464 PICRERING : THE COLOUR INTENSITY OFL.-The C'olour Intensity of I r o n and CopperCompounds. *By SPENCER UMFREVILLE PICKERING.THAT the so-called ferric salts of organic acids are re-ally ferri-compounds, similar to the cupri-compounds, with the metal in theelectronegative portion of the molecule, has been already provedby electrolysis (T., 1913, 103, 1359).One of the most marked features of the cupri-compounds is theirgreat colour-intensity, this being about nineteen times that ofcopper in inorganic salts. A similar feature exists in the case ofthe ferri-compounds, but since there are two atoms of iron in themolecule, two different intensity values are possible, one of themprobably about double the magnitude of the other, according asone o r both atoms are united with the carbon atom; the ferric salt,*C(OH),The facts established below bear out this view.Of the eightinstances available, four show a molecular colour-intensity' of(average) 20.2, and four, a value of (average) 10, (that of iron ina solution of the chloride containing 100 grams of iron per litrebeing taken as unity), and in three of the latter, a constant ofabout 20 is also recognised when the strength is increased beyondcertain limits.The ferri-compounds are found to be more stable than the corre-sponding cupri-compounds, owing t o quadrivalence not being anabnormal condition of the iron atom. Thus, i t is only in somecases, and when the solutions are concentrated, that the cupri-compounds contain the whole of the metal in this condition; dilu-tion (except in one or two instances) converting it into a cupricsalt, and, as will be shown below, the addition of the correspondingacid doing the same.With the ferri-compounds, however, thecolour-intensity is generally constant throughout a considerablerange of dilution, and is not affected by the addition of the corre-sponding acid.Where inorganic copper salts form blue solutions, the colour-intensity is independent of the nature of the acid radicle, of thestate of the dilution, and of the presence of excess of the corre-* For abstract, see P., 1913, 29, 192IRON AND COPPER COMPOUNDS. 465sponding acid, although, beyond a certain point, excess of suchacid alters the nature of the colour. With ferric chloride, nitrateand sulphate, the colour-intensity is nearly the same in the caseof concentrated solutions, and is approximately constant throughouta considerable range of concentration; there is, however, no realconstancy, and the values alter continuously on dilution, thusnecessitating the conclusion that a t least two forms of iron withdifferent colour-intensities are present.From the fact that excessof acid deprives the solutions more or less entirely of any yellowcolour, substituting for it a faint amethyst colour, i t is probablethat such is the real colour of electropositive iron, and that theyellow colour of ordinary solutions, and the variation of the colourintensity of these, are due to the presence of electronegative iron,that is, iron united directly with the chlorine, nitrogen, sulphur,etc., in the same way as i t is with the carbon in ferri-compounds.I n the rase of the sulphate, a constant value of 21 is attained ondilution, almost identical with that of Fez in the carbon compounds.The change of Fe into Fe, and vice versa, is immediate on alteringthe s t a b of dilution.With a large excess of acid added to the inorganic ferric salts,the yellow colour reappears, but is then of a lemon tint, and notcomparable with that of neutral solutions.When dilution is extended beyond a certain point (approximatelythe same in all cases), the solutions become unstable, and, frombeing nearly colourless when first prepared, they gradually changeto a deep yellow in the course of several days, the molecular colour-intensity with the weakest solutions attaining a constant value ofabout 137; this is due to the hydrolysis of the normal ferric saltinto free acid and colloidal ferric hydroxide.The latter, whenobtained by dialysis, gives a value of 134 for its colour-intensity.The transformation is not reversible.By heating solutions of certain concentr.ations, hydrolysis of theferri-compound present into another form of soluble hydroxideoccurs; this has a colour-intensity of about 280, and graduallychanges into the lighter hydroxide. Weak solutions of the organicferri-compounds are also hydrolysed by protracted heating intothis colloidal hydroxide.Solutions of the formate, acetate and propionate, whether con-centrated or dilute, appear to be similar to the most dilute solu-tions of the inorganic salts, and consist of acid phis colloidal ferrichydroxide; but by prolonged heating a t looo (if the solutions arenot too dilute), they are converted into the ferri-compounds witha colour-intensity of about 20.These are stable.+ 466 PICKERING) : THE COLOUR INTENSITY OFThe normal ferric organic salts, metameric with the ferri-com-pounds, have not yet been obtained as solids,* but are formed, insome cases (racemate, glycerate, malate, etc.) by the prolongedheating of fairly concentrated solutions of the ferri-compounds,the colour-intensity falling to what it is in inorganic salts. Theferric salts thus formed are unstable, and, in the cold, graduallyrevert to the ferri-compounds.I n other cases, the heating pre-cipitates a basic, or, possibly, sometimes, the normal salt.The ferri-compounds appear, in most instances, to dissolve excessof ferric hydroxide, just as the inorganic salts do, converting itinto colloidal hydroxide ; this has interfered with the examinationof the colour-intensity of the basic ferri-compounds of the typeR JI ,Fe2,F%03.Where the character of the colour is the same, and where con-stant values for the intensity are found to apply in a number ofdifferent cases, it seems certain that such const*ants must be con-ditioned by the character of the connexion of the iron with theother atoms in the molecule, that is, whether it is pseudo-tervalento r quadrivalent, whether it is in the electropositive or electro-negative portion of the molecule, etc.; but it appears equally certainthqat the character and number of the other atoms in the moleculemay affect the vibration of the iron atom, and hence producemodification in the intensity and quality of the colour. We may,therefore, expect that the constancy of the values will not alwaysbe absolute, and that instances will occur where differences oftint render comparison impossible. The oxalate, ferricyanides,thiocyanates, malate and ethylsuccinate are all instances wherethe tint is not comparable with that of the other compounds, nor,in most cases, with that of each other.Effect of Acids o n the COlOlLr of the Copper Compounds.The resalts entered in table I show that the addition of thecorresponding acid to an inorganic salt of copper is without effecton the colour-intensity, so long as the tint is unaltered.With thesulphate, this is so up to 2H,S0,:CuS04, and almost so up to12'5R,SO, : CuSO,;. with the chloride, there is no alteration upto 0.5HC1: CuCl,, but with 1.4HCl a difference in tint is notice-able, and with more acid, the well-known change to greenbecomes evident. With the nitrate, there is constancy up toO-THNO, : Cu(NO,),, beyond which a similar change begins. Thepoints a t which these changes start will vary, of course, with theproportion of water present.* Except the oxalatc, of which the ferri-form has not been obtainedIRON AND COPPER COMPOUNDS. 467TABLE I.-Effect o n the Colour-intenszty of Copper by the Additionof x Gram-molecular Proportions of the Corresponding Acidt o 64 Grams of Copper.Colour-intensity without extra acid taken a,s unity, in each case.The values in brackets are those compared with iron in the chloride.Grams ofcopperCompound.in 100 c.c.*Sulphate . ..................Chloride . . . . . . , . . . . . . . . . . . . . .Nitrate . . . . . . . . . . , . . . . . . . , . . .Formate . . . . . . , . . . . . . . , . . . . . .Acetate . . . . . . . . . . . . . . . . . . . . .Propionate . . . . . . . , . . . . . . . .Lactate . . . . . . . . . . . . . . . . . . , , .Glycollate . . . . . . . . . . . . . . . . . .Glycerate . . . . , . . . . . . . , . . . . . .¶, (cupriglycerate) . . . .Malate ...................... 0.51Maleate .. . . . . . . . . . . , . . . . . , . , 0.031 9Y Y9 9Y Y2.0.002 to 2.00.008 to 0.51-36 to 4.10.012 to 0.71-92 to 4.80.05 to 0.160-4 to 2.50.075 to 3.77.5 to 38.00.052.5 to 12.50.16 to 6.00.024 to 24.0’1.424.3 to 5.70.25 to 0.751.5 to 3.000.120.350.711.182.364.7200.11-02.03.56.011.000.51.02.34.0Relativecolour-intensity.11 (duller blue)10-95 to 0.80 tint altered10.9 tint altered1 (3.0)0-99 to 0.831 (6.0)0.89 to 0.861 (7-0)0-96 to 0-SOl(2.6)1 (2.7)0.88 to 0.840.840.860.620.400.370.360.331 ( > 6-0)0.820.620.520-510.420.391 (> 6.0)0.60.270.210.191 (2.5)1 (9.0)* The values apply to the strength of the solntion taken, and would be reduced inextreme cases by the dilution caused by the addition of t4e acid to four-fifths ofthose entered.With the organic salts, the addition of acid has generally someeffect, but not much in cases where the proportion of normal saltpresent is considerably greater than that of the cupri-form of thesalt ; thus, the formate, glycollate, crystallisable gl-ycerate andlactate, with colour-intensities of 3.0, 2.7, 2.5 and 2.6, respectively,of that of copper as sulphate, show, on the addition of the corre468 PICKERING : THE COLOUR INTENSITY O Fsponding acid, diminutions extending to 17, 16, 16 and 0 per cent.;whereas, with the malate, freshly-prepared glycerate and maleate,where the dour-intensity is 6, 9 and 6, respectively, and wherea large proportion of the salt must be present as the cupri-com-pound, the addition of acid reduces the colour-intensity to one-half, one-third, and to one-fif thy respectively.With the acetateand propionate, however, where the colour-intensity of 6 and 7indicates the presence of a considerable proportion of the cupri-compound (although this has not been isolated, as in the case of themalate and glycerate), the addition of acid produces only the smallreduction of 14 and 20 per cent., this exceptional behaviour beingdue, doubtless, to the weak character of these acids.Organic Iron Compounds.Instlances of organic compounds of iron being required for thisinvestigation, the behaviour of various acids towards freshly-precipitated ferric hydroxide in the dark was examined in the sameway as had been done previously in the case of tartaric, malicand citric acids, the acids being digested with the hydroxide atZOO until the action was complete (T., 1913, 103, 1364).I nprevious cases, it had been found that the hydroxide dissolvedentirely when the proportion taken did not exceed that requiredto form a normal ferric salt, thus affording evidence of the realityof the existence of such salts, or of their metameric ferri-compounds ;with more than one equivalent, some of the excess dissolved (exceptwith tartaric acid), but a certain amount of basic insoluble saltwas also formed, and, as the proportion of hydroxide taken wasincreased, the quantity of basic salt formed also increased,so that the amount of iron remaining in solution actuallydiminished with an increase in the amount taken.The course ofthe action was shown to be the formation of a soluble ferri-com-pound of the f ormu1.a [Rr/3R///,]Fe,,F%0,, and its subsequentgradual transformation into an ordinary insoluble basic salt ofsimilar composition.I n the first twelve instances quoted in table 11, one equivalentof hydroxide dissolves completely, or nearly so, in one of acid;with the ethylsuccinate, only 0.5 equivalent dissolves, and thesolution, therefore, must contain an acid compound. When morethan one equivalent is taken, there is generally some increase atfirst in the amount dissolved, followed by a reduction, and theformation of a larger proportion of insoluble basic salt.Theincrease, however, is inappreciable in the case of the tartrate,oxalate and ethylsuccinate, whilst the subsequent decrease doeIRON AND COPPER COMPOUNDS. 469not occur, so far as the present experiments show, with theglycerate, lactate and the compounds of the fatty acids.TABLE 11.-Ferric Hydroxide Dissolved at 20° b y 1 Equivalent ofA cid.Equivalents of F,O, taken ...Acid.Tartaric ..............................Malic .................................Citric .................................Oxalic ...............................GIycollic .............................Glyceric ..............................Lactic .................................Formic ...............................Acetfic .................................Propionic ............................Maleic .................................Racemic ..............................E thylsuccinic ......................Succinic ..............................Fumaric ...............................Aconitic ..............................Mucic .................................1.0 1-25 2.0 3.0Equivalents of Fe,O, dissolved. -1.001.001.001.001 -001.001.000.981.001.000.960.920.540-86 0.411-18 1.261.23 1.091.01 0.991.08 1-411-04 1.791.24 1.981.21 1.901.24 1.951.22 1-921-03 0.981.09 0.620.55 0.49nil.tracas.nil.action incomplete.-0.150-860.740.821-251-942.542-772.802.780.750.340.30The proportion of hydroxide dissolved in the case of oxalic,maleic, and ethylsuccinic acid, is nearly constant a t 1, 1 and0.5 equivalents, respectively, so long as thO proportion taken doeenot exceed 2 equivalents, and with the lactate the whole of thehydroxide taken is dissolved up to 2 equivalents, this givingevidence of the formation of a compound of such composition.With the last four acids entered in the table, little or no ironpasses into solution, but the hydroxide is converted into an in-soluble salt, a basic salt represented by RN,F%,Fe,O, (when drieda t looo) in the cwe of the succinate, and a normal salt, mixedwith a small proportion of basic salt, in the case of the fumarateand maleate, whilst with the mucate, the conversion of hhe hydr-oxide into a salt is very incomplete.Inorganic Ferric Salts.Solutions.of ferric chloride, each of half the concentration ofthe next more concentrated one, and prepared a t least one monthpreviously, were compared with each other. The values werechecked by Comparing together solutions differing considerably inconcentration, this being possible owing to the more dilutesolutions being as highly coloured as some of the concentrated ones.The molecular colour-intensity of the iron in a solution containing100 grams of the metal per litre was termed unity4'70 PICKERING : THE COLOUR INTENSITY OFIn other cases, a similar comparison was made between themembers of any given series, and these were then compared withthe ferric chloride solutions., Some variation of tdnt was oftennoticed a t certain concentrations, and where that was the case,those results were discarded,The number of instances in which the values remain constantFIG.1.Colour in,te?tsity of ferric ehloride solzltions.10 1 0.01 0'001 0'001Grams of iron per 100 C.C.throughout a series shows that no serious accumulation of errorscan exist.The results of two closely concordant series with the chlorideare summarised in col. I1 of table 111, and are plotted against thelogarithms of the strengths in Fig. 1, ABC and ABD. The valueTABLE III.-Colourjnteizsity of Inorganic FerricI n the split columns, the left half gives the values immediately after preparation,had been obtained.1.001.001-061.292.63291184218223257164134133-- -- -Gramsofironin100 C.C.I.20.913.910.810.06.96-02.51-250.630.310.160.080-040.020.010.0050.00250.00120.00060.00032.613398194140199138130133-Chloride.Colour-intensity .213251274229198178159156Series Iand 11.133224255190150148144138 - -6087185360Series111.111.-- -1.001.001-061.291-611.87173358100142142135-(3.77)(124) - -5680114150152Series 111. Afterheating.2-211.871.65 ---- - -'z: 86116136140(1462.802.31(1.71)- ---1.78GramsFe in100 C.C.VI .13.18.76.53.31.60.80.40.20.10.050.0260-01 30.0060.00030.00160.00080.0004clouds4511413013813713577Nitrate.Colour -intensit y.After heating.Labilesolu-tionsheated.IX.:548-35092125180214287194186185168 -- -1.546.3396782124181220167166161135 -472 PICKERING : THE COLOUR INTENSITY OFare apparently constant from about 12 to 3 grams of iron per100 c.c., but the rise with both more concentrated and more dilutesolutions (see the lower figure on a more open scale) renders i t im-probable that this constancy is real.After rising to' 2.9 at 0.08 percent., the values again fall t o 1.7 a t 0.005 per cent., when thesolutions become too pale for further measurement. The solutionson this falling branch of the curve, however, are not stable, andin about one hour they begin to darken, the change not becomingcomplete until after one to three weeks.The final values yielda curve, which attains and remains a t a maximum of about137. The temporary and permanent values are entered in the leftand right hand halves of col. 11. A t the point where the curvebifurcates-the critical point-the solution becomes opalescent afterseveral days, and sometimes a slight precipitate settles; in somecases, opalescence occurs in a minor degree in solutions of doubleand of half this concent.ration.The results of a third series, given in col. 111, differ slightlyfrom those in the other cases in that the critical point occursearlier, and th.at the subsequent rise in values is more rapid; butas regards general characteristics, and the initial and final values,it shows no- differences.Inappreciable impurities in the samplemay easily produce slight modifications in the actual valuesobtained.Besides the solid (doubtless a basic salt) which causes the opal-escence at the critical point, three or four other solids may separatefrom these golutions. (1.) On being heated to boiling, the more con-centrated solutions give a considerable precipitate of basic salt(approximately Fe : C1= 8 : l), form; ng in maximum quantity insolutions containing 1.2-0-6 per cent. of iron, and not a t all inmore concentrated (2.5 per cent.), nor in more dilute (0.1 per cent.)solutions. (2.) Solutions more dilute than that a t the critical pointoccasionally deposit, after some months, a small amount of basicsalt. (3.) Solutions a t and beyond the critical point, when boiled,deposit a perfectly transparent, brownish-yellow deposit of hydr-oxide on the glass.I n one o r two cases a similar deposit occurredin the cold: this was more frequent with solutions of the bromide.(4.) Solutions of extreme dilution occasionally give, after a time,a precipitate of flocculent ferric hydroxide.The action of heat on the solutions, besides causing precipitationin some cases, induces other changes. Taking, first, solutions whichhave attained a condition of stability in the cold, a considerabledarkening occurs when heated to boiling, but this disappears a tonce on cooling if the concentration is 1.2 per cent.of iron, or more.With more dilute solutions, it is persistent. The values obtaineIRON AND COPPER COMPOUNDS. 473when they are heated just to boiling in quantities of 5-10 c.c.,*and then rapidly cooled, are given in col. IV, those to the leftbeing the ones given immediately after cooling, and those to theright the ones obtained three months later. These are identical,+and the maximum is barely higher than the value for the most. diluteunheated solutions.On the other hand, solutions heated immediately after prepara-tion (labile solutions) give much higher values (left half of eol. V,and Fig. FB), but the intensity diminishes with time, the values inthe right half of the column being those obtained after threemonths, although, apparently, even then, they were still falling inthe case of the more concentrated solutions, and would, probably,in time, all become identical with those of the heated stablesolutions.The freshly prepared solutions near the critical pointdo not become cloudy on heating. At extreme dilution, the valuesfor the heated and unheated stable solutions, and for the heatedlabile solutions, are all identical.The significance of these results will be discussed below (p. 480).Bromide.--The results with this were substantially similar tothose with the chloride, and showed a bifurcation of the figurea t 0.07 per cent. of iron; but detailed examination was impossible,as the cloudiness appearing a t the critical point, appeared, also, inthree of the neighbouring solutions, and a deposit of transparenthydroxide formed in some of the others, whilst with the concen-trated solutions (4.4 t o 0.6 per cent.of iron) the red colour ofbromide predominates over that of the iron, although no freebromine could be detected. All the weaker solutions are yellow.I n the weakest solutions, the colour-intensity did not exceed 55,but the purity of the specimen examined was doubtful. The mini-mum value attained (in branch BC of the figure) was 1.2.Nitrate.-The results with the unheated solutions of this saltwere closely similar to those with the chloride (col. VII, table 111,and Fig. 2). With the heated solutions there is a difference,since both the stable and labile solutions (cols. VIII and IX)give practical!y the same values, and, in both cases, these show (aconsiderable reduction after three months (right hand halves ofthe columns).Sulphate.--This differs from the chloride and nitrate in thatthe values show a much greater rise up to the critical point, andThey form the curve ED in the figure.* The boiling may be protrazted in the case of the more dilute solutions withoutfear of precipitation: observations with solutions of concentrations of 0.06 to 0'16per cent.could not be made owing to their being, or becoming, cloudy. t In another series a slight reduction in intensity was observed during the firstfew days after the heating.VOL. cv. I 474 PICKERING : THE COLOUR INTENSITY OFremain constant a t 21 after that point. The latter solutions, asin the other cases, darken on keeping, but the extent of thisdarkening could not be measured, as they all became cloudy.Colloidal Ferric Hydroxide.-Freshly-precipitated ferric hydroxide250200150100 u *+5 u-+Lr-2 2 5060321FIG.2.Coloicr intensity of fwric nitrrttc so7ictions.10 1 0.1 0 '01 0-001Grains of iron per 100 C.C.was digested for some weeks with solutions of ferric chloride ornitrate, and the resulting liquids were dialysed until nearly freefrom acid. Szmples from the chloride gave the composition ofFe,C16 + 47Fe20, in one case, and Fe2C1, + 21Fe203 in another, andthe mean of series of coiour determinations with them arIRON AND COPPER COMPOUNDS. 475given in table IV. That from the nitrate was represented byF%(NO,), + 32Fe20,.In all cases the colour-intensity is constantthroughout, within the limits of experimental error (table IV), andit is practically the same (about 134) whichever salt is used in itspreparation. The solutions undergo no change on boiling, but, onprolonged heating a t looo, flocculent hydroxide is precipitated.TABLE IV.-Colour-intensity of Organic Ferri-compounds andColloidal Hydroxide.Grams ofiron per100 C.C. Citrate.3-0 21.92-3 19.81.5 17.31-0 12.10-5 10-30-25 10.4Race-mate. Lactate. - 23.8 - 25-5- 21.6- 18-69.1 10.28-2 9.8Glycerate. - A. B.Gly-Malate. Tar- col-trate. late.- -17.5 -15.6 -13.1 18.812.0 18-8- - -20.0 - 29.120.0 - 25.619-6 22.2 -16.8 22.2 22.4--0.1550.0620.0310.0160.0080.0040.0020.0010.0005Mean10.4 8.4 9.7 10-5 18-1 14.4 21.5 24.710.4 9.0 9.1 9.2 18-1 12.3 21.7 20.710.3 9.7 9-2 9.0 17.2 10.8 21.6 17.610.1 10.8 9-4 8-5 17.0 10.9 21.6 16.79.6 11.2 9.3 8.3 18.3 11.5 21.6 17.69.5 11.9 9-8 8.3 19.7 12-2 22.6 28.39.2 12.2 10.3 8.3 20.7 14.2 22.6 18.88.7 12.4 10.7 9.0 20.7 16.2 22.5 19.38-7 12.7 - - 20.7 - 22.6 19.99-7 11.5 9.7 9.1 19.0 {(%::) 22.1 19.6Oxalate, 0-5 to 0-004 per cent., constant at ............................ (?) 1.5Maleate, 0.5 to 0.001 per cent., constant at ........................... (?) 55Ethylsuccinato, 0-5 to 0.001 per cent., constant at .................. ( 9 ) 55Wormate, 0 5 to O.GOO1 per cent., constant at ........................ 159Acetate, 0.5 to 0.0001 per cent., constant at ......................... 169Propionate, 0.5 to 0.0001 per cent., constant a t ...................... 272Hydroxide from chloride, 0-5 to 0.0001 per cent., constant at...... 128Hydroxide from nitrate, 0.5 $0 0.0001 per cent., constant at ...... 139Organic Cornpouitds.The solutions (table IV) were prepared by dissolving moist,precipitated ferric hydroxide a t 20° in the amount of acid calcu-lated to form the normal salt.Dissolution was complete, or verynearly so, except in the case of the ethylsuccinate. Light had t obe excluded; the oxalate is rapidly decomposed by sunlight withthe formation of a ferrous salt; the racemate is similarly unstable,and, to a less extent, the tartrate, glycerate and lactate.At leasttwo preparations and series of determinations, were made in eachcase.The values for the colour-intensi ty are approximately constantI 1 476 PICKERING : THE COLOUR INTENSITY OF‘throughout a large range of concentration from 0.5 per cent. ofiron downwards. With the racemate, however, there is someincrease on dilution, with the citrate, some reduction, and withthe malate and glycollate, there is a reduction followed by anincrease.The average of the values for the weaker solutions of the citrate,racemate, lactate and glycerate A is 10*0, and that for thoae ofthe tartrate, glycollate and glycerate R is double this, namely,20-2.With the citrate, lactate, malate and glycerate A , where thelower constant obtains for dilute solutions, the higher one isindicated for concentrated solutions, although the values must beuncertain owing to the depth of colour of such concentratedsolutions. With the malate, the range throughout which the lowervalue extends would not justify its being termed a constant here.With the concentrat-ed solutions of the glycollate, the value isexceptionally high.Making allowance f o r the fact that some variation in the con-stants might naturally be produced by differences in the characterof the molecules with which the iron is combined, there can belittle doubt that the existence of two constants are here indicated,one of them double the other, as would be the caBe according asone or both the iron atoms had assumed a certain character (forexample, had become electronegative), and the tendency of dilutionis to reduce the higher value t o the lower.With the other compounds examined, the values were constantthroughout, .and the means only are given a t the foot of the table.With the oxalate, maleate and ethylsuccinate, the tint renderscomparison with the chloride impossible.With the first, the tintis greenish, and the intensity between 1 and 2, that is, approxi-mately that of the inorganic salts, to which, in nearly all otherrespects, it resembles. With the latter two, the tint is red, andthe intensity appeared t o be 50 t o 60, but judging by the factthat the colour became too faint for measurement a t about thesame degree of dilution as in the other cases, i t is probable thatthe real value is not above 20.Solutions of the formate, acetate and propionate are evidentlyvery different from those of the other compounds, and exhibit acolour-intensity as high as 159-172.These values are also main-tained when the iron is present. in excess; with 2Fez0, and 3Fez0,t o one equivalent of acetic acid, the values of 150 and 145 were* The two samples of glyceratlr were prepared from different specimens of acid,which may not have been of the same degree of purity, as the “acid” alwayscontains some anhydrideIRON AND COPPER COMPOUNDS. 477obtained, and with similar proportions in the case of propionicacid the values were 156 and 164.Determinations were made with potassium and sodium ferri-cyanides from a strength of 2 to 0*0001 per cent. of iron, and of4 to 0.0001 per cent.of iron, respectively; but the tints, althoughcomparable with each other, are not comparable with that of thechloride. The intensities ,are certainly greater than that of thelatter, estimates giving 6 with the potassium salt, and 18 withthe sodium salt. The intensity is unaltered by dilution.With the ferrocyanides the intensity is inuch less than withferric chloride, and the values alter with dilution. The followingdeterminations were made :Intensity{ Na-salt . . . . . . 0.062Per cent. Iron. 2 1.5 075 038 019 010 0.05K-salt ...... 0-064 0.064 0.064 0.067 0.088 0.101 0-1080.094 0.104 0.104 0.092 0.087 -EfJect of .Excess of 9 cid.With the nitrate, the addition of from 1.2 to 1.8 molecules ofnitric acid to each atom of iron discharges the yellow colour en-tirely, and the liquid becomes of a light amethyst tint, discernibleonly in the case of concentrated solutions.On the further additionof acid up t o 3 to 20 molecules, a strong lemon-yellow colourdevelops; this is of much lower intensity than the yellow of theneutral nitrate, but no exact comparison is possible. The amethystcolour is often noticeable in the solid crystallised nitrate, as it iswith iron alum; also in solutions of ferric salicylate.With the sulphate, the yellow colour does not entirely disappearon the addition of acid, hut becomes reduced t o about one-seventhof its original value, and is lemon-yellow in quality; this yellowincreases in intensity with further additions of acid, reaching toabout onshalf of the intensity of that in the neutral sulphate.With the more concentrated solutions, the coloration a t the pointof minimum intensity is apparently that of a mixture of amethystand yellow.With solutions containing 7, 1.8 and 0.4 per cent.of iron, the minimum coloration is attained with 27, 70 and 90molecules of sulphuric acid respectively. From very concentratedsolutions, acid precipitates the white solid sulphate.With the chloride the addition of acid produces a reduction ofintensity, and an alteration in quality t o the lemon tint, withoutshowing any minimum point or amethyst colour. From 0.2 to 2molecules of hydrochloric acid t o each atom of iron, where tliesolutions contain from 3 t o 0.04 per cent.of iron, the acid reducesthe intensity t o one-third or one-quarter of its original value.The behaviour of the organic ferric salts with excess of th478 PICKERING : THE COLOUR INTENSITY OFcorresponding acids is very different, the reduction in colour-inten-sity in those cases investigated being practically nil. The oxalate,however, is exceptional, behaving in this, as in other respects, likethe inorganic ferric salts, and showing a large reduction on theaddition of acid. The results may be summmised thus :Acetate ...... + 0.23 to 23.0 C,H,O to each Fe No changeMalate ...... +0*12 to 12.0 CJHliOS ,, Fe No changeTartrate ...... +0*09 to 27.0 C,H,O, ,, Fe No changeCitrate ...... +0.14 to 3.4 C6H807 ,, Fe No changeCitrate ......+ 6.8 to 13.4 C,H,O, ,, Fe Reduction extendingto 5 per cent.Oxalate . . ... . + 0.19 t o 16.0 C,H,O, ,, Fe Reduction throughout,extending to 79 percent.Effect of Heat on the Organic Compounds.P6an de Saint-Gilles (Ann. Chim. Phys., 1856, [iii], 46, 47)found that when the acetate was boiled for some hours it becameopaque through the formation of a brick-red ferric hydroxide,which remained in suspension. This has been confirmed as regardssolutions containing 1.5 t o 0.5 per cent. of iron. With a 0.25 percent. solution, however, it was found that only slight opalescenceoccurs, the colour-intensity is appreciably reduced, even in half-an-hour, and becomes constant in twenty-four hours, being reducedfrom 170 to (average) 21.9. This is identical with the value forthe f erri-compounds with 2pe.Such solutions remain unalteredafter the heating (which was done in sealed tubes), and alcoholgives a precipitate with them, which is not the case with theunheated solutions.The formate undergoes similar ch,anges, but the reduction incolour-intensity has not been traced beyond 55, which was reachedin twenty hours at looo. With the propionate, the reduction wastraced to 64 in twenty-four hours.I n all three cases there appears to be an increase of intensitypreceding the decrease. This was most marked with the propionate,where, with a 0.25 per cent. solution, the intensity increased a tfirst to 250. If the heating be then discontinued, the intensityreverts to its original value of 170 in two days.This indicates theconversion of the lighter form of colloidal hydroxide into the darkerform, prior to the production of the ferri-compound with 2Fe.With the other organic ferri-compounds, two diff went changesoccur on heating, according to the concentration of the solutions.When these are concentr.ated (1.5 t o 0.25 per cent. of iron), pro-tracted heating reduces the colour-intensity to, approximately, thevalue obtaining in the case of the inorganic salts, implying theconversion of the ferri-compound into the corresponding normaIRON AND COPPER COMPOUNDS. 479salt with the iron electropositive; in the first three instances quotedbelow, the intensity is reduced to nearly unity; in the other two,the reduction is less complete.The ferric salt formed, however, isunstable, and, in the course of a few days or weeks it reverts tothe original, dark-coloured ferri-compound. I n the case of thelactate, glycollate, maleate, and the strongest tartrate and racematesolutions, the heating produces a precipitate. This, in the oase ofthe racemate, is the basic salt, R,Fe,,Fe,O,; in the case of thetartrate, it is not the corresponding basic salt, but consists of abulky jelly which gradually redissolves in the cold. It may be thenormal salt. The oxalate darkens slightly, but becomes opalescent.More concentrated solutions. More dilute solutions.Intensity reduced. Intensity increased.Recemate ......... From 9 to 1.08 From 12 to 109Malate .............. 20 ,. 1.22 ,) 11 ,, 177Glycerate .........,, 17 ,, 1.30 9 7 8 9; 88Tartrate ............ ?, 22 ,? 7.5 9 9 22 9 , 275Citrate ............ ,, 17 ,, 14.2 19 9 s9 305Lactate ............ cloudy ,) 10 ,, 185With the more dilute solutions, hydrolysis occurs, forming col-loidal hydroxide, and when the heating is further protracted-insome cases more than twenty-four hours being required-this is con-verted into the ordinary, insoluble, flocculent hydroxide. The colloidalhydroxide is unstable, and, in the cold, gradually reverts into theferri-compound from which it was formed. I n most cases, theconversion of the whole of the iron into the colloidal hydroxidecannot be complete before precipitation begins, but it is probablyso in the case of the tartrate and citrate, where maximum valueshave been obtained, identical €or several different concentrations,and where no precipitation occurred after forty-eight, hours’ heating,The mean value in these two cases is 290.Some evidence was obtained that, with solutions of intermediateconcentration, the two actions characteristic of concentrated anddilute solutions, might proceed simultaneously, but a t differentrates, and this would indicate that these represent the changes ofthe two ferri-compounds present, with 2Fe and FeFe, respectively.Attempts were made in the case of the racemate and glycerate toobtain the lighter form of colloidal hydroxide by diluting therecently heated concentrated solutions : but no hydrolysis occurred.- - 480 PICKERING : THE COLOUR INTENSITY OFDiscussion of Results.The main features of the results with the inorganic salts may besuinmarised thus :Chloride.Nitrate. Sulphato. Bromide.Intensity with satur-ated solutions ......Minimum intensitywith concentratedsoh tions ............Position of minimumPosition of criticalpoint.. ................Final maximum in-tensity ............Intensity of colloidalhydroxide .........1.31 1.22 1.29 -0.98 1-01 1-04 -12 t o 3 10 to 2 10 to 5-]per cent.(0.16 to) 0.04 0.02 0.08 0.07- - 135 139;/137- - 139 128134The small rise in intensity with saturated solutions is, doubtless,due to the formation of hydroxide in some form; such solutions,especidly that of the nitrate, give off acid fumes when heated,*and the solid nitrate will do so in sunlight, even a t the ordinarytemperature, the fumes being re-absorbed in the dark.It is equailycertain that the iron in the most dilute solutions is in the formof colloidal hydroxide ; the colour-intensity is that of the hydroxideprepared by dialysis; it will not diffuse through parchment paper,is not affected except by prolonged boiling, and is precipitated byacid as ordinary flocculent hydroxide, just like the dialysisproduct. f-The more concentrated solutions constituting the curve AB mustcontain iron of, at least, two different colour-intensities, but thedarker one cannot be the colloidal hydroxide, f o r the two forms inquestion are convertible immediately into each other by dilution orconcentration, whereas the colloidal hydroxide is formed only verygradually, and its formation is irreversible. This is proved bythe fact that the same result is obtained by diluting a concentratedsolution, either with water or with any more dilute solution onthe curve AB, whereas, if diluted with a solution on the curveBD, instead of with water, the resulting mixture is much darker,and remains permanently so.+ With the sulpliste, only above 100".-f Such precipitation shows that the colloidal hydroxide is incapable of directreaction with acids, and seems to prove that in the insoluble flocculent hydroxidethe state of aggregation, and probably hydration, mnst be intermediatc betweenthat in the colloid and that of the nnimolecular condition which must precede itscombination with an acidIRON AND COPPER COMPOUNDS.481The most probable view is that the darker constituent of themore concentrated solutions (curve ABC) is a f erri-compound,similar to the organic compounds, with the iron united directlyto the nuclear element of the acid radicle. With both atomspresent in this condition, the colour-intensity should be about 20,and, in the case of the sulphate, this value is reached and main-tained from a strength of 0.014 to the limit of the observations(0.0017 per cent.). The lighter constituents of these solutionswould, of course, be electropositive or pseudo-tervalent iron ; whatthe colour-intensity of this is, is uncertain, but the probability isthat it is nil in respect of yellow, and that it would show, if pure,only a slight amethyst colour.That some form of iron possessingsuch characteristics exists, is certain from the results obtained onadding excess of acids to the solutions, and this almost colourlessiron is not the rmult of the formation of acid salts, as the effect ofstill more acid is to produce a, lemon-yellow colour.* On such aview, which is in harmony with the colourless nature of the solidsulphate and nitrate,+ the yellow of the solutions, even a t the mini-mum points, must be due to the presence of some electronegativeiron. It may be objected that this would necessitate the latterbeing present in approximately the same proportion, and to OCCUIa t approximately the same degree of concentration, in all cases.This is, however, not improbable, as it is seen that the formationof colloidal hydroxide (that is, the position of the critical point) isto a large extent independent of the nature of the acid present.The proportion of water may be the main factor in both cases.As the ferri-compound contains the elements of 2E20 over andabove that of the ferric salt, dilution would increase the proportionof the former, as it does in the figure A B ; but the effect of dilutionis two-fold, it favours hydration, but it also separates the dissolvedmolecules from each other, reducing them to the quasi-gaseouscondition.I€ the stability of the ferri-compound under such con-ditions is less than that of the ferric salt, as is most probably thecase, dilution beyond a certain point would begin to increase theproportion of the latter, and the fall in the colour-intensity in theregion BC would be accounted for.The ferric saIt, thus re-generated, not being stable, is gradually hydrolysed, forming col-loidal hydroxide or a basic salt, the latter only a t the critical pointIt seems difficult to explain thiecolour, especially if its intensity is the same in all cases, by the formation ofordinary acid salts, but it may be due to the formation of compounds analogous tothe a-cupricarbonates, wherein an acid radicle displaces the oxygen united to themetal (T., 1911, 99, 801).f The transparent amethyst crystals of the nitrate cannot be washed and driedwithout assuming a yellow tint.* This has not been fully investigated jet482 PICKERING : THE COLOUR INTENSITY OFor, exceptionally, a t some other concentrations.As the proportionof colloidal hydroxide eventually formed (curve BD) increases withthe proportion of ferric salt in the freshly-prepared solution (curveHC, so far as it can be followed), it is legitimate to conclude thatthe hydroxide is derived from the ferric salt, rather than from theferri-compound, which is present in decreasing proportions as thedilution is extended.It remains to1 account for the very deeply-coloured product ob-tained by heating solutions of the chloride from 0.6 to 0-005 percent. concentration. Doubtlas, this, also, is a form of ferric hydr-oxide produced by hydrolysis. As the ordinary colloidal hydroxideis not altered by heat, unless the heating is very prolonged, it isevident that this dark product cannot be formed from it.Now,the proportion of ordinary colloidal hydroxide in unheated solu-tions may be calculated from the curve BD, and thence the relativeproportions of the darker hydroxide in the heated solutions (curveF D ) ; these proportions, rising from nit? at a concentration of about1 per cent., are found to reach a maximum a t a concentration of0.04 per cent.-the exact point a t which the proportion of ferri-compound also reaches a maximum in the solutions before heating-and fall to nil again a t extreme dilution, where the proportionof ferri-compound originally present must, 80 far as the determina-tiom can show, be also nil; the obvious conclusion is that thishydroxide must be the hydrolysis product of the ferri-compound,just as the ordinary colloidal hydroxide is that of the ferric salt.On the assumption that this transformation by heat is complete,and taking the results a t points where the quantities dealt withare not too small for accuracy, the colour-intensity of the darkerhydroxide is given as 260; or, if the curve ED, instead of BD, betaken as giving the proportion of lighter hydroxide in the heatedlabile solutions, the value obtained is 270.Similar calculationswith the nitrate give 306; the mean is 280, which is just doublethat of tlie lighter colloidal hydroxide. This view is confirmed byfinding that the organic ferri-compounds are, as a matter of fact,hydrolysed by heat (although not by mere dilution) forming anunstable colloidal hydroxide, of which the colour-intensity is thesame as that obtained froin the nitrate and chloride, namely,Solutions of the formate, acetate, and propionate of all concen-trations appear to be similar to the most dilute stable solutions ofthe inorganic salts, containing the iron in the lighter colloidal form.The values, about 170, are somewhat higher than those given forthe colloidal hydroxide in other cases, and it is probable that acertain amount of the darker hydroxide is present in them; it275-305 (p.479)THE COLOUR INTENSITY OF IRON AND COPPER COMPOUNDS. 483appears, also, that the whole of the lighter hydroxide present canbe converted into the darker form by heat (250 being reached withthe propionate, p. 478), this being ultimately converted by heat intoinsoluble hydroxide if much water is present, or into the stablef erri-compound if the solutions are more concentrated.The nature of the other organic iron compounds and the trans-formations which they undergo on dilution or heating has beensufficiently noted above (p. 478).Summary.Iron in ferric compounds may exist in five different conditions:(1) Both atoms in the electropositive or hydroxyl position :colour-intensity very small, probably nil as regards yellow light.(2) Both atoms in the electronegative portion of the molecule,directly united to the nuclear element of this portion (ferri-com-pounds) : intensity about 20.(3) One atom in each of the two above conditions: intensity 10.(4) As ordinary colloidd hydroxide by the hydrolysis of electro-positive iron : intensity about 135.(5) As darker colloidal hydroxide, by the hydrolysis of electro-negative iron : intensity about 280.The chloride (and probably bromide) exists in the conditions(l), (4), (5), and either (2) or (3).The sulphate exists as (l), (2), and in one or both the colloidalconditions.The oxalate is a normal ferric salt, with its iron in the con-dition (1).The format-e, acetate and propionate exist normally as (4), butmay, in some cases, probably in all, be obtained in the conditions(2) and (5).The other organic compounds examined exist normally as (2)(tartrate, glycollate, and, probably, maleate and ethylsuccinate),or as (3) (racemate), or else in both these conditions, according tothe concentration of the solution (citrate, lactate, glycerate andmalate). When the more concentrated solutions of them areheated, they form, or tend to form, the normal ferric salt(1) (racemate, malate, glycerate, tartrate and citrate) ; when themore dilute solutions are heated, they hydrolyse into the condition(5) (racemate, malate, glycerate, tartrate, citrate and lactate).The only condition in which none of them has been obtainedis (4).HABPENDEN.VOL. cv. K