzene in a mixtiwe with petroleum spirit, by measuring the nitrobenzene produced, but the results have not been satisfactory, partly in consequence of tlie solubility of nitrobenzene in water and acid liquids. By employing 250 C.C. of water for dilution, allowing the nitrobenzene to settle completely, decantiiig tlie greater part of tlie water, and pouring the '' bottoms '' into a graduated tube, the nitrobenzene may be readily measured.If an allonranee of 1.5 C.C. be madc for solubility in the one-quarter litre of acid liquid, the measure of benzene present in the sample taken may be rouglily ascertained by multiplying the number of C.C. of nitrobenzene obtained by tlie factor 0.85.Thus if V be the volumc in C.C. of nitrobenzene, then the benzene in tlie amouiit of szuple taken mas (V+ 1.5) x ~ 8 5 . The method is not capable of giving actual results, but may be useful in some cases as a check 011 tlie deterininatioii of petroleum spirit by m2asurement of the layer insoluble iii nitric acid. With a view of learning something respecting tlie proportion of lieptane present in ordinary petroleum spirit, I madc a mixtiire in equal measures of fonr samples of commercial '' benzoline," such as is ascd for sponge lamps.This mixed specimen had a density of -7001 a t 15.5O C., and commenced to boil at abont 5 4 O C. It was distilled in the manner first described by Warren, in a flask furnished with an inverted condenser, filled ivit1-L water, inaiiitained at a temperatnre of 70° C., a second condenser being kept well cooled by a current of cold water.The distillation ceased when tlie temperature in tlie flask was 840 C., that is, 14" C. above that of ihe first condensei-. Tlie water in the first condenser was tlien raised to the boiling point, and the distillatioii continned till scarcely any more came over, by vhich time the contents of the flask wore at 114O C.The result of tlie distillation was as follows-for 100 measures of petroleum spirit t a ,1 {en :- Condensed Condensed Condensed between above $$; 700 and IOOQC. loop C. (reRidne). Percentage by measure . . . . . . . . . . . . . . . . 16 58 26 Deiisity at 15.5Q C., compared wit11 watcr at mine temperature .. . . *G67 ,707 .r ('4, 0 Tlie loss ~ v a s abont 2 per cent. of tlic original incns~irc. Another sl~cciirieii of petrolenm spirit gave 224 measures of distillate with the receiver a t 70° C., 42 per cent, betjveen 700 and 1000 C., and 32 per cent. of residue ; thc loss being 3+ per cent. Tile densities of the tllree prodncts jyere a h o s t identical with those previously obtained. Tl1e observed specific gravities of tlie first and second distillates correspond approximately with tlie recordcd deiisities of liesaiie niicl heptsne, and from this and tlie knowii boiling points of these liqiiids it is evident that tlie portion of petrolcim spirit not coiiileiised a t 70" C. will consist chiefly of hexane and lover hornolognes, nTllile the part condensed at 700 C., but distilling at 100" C., will be chiefly heptancTHE ANALYST. 189 and isoheptane. As, in tlle experiments described, this fraction measured from 42 to 56 per cent. of the entire spirit, it is evident that the proportion of heptane present equals if it does not exceed that of dl the other constituents.':