ANNUAL REPORTSON TEEPROGRESS OF CHEMISTRY.GENERAL AND PHYSICAL CHEMISTRY.THE output of research in this branch of science continues toincrease, and this fact is marked this year by the inclusion of aspecial Report on Chemical Kinetics (p. 314), a subject whichincludes two Meldola Medallists (1923 and 1926) amongst its workersin this country. No attempt has been made to report on thewhole field of progress, and topics not recently dealt with haveas far as possible been chosen for discussion.The Atomic Nucleus.Two papers of outstanding interest have been published thisyear, and it is considered that their importance justifies theirinclusion here in anticipation of next year’s Report on Radio-activity and Sub-atomic Phenomena.The construction of an improved form of mass spectrographhas been described by F.W. Aston in the Bakerian Lecture tothe Royal Society, The new instrument has a resolving powerfive times, and an accuracy ten times, as great as the original0118.2 These improvements have been effected, not by a change ofprinciple, but by doubling the angles of magnetic deflexion, andsharpening the lines by the use of finer slits placed farther apart.The dispersion varies from 1.5 mm. to 3 mm. for a change of massof 1%, but owing to the fact that the lines on the plate areirregulrtrly curved and change gradually in shape throughout thespectrum, it is necessary to compare masses, which must not differby more than 1%, by the accurate measurement of (a) the distancebetween the lines, and ( b ) the dispersion constant a t the mid-pointbetween them.These measurements are applied in different waysto suit different cases, the most generally applicable one beinga modification of the original bracketing method.1 PTOC. Roy. SOC., 1927, [ A ] , 115, 487; A,, 914.Ann. Reports, 1920, 17, 22112 ANNUAL REPORTS ON THE PROGRESS OF CHEMISTRY.The results are now so accurate that the loss of an electron(mass = 0.00054; H = 1) in the formation of a positive particle isgenerally significant and must be allowed for. Determinations ofmass are now made to within 1-2 parts in 10,000 parts, and, as aresult, it is found that most of the elements have atoms whosemasses deviate from the whole-number rule, although not, ofcourse, to the same extent as does hydrogen.I n fact, there arenow three fundamental numbers characteristic of every atom :(i) the mass number, giving the number of protons in theatom,(ii) the atomic number, giving the number of extranuclearelectrons,(iii) the packing fraction, which is an indication of the forcesbinding the nuclear protons and electrons, and is thusa measure of the instability of the nucleus.The packing fraction is 10,000 times the departure from thewhole-number rule (when 0 = 16-000) divided by the mass number,and has a value 77.8 for hydrogen. Since loss of mass may be takenas equivalent to release of energy due to close approach of protonsand electrons in the nucleus and consequent partial annihilation oftheir electromagnetic fields, it follows that high packing fractionsindicate looseness of packing and therefore low stability, and lowpacking fractions the reverse.When the packing fractions of the atoms are plotted against theirmass numbers, it is found that all but light atoms of even atomicnumber (helium, carbon, and oxygen) lie on a smooth, non-periodiccurve which descends steeply from hydrogen (+ 77.8) throughfluorine (& 0) to a minimum a t bromine (- 9), thereafter risingmuch more gently to cross the axis again at about mercury.Thelight atoms of even atomic number have packing fractions wellbelow this curve, and approximate to a branch rising much lesssteeply to helium (+ 5.4). The observed stability of the nuclei ofhelium, carbon, and oxygen (beryllium has unfortunately not yetbeen measured) is in accord with their position on the lower curve.Incidentally, the research has settled the isotopic constitutionsof mercury and xenon, and has recorded new isotopes of sulphur andtin, bringing the number of isotopes of the last-named element upto eleven.Sir E.Rutherford has put forward a theory of the structure ofthe nucleus of a radioactive atom. The nucleus is imagined toconsist of a central part around which revolve " neutrons "-a-particles plus two electrons (potential helium atoms)-in quant-Phil, Mag., 1927, [vii], 4, 680; A,, 1002GENERAL AND PHYSICAL CHEMISTRY. 13ised orbits. If the system should for some reason become unstable,the “ neutron ” is ejected as an a-particle, and its two electronscirculate close to the central nucleus with a velocity approachingthat of light.One of these may later be hurled from the atom asa @-ray. Either of these changes may be accompanied by arearrangement of the “ neutron ” orbits involving the emission ofy-rays. In all cases the changes are governed by quantumrelations.Refractivity and Refractive Dispersion.The departure from additivity of molecular refractivity has beencalculated on the hypothesis that the electron shells of the atomsin the molecule are displaced (polarised) by the proximity of otheratoms. The case for organic molecules is dealt with by K. Fajansand C. A. K n ~ r r . ~ E’or saturated hydrocarbons the problem issimple. According to the Lewis-Langmuir electronic theory ofvalency, the 8 valency electrons of the carbon shell in methane areregarded as equally distributed amongst four carbon-hydrogenbonds, so that one-fourth of the molecular refraction of methanerepresents the value for one such bond.The refractive equivalentof a carbon-carbon linking may be obtained by subtracting sixtimes the value of a carbon-hydrogen bond from the molecularrefractivity of ethane. With substituted hydrocarbons, however,the problem is not so simple. Methyl chloride, for example, con-tains three carbon-hydrogen bonds and one carbon-chlorine linking,and the refractivity for the chlorine atom in these circumstances isconsidered to be due to the combined influence of the bonding pairof electrons and the three lone pairs. It is thus nearly the sameas the refractivity of the chlorine atom in hydrogen chloride andis lower than the refractivity for the free chlorine ion.The refrac-tivities for a large number of groupings are worked out in this way.Somewhat similar considerations are applied by T. H. HavelockY5who treats atoms as isotropic resonators in fixed relative positionsin the molecule.The case for ions in solution has been worked out by K. Fajans,6who points out that the refractivity of an ion is a measure of theease of displacement of its electron shells with respect to thenucleus. An anion is rendered more rigid by the proximity ofpositively charged kations, since the positive charges tend tobalance the inward attraction of the positively charged nucleus.Its refractivity thus tends to fall.Conversely, the proximity ofnegatively charged anions tends to displace the electron shells of aBer., 1926, 59, [B7, 249; A., 1926, 336.Phil. Bag., 1927, [vii], 3, 158, 433; A., 189, 294.ti Trans. Paraday SOC., 1927, 23, 357; A., 102314 ANNUAL REPORTS ON THE PROGRESS OF CHEMISTRY.kation, since now the attractive force of the nucleus is reinforced bythe repulsive effect of the neighbouring anions. Its refractivitytherefore tends to rise. “The union of ions into molecules orcrystals will thus be accompanied by a net diminution of therefractivity whenever the consolidating effect of the kation uponthe anion outweighs the loosening effect of the anion on the kation,and vice versa.”K. I?. Herzfeld and K.L. Wolf and B. Davis 8 have attemptedto fit dispersion equations involving one or more frequency termsto the observed data for elements and compounds, but R. A. Mortonand R. W. Riding consider that no satisfactory two-term equationfor the variation of refractive index with wave-length can beobtained until further absorption data are secured in the short-wave ultra-violet region of the spectrum. They are of the opinionthat existing data are best fitted by equations of the type : lo(n - WV1 + v, + v3 + . .) = V1N1/(V,2 - v2) +72N,/(v22 - v2) + V3N3/(v,2 - v2) + . . .where n representa the refractive index; V,, V2, etc., the volumesof the respective molecular phases; l1 N l , N,, etc., constantsassociated with the respective molecular phases ; v the frequencya t which the refractive index is observed; and vl, v2, etc., the oscill-ation frequencies associated with the respective molecular phasesand are all integral multiples of a fundamental frequency in theinfra-red.The conclusions of H.Hunter l2 and others13 that the neglectof the dispersion factor is possibly responsible for the failure ofthe generally accepted methods of applying refractometric data t oproblems of chemical constitution receive experimental supportfrom the work of H. Voellmy.14 This author has examined themolecular refractivities of more than 30 organic liquids a t wave-lengths between 6560 and 2100 8., and has shown that the refract-ivity does, in fact, increase on the long wave-length side of anabsorption band and decrease on the other side in accordancewith theory.Workers on refractivity in the infra-red will do well to note thewarning of Sir R. Robertson and J.J. I ? o x , ~ ~ that the temperatureAnn. Physik, 1925, [iv], 76, 71; A., 1925, ii, 182.Physical Rev., 1925, [ii], 26, 232; A., 1925, ii, 933.Phil. Mag., 1926, [vii], 1, 726; A., 1926, 658.lo E. C. C. Baly and R. A. Morton, J . Physical Chem., 1924, 28, 659; A.,l1 Ann. Repom, 1915, 12, 6.l4 Z. physikal. Chem., 1927, 127, 305; A., 812.l6 Nature, 1927, 119, 818; A., 607.1924, ii, 714.l2 Ibid., 1923, 20, 15.F. R. Goss, C. K. Ingold, and J. F. Thorpe, J., 1924,125, 1927GENERaL AND PHYSICAL CHEMISTRY. 15coefficients of refractive index for rock salt and fluorite are importantand cannot be neglected.Molecular Volume.One of the outstanding achievements of the electronic theory ofvalency is the prediction of two kinds of double bond-the semi-polar double bond, occurring mainly but not exclusively in inorganiccompounds, and the non-polar double bond which is chiefly, butagain not entirely, to be found in carbon compounds.For sometime after the theoretical prediction of the existence of the semi-polar double bond there was no experimental method of detectingits presence ; now, however, there are three methods available.In order of priority we have : (a) the parachor,16 (b) resolution intooptical enantiomorphs, l7 ( c ) zero volume.18Of these, (b) is an absolute method, but is obviously limited inapplication to a very few compounds, although it can be appliedto solids as well as to liquids, (a) is of more extended applicability,but is limited to non-associated liquids, whilst (c) can be appliedto all liquids.Method (b) is outside the scope of this Report.The principleunderlying method (a) has already been described,l9 but hasrecently been greatly extended in application. The examination 20of a large number of double-bonded compounds by the method ofthe parachor has shown that such substances fall into two sharplydefined classes, one showing an increase in the parachor of 23.2units due to the presence of the double bond, and the other adecrease of 1.6 units. The evidence is clear that the first classrepresents the non-polar linking and the second the semipolar.Itis found that in every case where a carbon atom is concerned inthe double bond the linking is non-polar. T. M. Lowry’s suggestionto tthe contrary,21 improbable enough on other grounds, is thusruled out of court, a t all events as far as molecules not in a reactingstate are concerned.A later paper 22 lends support to the view that the double bondsin maleic-fumaroid geometrical isomerides are non-polar, andindicates that isomerism of this type has little or no effect on the16 S. Sugden, J., 1924, 125, 1185.1 7 H. Phillips, J., 1925, 127, 2552; P. W. B. Harrison, J. Kenyon, andH. Phillips, J., 1926, 2079; S. G. Clarke, J. Kenyon, and H. Phillips, J.,1927, 188.1* S. Sugden, J., 1927, 131.19 Ann. Reports, 1924, 21, 8.20 S.Sugden, J. B. Reed, and H. Wilkins, J., 1925, 127, 1625.21 J., 1923, 123, 822.23 S. Sugden and H. Whittaker, J., 1925, 127, 152516 ANNUAL REPORTS ON THE PROGRESS OF CHEMISTRY.parachor. A similar conclusion,% that 0-, m-, and p-isomerideshave identical values for the parachor, has also been reached. Theevidence of the parachor in every case supports the maintenanceof the octet as against shells of 6, 10, or 12 electrons. S. Sugdenand H. Wilkins have determined the effect of ring structure on thep a r a ~ h o r , ~ ~ and have correlated the values so found with the degreeof unsaturation measured as the quotient of the number of latentvalencies by the number of octets involved. The figures, givenbelow, are striking.LatentvalenciesStructure.(x).Triple bond .................. 4Double bond .................. 2Three-membered ring ...... 2Four-membered ring ...... 2Five-membered ring ...... 2Six-membered ring ......... 2Degree ofunsatur-Octets ation2 2.0002 1.0003 0.6674 0.5005 0.4006 0.333(n). (xln).Parachor,obs. calc.46.6 46.4(23.2) (23.2)17 15-511.6 11-68.5 9.36.1 7.7The view that structural constants affecting the parachor havevalues proportional (to a first approximation) to the degree ofunsaturation involved receives additional support from the valuefound for the semipolar double bond (- 1.6); this bond is notunsaturated and therefore we should expect it to have no influenceon the parachor. The slight negative value actually observed isprobably due to a small contraction in volume caused by electro-static attraction (rendered obvious when the semipolar double bondis written thus : S-0).Compounds of phosphorus and arsenic have been examined,25and Sugden considers that the formula suggested by E.B. R.Prideaux 26 for phosphorus pentachloride, which involves a single-electron bond (" singlet ") between the phosphorus atom and eachof two of the chlorine atoms linked to it, the other three beingheld in the normal manner by duplets, best represents the facts.This has the advantage that it preserves the rule of eight inviolate,but N. V. Sidgwick 27 considers another explanation preferable.Finally, F. B. Garner and S. Sugden have applied the method todecide between tautomeric ring and chain formulae for quinones,benzils, and succinyl and phthalyl chlorides, and find that allexcept the last-named have the normal structure and cannot inthe liquid form contain more than traces of the ring isomeride.In+ -23 S. Sugden and H. Wilkins, J . , 1927, p. 2517.Z4 Idem, ibid., 1927, 139.z5 Ibid., p . 1174.26 Chem. and Ind., 1923, 42, 672.2' " The Electronic Theory of Valency," Oxford Univ. Press, 1927, p. 130.a* F. B. Garner and S. Sugden, J., 1927, 2877GENERAL AND PHYSICAL CHEMISTRY. 17the case of phthalyl chloride, the high-melting form was shown tohave the unsymmetrical, and the low-melting form the symmetrical,structure.Sugden has also shown29 that the variation of density withtemperature from the freezing point to the critical point is repre-sented accurately for normal liquids by the equation :where D and d are the densities of the liquid and vapour a t To K.,T, is the critical temperature on the same scale, and Do is a con-stant representing the liquid density a t absolute zero. This equa-tion also holds for associated liquids at lower temperatures. Thezero volume,” Vo, obtained by dividing the molecular weight byDo, is found to be nearly proportional to the critical volume, thefactor of proportionality being about 0.27.The zero volume thus obtained is an additive function of thefollowing constants?O the observed values lying within 2% ofthose calculated for 236 out of 284 compounds considered.D - d = Do(1 - T/Tc)0’3,66Atomic constants.Structural constants./--p/L H = 6.7 I = 28.3 Triple bond = 15.5 c = 1.1 P = 12.7 Double bond = 8.0N = 3.6 S = 14.3 Three-membered ring = 4.80 = 5.0 0 (inalcohols) = 3.0 Four-membered ring = 3.2F = 10.3 N (insmines) = 0.0 Five-membered ring = 1.8C1 = 19.3 Six-membered ring = 0.6Br = 22.1 Semipolar double bond = 0.0The difference between the values for non-polar and semipolardouble bonds is noteworthy, and it will be observed that there isthe same connexion between the values for the structural constantsand their degree of unsaturation as in the case of the parackor,although the quantitative agreement is not so good.Other, less successful, relationships 31 have been put forwardrecently, notably Vopc/Tc = const.and modifications of this (pchere represents the critical pressure, the other symbols have theirprevious significance).W. Herz 32 has also correlated Vo with latentheat of vaporisation, the Poisson capillarity constant, molecularelevation of the boiling point, and the difference between thespecific heats at constant pressure and constant volume. A closeconnexion between molecular volume and molecular refra~tivity,~~29 F. B. Garner and S. Sugden, J., 1927, 1780.30 Ibid., p. 1786.31 R. Lorentz and W. Herz, 2. anorg. Chew., 1924,375; A,, 1925, ii, 25, 183; W. Hem, ibid., 1925, 149,A., 1926, 110, 778.32 Ibid., 1926, 153, 269; A., 1926, 786.33 €4. Lorentz and W. Herz, &id., 1925, 1142, 80; A.,140, 379; 1925, 141,270; 1926, 153, 339;1926, ii, 36618 ANNUAL REPORTS ON THE PROGRESS OP CHEMISTRY.and between molecular refractivity and the parachor 34 have alsobeen indicated.The Metastability of Matter.The view that elements and compounds, even when chemicallypure, may not be physically homogeneous, is gradually gainingground and seems destined to prove of great importance.Only afew years ago, the phenomena of allotropy and polymorphism wereregarded as the exception rather than the rule, yet to-day E.Cohen 35 is able to state his belief that “ every solid substance mayexist in two or more modifications ” and that “ many of the hithertorecognised physical or physicochemical constants of solid substancesare values which refer very often, if not always, to metastablemixtures which contain two or more modifications of that substancein unknown proportions ” so that “ no definite importance can beassigned to such constants.” This behaviour of solids-their dis-inclination to change a t once into stable modifications a t theappropriate transition point-is, of course, not an unmixed evil.Many of the special steels, to take a familiar example, depend fortheir properties on constituents deliberately added to prevent suchchange.Nevertheless, such wholesale doubt cast upon the accuracyof physical properties, determined without special precautions toensure physical as well as chemical purity, is disturbing. Nor isthe doubt confined to solids. Liquids and gases-as shown byexperiments on intensive drying-must also be considered, evenwhen chemically pure, as more or less complex mixtures.In the case of some substances, of course, it has long been knownthat metastability over a long period of time occurs. E.Cohen36has recently directed attention to the fact that as long ago as 1847,St. Claire Deville pointed out that the stabilisation of solid sulphurat the ordinary temperature was not complete even after 8 years,as indicated by a progressive change in density. In the classicalcase of tin, too, it has been shown that the physical properties aredependent on the previous thermal history of the sample,37 and itis only recently38 that the true densities and specific heats of thewhite and grey varieties have been determined. A similar uncer-tainty exists in the case of84 W. Herz, 2.anorg. Chem., 1937,159, 316; A., 189.35 “ Physico-chemical Metamorpbosis and Problems in Piezo-chemistry,”McGraw-Hill, 1926, p. 50.8 6 2. physikal. Chem., 1924, 109, 109; A., 1924, ii, 450. See also “ ThePhase Rule,” A. Findlay, Longmans, 1923.3 7 A. Travers and Huot, Compt. rend., 1927, 184, 162; A,, 194.38 E. Cohen and K. D. Dekker, 2. physikal. Chem., 1927, 12’7, 183; A.,39 D. Cannegieter, Chem. Weekblad, 1927, 24, 350; A., 818.818GENERAL AND PHYSICAL CHEMIS!FRY'. 19Compounds axe equally difiicult to deal with. The differentcrystal structures-cubic and hexagonal-assigned to silver iodideby different workers are considered40 to be due to the metastableexistence of one form in the stable region of the other. The heatsof solution of two forms of cadmium iodide have been measured41and earlier discrepant results shown to be due to physical hetero-geneity.Ammonium nitrate is a particularly glaring case in-vestigated by R. G . Early and T. M. Low~Y.*~ It can exist in nofewer than six solid modifications, and recent determinations ofthe transition temperature of, and the volume change accompany-ing, the I11 * IV transformation have been made,43 confirming thetemperature found by Lowry and Early, but differing by 8% fromthe volume change determined by Bridgman.44The question of the preparation of physically pure modificationsof a substance is a difficult one. Since, in any change of state,metastable modifications may be produced in preference to stableones (Ostwald's rule of the succession of phases) it is obvious thatcrystallisation, freezing, sublimation, distillation, etc.-the verymethods employed for chemical purification-will almost inevitablylead to the formation of metastable modifications.The question ofphysical purification therefore resolves itself into one of acceleratingthe stabilising change after the metastable modification has beenproduced, and it has been shown that repetition of the transitionprocess a number of times is one method of effecting this45-thepresence of a solvent for one form,46 or even of water 47 or electro-lytes, is another. The presence of adsorbed impurities48 is alsoeffective in many cases. The only evidence of physical purity isthe constancy of physical properties of different specimens-anegative test, but the only one available.The effect of intensive drying on the properties of liquids andsolids has been studied-particularly by H.B. Baker,49 A. Smits,60and S. B. MalL51 The results obtained by Smits on the self-40 E. Cohen and A. L. T. Moesveld, 2. physikal. Chem., 1924, 109, 97; A . ,1924, ii, 450.I1 E. Cohen, W. D. Helderman, and A. L. T. Moesveld, ibid., p. 100; A . ,1924, ii, 450.43 J., 1919, 115, 1387.43 E. Cohen and J. Kooy, 2. physikal. Chem., 1924, 109, 81; A., 1924, ii,44 Proc. Amer. Acad. Arts Sci., 1916, 51, 581.4 5 '' Physico-chemical Metamorphosis, etc.," p. 87.46 E. Cohen and A. L. T. Moesveld, 8. physikal. Chem., 1920, 94, 450;46 Amer. J . Xci., 1916, [iv], 16, 504.4n Inter alia, J ., 1922, 121, 668.6o Inter alia, J . , 1924, 125, 2573.61 2. anorg. Chem., 1925,149, 150; A., 1926, 117.449.A., 1920, ii, 611. 4 7 Vide inEra, Ref. 6820 ANNUAL REPORTS ON THE PROGRESS OF CHIEMISTRY.intensive drying of sulphur trioxide and phosphorus pentoxide havealready been described.52 The general effect of intensive drying onliquids has been to lower the vapour pressure, but that of nitrogentetroxide 53 was raised. H. B. Baker has now shown that changesin vapour pressure and molecular complexity in liquids can beproduced by means of catalysts (the author's term) such as char-coal, platinum-black, or th0ria.~4 His view is that " all liquidsmay be regarded as analogous to a dissociable gas such as nitrogentetroxide. . . . Liquids differ, however, in this respect, that dis-sociation and association are much slower in liquids than in gases.Equilibrium in liquids may be profoundly disturbed by even a com-paratively slight change of temperature, and complete recovery ofthe normal condition may be a question of weeks or even months.''These observations are confirmed by J.B. Peel, P. L. Robinson,and H. C. Smith,55 who report changes in density under similarconditions. The following figures for water with thoria as catalystgive an idea of the magnitudes of the observed changes :Molecular complexity. Density.After3 weeks ......... 3.125 x 18 After 1 day ............ - 0.000179 weeks ......... 3.866 x 18 4 days ......... + 0-000019 days ......... + 0.000155 weeks .........3-612 x 18 2 days ......... - 0.000028 days ......... + 0.00011Vapour pressure : + 2-4 111~. before heating; + 4.0 mm. after heating at80" for 48 hrs., then cooling to 20"; + 1.2 mm. 1 day later; + 0.9 mm.2 weeks later.In this connesion, the observations of G. Tammann 56 are ofinterest. This author, from an examination of specific volume andcompressibility data, deduces the presence of large molecules of" water-I," (H,O),, with the space lattice of ice-I. The concen-tration of these molecules falls with rise of temperature or pressure,and disappears a t 50" or 2500 kg./cm.2. With regard to the effectof intensive drying on the properties of liquids, D. Balarev 5' statesthat repetition of Baker's experiments gave liquids which invariablycontained phosphorus, and he suggests that the formation of volatilecompounds of phosphorus is the cause of the observed phenomena.Baker, however, has pointed out 57a that, apart from the factthat the large elevations of boiling point observed could not be dueto this cause, his published paper definitely states that all the52 Ann.Reports, 1924, 21, 30.68 A. Smits, W. de Liefde, E. Swart, and A. Claassen, J., 1926, 2657;64 J . , 1927, 949.5 6 Nature, 1927, 120, 814; A., 1019.5 6 2. anorg. Chem., 1926, 158, 1; A., 1927, 93.67 J . pr. Chem., 1927, [ii], ll$, 5 7 ; A., 613.57a J., 1927, 2902.J. W. Smith, J., 1927, 867GENERAL AND PHYSICAL CHEMISTBY. 21liquids were examined for phosphorus, with negative results. Hesuggests that Balarev’s results are possibly due to the presence ofvolatile phosphorus trioxide in the pentoxide used.A.Smits 58 considers that every phase of an allotropic substancecontains different molecular species which may be isomerides,polymerides, or dissociation products of the simple molecule.Normally, inner equilibrium between these forms is rapidly estab-lished, and the system behaves as a unary one, but in certain cir-cumstances (intensive drying) the establishment of this innerequilibrium may be rendered very slow, or even stopped altogether.Alternatively, the inner equilibrium may be fkst displaced and thenfixed in its new position. In the simple case where two componentsonly are concerned, three cases may arise :FIG. 1.dmJmoistA X . iComposition.FIG. 2.moistdryA X.Composition.(i) Both forms are more stable in the moist than in the dry state.Then two curves for the chemical potential (C) as a function of thecomposition (x) are obtained (Fig.l), and it will be a coincidence ifboth curves have minima (equilibrium points) a t identical valuesof x. The equilibrium position will therefore be shifted by intensivedrying, but the direction of the change cannot be predicted.(ii) A may be more stable in the moist, and B in the dry, state.We then get the condition of affairs represented in Fig. 2 and theconclusions are as before.(iii) Intensive drying may have no effect on the stability ofeither form-the <--x curves for moist and dry conditions are thenidentical, and no effect will be produced by intensive drying.It is obvious that these views can be applied to the action ofcatalysts other than water.68 J., 1926, 2655.Compare A. Smits, “ The Theory of Allotropy,” Long-mans, 192222 ANNUAL REPORTS ON THE PROGRESS OF CHEMISTRY.Strong Electrolytes.The Debye-Huckel theory still continues to attract considerableattention and much work has appeared on this subject during theyear. Hence, although this section was included in the Reportsfor the last two years, which stressed mainly the agreement of thetheory with observation, the opportunity is taken this year ofrecording some of the work which has gone to make up the otherside of the picture.W. D. Bancroft 59 has revived H. E. Armstrong’s conception ofthe complexity of water, although in a more comprehensible form.He suggests that many of the anomalies of strong electrolytes, inparticular the neutral-salt effect and the failure of the Ostwalddilution law, may be due to a disturbance of the equilibria betweenthe various polymerised or associated molecules of which liquidwater is generally considered to consist.He is of the opinion that“ there is very little to be gained by devising empirical formulaswhich ignore this factor.” His pronouncement,60 that physicalchemists in this field are ignoring reality and working with “ slightlypolluted water,” since their formula break down at concentrationsgreater than about O-OlM, will possibly pass into history.The bulk of the work recorded goes to demonstrate the essentialcorrectness of the theory and, where disagreement is expressed, itis generally on matters of detail.On the question of completeionisation, for instance, the Arrhenius theory leads to figures foruni-univalent strong electrolytes from conductivity measurementsof about 90-95% ionisation at about 0~01--0~001M, whereas thenew theory demands 100~o. The tendency now seems to be toadmit a compromise, in some cases up to a figure of 3% (or there-abouts) of ionic association. For instance, K. Fajans 61 has em-ployed refractometric means to test the theory and finds that hismeasurements indicate, in general, incomplete ionisation and, inparticular, that in solution a chlorine ion can approach more closelyto a lithium ion than to a sodium ion, a conclusion in direct conflictwith the theory.It is urged, therefore, that the parameter whichDebye calls the radius of the ionic atmosphere has not really thatphysical significance. D. A. MacInnes and I. A. Cowperthwaite,62from measurements of transport numbers, draw similar conclusions.L. 0nsagerG3 has modified the Debye-Hiickel equation for con-ductivity g4A, - A, = Ao(Klwl + K2b)d.%59 J . Physical Chenz., 1926, 30, 1194.6o J . Amer. Chem. SOC., 1926, 48, 94 (Jubilee No.).G3 Ibid., p. 341; A., 1031.This Report, p. 13. 62 Trans. Par&qSoc., 1927,23,400; A,, 1031.fie Ann. Reportp, 1925, 22, 36GENERAL AND PHYSICAL CHEMISTRY. 23by introducing a correction for the Brownian movement of theions. In the case where the mobilities of the ions of a uni-univalentelectrolyte are equal, the correction term becomes 2 - 4 3 , so thatthe equation becomesA, - A, = A,[K1(2 - 4 2 ) + K2b]d%since wl = &(l& + &/la), where la and I, are the mobilities andare equal in this case. For potassium chloride, with A, = 129.9,the original expression leads toA, - A, = 0*5471/%,whereas Onsager's givesA, - A, = 0*4332/2C,as against the experimentally observed value 0.4612/%.Eventhen, however, the results indicate association of the ions intomolecules a t quite low concentrations. H. B. Hartley and R. P.Bell,65 from conductivity data, and C. A. Kraus and R. P. Seward,60from solubility data, consider that incomplete ionisation must berecognised in solvents other than water and possibly methyl alcohol.Both G.Nonhebe1,G' from E.M.F. measurements, and C. W.Davies,68 from conductivity measurements, prefer the Milnercoefficient in the activity equationto express their results. (For uni-univalent electrolytes, the Debye-Huckel theory requires A = 0.5, and the MiJner theory at a i d c a n tconcentrations requires A = about 0.39.)The failure of the DebyeHuckel theory to deal with the case ofsmall ions of high valency has been shown by N. Bjerrum 69 andD. L. Chapman 70 to be due to the inapplicability of the approxim-ation sinh + / E t = +/kt under these conditions. W. Nernst andW. Orthmann 71 state that the heats of dilution of salts of thesame valency type a t low and very low concentrations are not thesame, and are in some cases even negative, and hence do not agreewith the theory, and their view receives partial support from thework of P.G r ~ s z . ' ~ N. B j e r r ~ m , ~ ~ however, asserts that these andlogf= - Ad;6 5 Trans. Paraday SOC., 1927, 23, 396; A,, 1032.O6 Ibid., p. 488; A., 1021.67 Phil. Mag., 1926, [vii], 2, 1086; A,, 1927, 21.O 8 Ibid., 1927, [vii], 4, 244; A,, 936.KgE. Damke VVidenskab. Selsk. math.-fys. Medd., 1926, 7, [9], 1; A,,1927, 314. '' TTan8. ParaChy SOC., 1927, 23, 443.71 Sitzungsber. Preuss. Akad. Wiss. Berlin, 1926, 51; 1927, 136; A., 1926,73 Trans. Faraday SOC., 1927, 23, 445; A., 1028.579; 1927, 733. 73 Monatsh., 1927, 48, 243; A,, 94024 ANNUAL REPORTS ON THE PROGRESS OF CHEMISTRY.other small discrepancies may be quantitatively explained byassuming a decrease of the effective dielectric constant of thesolution in the immediate neighbourhood of the ions.One clear-cut issue is the formulation of an equation to representthe conductivity of a strong electrolyte as a function of the con-centration.Four such equations have recently been proposed :(1) 74 1 - ~ , p , = o q V - i)[pco'5 - 2yc + 3 w 5 . . .I(2) 75 A, - A, = Bc"(3) 76 A, - A, = Aco5 - ABc( 4 ) 77 A, - A, = A / ( B + c-"')where A, and A, are the equivalent conductivities a t infinitedilution and concentration c, respectively, v is the number of ionsinto which the electrolyte dissociates, and p, y, 6, A , B, and n areconstants typical of the electrolyte. Of these, all except (2) aredevised to reduce to the formwhen c is small, in order to be so far in accord with the Debye-Huckel theory.Equation (2) has the advantage that it does notprejudge the issue as to the value for the index for c. By suitablechoice of variables, Ferguson and Vogel have established the facts,not only that equation (2) will represent the experimental figures,but also that n, when determined without any preconceived ideas asto its value, never has the value 0.5, and varies from electrolyteto electrolyte. Their view is that n and B are functions of theionic masses and that conductivity measurements do not supportthe DebyeHuckel theory, a t least in its present form. Theirconclusions are thatA, - A, = XCO'~(a) the constants B and n vary from electrolyte to electrolyte ina regular manner for related electrolytes ;(b) the extreme variations of n are from 0-3742 for potassiumchloride at 25" to 0.9687 for iodic acid at the same temperature,although most of the values lie within 20% of 0.5;(c) both B and n vary with temperature, although there are notsufficient data available to determine the precise mode of thisvariation ;( d ) the formula is applicable to uni-uni-, uni-bi-, and bi-bi-valentelectrolytes, and to water and (as far as data are available) methyl74 B.Szyszkowski, Bull. Acad. Polonaise, 1926, [A], 325; A., 1927, 415.75 A. Ferguson and I. Vogel, Phil. Mag., 1925, [vi], 50, 971; 1927, [vii],4, 1, 233, 300; Trans. Faraday SOC., 1927, 23, 404; A., 1925, ii, 1163; A,,936, 941; I.Vogel, Phil. Mag., 1928, [vii], 5, 199.7 g Inter alios, L. Onsager, Trans. Paraday SOC., 1927, 23, 341.7 7 R. T. Lattey, Phil. Mag., 1927, [vii], 4, 831GENERAL AND PHYSICAL CHERIISTRY. 25alcohol and nitromethane as solvents. It is also applicable tostrong acids such as hydrochloric and iodic.It is of course possible that their results are capable of other inter-pretations. The fact that an empirical equation will fit given setsof figures is no guarantee a t all that it is the correct equation toapply, and it is obvious that the good agreement obtained byFerguson and Vogel by variation of the coefficient and the exponentof c can equally well be obtained by varying A and B in equations(3) and (4). Indeed, equation (4) represents the facts for potassiumchloride in aqueous solution a t much higher concentrations thanany of the others, except equation (1) with its unlimited supply ofarbitrary constants, but in doing so it would seem to prove toomuch, since it applies in regions of concentration for which theDebye-Huckel theory itself is not valid.Nevertheless, Fergusonand Vogel have effectively refuted the claim that the simple square-root formula of Kohlrausch best represents the dependence of theconductivity of strong electrolytes on concentration, and they havedone yeoman service by their resolute appeal to the test of experi-ment at a time when theoretical formulae are being applied some-times with more enthusiasm than discretion.Equilibria between Gases.S. W. Saunders 78 has collected and analysed the available dataon the molecular heats of gases, heats of reaction, and chemical andequilibrium constants for ten well-known reactions involving carbon,hydrogen, oxygen, and nitrogen.The results obtained should be ofvalue in the study of fuel-gas production, reactions in the cylindersof internal-combustion engines, detonation of high explosives, etc.The method employed was to fit equations empirically to the mole-cular heat-temperature curves in each case, and then, by integratingthe van ’t Hoff isochore and applying the Nernst heat theorem, tocalculate the best values for the equilibrium constants. Since theavailable data are not critically accurate, the Nernst conventionalchemical constants were used for the calculation. Where chemicalconstants were not available, they were calculated either by thevan der Waals-Nernst equation or by the Trouton-Nernst rule.In general, the equations obtained for the equilibrium constantsagree well with the experimental figures when these are available,and the results agree on the whole with those published by G .N.Lewis and M. Randall.79 J. R. Partington and W. G. Shillings0have critically surveyed the figures for the water-gas equilibrium’* S. W. Saunders, J. Physical Chem., 1924, 28, 1151; A., 1924, ii, 836.7D “ Thermodynamics,’’ McGraw-Hill, 1923.80 J . SOC. Chem. Id., 1926, 44, 1 4 9 ~ , 242r26 ANNUAL REPORTS ON THE PROGRESS OF CHEMISTRY.and, by following substantially the same procedure except thatthey used experimental values and did not apply the Nemttheorem, obtained the following equation :log,, K =z - 2125/T + 1.077 loglo T - 0.000898T +0~000000133T2 - 0.5425where K = pco, .p ~ , / p c o ~ .p~,,. The agreement of Saunders'sfigures with those of Partington and Shilling is illustrated below :T" K. .................. 1000 1200 1400 1600 1800K (Saunders) ......... 0.71 1.29 2-04 2.95 3.72K (P. and S.) ...... 0.63 1-31 2-15 3-05 3-91K (exptl.) ............ 0.65 1.32 2-08 2.95 3.80All the above values refer, of course, to true equilibrium whichtakes some time to obtain. In actual combustion practice, thegases are rarely in contact for a sufficiently long period of time forthe attainment of true equilibrium. It has been found, however,by R.T. Haslam 81 from his own results and those of other workers,that in actual water-gas producers, a false equilibrium correspondingfairly well to K' = O.O96L, where L is the depth of the fuel bed infeet, is attained.R. C. Cantelo 82 has investigated the equilibriumC(amorph.) + 2H2 CH, + 21,730 cal.both theoretically and experimentally. In the theoretical in-vestigation he takes account of ten possible reactions involvingmethane, ethane, ethylene, acetylene, benzene, carbon, and hydro-gen, and, by an application of the Nernst approximation formula,shows that in all cases the final equilibrium system consists ofmethane, carbon, and hydrogen, with less of the first as the tem-perature rises. He confirms the equation deduced by Saunders(Zoc.cit.) and shows that the apparent disagreement between theearlier results of Mayer and A l t m a ~ e r , ~ ~ Bone and Coward,*4and Coward and Wilson85 may be reconciled by the use of theabove value for the heat of reaction with amorphous carbon, inplace of the value 18,500 cal./g.-atom for graphite. In order toattain equilibrium in a reasonable time, it is necessary to use anactive catalyst (nickel in this instance) and to pass the gaseousmixture repeatedly over it. From the results obtained, the free-energy decrease for the above reaction is calculated to be - AFZQ8 =14,500 cal., which, combined with the value - AFgg8 = 12,800 cal.I n d . Eng. Chem., 1924, 16, 782.82 J . Phyaical Chem., 1926, 30, 1641; 1927, 31, 124, 246, 417; A., 20, 204,321, 322.Ber., 1907, 40, 2134; A ., 1907, i, 457.*4 J., 1908, 93, 1197.06 J., 1919,115, 1380GENERAL AND PHYSICAL CHEMISTRY. 27given by Lewis and Randall86 for the same reaction usinggraphite, leads to C(amorph.) = C(graph.); - AF298 = 1700 cal.F. E. C . Scheffer, T. Dokkum, and J. Als7 obtain results by asimilar method which are in even better agreement with Saunders'sequation for the dissociation of methane and point out that, a tlower temperatures with a nickel catalyst, a carbide of nickel isformed which alters the equilibrium equation by changing theheat of reaction.R. W. Penning and H. T. Tizard 88 have investigated the dis-sociation of carbon dioxide at high temperatures and pressurescomparable with those obtained in internal combustion engines.The method employed was to explode standard mixtures of nitrogenand oxygen containing various amounts of carbon monoxide atconstant initial temperatures and pressures. In this way theywere enabled to determine the carbon monoxideoxygen ratiogiving the maximum pressure at any temperature, and could varythe temperature by varying the nitrogen-oxygen ratio.They findthat the carbon monoxideoxygen ratio has little effect on theexplosion pressure, thus indicating considerable dissociation ofcarbon dioxide under the conditions of their experiments (30 atm.pressure and 3000" K, approximately). Their results are expressedby the equationwhich gives results considerably lower than the accepted values,The usual procedure, the calculation of equilibrium data frommolecular heats and heats of reaction, has recently been reversedby W.G . Shilling,89 who has calculated the molecular heats ofnitrogen, oxygen, nitric oxide, carbon monoxide, carbon dioxide,and ammonia from considerations of various gaseous equilibria inwhich these gases participate. The results are in good agreementwith the accepted data.has investigated the thermal dissociation of carbony1 chloride at temperatures between 360" and 480" by twomethods, chemical analysis of the products of dissociation afterheating to a constant temperature, and physical measurement ofthe increase of pressure on dissociation under tho same conditions.It was found that the carbon monoxide and chlorine produced bythe dissociation were not equivalent when the carbonyl chloride washeated in glass vessels, and this was traced to the attack of the glasslog Kp = log p2co .p~,/p2co, = 8.46 - 28,60O/T,H. Ingleson88 " Thermodynamics," p. 672.87 Rec. trav. chim., 1926, 45, 803; A., 1927, 29.8 8 Proc. Roy. SOC., 1927, [ A ] , 115, 318.yo J . , 1927, 2244.Trans. Faraday Soc., 1926, 22, 377; A,, 1927, 1228 ANNUAL REPORTS ON THE PROGRESS OF CHEMISTRY.by the chlorine. Quartz vessels gave satisfactory results. A linearrelationship is found to exist between log K and 1/T, and a valueis deduced for the heat of dissociation at constant pressure of- 25,500 cal. in good agreement with the value obtained byThomsen. Little change in this value occurs with change oftemperature.P.G. Colin and H. V. Tartarm have investigated the form-ation of nitric oxide in the high-tension arc and agree with theview expressed by Daniels, Keene, and Manning that the " tem-perature " of an electric discharge has no thermodynamic significance,since some molecules become charged and accelerated by the fieldso that the Maxwell distribution of velocities is no longer obtained.In any case there is a region in the immediate vicinity of the dis-charge with a lower temperature, which, however, is still highenough to ensure rapid attainment of equilibrium. Provided thatthis temperature is taken as the thermodynamically significant one,the authors maintain that the law of mass action holds approx-imately for this reaction in the high-tension arc at pressures greaterthan half an atmosphere.It is considered that the equilibratedmixture is " frozen " so rapidly after leaving the arc that the useof water-cooled surfaces for this purpose is futile.Combustion and Flame.Considerable attention is being paid at present to radiation andionisation effects in combustion processes. W. A. Bone and hiscollaborators have greatly extended their work on gaseous explosionsat high pressures, following the course of the explosion by photo-graphically recorded tim*pressure curves. The marked differencebetween the rapid explosions of hydrogen-air mixtures as com-pared with the comparatively slow explosion of carbon monoxide-air mixtures at 50 atm. initial pressure was originally tentativelyexplained91 as being due to the fact that in the former case theenergy of reaction was released entirely in a kinetic (temperature)form, thus accelerating the reaction, which proceeded rapidly toequilibrium and was immediately followed by cooling by the wallsof the apparatus.In the second case, it was considered that thenitrogen absorbed some of this energy and stored it for a time in apotential form (activation), thus causing the reaction to proceedmore slowly, but that this energy was afterwards released in theform of heat, thus delaying the cooling. Subsequent work hasconfirmed this view in a striking manner. Mixtures of carbon90a J . Phyeical Chem., 1927, 31, 1539.QOb Trans. Amer. Electrochem. Soc., 1923, 44, 247.Dl Ann.Repom, 1923, 20, 20GENERAL AND PHYSICAL CHEMISTRY. 29monoxide and oxygen with no nitrogen present behave just likehydrogen-oxygen and the addition of small quantitiesof hydrogen to carbon monoxideair mixtures is suf6cient to changethe character of the explosion to the hydrogen-air type. Thislatter effect is explained by the theory that the hydrogen fist burnsto steam, which then oxidises the carbon monoxide so that none ofthe characteristic carbon monoxideoxygen radiations (which aloneare absorbed by the nitrogen) is emitted. This is further confirmedby experiments carried out by F. R. Weston93 on the spectra ofcarbon monoxide-air and carbon monoxide-hydrogen-air flames.The former show the characteristic blue colour and continuousspectrum of burning carbon monoxide, but as more hydrogen isadded the blue colour and the continuous spectrum disappear andsteam lines appear in the spectrum until, when equimolecular pro-portions of carbon monoxide and hydrogen are present, the appear-ance and spectrum of the flame are practically those of hydrogen.The effect of increasing initial pressure on the rapidity of theexplosion was further studied,94 and it was found that hydrogen-air and carbon monoxideoxygen mixtures show increasing rapiditywith increasing pressure, but that carbon monoxideair mixturesshow the reverse effect.This would be expected, since increase ofpressure increases the density and hence the absorbing capacity ofthe nitrogen present, and results in the fact that carbon monoxideair and hydrogen-air mixtures behave similarly a t lower pressures.Further evidence in support of the main hypothesis is obtainedfrom the observation 95 that, with excess air, carbon monoxide-airmixtures give quantities of oxides of nitrogen in excess of thethermodynamic yield calculated for nitrogen-oxygen mixturesalone. This would be expected if the carbon monoxide-oxygenreaction activates the nitrogen.Other diluents, except possiblyhelium,g6 have little or no effect on the carbon monoxide-oxygenreaction. Finally, direct confirmation of the theory was obtainedby a spectrographic examination of the ultra-violet radiationemitted during the explosion.97 Marked absorption was observedwhen nitrogen or excess carbon monoxide was present, and theabsence of bands due to oxides of nitrogen shows that the form-ation of these compounds is a secondary and later reaction.The92 W. A. Bone, D. M. Newitt, and D. T. A. Townend, Proc. Roy. SOC.,1923, [ A ] , 103, 205; d., 1924, ii, 398.93 Ibid., 1925, [ A ] , 109, 176, 523; A . , 1925, ii, 928; A., 1926, 8.94 W. A. Bone, D. M. hTewitt, and D. T. A. Townend, ibid., 1924, [A], 105,95 Idem, ibid., 1925, [ A ] , 108, 393; A., 1926, ii, 800.9 6 Idem, ibid., 1926, [ A ] , 110, 645; A., 1926, 480.406; A . , 1924, ii, 398.W. A. Bone and D. M. Newitt, ibid., 1927, [A], 115, 4130 ANNUAL REPORTS ON THE PROGRESS OF CHEMISTRY.experiments have been extended to the explosion of methane-oxygen rnixturesy98 and it is found that nitrogen is not activatedin this case.The results support the following scheme of reaction :-+ H,iC*OH 7- -+ H,:C:O + H,Ooxidation viaoxidation HGC: (OH),I +- 13.4 Oal.CO + H,IJ/- 22.8 Gal.CH4 TizxIJ.- 21*7 c + 2% CO + 2H,with the possibility of5 CH,*OH = 1 H, j + H, + COCH,*OH = j . .- :CH, - --___ I + H,Ooccurring a t lower temperatures. The main conc1usion.s obtainedfrom the above research have been confirmed by C. F. R. Harrisonand J. P. B a ~ f e r , ~ ~ who followed the course of the reactions bymeans of temperature measurements. W. E. Garner and C. H.Johnson1 have studied the effect of the addition of various sub-stances on the infra-red radiation emitted by the carbon monoxide-oxygen reaction. They find that small quantities of water, ethyliodide, or ethyl nitrate accelerate the explosion and depress theamount of infra-red emission, and that carbon tetrachloride andnitrogen peroxide have the opposite effect.They suppose that theacceleration of the reaction is caused by the conservation of theenergy within the system and that the dissipation of the energy byinfra-red radiation has the opposite effect. They propose the termenergo-thermic catalysis for this conserving phenomenon and suggestthe following scheme :infrct-red 2C0 4- 0, s internal += 2C02 + radiation2c0,[+energy 1+ X / \ +Y (energo-thermic catalyst)$2c0, 2c0, + infra-red radiation + thermal energyG. L. Wendt and F. V. Grimm employ Sir J. J. Thomson’s sug-gestion that an explosive flame is propagated by the emission ofelectrons from the reacting molecules and that the advance of theseelectrons before the flame front ionises and activates the unburntmolecules, ultimately causing detonation at high temperatures andpressures.They suggest that “ anti-knocks ” produce their effectss8 D. T. A. Townend, Proc. Roy. SOC., 1927, [A], 116,637; A., 1146.9s Phil. Mag., 1927, [vii], 3, 30.Ibid., p. 97.Ind. Eng. Chem., 1924, 16, 890GENERAL AND PHYSICAL CHEMISTRY. 31by removing these electrons. A considerable amount of work hasbeen done on this problem, and W. E. Garner 3 has summarised ourknowledge of explosion reactions from this point of view. Incollaboration with S. W. Saunders he has studied thermal ionis-ation in gas explosions and gas reactions.The Saha equationis applied to the ionisation of gases in explosions, and it is pointedout that, although the heat of ionisation (represented by the firstterm) may be as large as 350,000 cal. (corresponding to an ionis-ation potential of about 15 volts), yet, since the thermal energyliberated by chemical means during the explosion cannot be dis-tributed instantaneously amongst the molecules present, the Max-well distribution of energy may be momentarily disturbed andthere may be more molecules with large energy content than thelaw predicts. In effect, this implies a reduction of the energyrequired for ionisation, so that Kp will be larger than the Sahaequation predicts. In order to decide this point, measurementswere made of the electrical conductivity of exploding mixtures ofhydrogen and oxygen.6 It was found that the results agreed asclosely as could have been expected with the Saha equation, so thatlittle, if any, of the ionisation was due t o chemical energy suppliedas such and not thermally.The addition of anti-knocks was foundto diminish materially the electrical conductivity of such mixtures.S. W. Saunders finds that the increase of conductivity of anexploding mixture of hydrogen and oxygen in a spherical bomb isdirectly proportional to the distance travelled by the explosionwave in the firing tube. He suggests that this is caused by theincrease of ionisation consequent on the rise of temperature due tothe d u x of hot gases or flame from the firing tube.With K. Sat0 *he has studied explosions of carbon monoxideoxygen mixtures ina similar manner, and finds that the addition of hydrogen or waterto the dry gases materially increases the ionisation produced duringthe explosion, but that the duration of the ionisation is muchgreater when dry gases are used, this being in agreement with thelonger time of explosion (Le., the slower development of pressure)in this case. He has also investigated methane-oxygen andacetylene-oxygen mixture^,^ and finds that the maximum elec-trical conductivity occurs in the former case when the hydrocarbonis burned to carbon dioxide and water, and in the latter whenburned to carbon monoxide and water. J. A. J. BennettlO haslog Kp z= - 5048V.’/T + 2.5 log T - 6.56Trans. Famday Soc., 1926, 22, 263.Ann.Reports, 1923, 20, 5.6 Trans. Faraday Xoc., 1926, 22, 253.8 Ibid., p. 248.* Ibid., p. 281.’ Ibid., 1927, 23, 242.Ibid., p. 256. lo Ibid., p. 30732 ANNUAL REPORTS ON THE PROGRESS OF CHEMISTRY.measured the electrical conductivity of various flames, and hasstudied the effect of the addition of various substances thereon.He finds that, although ionisation accompanies detonation, it isnot related to it. This is confirmed by S. C. Lind,ll who agreeswith W. H. Charch, E. Mack, and C. E. Boord l2 that ionisationaccompanies, but does not cause, detonation.Ionisation has also been found to accompany slow combustion,13and J. A. J. Bennett and E. W. J. Mardlesl* have shown thatdetectable ionisation does not precede chemical change, and thatanti-knocks increase the temperature of first detectable ionisation.On this account they are inclined to favour the theory that per-oxides l5 are formed immediately prior to combustion, since thecomplete scission of the oxygen molecule required by the hydroxyl-ation theory might be expected to produce a more copious supplyof electrons.J. S. Lewis,16 however, finds that for paraffin hydro-carbon-air mixtures there exists a critical temperature at whichrapid chemical action takes place with an increase in the numberof molecules, and considers that his results support the hydroxyl-ation theory. A. Egerton and S. I?. Gates l7 find that anti-knocksdo not appreciably affect the detonation of acetylene-air or pentane-air mixtures a t ordinary or high temperatures and pressures (230”,10 atm.), but that they do retard the rate of slow combustion ofpentane.18 They point out that “knocking” is associated with avibratory form of combustion and not with detonation.SimiIarconclusions as to the effect of anti-knocks on the combustion ofhydrocarbons are reached by W. G. Lovell, J. D. Coleman, andT. A. Boyd.19The general conclusions appear to be that gaseous ionisation is athermal, and not a chemical effect, that there is little, if any, con-nexion between ignition temperature and ionisation, or betweenionisation and detonation, and that therefore Wendt and Grimm’shypothesis is a t fault.The law of flame speeds 20 has been reiterated by W. Paymanand R.V. Wheeler,21 but has been subjected to examination andl1 Trans. Paraday SOC., 1926, 22, 291.l2 Ind. Eng. Chem., 1926,lS, 334; B., 1926, 570.l3 J. A. J . Bennett, Trans. Paraday SOC., 1927, 23, 295.l4 J., 1927, 3156.l5 H. L. Callendar, Engineering, 1927, 123, 147.l6 J., 1927, 1555.l7 Proc. Roy. SOC., 1927, [A], 114, 137, 152; A., 318.l8 Ibid., 1927, [ A ] , 116, 516.2o Ann Repon%, 1922,19, 20.Ind. Eng. Chem., 1927, 19, 373.Trans. Famday SOC., 1926, 22, 301GENERAL AND PHYSICAL CHEMISTRY. 33severe criticism by W. A. Bone, R. P. Fraser, and D. A. Winter.22These authors conclude that it does not hold generally, since it isdemonstrably false for both slow- and fast-burning mixtures ofcomplex hydrocarbon, hydrogen, and oxygen (or air).In the samepaper, the validity of the suggestion23 that, with central ignitionin a spherical vessel, the maximum explosion pressure is developeda t the instant of contact of the flame front with the walls is ques-tioned. It would seem, however, that there is no doubt aboutthis fact, since 0. C. de C. Ellis and R. V. Wheeler 24 have pub-lished convincing photographs which show that, provided the flamefront travels quickly enough to eliminate convection effects, theinstants of maximum pressure and of flame contact with the wallsof a spherical vessel do coincide. With a cubical vessel undersimilar conditions, this result would hardly be expected, and it isfound in fact that pressure continues to be developed after con-tact. The same authors have studied ignition of gases in cylindricaland spherical vessels 25 and find in all cases a luminous region behindthe flame front.They ascribe this to the " after-burning " of com-bustible gas left behind by the flame front, and conclude that,although the explosion is complete at the instant of maximumpressure, the combustion process continues for some time afterwards.The Catalytic Catenary.Continuing their previous work 25a on the catalytic minimum-velocity point in the iodineacetone reaction, H. M. Dawson andhis co-workers 26 have obtained results supporting the " dual "theory of catalysis in an extended form. They have shown thatthe velocity of this reaction, when occurring in acetic acid-sodiumacetate b d e r solutions, is influenced, not only by the hydrogen-ionconcentration, but also by the concentrations of other ions andmolecules in the solutions. (The word concentration in this sectionmeans molar concentration uncorrected by any thermodynamicactivity factor.) The velocity equation then becomes :V = v h + Urn Vu + VOH = h[H'] -I- k m [ m ] + kJA-1 +kla[OH-] (1)terms for the hydrogen and hydroxyl ions, the acid anion, and theundissociated acid molecule entering the velocity equation, but the22 Proc.Roy. Soc., 1927, [A], 114, 420.28 J., 1923, 123, 1257.24 J., 1927, 153; compare J., 1925, 127, 760, 764.26 J., 1927, 310.26 H. M. Dawson and C. R. Hoskins, J., 1926, 3166; H. M. Dawson, J.,1927, 213, 458, 756, 1146, 1290; H. M. Dawson and W. Lowson, J., 1927,2107, 2444; H.M. Dawson and C. R. Hoskins, Proc. Leeda Phil. Lit. SOC.,1926,1, 108; A., 1927, 117.Ann. Reports, 1926,23, 86.REP.-VOL. XXIV. 34 ANNUAL REPORTS ON THE PROGRESS OF CHEMISTRY.metallic kation of the salt of the acid is without effect. If theacid concentration (c) in the buffer solution is kept constant and thehydrogen-ion concentration of the solution is sficiently great toallow the effect of the hydroxyl ion to be neglected, and assumingthat dissociation of the salt in the buffer mixture is complete and thatthe ionic dissociation of the acid (dissociation constant = R) followsthe simple mass law (again without activity factors), it can readilybe shown that the velocity is a minimum whenandfrom which it follows that the velocity-pH curve is symmetricalabout pHi.As the acid concentration is constant, the velocity dueto the ions alone is given byThis may be transformed to a general form by an application ofthe method used to obtain a reduced equation from a specificequation of state. Expressing [Hf] in terms of [H+]i([H+] = n[H+]i)and u in terms of ui (u = ruJ, an equation is obtained, r = u/Ui =&(n + l / n ) , between the reduced hydrogen-ion concentration andthe reduced ionic velocity, which, since log n = log [H+] -log [H+]i = pH8 - pH = A p H , becomes r = &(lod% + theequation to a catenary.This appears to be independent of the nature and concentrationof the catalysing acid, the nature of the solvent, and the nature ofthe reaction. Exactly the same equation is obtained by similarconsiderations applied to the catalytic effect of water and its ionsalone and, since in this case km is possibly very small compared withkoI1 and kh, the ionic velocity, u, is identical with the measuredvelocity v.The isocatalytic data in this case are obtained from( 2 ) and ( 3 ) by equating terms relating to HA to zero, and replacingterms in A- by terms in OH-, giving :andIn fact, there appears to be no valid reason for treating the hydroxylion in a different manner from other acid anions as regards itscatalytic properties.[H+J = d k , K ~ / ( k h - km) . . . . (2)Vi = 2 d ( k h - k,)kaKc + k,C . . . (3)u = ~h + VU = (kh - km)[H+] + kuKc/[H+][H+]i = dko,K,ja . . . . . (4)V i = SdkhkOHK, . - * ( 5 )If the ionic velocity equation be writtenit will be seen that u depends on [H+], ui, and [H+]i.Since thGENERAL AND PHYSICAL CHEMISTRY. 35last two are functions of the nature and the concentration of theacid alone, a complete representation of the equation is possible inthree dimensions. Taking u, pa, and c as the variables, and plottingrising values of ZL from bottom to top, rising values of p , from leftto right and rising values of c from front to back, the catalyticcatenary surface appears as a U-shaped valleyz7 with a definiteboundary to the left and in the front. The left-hand boundary ofeach section of the surface parallel to the u-pH plane (sections ofconstant c) terminates a t a point corresponding to salt-free acid,and at higher values of c ends obviously at lower pH values.Theprojection of these terminal points on the u-c plane is a paraboliccurve, whereas on the pH-c plane it is logarithmic. The valleybecomes narrower towards the back (as c increases), but the widthsof the valley sections measured at their left-hand terminal pointsare constant and depend only on the nature of the acid. Loci ofpoints of equal reduced hydrogen-ion concentration (reducedisohydric curves) run along the valley approximately from frontto back. The front termination of the valley is obviously thecatenary (c = 0) characteristic of the catalytic effects of hydrogenand hydroxyl ions. alone. Finally, the u-pE-c surface may betransformed into a v--rpH-c surface by lifting it vertically at each left-to-right section by an amount equal to knLc.If the conditions are such that the catalytic effect of the hydroxylions cannot be ignored, the mathematical treatment is the same,amounting in effect to the introduction of terms involving OH-and H20 corresponding to the terms in (2) and (3) which involveA- and HA. A compound catenary is then obtained with theisocatalytic dataandThere is thus a continuous series of compound catenaries of thisthird type between the H+-A- catenaries of the first type and theH+-OH- catenary, i.e., equations (2) and (3) represent the limitingcase for large acid concentrations, (4) and ( 5 ) for c = 0, and (6)and (7) for intermediate conditions.These considerations have been successfully applied to velocitymeasurements, not only of the iodinoacetone reaction, but also ofthe hydrolysis of ethyl acetate in acetic acid-sodium acetate buffers.T. M. Lowry and G. F. Smith z8 have applied them to the muta-rotation of dextrose and have found a small but measurable catalyticactivity in the kation of a weak nitrogenous base. T, M. Lowry 29[H+], = 2/(kaKc + hmKw)/(k> - L) (6)vi = 22/(k,' - Jcm)(kaKc + konliw) + krnc + Kwcw (7)2 7 J . , 1927, 756. 28 Ibid., p. 2539. 29 Ibid., p. 256436 ANNUBL REPORTS ON THE PROGRESS OF CHEMISTRY.has explained the catalytic effects of various ions and molecdes interms of an electrolytic theory. H. M. Dawson30 has pointed outthat the experimental results obtained by him and his collaboratorsare entirely at variance with the protion theory of catalysis advancedby F. 0. Rice,3l according to which the catalytic minimum of reac-tions affected by both the hydrogen and hydroxyl ions should liea t about pH = 5. Equation (4) shows that at a given temperature,and therefore at a fixed value of K,, the hydrogen-ion concentrationof the minimum point is determined by the ratio of the velocitycoefficients koH and kh. This ratio varies widely according to thenature of the reaction, and a variation of 1 : lo6 is by no meansextreme. This corresponds to a change in [H+]i of 1 : 1000, and toa change in the pH value of the isocatalytic point of 3 units. Theprotion theory has also been criticised on somewhat similar lines byM. Bergstein and M. Kilpatrick, j ~ n . , ~ ~ and by M. Bergstein.33HAROLD HUNTER.30 J . Physical Chem., 1927, 31, 1400; A., 1033.a1 Ann. Reports, 1926, 23, 36.J . Physical Chem., 1936,30, 1616; A., 1927, 214.83 Ibid., 1927, 31, 178; A., 321