102 PERKINS: THE POROSITY OF IRON.XII1.-- The Poi-osity of I r o n .By WILLIAM HUGHES PERKINS.IRON which has been immersed for some time in a solution ofalkali hydroxide and then thoroughly washed with water is statedby Dunstan and Hill (T., 1911, 99, 1853) to exhibit a certain‘‘ passivity ” towards nitric acid or copper sulphate and towardsatmospheric corrosion. Friend (T., 1912, 101, 50) believes thisphenomenon to be different from ordinary passivity, and assertsthat it is caused by the retention in the pores of the metal of asmall quantity of alkali, which is sufficient to prevent corrosion.He does not explain why it also renders the iron inactive towardscopper salts and nitric acid. The statement that alkali is absorbedis criticised by H. B. Baker (Annual Reports, 1911, 8, 31), whPERKINS: THE POROSITY OF IRON.103failed to confirm E’riend’s results, using rather more dilute solutionsof the alkali hydroxides. Before Baker’s criticism appeared thepresent author had been led to repeat Friend’s experiments, onaccount of what appeared to him to be a grave defect in the methodused to extract the retained alkali from the metal. After soakingKahlbaum’s iron foil in concentrated (6N) sodium or potassiumhydroxides and then thoroughly washing in a stream of distilledwater, Friend placed his metal in a shallow porcelain dish contain-ing distilled water. It is difficult to see how this procedure couldlead, in the case of sodium a t any rate, to satisfactory blank teste.To obtain distilled water which is free from sodium is no easymatter, and contact with the glaze of a porcelain dish will alwaysproduce a distinct flame reaction in quite a short time.It was feltdesirable, therefore, that the experiments should be repeated inplatinum dishes, and that an attempt should be made to obtainmore conclusive evidence, if possible, of a roughly quantitativenature.The first problem was a careful repetition of Friend’s experimentswith sodium hydroxide. His procedure was followed in every detail,except that platinum dishes were used for the extraction in placeof porcelain. It was possible inmost cases to distinguish between the test and the blank, but thedifference was not a t all striking. A satisfactory blank experimentwas not obtained, even when the distilled water was collecteddirectly from the tin worm of the condenser in the platinum basin.I n the case of potassium the flame test and spectroscopic test aremuch less delicate under ordinary conditions, but it was found that,on carefully concentrating the solution to about 0.1 c.c., there wasa distinct indication of potassium.It might be possible t o arriveat a more definite solution of the problem as far as it concernssodium and potassium hydroxides by the use of more highly refinedmethods and more complicated apparatus, but it appeared moreprofitable at this &age to extend the inquiry to other substances.Baker (Zoc. cit.) mentions the use of barium hydroxide, and thisalkali was chosen for the next experiments. The iron (Kahlbaum’siron foil) was immersed for three months in nearly saturatedbaryta water in an atmosphere free from carbon dioxide.It waethen well washed with distilled water until after soaking for fiveminutes, the solution gave no turbidity with sulphuric acid and noflame coloration. The metal was then immersed in dilute hydrechloric acid for an hour, the liquid being then poured off througha filter and tested for barium with sulphuric acid and by the flamecoloration. A distinct cloudiness and a green flame colorationwere always obtained from the test and none from the blankThe result was very uncertain104 PERKTNR: THE POROSITY OF IRON.solution. It is possible, however, that the barium may have beenretained on the surface of the iron as insoluble carbonate formedfrom the traces of carbon dioxide which could not be asmmed tobe absent from the metal.Further experiments were thereforecarried out xith lithium hydroxide. A saturated solution of thisalkali was mixed with about one-fifth its volume of boiled distilledwater. Pieces of iron 5 x 4 cm. in area were immersed in it forperiods varying from three weeks to six months, the vessel beingkept well stoppered. When the metal was removed it was wellwashed, first under the tap and then in a stream of distilled water,being subjected at the same time to vigorous rubbing either withthe fingers or with cotton-wool. This process occupied from ten totwenty minutes in each case. After a final thorough rinsing it wasplaced in a platinum dish containing about 5 C.C. of distilledwater and left for about twenty-four hours.The water was thenpoured off into a clean platinum crucible, evaporated down to aboutone-fifth of a c.c., and then tested by the flame test on a cleanwire. There was in all caaes a distinct coloration, which was notobtained in any of the blank tests. For the blank tests a pieceof the same iron was treated in exactly the same way, except thatit was not immersed in the lithium hydroxide. It is, of course,possible, on account of the relative insolubility of lithium carbonate,to advance against these results arguments similar to those usedin the case of barium. The solubility of lit’hium carbonate, however,is so distinct (more than 1 per cent. a t 1 5 O ) that i t is not likelyto have been retained as a surface deposit during such thoroughwashing.The substitution of electrolytic iron (Schuchardt) forthe iron foil did not modify the results obtained. A gold cruciblewhich had contained lithium hydroxide for some months, afterwashing well for five or ten minutes with running water, requireda daily change of water for more than four weeks before thespectroscopic test for lithium failed to show its presence in thewater. It is clear, therefore, that traces of alkalis, and presumablytherefore other solutions, are retained by metals in such a way thattheir extraction is a slow process of simple diffusion, and cannot behastened by shaking or even by gentle rubbing. Whether this isdue to actual porosity or to the formation of a surface layer isnot quite clear, but the “ absorption ” is obviously very slight.Toobtain some estimate of the absolute quantity an attempt was madeto obtain approximate figures, using ammonium hydroxide as thealkali and Nessler’s reagent as a quantitative indicator. A largepiece of iron foil about 500 sq. cm. in area (reckoning both sides)was well polished and cleaned, and then immersed for six weeksin concentrated ammonia solution. The washing was carried outfirst with distilled water, and $hen at the end wikh four changes oPERKINS: THE POROSITY OF IRON. 105ammonia-free water (500 C.C. contain less than 0.000002 milligramof ammonia). As a rule, the washing, after three or four prelimin-ary rinsings, required about fifteen minutes and about ten changesof water, the last four each remaining in contact with the metalfor two minutes.The metal was then covered with 500 C.C.ammonia-free water in another vessel and left for three days. Atthe end of this time all the water was distilled through anammonia-free condenser, and the distillate tested. Four experi-ments gave quantities of ammonia varying from 0*00002 t o0*00003 gram for 500 sq. cm. of metal. A quantity of this order ofmagnitude may or may not be sufficient to account for the anoma-lous behaviour of the iron which retains it, but it is doubtfulwhether one is justified in assuming that it is retained by actualcrpores’7 in the metal. I n connexion with ammonia one difficultywas kindly pointed out to the author by Professor Smithells.Itwas a t one time believed that ammonia was formed during therusting of iron, and this view has not, to the knowledge of thewriter, ever been confuted. The German edition of Berzelius’s“ Lehrbuch der Chemie ” (1834) contains the following paragraph(Vol. III., p. 427):“ I n trockner Luft oxydirt sich das Eisen nicht, um so rascheraber in feuchter und besonders bei Gegenwart von vie1 Kohlen-saure. Es entsteht hierdurch der sogenannte Rost, welcher einGemenge von kohlensaurem Eisenoxydul mit Eisenoxydhydrat ist.Das Eisen oxydirt sich dabei nicht bloss auf ICosten der Luft,sondern zugleich wird auch Wasser zersetzt, dessen Wasserstoff sichim Entstehungszustande mit Stickstoff aus der Luft zu Ammoniakverbindet.I n dieser Reaction besteht zwar nicht hauptsachlichdie Oxydations-Erscheinung, sie findet aber doch stets so unverkenn-bar staat, dass ein schwach gerothetes Lackmuspapier, welchm manin eine verkorkte Flasche aufgehangt hat, auf deren Boden sich mitWmser angefeuchtete Eisenfeilspahne befinden, nach wenigenStunden geblaut wird. Ein Theil des sich bildenden Ammoniaksverbindet sich rnit dem Eisenoxyd, und so enthalt auch sonderbarerWeise alles mineralisch vorkommende Eisenoxyd, 60WOhl das ausden Urgebirgen, als das aus jungeren Formationen, Spuren vonAmmoniak, welches in Destillations-gefiissen ausgetrieben werdenkann.* Diese Verhaltnisse sind zuerst von Chevallier beobachtetworden.”It appears, therefore, if this view is correct, that a quantity ofammonia may be produced in the period during which the ironis immersed in the distilled water when it undergoes a, gooddeal of corrosion. Carefully conducted experiments have shownThis statemeut is repeated by MeiidelBev (Principles, 1891, Vol. II., 318)106 SEGALLER THE RELATIVE ACTIVITIES OFthat the ammonia produced during the rusting of large pieces ofiron foil, which lost from 0.4 to 0.6 gram in weight in the process,was less than 0-000002 gra.m, as determined by Nessler’s reaction.The experiment with litmus paper as indicator could not berepeated, whilst the geological evidence, bearing in mind the natureof iron oxide, is not very surprising.It may be remarked, in conclusion, that in all the author’s experi-ments the iron which had been soaked in alkali and then wellwashed, rusted much more irregularly, and was more liable topitting than iron which had not been so treated.THE UNIVERSITY, LISEDS