2’7’16 FRIEND AND MARSHALL : THE CORROSION OF IRON AND ITSCCIIX.-l’he CoryosiorL of Iron cind its AppLicutiwto Determine the Relc&v Stwiqths of Acids.By JOHN ALBEI~T NEWTON FRIEND and CHARLES WILLIAMMAR SHALL.IT is now a matbr of common knowledge that fairly concentratedsolutions of sodium hydroxide or carbonate will inhibit the corro-sion of iron. I n 1911 attention was drawn to the fact that thecorrosive action of aqueous solutions or” metallic salts of any con-centration may be completely inhibited by the addition of sufficientalkdine hydroxide (Friend, J . Iroth and Steel Inst., 1911, 111).I n view of the close connexion between corrosion and ionisation itseemed of interestl to determine the minimum quantities of alkalirequired to inhibit the corrosive actions of different salts and todiscover whether or not any relationship exists between these quan-tities and the relative strengths of the acids and bases constitutingthe salts.To this end numerous experiments were carried out by exposingsinall pieces of iron to the action of different salt solutions mixedwith varying quantities of sodium or potassium hydroxide; but theresults were too uncertain and irregular t o be of any value.Thiswas ultimately traced to the presence of carbon dioxide in the air,which was readily absorbed in irregular quantities by the alkalinehydroxides. With the carbonates and borates of the alkali metals,however, very trustworthy results were obtained, capable of exactrepetition an indefinite number of times. Ultimately, therefore,these were used as inhibitors, the carbonates proving the moreconvenient both on account of their simpler constitution and theirgreater solubility in waterAPP1,TCATlON TO DETERMINE RELATIVE STRENGTHS OF ACIDS.27.7 7The procedure was as follows: Five C.C. of ;I N/lO-solution of asodium salt were introduced by means of a pipette into each oneof a series of five hard-glass tubes containing 1, 2, 3, 4, and 5 C.C.of standard sodium carbonate solution respectively. The tubes hadbeen previously steamed out in order to remove any traces ofsoluble matter-a precaution that was found to be absolutelynecessary. The volume of each solution was now made up to10 C.C. by the addition of freshly distilled wates. Finally, pieces ofpure iron foil, measuring 1 cm.square, which had been wellscrubbed with old emery paper and not touched with the fingers,were added, one to each tube, and the latter was sealed with awaxed cork. It was found important to add the iron last of allafter thorough mixing of the solutions employed, as contact withthe different solutions before mixing appeared to exert some influ-ence on the surface of the metal rendering the results uncertain.For the same reason the pieces of iron were never used twice(Friend, T., 1912, 101, 50). The sealed tubes wer0 kept in acupboard a t unifprm temperature, and subject in the daytime toweak, diffused sunlight. After two or three days they wereexamined. The iron in the first one, two, three, or four tubes wasthen usually found t o be corroded, but that in the remaining tubesor tube was not.The tubes could now, as a general rule, be kept for months with-out any further pieces of iron undergoing oxidation.The corrodedmetal in cases distant from the end-point gave signs of furthercorrosion; but those close to the end-point appeared to undergono change. This is remarkable in view of the fact that only asmall fraction of the oxygen in the air in the tubes had beenabsorbed during corrosion. Sulphides yielded a very clear end-point during the first few days, but after prolonged keeping ironsthat had previously shown no tendency to corrode were found tobecome oxidised, thus altering the end-point very considerably.This was ultimately traced t o slow oxidation of the sulphite tosulphate.By repeating tho experiments with intermediate quantities ofcarbonate it was found possible to determine to within about 5 percent. the amount of carbonate required to inhibit corrosion underthe particular conditions of the experiments.The presence of rustwas always easy to detect after a little practice, and it is notimprobable that results of still higher accuracy could be obtainedby the use of more finely graduated instruments.The best method of observation was found to consist in exposingt o a, powerful light, the tubes being continually rotated in orderthat the spots of rust should not be overlooked. These spots wer27’78 FRIEND AND MARSHATAT, : THE CORROSION OF IRON AND ITSoften minute, particularly near the end-points in the case of thestronger acids, frequently occurring a t corners of the inetal iiicontact; with the glass.I n the case of the weaker acids the voliinleof rust produced was, as a general rule, considerably greater,rendering the end-point more easy t o detect.With phos-phates it was white, with an under-layer of green in very dilutesolution. Sulphites gave a green rust, which became red on longkeeping. Chlorides, nitrates, and sulphates yielded reddish-brownrust ; iodides, black.The results obtained with sodium salts, using sodium carbonateas inhibitor, are given in tables I and 11. The concentrations ofthe various salts are expressed in terms of a normal solution, whilstthe amounts of sodium carbonate required t o be present in 10 C.C.of such solutions in order just t o inhibit corrosion are, for the sakeof convenience, expressed both as C.C.of a molecular solution andrelatively t o one another, the highest amount being taken as 100.In actual practice; in the #/lo0 tests, 1 C.C. of a N/lO-solution ofthe salt was taken, varying quantities of the carbonate solutionadded, and the volume made up to 10 C.C. T ~ U S a better comparisonwits obtainable than by making a new solution of the salt.The rust formed varied very much in appearance.TABLE I.Sodium Carbonate asSodium salts, ConcentrationN/20. of Na,CO,.Chloride. ........................... 1-35Iodide .............................. 1.20Bromide ........................... 0.975Nitrate ............................0.725Sulphate ........................... 0-700Fluoride ........................... 0.525Acetate ............................ 0.120Sulphite.. .......................... 0.025Inhibitor.Relativeconcentrationof N+CO,(NaC1=100).10088.972.253.761.8538.98.91.9Relativestrengths offree acids(HC1= 100).1009898987010.5268TABLE 11.Sodium Carbonate as Inhibitor.Sodium salts,NI100.Chloride ............................Iodide ..............................Bromide ...........................Sulphate ...........................Nitrate ............................Fluoride ...........................Acetate ............................Sulphite ............................RelativeconcentratioiiConcentration ofof Na,CO,. Na,CO,.0.526 1000.500 95.20.475 90.50.400 76.20.376 71.40-376 71-40.070 13.30.035 6.7Relativestrengths offree acids inN/lOO-solution.1009897.582.5981748APPLICATION TO bETERMINE RELATIVE STRENGTHS OF ACIDS.2’979Consideration of the above tables reveals the interesting fact thatwhen the salts are arranged in descending order of inhibitingcarbonat’e concectrations, not only are they in the order of thedecreasing electrica1 conductivity of their acids, but the relativequantities of carbonate solution bear a general relationship to thenumerical values found for the strengths of the acids by electricalconductivity and hydrolysis methods.It will also be observed that on increasing the dilution fromN / 2 0 to N/lOO the relative aniounh of the carbonate required bythe different salts steadily approach that required by the sodiunichloride, which is taken as the standard. I n other words, therelative strengths of the acids tand to approach equality withdilution-as the ionic theory requires.It is interesting to note thatin the more dilute solutions the nitrate and sulphate exchangepositions. The low position of the nitrate in the N/20-solution isremarkable, and one is led to ask whether or not passivification hasinfluenced the results, nitrates being well known passivifiers of iron.Still more exceptional are the values found for the sulphite iniV/ZO- and N/lOO-solution. The figures given in the final coluinnsof tables I and 11, however, are based on freezing-point measure-ments, and are therefore less strictly comparable.The results obtained with sodium borate as inhibitor were, onthe whole, very similar. The concentration of the sodium saltswas N/lOO, the’ inhibiting quantities of borax being expressed asC.C.of a molecular solution present in 10 C.C. of the inixed solixtions(see table 111).Concentrat,ion RelativeSodium salts, of concentrationN/100. borax. of borax.Chloride. ........................... 0-36 100Bromide ........................... 0.3 15 87.5Sulphate ........................... 0.3 1 86.1Iodide .............................. 0.33 91.7Nitrate ............................ 0.195 54.2Fluoride ........................... 0.100 27.SAcetate ............................0.070 19.4Sulphite ............................ 0.050 13.9Carbonate ........................ 0.045 12.6Relativestrengths offree acids inN / 100-solution.10097.582.5.089s174830.8”130tThe relatively low solubility of borax in water renders i t lessconvenient ss an inhibitor than sodium carbonate, by restricting* Assnming a dissociation constant Jc= 1.8 x 10W5 (Wdkcr and Corniack, T., 1900,77, 5 ) .t Assuming k=5 x (Thiel and Strohecker)2’780 FRIEND AND MARS€€AT,L : THE CORROSION OF IRON AND ITSthe range of coiiceutrations. When the above resiilh are compare( 1with those in table 11 of like concentration i t will be seen that,a reasonably close similarity obtains. With the exception of theiodide and bromide, which now interchange positions although thedifference between them is very slight, all the other acids retainthe same positions.It is interesting t o note that by this method avalue for carbonic acid is obtainable, which, however, is very high.Thiel and Stlrohecker (Ber., 1914, 47, 945) have recently adducedevidence in favour of the supposition that the true dissociationof carbonic acid is much greater than that usually ascribed to it,since only a small percentage of dissolved carbon dioxide existka ascarbonic acid in aqueous solutions of the gas.Assuming this to be correct, the high value for sodium carbonatein table I11 is readily understood, because practically the whole ofthe carbon dioxide is “ fixed,” and therefore ionised in the solution.Experiments with potassium salts, using potassium carbonate asinhibitor, yielded closely similar results t o those detailed in table I,in so far as the relat’ive quantities of inhibitor were concerned;but the absolute quantities were greater.The concentrations of the potassium salts were N / 2 0 , and intable I V the amount of potassium carbonate required to produceinhibition is expressed as C.C.of a molecular solution in 10 C.C.of the mixed solution.TABLE IV.Potassiitm Carbonate as Inhibitor.RelativeNj20. of K,C03 of K,CO::.Potassium salts, Concentration concentrationChloride ................... 1.85 100Bromide .................. 1.70 92Iodide ..................... 1.70 92Nitrate ................... 0.85 46Sulphate ..................0.775 42Fluoride .................. 0-675 31Acetate ................... 0.225 12Sulphite. .................. 0.050 2.7Relativestrengths offree acids inN/20-solution10097.59897.57010-5258Many salt solutions, such as sodium acetate, borate, carbonate,and sulphide, whilst capable of inhibiting corrosion when fairlyconcentrated, cannot do so if fairly dilute. It was of interest t odetermine accurately the concentrations a t which this auto-inhibi-tion just begins in the case of the above-mentioned salts, in orderto discover whether or not the same ionic relationship holds, as inthe experiments with an added inhibitor. This was accomplishedby immersing small pieces of iron foil in 10 C.C.of varying concen-trations of the different salts, and noting the lowest concentrationAPPLICATION TO DETERMIX E RE1,ATIVE STRENGTHS OF ACIDS. 2781a t which corrosion was pertiianently iiiliibiterl. Tlie results obtainedare given in t,aJJle V.TABLE V.A ut o-inhib i tion.Inhibitingconcentrat,ionof salt.Sodium salt. (i. 1Acetate ............. 0.30 NSulphite ............. 0.06 ,,Arsenate ............ 0-06 ,,I'liospliatc .......... 0- 05 , ,Carbonate ......... 0.02 1 ,,Borate ............. 0.009 ,,Relativeconcentrationofsalt takingacetic acid as8.9 (Table I).(ii.)8.91.81.81.50.62 a. 27Percentage Relativeionisation strengths ofof free acids free acidsat dilutions at dilutionsgiven in given in Col.1Col. 1 at 18". (HCl= 100.)(iii. ) (iv. )0.77 0.5755 5842 4450 530.55 0.570.07 0.07It is not strictly logical to compare column i with column iv,because in the former acetic acid is taken as the standard althoughits concentration is a t least five times as great as that of any otheracid, whereas in column iv _due allowance is made f o r the varyingdegrees of dilution. In the circumstances, however, it is the onlymethod possible, and since weak acids are being dealt with theerror will not be unduly great.It will be observed that the sulphite gives the same r e d t as bythe carbonate method (table I), and the concentration is practicallythe same. The result for sodium carbonate is much lower thanbefore, (table II), and approaches more nearly the usually acceptedvalue for carbonic acid.The difference in the concentrations in thetwo cases (tables I11 and V) is far too small to account for thealteration. The values for the arsenate and phosphate are inkeeping with that for the sulphite, the three free acids beingsimilar in strength according to conductivity measurements. Thelow figures given in table V, however, iudicate that the acids aremuch weaker even than acetic acid, which is not really the case. Onthe other hand, the valus obtained for the borate is four timesgreater than that f o r the free acid. It would appear, therefore,that' these results are not comparable with those obtained by theaddition of inhibitors.Series of experiments have also been carried out with the objectof determining the quantities of sodium carbonate required t$oinhibit corrosion in the presence of varying quantitiee of the sodiumsalts.The' results are given in table VI, the concentrat4ions of thesalts heing expressed in terms of normality, whilst the inhibitingquantities of sodium carbonate in 10 C.C. of the salt solutions are,f o r the sake of convenience, expressed as C.C. of a normal solution.VOL. cv. 8 2782 PEACOCK : ROTATORY POWER AND REFRACTIVITP. PART I.TABLE VI.Sodium salt. N . NJ2. NJ2.5. Nl3.3. Nf4. N/5. N/10. N/17. N/20. N/100.16.0 - - 8.0 - - 2.7 1.05 Chloride . . . . . . .Nitrate ....... 10.0 7.6 5.6 - - 3.6 - - 1.45 0.75Sdphete ...... 4.8 3.2 2.8 - - 2.4 - - 1.40 0.80Fluoride ...... - 1-4 - - 1.4 - 1-4 - 1.05 0.750-O* 0.05 0.07 Sulphite . . . . . . .0.24 0.14 - - - o.o* - - - - Acetate . . . . . . .I -- - - - - - -* Aut~J-inhibition.If these results are plotted diagrammatically a distinct rwem-blance can be traced to the specific conductivity curves of the freeacids when drawn from the data published by Kohlrausch. Unfor-tunately, owing to the limited solubility of the salts employed it isnot possible to employ high concentrations, and thus to pursue theanalogy further. It is interesting t o observe, however, that theacetate curve indicates the existence of a maximum effect,* similarto that observed by Kohlrausch, but a t a different concentration.No attempt is here made to explain these results. The authorswish, a t this stage of their work, merely to draw attention to theinteresting ccmnexion between corrosion and ionisation.In conclusion, the authors have pleasure in expressing theirindebtedness to the Research Fund Committee for a grant that hasdefrayed the greater part of the expense entailed by this work.YICrORIA IKSTITUTE,WORCESTER