SATISE OP ELECTRO-CHE] 
 
 Edited by BERTRAM BLOUNT, F.IC, 
 
 OZONE 
 
 E. K. Ridea! M.B.E.. M,A.(Cantab.), Ph. D, 
 
A TREATISE OF ELECTRO-CHEMISTRY. 
 EDITED by BERTRAM BLOUNT, F.I.C., ETC. 
 
 OZONE 
 
A TREATISE OF ELECTRO-CHEMISTRY. 
 Edited by BERTRAM BLOUNT, F.I.C. 
 
 THE MANUFACTURE OF CHEMICALS 
 BY ELECTROLYSIS. By ARTHUR J. 
 HALE, B.Sc., F.I.C. 
 
 OZONE. By E. K. RIDEAL, M.B.E., M.A., 
 Ph.D. 
 
 Other volumes in preparation. 
 
A TREATISE OF ELECTRO-CHEMISTRY. 
 EDITED by BERTRAM BLOUNT, F.I.C., ETC. 
 
 OZONE 
 
 E. K. RIDEAL, M.B.E., M. A. (CANTAB.), Pn.D. 
 
 t ' 
 
 PROFESSOR OF PHYSICAL CHEMISTRY, UNIVERSITY OF ILLINOIS 
 
 NEW YORK 
 D. VAN NOSTRAND CO. 
 
 TWENTY-FIVE PARK PLACE 
 1920 
 
EDITOR'S PREFACE. 
 
 THE idea of a series of books on Electro-Chemistry emanated 
 not from me, but from Messrs. Constable. Some years back 
 I wrote for them a book called "Practical Electro-Chemistry," 
 intended to cover a great part of the ground of knowledge 
 then extant. Fortunately, knowledge has a habit of growing 
 and of propagating its kind, and my book, in consequence of 
 this, became a "back number". 
 
 The subject of Electro-Chemistry is so ramified and 
 specialised that it was impossible for one man to make a 
 survey of the whole field. This fact is the genesis of the 
 present series in which those who have accurate and intimate 
 knowledge of the various branches of electro-chemistry have 
 undertaken the work for which they are particularly qualified. 
 It will be readily understood that, as the series of books was 
 started at an early period of the war, many contributors were 
 engaged in work of national and primary importance, and 
 were unable, how r ever willing, to apply themselves at the 
 moment to exacting literary work. But this difficulty was 
 gradually overcome, as some prospect of a period to the 
 struggle came within view, with the result which the reader 
 will judge with consideration for the onerous conditions 
 under which my contributors have wrought. 
 
 The monographs resulting from their labours speak for 
 themselves, and if the educational advantages which I have 
 obtained from reading them during their passage through 
 the press is shared by the public, I believe that the thorough 
 and modern work of my friends and collaborators will be 
 appreciated, and such faults as there be will be attributed to 
 the person ultimately responsible the Editor, 
 
 434859 
 
AUTHOR'S PREFACE. 
 
 EVER since the time of its discovery Ozone has attracted the 
 attention of chemists, physicists, and industrialists alike. 
 To the former it presented the first example of a gaseous 
 allotrope of an element, differing from oxygen in many 
 remarkable ways. The physicist frequently came in contact 
 with the substance in his investigations on the conduction 
 of electricity through air, whilst the industrialist was not 
 slow to avail himself of an oxidising agent, unsurpassed in 
 strength, leaving no objectionable material in its wake, and 
 at the same time easy, if indeed somewhat expensive, to 
 manufacture. 
 
 The angle from which Ozone and its modes of prepa- 
 ration was regarded by these three different sets of investi- 
 gators naturally varied, and an endeavour has been made in 
 the following pages to summarise and correlate the many 
 different references which are to be found scattered over a 
 wide field of literature. The merest survey, however, was 
 sufficient to indicate that our knowledge of Ozone, its pro- 
 perties and modes of formation, is exceedingly scanty. The 
 industrialist is ever at hand with extravagant claims as to 
 the utility of " electrified oxygen " ; the evidence as to the 
 chemical behaviour and properties of ozone is somewhat 
 meagre and frequently conflicting, for example, the existence 
 of the ozonates and of oxozone still awaits confirmation ; 
 whilst the hypotheses advanced to explain the mechanism 
 of its formation, either chemical, thermal, electrolytic, or 
 photo-chemical, are purely speculative. Ozone is generally 
 produced by means of the silent electric discharge, the 
 
viii AUTHOR'S PREFACE 
 
 Aladdin's lamp of synthetic chemistry, for which no satis- 
 factory "modus operandi " has been suggested, synthesis 
 appearing to result from a combination of photo-chemical 
 action and electron emission. 
 
 A study of the ultra-violet spectrum of oxygen and its 
 allotropes gives us an insight into the various photo-chemical 
 actions involved, and quantitative relationships may be ob- 
 tained by an application of the quantum theory ; at the same 
 time the study of the disintegration or synthesis of the 
 molecules by electron emission is as yet in its infancy. 
 
 The work of Sir J. J. Thomson at the Cavendish 
 Laboratory on the subject of thermionics has opened up a 
 new vista of electro-chemical research, for it would appear 
 that the elements, including oxygen, can exist not only in 
 the form of allotropes, but also as allotropic modifications 
 possessing electrical charges. It remains for the future to 
 reveal the influence of these charges on chemical reactivity. 
 
 Thanks are due to those who have been kind enough to 
 place material dealing with the applications of ozone at my 
 disposal, and if the following pages can assist in stimulating 
 research both scientific and technical in this, one of the most 
 interesting branches of electro-chemistry, the object of the 
 writer will be fully attained. 
 
 EKIC K KIDEAL. 
 
 UNIVERSITY OP ILLINOIS, 
 ILLINOIS, U.S.A., 14th November, 1919. 
 
CONTENTS. 
 
 CHAPTER I. 
 
 PAGB 
 
 EARLY HISTORY OP OZONE AND ITS GENERAL PROPERTIES 1 
 
 CHAPTER II. 
 
 THE NATURAL OCCURRENCE OF OZONE ..... 16 
 
 CHAPTER III. 
 CHEMICAL PRODUCTION 28 
 
 CHAPTER IV. 
 THERMAL PRODUCTION 44 
 
 CHAPTER V. 
 
 THE ELECTROLYTIC PREPARATION OF OZONE ... 57 
 
 CHAPTER VI. 
 
 PRODUCTION BY ULTRA-VIOLET RADIATION AND BY IONIC 
 
 COLLISION 70 
 
 CHAPTER VII. 
 
 PRODUCTION BY MEANS OF THE SILENT ELECTRIC DIS- 
 CHARGE 91 
 
 CHAPTER VIII. 
 
 THE CATALYTIC DECOMPOSITION OF OZONE .... 133 
 
 CHAPTER IX. 
 INDUSTRIAL APPLICATIONS 142 
 
 CHAPTER X. 
 METHODS OF DETECTION AND ANALYSIS 179 
 
 NAME INDEX 191 
 
 SUBJECT INDEX , 195 
 
CHAPTEE I. 
 
 OZONE. 
 
 EAELY HISTOEY. 
 
 IN 1783 Van Marum, a Dutch philosopher, noticed that the 
 air in the neighbourhood of his electrostatic machine (now 
 in the museum at Haarlem, Holland) acquired a marked and 
 characteristic odour when subjected to the passage of a series 
 of electric sparks. Cruickshank in 1801 likewise drew at- 
 tention to the fact that the oxygen gas produced by the elec_ 
 trolytic decomposition of dilute acids under certain conditions 
 was possessed of a similar odour. 
 
 These two investigators merely chronicled the results of 
 their experiments, and did not pursue their inquiries to 
 elucidate the origin of the odoriferous substance. Schonbein, 
 in a memoir presented to the Academy at Munich in 1840, 
 recognised that the smell noted in air subjected to the spark 
 discharge, and in the oxygen generated by electrolysis, was 
 due to the presence of a new gas, to which he gave the name 
 " ozone " (oo> to smell), he also showed that ozone was 
 formed in certain processes of. autoxidation, notably by the 
 action of air on phosphorus, but failed to establish the exact 
 nature or composition of this new substanoe. 
 
 We shall have cause to observe, when discussing the pro- 
 cesses of autoxidation, the development of Schonbein's hypo- 
 thesis in that ozone or active oxygen is produced with its 
 
 1 
 
('ZONE 
 
 electrical isomer * ' antozone " by the disruption of the neutral 
 oxygen molecule 
 
 2 - (X ozone + antozone. 
 
 This hypothesis naturally led to the division of peroxides 
 into two groups, the ozonides and the antozonides, and to 
 an extended search for the two active electrically charged 
 forms of the oxygen atoms. 
 
 Various other speculative hypotheses were made as to 
 the composition of ozone, all unsupported by experimental 
 evidence, thus, Williamson suggested that it might be gase- 
 ous hydrogen peroxide, and Baumert considered ozone to be 
 an oxidised form of hydrogen peroxide, i.e. H 2 3 . 
 
 Becquerel and Freny first showed that oxygen could be 
 completely transformed into ozone, thus proving that ozone 
 was an allotropic modification of this element. 
 
 These experimenters effected the conversion of oxygen 
 into ozone by the passage of a stream of electric sparks 
 through the gas, the ozone formed being continuously re- 
 moved by means of a solution of potassium iodide. In this 
 way all the oxygen originally in the tube ultimately disap- 
 peared. 
 
 Andrews, Tait and Soret (" C.E.," 1876) took up the in- 
 vestigation at this stage, and by the following experiments 
 proved that the allotrope was actually a condensed form 
 of oxygen : 
 
 A tube of volume V connected to a sulphuric acid ma- 
 nometer and containing oxygen gas was submitted to the 
 action of the spark discharge when a contraction in volume v 
 was recorded on the manometer. On heating up the tube to 
 
OZONE 
 
 270 C. the ozone was destroyed and the gaseous mixture then 
 occupied its original volume V. 
 
 Soret showed that no change in the volume of the ozonised 
 oxygen (V - v) took place when the gas was exposed to 
 potassium iodide or metallic silver, nevertheless the ozone 
 was destroyed. 
 
 When, however, the gas mixture was exposed to turpen- 
 tine a further contraction in volume was observed, the final 
 
 C 
 
 Sulphuric Add- 
 in Manometer. 
 
 FIG. 1. 
 
 volume being V - 3v where v was the volume contraction on 
 ozonisation. 
 
 As a result of the experiments Soret came to the conclusion 
 that the molecule of ozone consisted of three atoms of oxygen, 
 three volumes combining to give two volumes of ozone 
 
 30 2 = 2O 3 , 
 
 since the volume pontraction v on ozonisation is clearly equal 
 to one-third of the oxygen converted into ozone (or one-half 
 of the resulting ozone), which is subsequently absorbed by 
 the turpentine. Further, that when ozone reacted with 
 
4 OZONE 
 
 potassium iodide or metallic silver it liberated an equal 
 volume of oxygen : 
 
 3 + 2Ag = Ag 2 + 2 . 
 Soret ascribed the structural formula to the tri- 
 
 
 
 atomic allotrope of oxygen, and confirmed the existence of 
 ozone by a determination of its density. The theoretical 
 density of ozone at N.T.P. should be equal to one and a half 
 times that of oxygen, and this value was obtained by Soret 
 and Otto by several different methods, which may be briefly 
 described : 
 
 A glass globe of about one litre was filled with pure dry 
 oxygen at a determined temperature and pressure ; and 
 subsequently weighed; the oxygen was then displaced by 
 ozonised oxygen reweighed, and the weight of ozone in the 
 flask determined by titration with iodine and sodium thio- 
 sulphate. 
 
 If V be the volume of the globe, containing w grams of 
 oxygen of density J, and w + w be the weight of the ozonised 
 oxygen in the globe, where S is the density of ozone, and v 
 and w" the actual volume and weight of ozone in the globe, 
 then 
 
 (i) w + w' = $v + (V - v)A t 
 
 (ii) w" = vS, 
 
 (iii) w = AV. 
 
 From (i) v(8 - A) = w + w' - A~V = w' y - 
 
 hence : 3 
 
 or 
 
OZONE 5 
 
 From two determinations Otto obtained the values for the 
 
 ratio A 
 A 
 
 = 1-5 - 0-0034 and 1*5 + 0'0035, 
 
 or the density of ozone was practically one and a half times 
 that of oxygen. 
 
 DENSITY BY DIFFUSION. 
 
 Soret showed that the rate of transpiration through a 
 small aperture of the purest ozone which he could obtain was 
 intermediate between the values obtained for chlorine and 
 carbon dioxide. By applying Graham's law to the figures 
 obtained for the time of transpiration of ozone and carbon 
 dioxide, taking t' as the time of transpiration for a volume of 
 carbon dioxide of density A, and t. 2 for an equal volume of 
 ozone of density S, 
 
 +'i A 
 
 then L- -I 
 
 #2 
 
 the value T554 was obtained, taking oxygen as unity. Laden- 
 burg (" Ber.," 1901) at a later date obtained the value T3698 
 for an ozonised oxygen containing 86 per cent, ozone. 
 
 We have already referred to Soret's early experiments on 
 the comparison of the volumes occupied by equal weights of 
 oxygen and ozone, in which the ozone formed from a known 
 quantity of oxygen was removed by absorption in turpentine. 
 From the results of seven experiments Soret obtained a mean 
 value differing by only 2'7 per cent, from the theoretical. 
 
 PHYSICAL PROPERTIES OF OZONE. 
 
 Ozone possesses a strong penetrating and characteristic 
 odour which can be detected in concentrations of one part in 
 
6 OZONE 
 
 a million of air. We may note that there is no unanimity in 
 describing this odour, since it has been likened to sulphur, 
 chlorine and phosphorus (presumably undergoing oxidation 
 when ozone itself would actually be present) ; other observers 
 have compared it to dilute oxides of nitrogen, whilst De la 
 Coux likens it to lobster. 
 
 Dilute ozone is practically colourless, but when viewed 
 through a tube five or six feet long it is found to give a sky 
 blue tint to the column of air. 
 
 Hautefeuille and Chappuis ("C.B.," 94, 1249, 1882) ob- 
 tained liquid ozone by compressing ozonised oxygen to 125 
 atmospheres at a temperature of - 103 C. Liquid ozone is 
 soluble in liquid oxygen, and Ladenburg ("Ber.," 31, 2508, 
 1898) obtained a mixture of liquid ozone and oxygen, con- 
 taining 84'4 per cent, ozone by passing a current of ozonised 
 oxygen through a tube cooled in liquid oxygen, whilst E. 
 Goldstein (" Zeit. Elektrochem.," 50, 972, 1903) obtained 
 pure liquid ozone by immersion of a double walled quartz 
 mercury vapour lamp in liquid oxygen. When a small 
 quantity of oxygen was admitted into the vacuous space it 
 was rapidly ozonised by the ultra-violet light emitted by the 
 mercury vapour and condensed in the form of small drops, 
 the pressure rapidly fell and fresh oxygen could then be 
 admitted. Dewar likewise obtained practically pure liquid 
 ozone by the careful fractionation of liquefied ozonised oxygen. 
 Liquid ozone, which is very liable to explode if accidentally 
 brought into contact with a trace of organic matter or if the 
 temperature be allowed to rise, is a dark blue liquid, opaque 
 in thickness exceeding 2 mm. Olszewski (" Monatsh.," 8, 109, 
 1887 ; " Ann. der Physik," 3, 31, 1887) gave the boiling-point 
 
OZONE 7 
 
 at - 106 C. to - 109 C., whilst Troost (" C.K.," 126, 1751, 
 
 1898) determined it at - 119 C. 
 
 The formation of ozone from oxygen is accompanied by 
 the absorption of heat and the instability of liquid ozone 
 and the gas at ordinary temperatures is doubtless occasioned 
 by its strongly endothermic nature. 
 
 20 3 = 30 2 + 2Q. 
 
 The lowest value of Q, the heat of decomposition per gram, 
 mol. of ozone, is recorded by Hollman in 1868 as 17,064 
 calories, later determinations by Berthelot (1876) gave 29,600, 
 Van de Meulen obtained values between 32,600 and 36,000, 
 whilst Eemsen gives the highest value of 36,600. The most 
 recent observations of Jahn (" Zeit. Anorg. Chem.," 60, 357, 
 1908, and 68, 250, 1910) give 34,000 (see p. 45). 
 
 Ozone is soluble in water, but wide discrepancies are found 
 in the published figures, doubtless occasioned by partial de- 
 composition during solution. 
 
 Schone (" Ber.," 6, 1224, 1873) obtained the value for the 
 solubility coefficient at 18 C. of 0'366, McLeod at 14 C. 
 0-2795, Carius ("Ann.," 174, 30, 1874) at 1 C. 0'834, Laden- 
 burg ("Ber.," 31, 2510, 1898) gave the solubility at 12 C. as 
 O'Ol per cent, by volume, whilst Mailfert (" C.K.,." 119, 951, 
 1894) gives the following values for the coefficient : 
 
 Temperature. Coefficient of 
 
 Solubility. 
 0C 0-64 
 
 11-8 0-5 
 
 15 0-456 
 
 19 0-381 
 
 27 . . . . . . . . . 0-27 
 
 40 0-112 
 
 55 0-031 
 
 60 . 
 
8 OZONE 
 
 about fifteen times the values obtained for oxygen. Mouf- 
 gang ("Woch Brauerei," 28, 434, 1911) determined the 
 following values : 10 mgm. per litre at 2 C. and 1'5 mgm. at 
 28 C. 
 
 In dilute solutions of sulphuric acid (0'03 - 0*09 per cent.) 
 the coefficient of solubility is somewhat higher, as is indicated 
 by the following figures : 
 
 Temperature. Coefficient of 
 
 Solubility. 
 
 30 C 0-240 
 
 33 0-224 
 
 49 0-156 
 
 57 0-096 
 
 Eothmund (" Nernst Festschrift," 391, 1912) has indicated 
 that the above figures are in all probability too low owing to 
 the decomposition of ozone occurring during the estimation 
 of the solubility. He found that this decomposition was 
 remarkably small in O'l N. sulphuric acid at C. and obtained 
 a value 0*487 for the absorption coefficient at this tempera- 
 ture ; when corrected for the salting out action of the sul- 
 phuric acid the coefficient in water would be equal to 0*494. 
 
 Ozone is soluble in acetic acid, acetic anhydride, ethyl 
 acetate, chloroform, and carbon tetrachloride (Fischer and 
 Tropsch, " Ber.," 50, 765, 1917), forming blue solutions which 
 are fairly stable. Solutions of ozone in carbon tetrachloride, 
 in which the solubility is seven times that in water, do not 
 undergo decomposition for twenty-four hours. 
 
 The decomposition of ozone (see p. 133) is frequently ac- 
 companied by a phosphorescence noted by Dewar when pass- 
 ing ozonised air through a capillary opening, and by Otto in the 
 action of ozone on water containing traces of organic matter. 
 
OZONE 9 
 
 A vivid phosphorescence is likewise obtained when a hot 
 glass rod is brought near the surface of liquid oxygen contain- 
 ing ozone ("Beger. Zeit. Elektrochem., 3 ' 16, 76, 1910). 
 
 R S. Strutt ("Proc. Koy. Soc.," 85, 10, 1911) has ex- 
 amined a number of cases of phosphorescent combustions, 
 especially marked in vacuum tubes containing ozonised air 
 under low pressures. Phosphorescence was noticed during 
 the oxidation of a number of substances by ozone, amongst 
 the more important being nitric oxide, sulphur, hydrogen 
 sulphide, ethylene, and iodine. The spectroscopic examina- 
 tion revealed a banded spectrum in the majority of cases, but 
 occasionally continuous spectra were obtained. 
 
 The spectrum of ozone is exceedingly complex and has 
 been the subject of numerous investigations. 
 
 Chappuis ("C.K.," 94, 858, 1882) found eleven lines lying 
 in the region X = 628'5 pp and X = 444 /*/* in the visible 
 spectrum, those lying on either side of the sodium lines being 
 particularly distinct and characteristic, X = 609*5 - 595'5 pp 
 and X = 577 to 560 pp. Schone (" Zeit. Anorg. Chem.," 6, 333, 
 1894) added two to the above number, whilst Ladenburg and 
 Lehmann ("Ber.," 4, 125, 1906) noticed a line in the red 
 portion of the spectrum. 
 
 J. Stark (" Ann. der Physik," 43, 2, 319, 1914) has shown 
 that the ozone molecule gives rise to many bands lying be- 
 tween the visible green and the ultra-violet, X = 210 pp. The 
 
 bands of long wave lengths were found to be resolvable into 
 
 ,. 
 
 line series. 
 
 Certain lines attributed to the ozone molecule are fre- 
 quently caused by other allotropes of the element, either of 
 elementary oxygen in the monatomic or diatomic form, or 
 
10 OZONE 
 
 2 , or of those substances when charged. Thus Stark (" Phys. 
 Zeit.," 14, 720, 1913) has shown the existence of two distinct 
 arc spectra of oxygen attributable to the substances 2 and 
 
 6,. 
 
 The line in the visible red of the spectrum noticed by 
 Ladenburg and Lehmann (loc. cit.) is possibly not due to ozone 
 but to another allotrope of oxygen, viz. oxozone, 4 ; whilst 
 the existence of a band in the infra red or thermal region at 
 X = 1040 w has been claimed for ozone but has not received 
 confirmation. 
 
 The lines of the oxygen spectrum at the negative electrode 
 of a discharge tube were examined by Schuster, Steubing and 
 F. Croze ("C.B:," 153, 680, 1916) who gives the following: 
 X = 685'3 pp, 662-5, 603'2, 564'6, 529'6 and 498. Schuster's 
 two negative bands X = 570 - 584 yu/z and X = 601 - 596 JJL/J, 
 could not be resolved. 
 
 The lines of atomic oxygen are found in the examin- 
 ation of water vapour as well as in oxygen submitted to 
 intense electrical discharges. Fowler and Brooksbank (" Boy. 
 Astron. Soc.," 77, 511, 1917) have likewise shown the pres- 
 ence of lines of this series, the third line spectrum of oxygen 
 X = 559'2 pp, and 39618 in stars of the ft type as well as in 
 Wolf Bayet stars. 
 
 The ultra-violet spectra of oxygen and its allotropes are 
 of special significance in the consideration of their photo- 
 chemical interconversion (see p. 70). 
 
 That of ozone has been examined by Lenard (" Ann. der 
 Physik," i, 480, 1900), Goldstein (" Ber.," 36, 304, 1903), 
 and more especially Begener (" Ann. der Physik," 20, 1033, 
 1906). 
 
020NE 11 
 
 The 2 molecule gives short wave length bands resolvable 
 into lines between the region X 200 and X 188 /*/* correspond- 
 ing to the ultra-violet fluorescence of oxygen. Steubing 
 noticed five bands between X = 183*1, and 191*1 /I/A, whilst 
 L. and E. Bloch (" C.K.," 158, 1161, 1914) isolated two new 
 ones conforming to the Delandres formula at X = 192'3 to 
 193'6, and 194'6 to 195'7 pp. The ultra-violet oxygen atom 
 band in the region X = 230 ftp and X = 340 ^ is observed 
 in the positive column in pure rarified oxygen, and in the 
 decomposition of dissociation of many oxygen-containing 
 compounds. The strongest band (see Meyerheim, Grebe, 
 Holtz and Fowler, " Proc. Koy. Soc.," 94, 472, 1918) is found 
 at X = 306 '4 pp, and is usually attributed to water vapour. 
 
 Investigations on the carriers of positive electricity by 
 Sir J. J. Thomson and his co-workers (" The Carriers of 
 Positive Electricity") have revealed the presence of a great 
 number of allotropes of oxygen which give rise to their re- 
 spective band spectra. F. Horton (" Phil. Mag.," 22, 24, 1911) 
 has shown the existence of carriers of positive electricity in 
 oxygen of electric atomic weights, 8, 16, 32, 48 and 96. 
 
 Becquerel has shown that the magnetic susceptibility of 
 ozone exceeds that of oxygen, and that the ratio of the 
 specific magnetic susceptibilities exceeds that of the ratio of 
 their densities. 
 
 CHEMICAL PROPERTIES. 
 
 Chemically, ozone is a strong oxidising agent, capable of 
 effecting the oxidation of all the elements with the exception 
 of gold and some of the metals of the platinum group. 
 
 It liberates iodine from potassium iodide and brings about 
 
12 OZONE 
 
 the oxidation of numerous substances such as lead sulphide, 
 manganous salts and ferrocyanides, reactions which form the 
 basis of its qualitative and quantitative detection and esti- 
 mation. 
 
 The general reaction may be expressed by the equation : 
 
 M + 3 = MO + 2 . 
 
 In some cases, however, oxygen is not liberated, but the whole 
 of the ozone reacts and no free oxygen is evolved. Thus 
 sulphur dioxide is oxidised to sulphuric anhydride by ozone 
 according to the reaction : 
 
 3S0 2 + 3 = 3S0 3 
 
 (seeBrodie, "Phil. Mag.," 1894, and Kiesenfeld, "Zeit. Elek- 
 trochem.," 17, 634, 1911). In the combustion of the organic 
 matter in water during the process of sterilisation by ozonised 
 air this reactivity of the ozone molecule as a whole is likewise 
 noted. 
 
 Eiesenfeld (" Zeit. Anorg. Chem.," 85, 217, 1914) observed 
 a similar series of reactions in the action of ozone on sulphur 
 compounds. Three atoms of oxygen in the ozone molecule 
 react with sodium hydrogen sulphite, whilst with neutral 
 sulphites and alkaline thiosulphates only two atoms react, 
 the third being liberated as oxygen gas. 
 
 With certain peroxides, such as hydrogen peroxide, it 
 undergoes decomposition as follows : 
 
 3 + H 2 2 = H 2 + 20 2 , 
 
 Kothmund ("Monatsh.," 38, 295, 1917) showed that the 
 reaction was unimolecular in excess of hydrogen peroxide, 
 but in dilute solutions the ozone underwent catalytic decom- 
 position. 
 
OZONE 13 
 
 It has found many uses industrially as an oxidising agent, 
 which will be detailed in a subsequent section of this volume. 
 Keference, however, may be made to the deodorising of air, 
 the conversion of manganates into permanganates, of chlo- 
 rates into perchlorates, and the " drying " of oils in the pre- 
 paration of linoleum and varnishes. 
 
 At suitable temperatures selective oxidation of undesirable 
 substances which give an objectionable colour or odour to 
 many fats and waxes may be obtained, and such processes of 
 bleaching are receiving extended application. Attempts have 
 also been made to accelerate the ageing of spirits and wine 
 by fractional oxidation with ozone. 
 
 Ozone is a powerful germicide, as was first indicated by 
 Frohlich. Its high germicidal activity is doubtless due to its 
 oxidising power, and as a dual agent of this character it has 
 been fairly extensively employed for the sterilisation of public 
 water supplies, for the treatment of wounds in hospitals, and 
 for various purposes of sterilisation and preservation in in- 
 dustries, such as hide preservation, cold meat storage and 
 the like. Although ozone in high concentrations will effect 
 the sterilisation of air, yet such concentrations as are neces- 
 sary (ca. *05 per cent.) are not capable of respiration without 
 damage to the tissues, consequently its chief function is as a 
 deodoriser and " freshener " for air in confined and crowded 
 spaces. 
 
 In the realm of organic chemistry ozone has received ap- 
 plication in two directions, firstly as an oxidising agent of 
 great strength which introduces no foreign matter, and 
 secondly as a reagent for the ethylene linkage - C = C - . 
 
 As an oxidising agent it is employed for the preparation 
 
14 OZONE 
 
 of vanillin on an extremely large scale. The production of 
 other substances, such as heliotropine, piperonal, and anisalde- 
 hyde, can also be accomplished with its aid (see chap. ix.). 
 Apart from its powerful oxidising properties, ozone will react 
 with certain substances in two definite and characteristic 
 ways to form ozonates and ozonides. 
 
 THE OZONATES. 
 
 Baeyer and Villiger ("Ber.," 35, 3038, 1908) state that 
 strong ozonised air fumes in moist air colours blue litmus 
 red, and causes an increase in the conductivity of distilled 
 water when passed through it. They therefore regarded 
 ozone as the anhydride of an unstable ozonic acid, H 2 4 . 
 According to these authors, if due precautions are taken, 
 highly coloured ozonates may be prepared by the interaction 
 of ozone and moist solid alkali hydroxides or concentrated 
 solutions of the same at low temperatures. 
 
 The ozonates are usually orange or brown. If ozone be 
 passed into a cold ammonia solution, it acquires a dark red 
 colour attributed by these investigators to the formation of 
 ammonium ozonate, NH 4 H0 4 . Lithium ozonate was found 
 to be least, and that salt of caesium most stable. 
 
 A white granular precipitate of calcium peroxide is formed 
 on the passage of ozonised air into cold lime water. 
 
 According to W. Manchot (''Ber./' 41, 47, 1908), Baeyer 
 and Villiger's results are to be attributed to the presence 
 of small quantities of oxides of nitrogen in their ozonised 
 air, since he found that ozone possessed no acidic quali- 
 ties. 
 
OZONE 15 
 
 THE OZONIDES. 
 
 The ozonides are formed by the interaction of ozone with 
 organic compounds containing unsaturated ethylene linkages 
 according to the general equation : 
 
 _c c o x 
 
 !l + 3 -> | J>0 
 
 _c c cr 
 
 Discovered by Harries ("Ann.," 343, 311, 1905; " Ber.," 38, 
 1195, 1905), this reaction was successfully employed by him 
 to elucidate the composition of rubber (see p. 170), and has of 
 recent years been frequently utilised to identify the presence 
 of ethylene linkages. 
 
 Where compounds containing ethylene linkages are 
 treated with strongly ozonised oxygen (ca. 40 per cent. 3 ) 
 the ozonides thus formed on analysis yield more oxygen than 
 is to be expected by the assumption of simple saturation of 
 the ethylene linkage according to the equation : 
 
 II + 3 -> | 
 
 c c cr 
 
 According to Harries, oxozonides are formed by interaction of 
 the organic compound with oxozone present in the gas : 
 
 _C C 
 
 11 + 4 -> | | 
 
 C C 
 
 More recent experiments, however (see p. 184), have failed 
 to establish the existence of oxozone in ozonised air or oxy- 
 gen and consequently some other explanation for the forma- 
 tion of oxozonides must be advanced. 
 
CHAPTEE II. 
 
 THE NATURAL OCCURRENCE OF OZONE. 
 
 THE occurrence of ozone in ordinary atmospheric air has 
 long been a matter of dispute. C. Schonbein (" J. f. Prakt. 
 Chemie," 73, 99, 1858), as early as 1858, showed that starch 
 iodide paper, when exposed to the air, slowly turned blue, 
 thus demonstrating the existence of some oxidising agency. 
 He noted that the rate of liberation of iodine varied from day 
 to day and attributed this to an alteration in the ozone con- 
 tent of the air. Cloez and Bineau pointed out that the 
 liberation of iodine from starch iodide could likewise be 
 caused by the presence of oxides of nitrogen naturally present 
 in atmospheric air. 
 
 Houzeau ("Ann. Chem. Phys." IV, 27, 5, 1872, and 
 " O.K.," 74, 712, 1872), as a result of over 4000 determina- 
 tions with neutral litmus starch iodide paper, came to the 
 conclusion that ozone was present in atmospheric air in ad- 
 dition to the frequent occurrence of oxides of nitrogen. As 
 a maximum ozone content Houzeau recorded 2*8 mgm. per 
 cubic metre. Houzeau's views were supported by Hartley 
 ("Trans. Chem. Soc.," 39, 10, 111, 1881, and "Nature," 39, 
 474, 1889), who noted that many of the dark lines of the solar 
 spectrum were coincident with those that would have been 
 produced on the assumption that atmospheric ozone absorbed 
 
 light of these particular wave lengths emitted from the sun, 
 
 (16) 
 
THE NATUKAL OCCURRENCE OF OZONE 17 
 
 thus exhibiting the Frauenhofer lines ; which conclusions 
 were confirmed by Meyer ("Ann. der Physik," IV, 12, 
 849, 1903). 
 
 The vivid blue colour of ozone was asserted by Hartley 
 to give the characteristic coloration to the summer sky, an 
 alternative theory to the one first propounded by Lord Kay- 
 leigh in 1871 (Hon. S. W. Strutt, " Phil. Mag.," 11, 107, 1871) 
 and extended by Schuster (" Theory of Optics," p. 325) and 
 King ("Trans. Phil. Boy. Soc.," A, 212, 375, 1913). 
 
 Eayleigh showed that the intensity of the light scattered 
 by small particles of dust in the atmosphere would vary in- 
 versely as the fourth power of the wave length, i.e. the light 
 in the ultra-violet and blue end of the spectrum being of the 
 shortest wave length would be most intensely scattered and 
 thus made visible. It may be noted that in Lord Rayleigh's 
 experiments the sky light appeared somewhat bluer than 
 anticipated by this theory, and thus indicated that absorption 
 by ozone may be a contributary cause to the colour of the 
 sky. C. Fabry and H. Buisson (" C.R.," 156, 782, 1913), as 
 a result of a series of experiments on the absorption coefficient 
 ozone for light of the wave lengths X = 255 //.//, to 330 //,/*, 
 showed that a thickness of only 25 yu, of ozone reduces the 
 incident light intensity by over 50 per cent. For a wave 
 length of X = 300 IJL/JL the proportion of transmitted light for 
 a thickness of 5 mm. of ozone was only 1 per cent., approxi- 
 mating to the conditions of the terrestrial atmosphere exposed 
 to solar radiation. If uniformly distributed this would equal 
 0'6 c.c. or 1*4 mgm. per cubic metre ; this concentration is 
 somewhat high for air at low altitudes, hence it may be 
 
 argued that with increasing altitudes the ozone content rises. 
 
 2 
 
18 
 
 OZONE 
 
 E. Kron (" Ann. der Physik," 45, 377, 1914) records X = 
 325 pp as the limit of the effective solar spectrum at sea-level 
 on the clearest days. Fabry and Buisson's results between 
 the wave length X = 200 and X = 350 /i/* are shown in the 
 following graphical form : 
 
 ISO 
 
 200 
 
 250 
 
 300 
 
 350 
 
 FIG. 2. 
 
 More recently Fowler and Strutt ("Proc. Roy. Soc.," 93, 
 577, 1917) showed that the Frauenhofer lines between the 
 wave lengths X = 319*9 pp and X = 333*8 pp, the ultra-violet 
 lines shown by Ladenburg and Lehmann to be present in the 
 ozone absorption spectrum, were present in the greatest in- 
 tensity in the solar spectrum at low altitudes, or when the 
 terrestrial air stratum through which the light had to pass 
 was greatest, thus again supporting Hartley's contention that 
 the atmospheric ozone was an effective agent in fixing the 
 extension of the solar spectrum in the ultra-violet. 
 
 Stratt (" Proc. Eoy. Soc.," 114, 260, 1918) likewise showed 
 that the limitation of the solar spectrum to the lower wave 
 
THE NATURAL OCCUEEENCE OF OZONE 19 
 
 length of X = 294*8 //,//, was due to the absorption by atmos- 
 pheric ozone. By long distance experiments on absorption 
 of the light from a cadmium spark and mercury vapour lamp, 
 the lower air, mass for mass, was found more transparent 
 than the upper air, and that if the absorption was not due to 
 dust, the ozone content would not exceed 0'27 mm. at normal 
 pressure, in four miles of air. 
 
 Engler and Wild ("Ber.," 29, 1940, 1896) likewise con- 
 firmed the presence of atmospheric ozone by the action of 
 air on manganous chloride paper, whilst Schone in 1897 
 (" Brochure," Moscow, 1897) obtained as maxima and minima 
 the following values : 
 
 Maximum, 100 mg. per cubic metre. 
 Minimum, 10 ,, ,, ,, 
 
 In the same year, Thierry ("C.R," 124, 460, 1897) made 
 the important observation from experiments conducted on 
 Mont Blanc, that the ozone content of the atmosphere rose 
 with increasing altitude, thus, at 1000 metres height he ob- 
 tained 0*039 mg., and at 3000 metres, 0'094 mg. of ozone per 
 cubic metre of air. Similar figures were observed by H. de 
 Varigny (Smithsonian College, " Proc.," 39, 27), viz. a maxi- 
 mum and minimum of 0'03 and O'Ol mg. per cubic metre. 
 
 These observations were continued by Hatcher and Arny 
 (" J. Amer. Pharm.," 72, 9, 1900), who determined the amount 
 of ozone in the air by two different methods, viz. the iodide 
 and arsenitic titration processes, as maxima and minima, 
 they observed the following values : 
 
 Method of Minimum. Maximum. 
 Estimation. 
 
 Iodide 158 316 mg. per cubic metre. 
 
 Arsenite 34 80 
 
20 OZONE 
 
 Henriet and Bonyssy ("C.R," 146, 977, 1908) showed that 
 the ozone content of the air at ground level varied approxi- 
 mately inversely with the carbon dioxide concentration. 
 
 Hayhurst and Pring ("J.C.S.," LXII, 868, 1910) drew 
 attention to the wide variation of the results obtained by 
 numerous investigators, and conducted a series of observa- 
 tions on Glossop Moor in Derbyshire. They showed, utilising 
 Houzeau's original method of estimating both iodine and 
 alkali liberated from potassium iodide solutions, a procedure 
 which was found to give extremely accurate results, both for 
 ozone and mixtures of ozone and nitrogen dioxide, that in 
 this neighbourhood at least, oxides of nitrogen were always 
 present in the air up to an altitude of 8000 ft., and that the 
 quantity of ozone present, if any, was too small to be de- 
 tected. With an increase in the altitude small quantities of 
 ozone were obtained up to a height of 10 miles. Concen- 
 trations of the order of 0'12 to 0*4 mg. per cubic metre of 
 ozone, and smaller quantities of oxides of nitrogen were ob- 
 tained. 
 
 H. N. Holmes (" J. Amer. Chem. Soc.," 47, 497, 1909) 
 has shown that the maximum amount of ozone is formed in 
 moving areas of air under a high barometric pressure when 
 the conditions are favourable for bringing air of high altitudes 
 close to the earth's surface. 
 
 It may be concluded that ozone is a normal constituent 
 of pure air, and that the quantity of ozone in the air increases 
 with the altitude. Seasonal variations in the ozone content 
 have likewise been obtained. Thus, Houzeau, as a result of 
 eight years observation with neutral and alkaline iodide test- 
 papers (loc. cit.\ noted that the ozone content of the atmos- 
 
THE NATURAL OCCURRENCE OF OZONE 21 
 
 phere rose in spring and summer, but sank to very small 
 proportions in autumn and winter. Berigny observed a 
 maximum in the month of May, and a minimum in Novem- 
 ber, and gives the following order for decreasing ozone con- 
 tent, May, March, April, June, August, July, September, 
 January, October, February, November. 
 
 Pring ("Proc. Eoy. Soc.," 96, 204, 1914) extended his 
 investigations to the air in the high Alps ; at 2100 metres, he 
 obtained the value 4'7 mg. per cubic metre, and at 3580 
 metres 8'8 mg. per cubic metre. No considerable increase 
 in the ozone content was observed at altitudes up to 20 km. 
 Oxides of nitrogen and hydrogen peroxide were absent in the 
 air at the higher altitudes. 
 
 Usher and Eao (" J.C.S.," in, 779, 1917) conducted a 
 series of estimations in India on the ozone (by manganese 
 dioxide), hydrogen peroxide (by chromic acid), and oxides of 
 nitrogen content of the atmosphere. They could not detect 
 the presence of ozone although oxides of nitrogen in con- 
 centrations of from 1 to 5 parts per million were frequently 
 obtained. 
 
 The ozone concentration in the lower air strata likewise 
 increases during periods of storm or after heavy rain storms, 
 and the south and south-west winds are said to be richer in 
 ozone than the northern ones. 
 
 K. Nasini ("Atti. d. K. Accad. Lincei," 21, 740, 1912) 
 records interesting cases of naturally occurring ozonised 
 water ; he states that the acid waters of Bagnone, Monte 
 Annata are highly ionised, and contain 1*25 of oxygen, and 
 0*135 c.c. of ozone per litre. 
 
22 OZONE 
 
 SOURCES OF NATURAL OZONE. 
 
 Various alternative hypotheses have been advanced to 
 explain the mode of formation of this small ozone concentra- 
 tion in atmospheric air. It is at once evident that even this 
 minute quantity exceeds the normal thermal equilibrium 
 amount, and consequently there must be a continuous source 
 of ozone. We may classify the various hypotheses as to this 
 source under three groups : 
 
 (a) chemical ; (6) photo-chemical ; (c) electrical. 
 
 NATURAL CHEMICAL PROCESSES. 
 
 The earlier investigators such as Schonbein, Houzeau, 
 Berigny, Peyrou, and Marie Davy, were of the opinion that 
 the green vegetation of plant life was responsible for the 
 production of ozone, thus accounting for the observed maxi- 
 mum and minimum ozone content in the months of May and 
 November respectively. It was shown, however, that coloured 
 plants yielded no volatile oxidising substances whatever, and 
 more recent experiments have shown that the oxidising agent, 
 which can always be detected in green plant growth, is hy- 
 drogen peroxide. According to Priestly and Usher ("Proc. 
 Phys. Soc.," 78, 3, 38, 1906), the plant chlorophyll serves 
 merely as a light sensitiser to bring about the reaction 
 3H 2 O + C0 2 + light energy = HCHO + 2H 2 2 , 
 
 the formaldehyde thus formed is subsequently polymerised 
 to formose (d.l. glucose) by the protoplasm of the cell chloro- 
 plast. 
 
 The hydrogen peroxide is usually destroyed by one of the 
 numerous enzymes, termed oxidases, present liberating mo- 
 lecular oxygen 
 
THE NATUEAL OCCUEEENCE OF OZONE 23 
 
 2H 2 2 -> 2H 2 + 2 
 
 (see Bach and Choat, "Arch. d. Sci. Phys. et Nat. Geneva," 
 17, 4771, 1909), but many investigators suspect that during 
 the decomposition of the hydrogen peroxide small quantities 
 of ozone may be produced. 
 
 The production of ozone by the atmospheric oxidation of 
 various gums and essential oils exuded by trees and plants, 
 such as turpentine, sandal-wood oil, or oil of lavender, has 
 long been suspected, and undoubtedly the rapidity with which 
 starch iodide slips are turned blue in a pine forest is, in some 
 measure, due to the ozone present in the surrounding air, 
 although the formation of hydrogen peroxide under these 
 conditions is without doubt the more important natural pro- 
 cess contributing to the freshness of the air. 
 
 PHOTO-CHEMICAL PEOCESSES. 
 
 We shall have occasion to refer to the interesting fact that 
 oxygen is ozonised by exposure to ultra-violet irradiation of 
 wave length X = 120 - 180 /Z/A whilst ozonised oxygen is re- 
 solved into its original form by light of somewhat longer wave 
 length, viz. \ = 330 P/JL. 
 
 Hartley ("Trans. Chem. Soc.," 39, 10, 111, 1881) noted 
 the presence of Frauenhofer lines in the visible solar spectrum 
 corresponding to those which would be absorbed by ozone. 
 These conclusions in the visible part of the spectrum were 
 confirmed by Meyer ("Ann. der Physik," IV, 12, 849, 1903) 
 and extended by C. Fabry and H. Buisson (" O.K.," 156, 
 782, 1913) and Fowler and Strutt ("Proc. Eoy. Soc.," 
 93, 77, 1917) to the ultra-violet portion of the spectrum. 
 Furthermore, all experimental evidence indicates that the 
 
24 OZONE 
 
 ozone concentration is greatest in the upper portion of the 
 atmosphere, where the intensity of the ultra-violet radiation 
 would naturally be greatest (see also K. Birkeland, " Cairo 
 Soc.," 8, 287, 1916). It would appear that the presence of 
 ozone in atmospheric oxygen is largely due to the synthetic 
 operation of solar radiant energy of short wave length 
 (\ = 120 - 180 pp), whilst the limitations in the amount in 
 the upper parts of the atmosphere is caused by the destructive 
 action of light of longer wave length (X = 300 /x/z), a dynamic 
 equilibrium being finally established between the rate of 
 formation and the rate of decay. Near the earth's surface, 
 as we have seen, smaller ozone concentrations are obtained, 
 partly owing to the fact that the light of longer wave length 
 penetrates somewhat further into a dusty atmosphere than 
 that of short wave length, but more especially to the reducing 
 action of easily oxidisable substances both on the earth's 
 surface, and carried to low altitudes by the wind. Country 
 air, according to Houzeau, contains more ozone than that 
 around villages, whilst its presence can rarely be detected in 
 towns. This observer, in fact, records the disappearance of 
 ozone in the air after the passage of a crowd on a public fete 
 day, and its gradual reappearance when the normal conditions 
 had been re-established. 
 
 ELECTEICAL PEOCESSES. 
 
 The increasing attention which during the last few years 
 has been paid to a study of atmospheric ionisation and electri- 
 fication has not only clearly demonstrated that the potential 
 difference between different parts of the atmosphere and be- 
 tween earth or water and the air may reach extremely high 
 
THE NATUEAL OCCUBBENCE OF OZONE 25 
 
 values during periods of atmospheric disturbances such as 
 electrical storms, but even during periods of fair weather, 
 local potential differences of high magnitude may result. 
 
 Evidence for the ozonisation of oxygen during periods of 
 intense electrical discharge, either silent as in the aurora, 
 natural corona, and the remarkable Andes glow occasionally 
 observed in S. America (see " Knoche Meteor. Zeit.," 29, 329, 
 1912), or violent as in lightning and the so-called thunderbolt 
 or globular discharge, is somewhat conflicting. Undoubtedly 
 oxides of nitrogen are present, since these can always be 
 detected during periods of heavy discharge, and in many cases 
 it appears probable that ozone is formed either without or 
 more probably in conjunction with the oxides of nitrogen. 
 
 Thornton (" Phil. Mag.," 21, 630, 1911) has advanced the 
 view that the globular discharges themselves are purely 
 gaseous bodies and consist of ozone in active combination. 
 
 In a subsequent chapter we shall observe that the condi- 
 tions for the possible ozonisation of oxygen by means of 
 ionisation are established when a discharged electron or a gas 
 ion strikes an oxygen molecule with sufficient violence so as 
 to permit the temporary distortion of, or the actual removal 
 of one of the valency electrons circulating round the oxygen 
 molecule from its orbit, and a rough computation of the volt- 
 age of discharge which is necessary to give the emitted 
 electron this requisite energy is but nine volts, a relatively 
 low figure. 
 
 The potential difference between strata of air during 
 periods of fine weather is frequently extremely great and quite 
 sufficient to produce atmospheric ionisation, with the conse- 
 quent possible production of ozone, thus P, Mercanton 
 
26 OZONE 
 
 (" Terrest. Magn.," 22, 35, 1917) obtained a P.D. of 1200 volts 
 per metre on the top of a tower at Lausanne, 930 metres 
 above sea-level. 
 
 C. Chree (" Phil. Trans.," 215, 133, 1915) gives 304 volts 
 per metre as the average potential gradient in the atmosphere 
 at Kew for the last fifteen years. 
 
 McLennan (" Nature," 92, 424, 1913) obtained the follow- 
 ing values for the number of gas ions formed per second per 
 cubic centimetre of air, nine ions in the air over the land and 
 four over sea water. He ascribed ionisation due to the influ- 
 ence of the ultra-violet light itself; a view supported by the 
 experiments of G. Simpson (" Monthly Weather Eeview," 
 44, 115, 1916), who found that at a height of 6000 metres, 
 over thirty times as many ions were formed per second as at 
 sea-level. W. Swann ("Terrest. Magn.," 21, 1, 1916) like- 
 wise showed that the upper air was a region of high electrical 
 conductivity, the source being the ultra-violet light of ampli- 
 tude X<135 pp, a fraction only T61 x 10 " 5 of the total 
 radiant energy derived from the sun. Production of ozone 
 by natural ionisation is thus chiefly a secondary effect of ultra- 
 violet irradiation, which, as we have already noted, is one of 
 the chief ozonising agencies in the atmosphere. Natural ion- 
 isation and consequent ozonisation is, however, not entirely 
 derived from ultra-violet radiation, since ionisation and small 
 quantities of both ozone and hydrogen peroxide are formed 
 by the evaporation of water in air, especially in the neighbour- 
 hood of fountains and waterfalls, where conditions of spray 
 formation obtain. 
 
 A. Besson (" C.E.," 153, 877, 1911) noted that the maxi- 
 mum concentrations of ozone and hydrogen peroxide were 
 
THE NATUEAL OCCUEEENCE OF OZONE 27 
 
 formed in air during the fall of heavy drops of rain from a 
 previously clear sky in hot summer weather. 
 
 C. Oddo (" Gaz. Soc. Chim. Ital.," 45, 395, 1915) ascribes 
 the formation of gas ions under these conditions to the 
 spontaneous ionisation of water vapour when rarified. One 
 kilogram of moist air (773*4 litres when dry at N.T.P.) 
 contains 89 x 10 ~ 20 hydrogen and hydroxyl ions at 15 C. 
 and 760 mm. pressure ; he shows that a fall in temperature 
 naturally diminishes the content of water vapour in the air 
 but also increases the degree of ionisation below 32 C., where 
 it is practically zero. The optimum temperature range for 
 maximum ionisation was found to be 5 to 20 C., which, it 
 may be noted, is the optimum for animal and vegetable life 
 in the temperate zones. 
 
 Lenard, in a series of researches on the Electricity of 
 Waterfalls (" Ann. der Physik," 45, 7, 100, 1914), showed that 
 ionisation was effected not only by the impact of suspended 
 drops upon obstacles such as rocks or stones, but by impact 
 of drops against each other resulting in the production of 
 secondary drops. 
 
CHAPTEK III. 
 
 CHEMICAL PRODUCTION. 
 
 AN indication of the production of ozone can be observed in 
 a great variety of chemical reactions such as the decomposi- 
 tion of certain peroxides, in processes of autoxidation, and in 
 many cases of combustion of gaseous fuels. In the latter 
 case the ozone is doubtless of a purely thermal origin and a 
 consideration of the mechanism of production by this means 
 will be deferred to a subsequent section. 
 
 Ozone can nearly always be detected in oxygen resulting 
 from chemical decomposition. The temporary existence of 
 atomic oxygen liberated according to the equation : 
 
 M"0 2 + H 2 S0 4 - MS0 4 + H 2 + O 
 
 has not yet been definitely proved, although the evidence for 
 the formation of atomic hydrogen by similar processes is now 
 extremely strong. In any case during the decomposition of 
 the peroxides, the atomic oxygen polymerises with great 
 rapidity to the molecular form : 
 
 + O -> 2 . 
 
 C. Brodie ("Phil. Trans.," 141, 759, 1850) first advanced the 
 view that ordinary oxygen during processes of chemical 
 action was split up into two parts termed ozone and antozone. 
 
 + 
 O 2 - 0' (ozone) + O (antozone). 
 
 As we shall have occasion to note in discussing processes of 
 
 (28) 
 
CHEMICAL PEODUCTION 29 
 
 autoxidation, Brodie's hypothesis was strongly supported by 
 the experimental work of Schonbein. 
 
 E. Clausius ("Zeit. Phys. Chem.," 103, 644, 1858) sug- 
 gested that Brodie's so-called " ozone " and " antozone " were 
 identical with atomic oxygen, possessing opposite electric 
 charges 
 
 2 -> 6 + 0'. 
 
 This view was further enlarged upon by van't Hoff, 
 (" Zeit. Phys. Chem.," 16, 411, 1895), who, as a result of his 
 studies on the autoxidation of phosphorus, came to the con- 
 clusion that there exists a definite equilibrium in normal 
 gaseous oxygen between the molecular and atomic form, the 
 atomic being charged : 
 
 2 ^:=t 6 + 0'. 
 
 Nernst ("Zeit. f. Elektrochem.," 9, 891, 1903) showed, from 
 a series of observations on the electromotive force of ozone- 
 oxygen cells, that if the three allotropes of oxygen were as- 
 sumed to exit in equilibrium with each other under normal 
 conditions, according to the reversible equations : 
 
 3 2 2 + 
 2 ^ + 0, 
 
 none of the allotropes possessing an electrical charge, then 
 the equilibrium concentration of the atomic oxygen would be 
 only 1/10 23 of the normal ozone concentration, which we shall 
 see is of the order 10~ 5 per cent. The normal concentration 
 of atomic oxygen is, therefore, so small as to render its exist- 
 ence as a chemical substance, to which the ordinary methods 
 of statistical calculation of its concentration and properties in 
 bulk can be applied, extremely doubtful. 
 
30 OZONE 
 
 Nevertheless, the formation of ozone in small quantities 
 may be expected to occur in the decomposition of the per- 
 oxides, since, on the above assumption, the following sequence 
 of chemical reactions may be assumed to occur : 
 (i) Ba0 2 -> BaO + O. 
 (ii) + O -> 2 . 
 (iii) 2 + 0-0 3 . 
 
 We have already noted that reaction (ii) proceeds with great 
 rapidity, and that reaction (iii) is merely a side reaction, 
 which will only proceed during the evolution of oxygen. 
 Any ozone formed may, of course, be subsequently decom- 
 posed by catalysis at the surface of the decomposed peroxide, 
 or by the somewhat elevated temperature necessary to cause 
 decomposition of the peroxide. 
 
 Houzeau, in fact, was able to obtain concentrations as 
 high as 28 gms. of ozone per cubic metre of oxygen evolved, 
 by gently heating small quantities of powdered barium per- 
 oxide in eight times its volume of concentrated sulphuric 
 acid. Hydrogen peroxide is likewise formed in small quan- 
 tities under these conditions : 
 
 4H 2 S0 4 + 4Ba0 2 -> (i) 4BaS0 4 + 4H 2 + 20 2 
 
 (ii) 4BaS0 4 + 4H 2 + 3 + 
 (iii) 4BaS0 4 + 4H 2 2 . 
 
 The same investigator showed that similar results were 
 obtained with other peroxides, notably those of magnesium, 
 zinc, sodium, and potassium. 
 
 Even better results can be obtained by the decomposition 
 and gentle dehydration of permanganic acid or potassium 
 dichromate, 
 
 Mn 2 7 -> 2Mn0 2 + 3 . 
 
CHEMICAL PRODUCTION 31 
 
 As dehydrating agent, sulphuric acid is most conveniently 
 employed in the proportions of one of potassium perman- 
 ganate to two of sulphuric acid. De la Coux (" L'Ozone," p. 
 67) states that oxalic acid can be likewise employed in the 
 proportion of 10 gms. of permanganate to 15 gins, of oxalic 
 acid, and that 90 c.c. of oxygen containing 3 mgm. of ozone 
 can be obtained from this mixture. 
 
 Satisfactory yields of ozone may also be obtained by the 
 cautious addition of barium peroxide to a solution of potas- 
 sium permanganate in sulphuric acid, of density 1*85. 
 
 By the thermal decomposition of the persulphates, small 
 quantities of ozone are likewise disengaged, Malaquin (" J. 
 Pharm. Chem.," VII, 3, 329, 1911) gives the following details 
 for the preparation of ozonised oxygen by this means. 
 Twenty gms. of dry and freshly prepared ammonium persul- 
 phate are mixed with 15 gms. of nitric acid in a small flask ; 
 the air is subsequently displaced by carbon dioxide, and the 
 mixture cautiously raised to 65 to 70 C. The reaction, 
 which is strongly exothermic, proceeds somewhat vigorously 
 when once started, and the resulting oxygen, after removal 
 of the carbon dioxide, contains 3 to 5 per cent, of ozone and 
 small quantities of nitrogen. 
 
 Moissan, in his researches on the properties of fluorine, 
 which he isolated by the electrolysis of fused potassium 
 hydrogen fluoride, noted that appreciable quantities of ozone 
 were produced when a few drops of water were introduced 
 into an atmosphere of fluorine. 
 
 The formation of ozone proceeding according to the 
 equation : 
 
 3F 2 + 3H 2 = 6HF + 3 , 
 
32 OZONE 
 
 is especially marked at low temperatures, when the rate of 
 thermal decomposition of any ozone formed is considerably 
 reduced. 
 
 An ozone content of upwards of 14 per cent, in the oxy- 
 gen disengaged by means of this reaction may be obtained, 
 if the temperature be maintained at C. De la Coux 
 (" L'Ozone," p. 70) suggests that the preparation of strongly 
 ozonised oxygen, by this method, offers some hope of techni- 
 cal application. 
 
 Small quantities of ozone may likewise be obtained by 
 the thermal decomposition of other oxygen-containing salts, 
 but the quantity of ozone in the liberated oxygen is far 
 smaller than in the cases alluded to above. Thus Eammels- 
 berg noted that ozone may be detected in the oxygen evolved, 
 on heating crystallised periodic acid up to 135 C. 
 
 Periodic acid is formed by the action of iodine on an 
 aqueous solution of perchloric acid, and can be obtained as 
 crystals containing two molecules of water. When heated 
 carefully, periodic anhydride is formed. 
 
 2(HI0 4 . 2H 2 0) - I 2 7 + 5H 2 0, 
 
 which on continued heating, loses oxygen to form iodic an- 
 hydride : 
 
 I 2 7 ->I 2 5 + 2 . 
 
 The iodic anhydride itself suffers decomposition into its ele- 
 ments at 300 C., consequently the liberation of ozonised 
 oxygen by decomposition of the crystallised periodic acid 
 only takes place within a somewhat narrow temperature 
 range. Aqueous solutions of the acid and its sodium salt 
 likewise gradually acquire the smell of ozone. 
 
CHEMICAL PRODUCTION 33 
 
 O. Brunck has shown that commercial samples of potas- 
 sium chlorate liberate ozonised oxygen during thermal de- 
 composition, although purified samples fail to yield any ozone. 
 The yield of ozone is sensibly increased by the addition of 
 manganese dioxide, thus equal weights of manganese dioxide 
 and potassium chlorate liberate 0'3 per cent, of the weight 
 of chlorate employed in the form of ozone. With twenty- 
 five times as much manganese dioxide, over 1*5 per cent, of 
 the weight of chlorate can be recovered in this form. Other 
 oxides, such as those of copper, iron and zinc do not exhibit 
 this behaviour, which appears to be characteristic of man- 
 ganese dioxide, although slight activity is noted in the cases 
 of the oxides of nickel and cobalt. This is doubtless associ- 
 ated with the property of forming unstable peroxides, which 
 undergo secondary decomposition, liberating atomic oxygen, 
 which can secondarily react with the molecular form to pro- 
 duce ozone. 
 
 In the thermal decomposition of many metallic peroxides 
 the presence of ozone may be detected in the oxygen evolved, 
 the yield of ozone being naturally greater in the case of those 
 peroxides which undergo thermal decomposition at relatively 
 low temperatures, such as silver oxide, yielding oxygen 
 containing 4 to 5 per cent, of ozone. Lead peroxide and 
 mercuric oxide are likewise capable of yielding small quantities 
 of ozone. 
 
 If the peroxide of manganese, or cobalt, or nickelic oxide 
 be subjected to thermal decomposition in a current of oxygen, 
 appreciable quantities of ozone are stated to be formed. 
 
 All these oxide decompositions, resulting in the formation 
 
 of small quantities of ozone, may be referred to chemical 
 
 3 
 
34 OZONE 
 
 processes of activating atmospheric oxygen, whilst in the 
 case of the decomposition of chlorates and iodic anhydride 
 these salts may be regarded as convenient sources of oxygen. 
 
 In the case of the elements of the first group of the peri- 
 odic table, namely, copper, silver and gold, the sub and 
 normal oxides of copper, Cu 4 0, Cu 2 0, and CuO, are somewhat 
 too stable, cupric oxide possessing only a small dissociation 
 pressure at very high temperatures. The oxides of both silver 
 and gold, on the other hand, dissociate much more readily, 
 silver oxide possessing a dissociation pressure equal to that 
 of atmospheric oxygen at 250 C. Silver peroxide, Ag 2 2 , 
 readily liberates hydrogen peroxide and oxygen containing 
 ozone when dissolved in acids. Mercuric oxide closely re- 
 sembles silver oxide in its chemical properties. 
 
 The general reactions involved may be expressed by the 
 following sequence of reactions : 
 
 (i) 2M + 2 = 2MO ; 
 (ii) 2MO = 2M + 20 ; 
 (iii) 20 ^0 2 ; 
 (iv) + 2 ->0 3 ; 
 
 in which by the operation of the first two reactions the oxygen 
 molecule is temporarily split up into its atoms, the necessary 
 energy to perform this operation being supplied by heating or 
 cooling the metal to form or decompose the oxide. The 
 atomic oxygen so formed may then instantaneously recom- 
 bine to form molecular oxygen or combine with molecular 
 oxygen to form ozone. 
 
 OZONE PRODUCTION BY AUTOXIDATION. 
 It had long been known that many substances when ex- 
 posed to the air undergo a process of slow oxidation. 
 
CHEMICAL PRODUCTION 35 
 
 Exemplifications are found amongst the most diverse 
 types of substances such as the corrosion or rusting of metals, 
 e.g. zinc, lead and iron, of certain non-metallic elements such 
 as sulphur and more especially phosphorus, and in many 
 organic substances, such as benzaldehyde, turpentine, linseed 
 oil and various essential oils, such as oil of cinnamon, lavender 
 or citronella. 
 
 It was formerly thought that these reactions were com- 
 parable to the ordinary processes of oxidation or combustion 
 except in so far as the reaction velocity was exceedingly low. 
 In 1858, however, C. F. Schonbein (" J. f. Prakt. Chemie," 
 73, 99, 1858, et seq., to 1868) opened a new and interesting 
 chapter in the theory of oxidation by showing that in these 
 cases of slow oxidation, for every molecule of oxygen consumed 
 by the substance undergoing oxidation a molecule of oxygen 
 was simultaneously transformed to a more active state. This 
 activated oxygen would then secondarily react to form a 
 fresh series of new substances. 
 
 Thus in the presence of oxygen, ozone could be produced ; 
 in the presence of water as in the wet oxidation of the metals, 
 an amount of hydrogen peroxide was produced equivalent to 
 the quantity of metal oxidised. In the presence of other 
 oxidisable substances the active oxygen can oxidise them, 
 frequently bringing about oxidations which cannot be accom- 
 plished by ordinary atmospheric oxygen ; thus indigo is 
 converted into isatin during the autoxidation of palladium 
 hydride or benzaldehyde. 
 
 The quantitative relationship between the production of 
 active oxygen and the quantity of substance undergoing the 
 process of slow oxidation was shown by Schonbein to hold in 
 
36 OZONE 
 
 the case of the wet oxidation of the metals by an estimation 
 of the quantity of hydrogen peroxide simultaneously produced. 
 An interesting confirmation of Schonbein's views was 
 afforded by A. Genthe's investigations on the drying of linseed 
 oils ("Zeit. Angew. Chem.," 19, 207, 1906). It had been 
 previously shown by Lippert (" Zeit. Angew. Chem.," n, 
 412, 1898) and Wegen ("Chem. Eev. f. fett. u. Harz.," 
 4, 345, 1899) that the drying of linseed oil was virtually 
 
 Time in Hours 
 Fio. 3. 
 
 a process of atmospheric oxidation. Genthe examined the 
 reaction velocity of this process of oxidation and found that 
 the time-increase of weight curves for the drying of a thin 
 film of linseed oil exhibited the sinuous character of an auto- 
 catalytic reaction. 
 
 It will be observed that the initial rate of dryings increases 
 somewhat slowly with the time ; as, however, the quantity of 
 autocatalyst increases simultaneously the reaction proceeds 
 at an ever-increasing velocity and only begins to sink when 
 
CHEMICAL PBODUCTION 37 
 
 the quantity of oil remaining to be oxidised diminishes in 
 amount. 
 
 It had therefore to be assumed that in the process of dry- 
 ing, a catalyst was simultaneously formed, thus if a and b be 
 the initial concentrations of the linseed oil and catalyst, then 
 the rate of oxidation of the oil after a time t will be given by 
 the equation : 
 
 f t = K(a - x)(b + x). 
 
 Genthe, in fact, showed by his experiments on reaction velocity 
 that there was a quantitative relationship between the quantity 
 of linseed oil oxidised and the quantity of autocatalyst simul- 
 taneously produced. 
 
 Houzeau (1860), Genthe (loc. cit.\ Hazura(" Zeit. Angew. 
 Chem.," i, 312, 1888), Kissling ("Zeit Angew. Chem.," 4, 
 395, 1891) and Friend (" Proc. Paint and Varnish Soc.," 1914) 
 all showed that the autocatalyst was an unstable peroxide, 
 since it liberated iodine from potassium iodide and showed 
 the other reactions of a peroxide and a similar catalytic 
 acceleration could be produced by the addition of ozone, 
 benzoyl peroxide, oxidised turpentine or ether, to the linseed 
 oil. It is still a matter of uncertainty as to the nature of 
 this catalytic peroxide. Houzeau was of the opinion that it 
 was dissolved ozone, whilst other investigators support the 
 theory of an unstable peroxide of linoleic acid, similar in 
 character to benzoyl peroxide. It appears probable that small 
 quantities of ozone can be isolated from turpentine, oil of 
 cinnamon and other essential oils, undergoing atmospheric 
 oxidation, but that most of the activated oxygen is absorbed 
 or combines with part of the substance to form a peroxide. 
 
38 OZONE 
 
 Jorrisen and Eeicher ("Ber.," 30, 1451, 1897; " Zeit. 
 Angew. Chem.," 22, 6829, and "Chem. Zeit.," 26, 99, 1902) 
 showed that ozone could be formed during the reduction of 
 certain oxidising acids, such as chromic acid, attributed to 
 the intermediary formation of an unstable peroxide with its 
 subsequent decomposition : 
 
 0. C 
 
 Cr0 3 + (COOH) 2 -> H 2 + >Cr< | 
 
 (T X O C 
 
 Relatively large quantities of ozone, however, are produced 
 in the autoxidation of phosphorus, and in view of the con- 
 veniences of this method of preparation the following details 
 may be given : A rapid current of air is passed through a 
 bottle containing sticks of yellow phosphorus, moistened with 
 a dilute sulphuric acid acidified solution of potassium per- 
 manganate or bichromate. The reaction proceeds but slowly 
 at 6 C., whilst the optimum temperature is stated to be 24 C. 
 Under reduced pressure the reaction still proceeds at C. 
 As is well known, pure oxygen reacts but slowly with phos- 
 phorus except under reduced pressure. A 20 per cent, mix- 
 ture of oxygen in hydrogen is particularly efficacious for the 
 production of ozone, but the phosphorus is liable to become 
 extremely hot, with the attendant risk of explosion. Small 
 quantities of hydrogen peroxide are simultaneously produced. 
 From time to time the stale phosphorus should be re-fused 
 in order to remove the superficial layer of phosphoric acid 
 which causes a diminution in its activity. 
 
 We have already noted that the theory of Brodie, developed 
 by Clausius and van't Hoff, postulating the existence of two 
 forms of oxygen : 
 
CHEMICAL PRODUCTION 39 
 
 fO (ozone), 
 
 o 2 + 
 
 \0 (antozone), 
 
 was supported by Schonbein as a result of these researches. 
 According to this hypothesis all processes of autoxidation are 
 dual in character, since two substances must simultaneously 
 undergo oxidation. Engler (" Kritische Stiidien liber die 
 Autoxydationsvorgange, Braunschweig," 1903) has attempted 
 to distinguish between these by terming the substance under- 
 going oxidation the autoxidiser, and the substance simultan- 
 eously oxidised the acceptor. Clearly, either the ozonic or 
 antozonic form of active oxygen may react with the autoxi- 
 diser or the acceptor to produce " ozonides " or " antozon- 
 ides " ; thus ozone is an " ozonide," and phosphoric acid the 
 " antozonide " produced in the autoxidation of phosphorus. 
 Van't Hoff (loc. cit.) noted that the presence of excess of 
 "ozonide" prevented the formation of the antozonide, and 
 
 thus it necessarily followed that the primary reaction, 
 
 + 
 
 2 ^ O + was reversible in character, Since the antozon- 
 ide, viz. phosphoric acid, is not volatile the escaping ozonic 
 form of active oxygen or ozone should be electrically charged. 
 A search for this electrically charged form of oxygen in air 
 which has been passed over phosphorus has yielded conflicting 
 results. Elster and Geitel ("Phys. Zeit.," 16, 321, 1890; 
 "Wied. Ann.," 39, 457, 1903) noted that air thus treated 
 was electrically conducting (see also Matteuci, " Enc. Brit.," 
 VIII, 622, 1855; Naccari, "Atti. della Scienze de Torino," 
 XXV, p. 252 ; J. Joubert, " These sur la Phosphorescence 
 du Phosphore," 1874; T. Evan, "Phil. Mag.," 5, 38, 512, 
 1897; J. Chappuis, "Bull. Soc. Chem.," 2, 35, 419, 1881). 
 
40 OZONE 
 
 However, Goekel ("Phys. Zeit.," IV, 1903) showed that this 
 conductivity was not due to the presence of ozone which 
 could be absorbed without destroying the conductivity. 
 
 Barus (" Washington," 1901), Harms (" Phys. Zeit.," 
 IV, in, 1902), and Bloch ("Ann. de Chemie et de Phys.," 
 u, 25, 1905) likewise showed that the conductivity was not 
 due to the presence of ozonic oxygen or charged ionic oxygen, 
 but to oxides of phosphorus collected round charged nuclei, 
 forming aggregates of fairly large dimensions (r = 10~ 6 cm.), 
 while the actual number of charged gas ions observed fell far 
 short of the stoichiometric ratio, oxygen absorbed oxygen 
 activated : 1 : 1, as postulated by the hypothesis. A. Blanc 
 (" O.K.," 95, 2, 1170, 1911) showed the existence of both 
 positive and negative ions, the production of which was ac- 
 companied by the formation of white fumes. The production 
 of these gas ions was accelerated by allowing the process of 
 oxidation to take place in an electric field. 
 
 K. Przibram (" Akad. Wiss. Wien., Ber.," 126, 247, 1912) 
 showed that the charge on each gas ion was approximately 
 6 x 10- 10 E.S. units, and that 1*43 x 10~ 6 gms. of phosphorus 
 in the form of phosphoric acid was associated with each E.S. 
 unit, and 1-3 x 10~ 16 gms. of phosphorus in each particle. A. 
 Blanc ("C.K.," 158, 1492, 1911) claims to have discovered 
 the existence of a radiation emitted during the autoxidation 
 of phosphorus like 7 rays, extremely soft and not corpuscular 
 in character. They are easily absorbed by air. E. Hoppe 
 Segler ("Zeit. Physiol. Chem.," 2, 23, 1878), and Baumann, 
 adopted the same hypothesis as Schonbein, but substituted 
 the somewhat less confusing term of "nascent" oxygen for 
 Schonbein's "ozone" and " antozone ". It is, however, 
 
CHEMICAL PKODUCTION 41 
 
 evident that the case for the existence of charged ions of 
 atomic oxygen of opposite electric sign is not strongly sup- 
 ported by the investigators cited above, although, as we have 
 observed, the existence of uncharged atomic oxygen is a 
 plausible hypothesis. 
 
 M. Traube (" Ber.," 15, 663, 1882, and 1471, 1843 ; " Ges- 
 ammelte Abhandlungen," Berlin, 1899), A. Bach (" C.K.," 
 126, 2957, 1897), and C. Engler and V. Wild ("Ber.," 30, 
 1667, 1897), and others, on the other hand, developed the 
 theory of an intermediate compound. 
 
 Thus, according to Traube, the presence of water is 
 necessary for all these processes of slow combustion, a point 
 of view strongly supported by the researches of Mrs. Fulhame 
 ("An Essay on Combustion," London, 1794), B. Baker, 
 H. B. Dixon ("Phil. Trans.," 175, 630, 315, 4795, 1896), and 
 H. E. Armstrong (B.A. Eeports, " Proc. Eoy. Soc.," 40, 287, 
 1886) ; the primary reaction taking place is the formation of 
 an oxide and hydrogen peroxide according to the equation : 
 
 M + 2 + H 2 = MO + H 2 2 . 
 
 The formation of Schonbein's ozonides must thus be con- 
 sidered as due to secondary reactions between the hydrogen 
 peroxide and the acceptor, in some cases exceedingly improb- 
 able reactions. Thus, it is difficult to imagine the formation 
 of ozone by the action of oxygen in a dilute solution of hy- 
 drogen peroxide according to the following equations : 
 
 P 2 + 2 + H 2 = P 2 + H 2 2 
 H 2 2 + 2 = H 2 O + O 3 , 
 
 although it is stated that by the distillation of strong solutions 
 of hydrogen peroxide in vacuo ozone can be obtained. 
 
42 OZONE 
 
 Bach's modification of the hypothesis embodied the con- 
 ception of the formation of an unstable intermediary peroxide 
 prior to decomposition into an oxide with simultaneous oxi- 
 dation of the acceptor thus : 
 
 / 
 
 M + 2 -> M< | 
 
 X 
 
 O 
 
 M/ \ + A - MO + AO 
 \0 
 
 Engler and Wild ("Ber.," 30, 1669, 1897), and Ostwald 
 ("Zeit. Phys. Chem.," 30, 250, 1900) applied Bach's concep- 
 tion of the mechanism of processes of autoxidation to the 
 case under consideration, i.e. the formation of ozone by the 
 autoxidation of phosphorus. 
 
 Engler and Wild suggested the following sequence of 
 reactions : 
 
 2P + 2 = P 
 
 X 
 
 2 \) 
 
 whilst Ostwald suggested that a still higher oxidation form 
 of phosphorus was produced as an unstable intermediate 
 product : 
 
 ,0 
 
 2P + 20 
 
 3 , 
 
 N 
 
 thus giving the stoichiometric ratio, P : 3 : : 2 : 1 which was 
 actually obtained by van't Hoff. 
 
CHEMICAL PRODUCTION 43 
 
 The fundamental difficulty inherent in the peroxide theory 
 was raised many years ago in a remarkable essay by G. Live- 
 ing (" Chemical Equilibrium, the Kesult of the Dissipation 
 of Energy," Cambridge, 1885). It is evident that the per- 
 oxide formed must be endowed with available energy 
 greater than that possessed by atmospheric oxygen, and it is 
 thus difficult to explain its formation as the result of an exo- 
 thermic reaction from phosphorus and air. It is usually 
 assumed that the chemical energy of one system is not avail- 
 able for another totally different system, i.e. that the energy 
 liberated during the oxidation of phosphorus is dissipated 
 through the system in the form of heat. Liveing introduced 
 the interesting hypothesis, that in certain cases, the liberated 
 energy was not dissipated in this form, but stored up in one, 
 or at least a very few, neighbouring molecules, which would 
 thus be endowed with a great deal of energy. Thus we can 
 imagine a simple transfer of energy from one set of reacting 
 molecules to another set, molecule to molecule, and thus ex- 
 plain the simultaneous equivalent formation of an endo- 
 thermic compound, during a strongly exothermic reaction. 
 
CHAPTER IV. 
 
 THERMAL PRODUCTION. 
 
 SINCE the formation of ozone is a strongly endothermic re- 
 action, we would expect, as pointed out by Nernst (" Zeit. 
 Elektrochem.," 9, 891, 1903), that the equilibrium: 
 
 30 2 ^ 20 3 
 
 would shift over from left to right with elevation of the tem- 
 perature. An approximate idea of the ozone concentration 
 in equilibrium, with oxygen at various temperatures, can be 
 obtained by two independent methods ; from a calculation of 
 the value of K, the equilibrium constant by means of the 
 Nernst heat theorem, as well as from the observed measure- 
 ments of the electromotive force of the ozone/oxygen cell. 
 
 According to the Nernst heat theorem (" Applications of 
 Thermodynamics to Chemistry," Sillman Lectures, 1906), 
 Gruneisen (" Ann. Phys.," 26, 401,1912), Pollitzer ("Berech- 
 nung Chemischer Amnitaten nach dem Nernstchen Warme- 
 theorem. Ahrens Sammlung. En eke," 1912), a simple ex- 
 pression for the equilibrium constant K, in homogeneous gas, 
 reactions can be obtained in terms of known quantities, 
 provided two basic assumptions are made, firstly, that the 
 entropy of a condensed chemically homogeneous system 
 vanishes at the absolute zero, and secondly, that the specific 
 molecular heat of a gas can be approximately evaluated 
 
 from a simple expression : 
 
 (44) 
 
THEEMAL PEODUCTION 45 
 
 C p = 3-5 + 2/3T, 
 
 where G p is the molecular specific heat and /3 a constant. 
 Making these two assumptions (and much experimental evi- 
 dence has been adduced to prove the validity of the Nernst 
 heat theorem), it is easily shown (loc. cit.} that the equilibrium 
 constant can be obtained from the following equation : 
 
 log 10 K = + 1-75 Sv log T - S 
 
 where Q is the heat of reaction, 
 and v a a + vjb = v c c + v d d 
 
 v = v a + v b - v c - v d , 
 
 v a) v b) v c , v d being the number of molecules of such species, 
 a, by c, d, reacting, C a , C 6) C c , C d being the so-called chemical 
 constants of each element or compound reacting, 
 
 and 
 
 applying this equation to the case under consideration, viz. : 
 
 20 3 = 30 2 + 68,000 calories. 
 5,, = [2] - [3] = - 1. 
 
 Svc = (2 x 3) - (3 x 2-8) = - 2-4. 
 
 Information as to the specific heat of ozone is at present not 
 available, but with Pollitzer, we may assume that its value is 
 not very different from that of the other triatomic gases, such 
 as sulphur dioxide, which has a molecular specific heat of 
 10'5, then 
 
 2 x I 
 Hence 
 
 0-005. 
 
 , v G8,000 n _ r . m 0-005T A 
 
 log K p = - - 1-75 log T + - - 2-4. 
 
46 OZONE 
 
 If x be the fraction of oxygen converted into ozone at equi- 
 librium, then since : 
 
 = J ^03 = ^ 
 p * p 
 
 when x is small, and p is the total gas pressure, 
 oo 000 
 
 log * = ~ ~ ' 875 Io T + ' 005T - 1 
 
 from this equation the values of x, and thus the percentage 
 of ozone present in oxygen at equilibrium at various tempera- 
 tures, can be calculated thus: 
 
 T centigrade. \QQ^ P = 10 000 atmospheres. 
 
 1000 10- 8 10- 6 
 
 2000 10- 5 10- 3 
 
 3000 10- 3 10- 1 
 
 It will be noted that increase of pressure greatly favours high 
 equilibrium amounts of ozone. 
 
 Somewhat higher values for the equilibrium amounts at 
 various temperatures are arrived at by means of evaluating 
 the magnitude of the potential difference between the ozone 
 and oxygen electrode (see p. 63). 
 
 The potential difference between two platinum electrodes 
 immersed in the same electrolyte, one saturated with oxygen 
 under a pressure IT, and the other with ozone at the same 
 temperature and pressure ir l9 is given by the equation : 
 
 TT TT BT 1 7T 
 
 V-V.-^rlog-, 
 
 where V represents the value observed of the potential differ- 
 ence 2 /0 3 , at one atmosphere, under conditions of reversi- 
 bility. 
 
THEEMAL PRODUCTION 47 
 
 There exists considerable uncertainty as to the values of 
 V , thus Luther and Inglis (" Zeit. Phys. Chem.," 43, 203, 
 1903) obtained the value V = - 0'736 volts ; Nernst (" Zeit. 
 Elektrochem.," 9, 891, 1903), V = - 0'57 ; Fischer and 
 Brauner (" Ber.," 39, 3631, 1906) - 0'64, and - 0*46 volts. 
 
 Calculation from the value of K obtained by the Nernst 
 heat theorem as follows, yields the value - 0'83 volts : 
 
 68,000 0-005T . 
 
 - 1-75 log T + -- 
 
 273 x 4-571 I"/ 34,000 
 
 ' ' 273 4 x 23,046 
 
 - 0-0005T + 1'2)] = - 0-83 volts. 
 
 It is evident that if the pressures of oxygen and ozone are 
 so adjusted that the cell shall have zero E.M.F., this will 
 represent the equilibrium conditions between oxygen and 
 ozone. 
 
 KT 
 
 V ' = 2F 
 
 TT 
 
 , 7T A 
 
 log - = f , 
 where A is a constant 
 
 E 
 
 The values of the percentage of ozone in equilibrium with 
 oxygen under one atmosphere pressure at various tempera- 
 tures, as calculated from the above equation for the two 
 extreme values of V , i.e. V = - 0'83 volts and V = - 0'46 
 volts, are given in the following columns : 
 
48 
 
 OZONE 
 
 Percentage of 
 Ozone. 
 
 Equilibrium 
 Temperature. 
 V = - 0-83 volts. 
 
 10 . 
 
 7900 
 
 1 
 O'l . 
 
 3950 
 2630 
 
 o-oi . 
 
 1970 
 
 o-ooi . 
 
 1580 
 
 Equilibrium 
 Temperature. 
 V = - 0-46 volts. 
 
 . 4400 
 . 2200 
 . 1460 
 . 1100 
 
 880 
 
 It will be noted that in this calculation the change in 
 specific heats of the gases with alteration in the temperature 
 have been neglected, consequently the values are probably 
 somewhat too high, and in view of the wide discrepancies 
 between the two values an experimental redetermination of the 
 oxygen ozone electromotive force would be very desirable. 
 
 From the foregoing considerations we must conclude that 
 the quantity of ozone in equilibrium with atmospheric oxygen 
 at normal temperature and pressures is scarcely detectable, 
 and that if present in measurable concentrations under these 
 conditions true equilibrium does not obtain. Further ap- 
 preciable quantities of ozone may be formed at high tempera- 
 tures and should be capable of detection and estimation. 
 
 The estimation of ozone in gases which have been heated 
 up to a high temperature, is somewhat difficult owing to the 
 fact that ozone rapidly undergoes decomposition to its 
 equilibrium concentration during the cooling of the gas mix- 
 ture. Dewar (" Year Book E.I," 559, 1887) inferred that 
 ozone had two centres of stability, one above the melting-point 
 of platinum and the other at ordinary temperatures, whilst 
 between these temperatures ozone is decomposed. Chapman 
 and Jones ("Trans. Chem. Soc.," 97, 2463, 1913, and 99, 
 1811, 1911) showed that at 100 C. nearly 75 per cent, of 
 
THERMAL PRODUCTION 49 
 
 the ozone in excess of the almost undetectable equilibrium 
 amount is destroyed in half an hour, whilst at 300 C. it is 
 practically instantaneous. 
 
 The explanation of these observations of Dewar is that 
 the velocity of decomposition of ozone from high temperatures 
 down to 100 C. is extremely rapid, whilst below 100 C. the 
 velocity of decomposition becomes markedly slower and the 
 ozone appears to be stable on account of the extremely low 
 velocity of decomposition, the equilibrium being "frozen". 
 
 The earlier experiments of Schonbein (Engler, "Hist. 
 Kritic Studien iiber Ozon," Halle, 1879), Bottger (" Ann. der. 
 Chem.," 125, 34, 1861), Pincus (" Pogg. Ann.," 144, 480, 1871), 
 Struve (" Jahresber. f. Chem.," 199, 1870), and Traube (" Ber.," 
 1 8, 1894, 1885), all indicated that small quantities of ozone 
 were formed during the combustion of hydrogen. Similar 
 results were obtained by the combustion of coal gas, notably 
 by Than ("Jour. f. Prakt. Chem.," 2, 1415, 1870), Loew 
 ("Zeit. f. Chem.," 65, 1870), Ilosvay ("Bull. Soc. Chem.," 3, 
 2, 360, 1881), whilst Zenghilis demonstrated the presence of 
 ozone (" Zeit. Phys. Chem.," 46, 1903) in the oxygen which 
 had been raised to a high temperature by the combustion of 
 aluminium powder. 
 
 Contemporary with these investigations, others were 
 carried out on the synthesis of ozone by merely heating air 
 or oxygen by means of an independent source of heat, as the 
 objection may be raised to the former experiments that the 
 ozone may have been formed by chemical activity (see Ch. III). 
 
 As catalytic agent hot platinum or silver was generally 
 employed, notably by V. der Willigen (" Pogg. Ann.," 98, 
 
 511, 1831), Meissner (" Neue Untersuchungen iiber Elekt. 
 
 4 
 
50 OZONE 
 
 Sauerstoff," Gottingen, 1863), Leroux (" C.R," 50, 691, 1860), 
 Troost and Hautefeuille ("C.R," 84, 946, 1877), Helmholtz 
 (" Wied. Ann.," 32, 18, 1887), and Elster and Geitel (" Wied. 
 Ann.," 39, 912, 1890). Troost and Hautefeuille (Zoc. cit.) 
 detected the presence of ozone in oxygen which had been 
 heated up to only 1400 C. The oxygen was heated by passage 
 through a porcelain tube maintained at 1400 C. In order to 
 effect the rapid cooling of the gas a water-cooled silver tube 
 passed along the axis of the porcelain tube. Samples of oxygen 
 were drawn from the annular space between the porcelain 
 and the silver tubes by aspiration through a small side tube 
 which passed into the silver tube itself. 
 
 J. Clements, at Nernst's instigation in 1904 (" Ann. Phys.," 
 14, 334, 1904), reviewed the whole subject and came to the 
 conclusion that many of the previous observers had mistaken 
 oxides of nitrogen or hydrogen peroxide for ozone. By the 
 use of Arnold and Mentzel's tetramethyl base paper (" Ber.," 
 35, 1324, and 2902, 1902), which is diagnostic for ozone, 
 Clements showed that ozone could be detected in the hot 
 gases from a Bunsen burner, but only in very small quantities. 
 (Tetramethyl base paper is stated to be sensitive to O'OOl per 
 cent, ozone.) 
 
 By the passage of ozonised air containing 1 per cent, of 
 ozone over a glowing Nernst filament maintained at 1000 C. 
 at various speeds up to 80 cms, per second, he showed that 
 the rate of decomposition of ozone was extremely rapid, 1 
 per cent. 3 sinking to O'OOl per cent, in 0*007 seconds. 
 
 In passing air over a Nernst glower even when heated up 
 to 3000 C., only oxides of nitrogen were obtained, a result 
 which was confirmed by Bossi (" Gaz. Chim. Ital.," 35, 1, 89, 
 
THEEMAL PEODUCTION 51 
 
 1905). Clements, however, confirmed the formation of ozone 
 by spark discharge, and Erode (" Zeit. f. Elektrochem./' 
 n, 754, 1905) observed the formation of ozone in the high 
 voltage arc at 4000 C. Ozone formation in these latter 
 cases may, however, be attributed to the action of ultra- 
 violet light (see p. 79) or electrical ionisation, and not to the 
 result of the establishment of a purely thermal equilibrium. 
 
 Fischer and his co-workers Brauner ("Ber.," 39, 940, 
 1906), Marx (" Ber.," 39, 3631, 1906, 40, 443, 1907), and Wolf 
 (" Ber.," 44, 2956, 1911) realised from Troost and Hautefeuille 
 and Clements' experiments that rapid cooling was essential to 
 preserve any ozone which might be formed, from secondary 
 thermal decomposition, during the process of cooling to the 
 point where the reaction of decomposition was negligibly 
 small. They showed by a series of interesting researches 
 that ozone could be formed by thermal methods provided 
 that the right conditions were obtained. 
 
 It was shown that ozone could be produced by plunging 
 a jet of burning hydrogen or acetylene into liquid aiy, the 
 ozone formed by the local heating being thus preserved by 
 rapid cooling. 
 
 When liquid oxygen was substituted for liquid air, large 
 quantities of ozone were formed and the liquid rapidly be- 
 came dark blue in colour, similar to ammoniacal solutions of 
 copper salts. When an electrically heated platinised wire 
 was immersed in liquid oxygen, practically no ozone was 
 formed. This somewhat unexpected result was shown to be 
 due to the fact that the platinum wire in the liquid oxygen 
 underwent dispersion into colloidal particles which catalytic- 
 ally accelerated the decomposition of the ozone. A bright 
 
52 OZONE 
 
 platinum wire, protected from dispersion by a coating of the 
 oxides of zirconium and yttrium, gave a uniform yield of 
 ozone. Using a glowing Nernst filament in liquid air and 
 oxygen, ozone was produced and no oxides of nitrogen. The 
 yield of ozone rose steadily with increasing temperature, the 
 maximum equilibrium amount being 1 '5 per cent, by weight 
 at an approximate temperature of 2200 C., a figure which 
 bears a strikingly close agreement to that obtained by calcula- 
 tion from the electrornetric force of the oxygen ozone cell. 
 Taking V = - 0'46 volts, this corresponds to an equilibrium 
 amount of 1*5 per cent, by weight at 2048 C. 
 
 Utilising an arc in liquid air a mixture of ozone and nitro- 
 gen peroxide was obtained which frequently exploded when 
 attempts were made to separate the residual oxygen 
 (B.P.0 3 - 120C.,0 2 - 182-7). 
 
 The maximum yield of ozone obtained by means of a 
 glowing Nernst filament in liquid oxygen was 40 mgms. in 
 twenty-five minutes with a current consumption of 0'25 
 amperes at 100 volts equal to a yield of 3*5 grams of ozone 
 per kilowatt hour. 
 
 Fischer having thus demonstrated the thermal production 
 of ozone with the aid of liquid air, proceeded to extend 
 Clements' experiments on the production of ozone by passing 
 air at a high flow rate over a glowing Nernst filament. We 
 have already noted that Clements was not able to detect the 
 thermal synthesis of ozone with air-flows of linear speeds, 
 up to 80 cms. per second. Fischer and Marx, using much 
 higher velocities, showed that ozone was formed under these 
 conditions and obtained a series of interesting results by 
 studying the conditions of oxidation. 
 
THEEMAL PRODUCTION 53 
 
 When dry air is passed over a glowing Nernst filament 
 two endothermic compounds may be formed, viz. nitric oxide 
 and ozone. If moist air be employed tbe presence of hydro- 
 gen peroxide- may likewise be demonstrated. The thermal 
 equilibrium concentrations of nitric oxide, formed according 
 to the reaction 
 
 N 2 + 2 ^ 2NO, 
 
 have been obtained by Nernst (" Gottingen, Nachricht.," 
 p. 261, 1904) and Jellinek and Finck (" Zeit. Anorg. Chem.," 
 49, 212, 224, 1906) and are given in the following table : 
 
 Temperature. Per Cent. Concentration 
 
 C. of NO in Air. 
 
 1811 0-35 
 
 2033 0-67 
 
 2580 2-02 
 
 2675 2-35 
 
 3200 5-0 
 
 Jellinek (loc. cit.) likewise calculated the rate of decomposition 
 of nitric oxide to its equilibrium value at various temperatures. 
 Nitric oxide in this respect is markedly different from ozone 
 since it is relatively much more stable at high temperatures ; 
 the times for the decomposition of half a given volume of NO 
 to nitrogen and oxygen at atmospheric pressure are as 
 follows : 
 
 Temperature. Time in Minutes to Effect 
 
 C. 50 Per Cent. Decomposition. 
 
 900 7-35 x 10 3 
 
 1100 5-80 x 10 2 
 
 1300 4-43 x 10 
 
 Similar calculations can be made for hydrogen peroxide. We 
 should therefore expect that with relatively low velocities of 
 
54 
 
 OZONE 
 
 air-flow over the glowing filament, only oxides of nitrogen 
 should be obtained ; with higher velocities mixtures of ozone 
 and nitric oxide, and with very high velocities only ozone, 
 since the rate of formation of nitric oxide as well as its rate of 
 decomposition is sensibly less than that of ozone. As will be 
 observed from the following figures obtained by Fischer and 
 Marx, this theoretical deduction is amply confirmed by ex- 
 periment : 
 
 Flow Rate in 
 Metres per Sec. 
 
 2-8 
 
 Reaction to Tetra- 
 methyl Base Paper. 
 
 N0 2 
 
 5-2 N0 2 + 3 
 
 5-5 3 + little N0 t 
 
 6-2 3 + trace N0 2 
 
 At flow rates exceeding 30 metres per second no oxides of 
 nitrogen could be detected in the air but ozone was always 
 present. The yield of ozone was influenced both by the 
 temperature of the glowing filament as well as by the 
 linear velocity of the gas flow, as shown in the appended 
 tables : 
 
 Temperature of 
 Nernst Glower 
 inC. 
 
 Weight Per Cent. 
 O 3 in Air. 
 
 Weight Per 
 Cent. O 3 in 
 Oxygen. 
 
 Air -flow 44 Meters 
 Per Sec., Oms. O 3 
 Per Kw. Hr. 
 
 1479 
 
 0-0029 
 
 0-0126 
 
 0-34 
 
 1698 
 
 0-0088 
 
 0-0382 
 
 0-80 
 
 1667 
 
 0-0118 
 
 0-0512 
 
 0-90 
 
 1772 
 
 0-0166 
 
 00720 
 
 1-07 
 
 1822 
 
 0-0218 
 
 0-0916 
 
 1-15 
 
 1889 
 
 0-0238 
 
 0-1032 
 
 1-19 
 
 1930 
 
 0-0293 
 
 0-126 
 
 1-30 
 
THERMAL PRODUCTION 
 
 55 
 
 
 Weight Per Cent. 
 
 
 Velocity of Air 
 
 O 3 in 
 
 Gms. O 3 
 Per Kw Hr 
 
 
 Air. 
 
 Oxygen. 
 
 
 
 at 1800 C. 
 
 
 30 
 
 0-011 
 
 0-049 
 
 0-29 
 
 44 
 
 0-019 
 
 0-082 
 
 0-85 
 
 57 
 
 0021 
 
 0-091 
 
 1-15 
 
 63 
 
 0-019 
 
 0-080 
 
 1-28 
 
 75 
 
 0-012 
 
 0-052 
 
 1-07 
 
 In the presence of water vapour, the yields of ozone are 
 considerably lower, and hydrogen peroxide is formed. The 
 influence of the water vapour was likewise investigated by 
 Fischer with the following results : 
 
 Water Vapour 
 
 Pressure in 
 
 mm. Hg. 
 
 . 
 
 0-002 
 
 5-0 . 
 
 10-0 . 
 
 441 . 
 
 149-2 . 
 
 Wt. Per Cent. 
 of0 3 . 
 
 0-0219 
 
 0-0176 
 
 0-00205 
 
 0-00136 
 
 0-00105 
 
 0-00076 
 
 Wt. Per Cent. 
 
 . 2 z n 
 10 Litres. 
 
 0-011 
 0-031 
 0-074 
 
 According to this investigator, the function of the water 
 vapour when present in but small quantities is purely cata- 
 lytic in depressing the yield of ozone by accelerating the 
 normal decomposition according to the equation 
 
 20 3 -> 30 2 . 
 
 When present in large quantities a reversible equilibrium 
 obtains as follows : 
 
 3 + H 2 2 ^ H 2 + O 2 + 20. 
 
 Nernst (loc. cit.) has calculated that the ratio of the equili- 
 
56 OZONE 
 
 brium concentrations of 3 and formed according to the 
 reversible equations 
 
 3 ^ 2 + 0, 
 2 ^ O + 0, 
 
 is as 10 23 to 1 or atomic oxygen is present in almost vanish- 
 ingly small concentrations ; consequently, in the above reaction 
 the yield of ozone and of hydrogen peroxide is extremely 
 small. Fischer was thus able to prepare ozone, nitric oxide, 
 or hydrogen peroxide, all endothermic compounds, from air 
 and water vapour at will by controlling the conditions so as 
 to take advantage of the different rates of formation and 
 decomposition of these substances at definite temperatures. 
 
CHAPTEK V. 
 
 THE ELECTROLYTIC PREPARATION OF OZONE. 
 
 As early as 1801 Cruickshank drew attention to the fact that 
 electrolytic oxygen, generated by the electrolysis of dilute 
 sulphuric acid at insoluble anodes, frequently contained 
 ozone. 
 
 Schonbein (" Ann. Phys. Chem.," 50, 616, 1840) showed 
 that the optimum yield of ozone was obtained when the sul- 
 phuric acid electrolyte contained 23*5 to 26*9 per cent, of 
 sulphuric acid ; solutions of phosphoric acid when submitted 
 to electrolysis likewise yielded small quantities of ozone in 
 the anodic oxygen. 
 
 De Marignac (" C.K.," 20, 808, 1845) appears to be the 
 first to point out the necessity of using cool electrolytes for 
 the production of ozone ; similar observations were made by 
 Williamson (" Mem. Chem. Soc.," 2, 395, 1845), H. Mei- 
 dinger (" Ann.," 88, 57, 1853), and Baumert ("Phil. Mag.," 
 4, 6, 51, 1853). 
 
 The next advance to be recorded was the observation of 
 H. Meidinger (" J.C.S.," 7, 151, 1854) that small anodes 
 were essential for the economic production of ozone. With 
 the aid of an electrode only 20 mm. long by 0'5 mm. wide, in 
 a sulphuric acid electrolyte of density T9, he obtained 0*3 
 per cent, ozone in the anodic oxygen. 
 
 Soret ("Pogg. Ann.," 92, 504, 1854), showed that the 
 
 (57) 
 
58 OZONE 
 
 quantity of ozone liberated in the oxygen was determined fcy 
 various factors. As electrode material, bright platinum, 
 gold or platinum iridium were found most suitable, since 
 other electrode materials, such 'as silver, black platinum, 
 or oxide anodes, such as lead, iron or manganese, exert a very 
 considerable activity in the catalytic decomposition of any 
 ozone which might be formed at the surface. 
 
 The temperature of the sulphuric acid electrolyte and also 
 of the anode itself plays an important part in obtaining 
 relatively large yields of ozone. Soret (loc. cit.) obtained the 
 following ozone concentrations when using constant currents 
 and electrolyte composition : 
 
 Temperature. Gms. O^per 
 
 Cubic Metre. 
 
 - 21 C 4-4 
 
 - 13 C 2-7 
 
 6C 0-9 
 
 De la Coux (" L'Ozone," p. 79) gives the following values for 
 the volume percentage of ozone liberated in the oxygen at 
 different temperatures : 
 
 Electrolyte : H 2 S0 4 : H 2 : : 1 : 5. 
 
 Temperature. Volume Per Cent. 
 
 of Ozone. 
 
 Boom 0-4 
 
 5-6C 1 
 
 Ice and salt freezing mixture 2 
 
 Andrews ("Phil. Trans.," i, 1850), utilising 20'81 per 
 cent, sulphuric acid as electrolyte and a bunch of platinum 
 wires as anode, kept cool during electrolysis by immersion of 
 the cell in ice water, obtained 0*85 per cent, ozone. Schone 
 (" Ber.," 6, 1274, 1873) claimed the production of 3'29 to 
 
THE ELECTROLYTIC PEEPARATION OF OZONE 59 
 
 8*6 per cent, ozone, and Carius (" Ber.," 174, 1, 1874), 3*44 
 per cent, of ozone by similar means. 
 
 Berthelot in 1878 (" C.R," 86, 74, 1878) observed the 
 formation both of ozone and hydrogen peroxide in sulphuric 
 and other electrolytes and that high anodic current densities 
 were essential for the production of ozone. The conception 
 of anodic current density as distinct from the utilisation of 
 small anodes marked a fundamental advance in the electro- 
 lytic synthesis of ozone. 
 
 Persulphuric acid (H 2 S 2 8 ) is simultaneously produced 
 when very concentrated electrolytes are employed. 
 
 The investigations of Berthelot were continued by 
 Eicharz (" Wiedemann Annalen.," 24, 183, 1885; 31, 912, 
 1887), who determined the yields of ozone, persulphuric acid 
 and hydrogen peroxide respectively with various current 
 densities and varying sulphuric acid concentrations at C. 
 It will be noted from the following tables that Kicharz con- 
 firmed the previous observer's results as to the necessity of 
 high anode current densities and relatively concentrated 
 electrolytes : 
 
 Calculated Volume 
 of O 2 Liberated 
 during a Definite 
 Time. 
 Litres. 
 
 Yield of Oxygen 
 
 Ozone. 
 Litres. 
 
 in the Form of : 
 
 Persulphuric Acid. 
 Litres. 
 
 2-4 . 
 
 
 
 . 0-03 
 
 3-74 . 
 
 . 
 
 . 0-40 
 
 7-47 . 
 
 . 
 
 . 2-32 
 
 17-12 . 
 
 . 0-04 
 
 . 8-08 
 
 30-0 . 
 
 . O'll 
 
 . 16-25 
 
 45-4 . 
 
 . 0-26 . 
 
 . 24-70 
 
 65-7 . 
 
 . 0-61 . 
 
 . 39-00 
 
 95-0 
 
 .1-1 . 
 
 , 45-60 
 
60 
 
 OZONE 
 
 Per Cent. 
 Sulphuric Acid. 
 
 Yield of Ozone. 
 
 Yield of Per- 
 sulphuric Acid. 
 
 Yield of Hydro- 
 gen Peroxide. 
 
 
 
 In Terms of Oxygen. 
 
 
 Litres. 
 
 Litres. 
 
 Litres. 
 
 10-1 
 
 0-11 
 
 0-62 
 
 
 
 19-8 
 
 0-18 
 
 6-79 
 
 
 
 28-3 
 
 0-11 
 
 16-75 
 
 
 
 39-5 
 
 0-11 
 
 22-01 
 
 
 
 50-7 
 
 0-15 
 
 18-76 
 
 
 
 60-0 
 
 0-06 
 
 4-85 
 
 2-54 
 
 69-4 
 
 0-05 
 
 3-49 
 
 3-43 
 
 77-6 
 
 0-07 
 
 2-55 
 
 4-17 
 
 89-4 
 
 0-07 
 
 1-21 
 
 2-61 
 
 McLeod ("Trans. Chem. Soc.," 44, 54, 1886) conducted a 
 very thorough investigation into the electrolytic preparation 
 of ozone, he showed the importance of the various factors, 
 viz. acid density, temperature of the solution and current 
 density to which attention had been drawn by previous 
 investigators. 
 
 From the following figures the extraordinary good yields 
 obtained by McLeod are evident : 
 
 Electrode Material. 
 
 Density of 
 Acid. 
 
 Per Cent. 
 Ozone by 
 Volume. 
 
 Current Density 
 Amps, per Sq. Cm. 
 
 6 platinum wires 
 
 ( 1-025 
 
 13-96 
 
 30-76 
 
 1 mm. long by 0-045 mm 
 diameter 
 
 4 1-075 
 U-25 
 
 16-7 
 9-5 
 
 50-6 
 50 
 
 6 platinum wires 
 6 mm. long by 0*045 mm 
 diameter 
 
 r 
 
 U-6 
 
 16-7 
 0-6 
 
 50 
 50 
 
 The influence of the acid density on the production of 
 available oxygen in the electrolyte in the form of persulphuric 
 acid and hydrogen peroxide was likewise investigated, the 
 optimum production occurring with an acid density of 
 specific gravity 1*20 with a current density of 50 amps, per 
 
THE ELECTEOLYTIC PREPAEATION OF OZONE 61 
 
 sq. cm., as is evident from the following figures taken from 
 McLeod's data : 
 
 Current Density Mols. active 2 
 
 Amps.jSq.Cm. Acid Density. Per 100 Mols. H.^ 
 
 evolved. 
 
 51 . . . . 1-05 .... 11-08. 
 
 53 . . . . 1-10 .... 20-80 
 
 54 . . . . 1-15 .... 25-8 
 53 . . . . 1-20 . . . .34 
 50 . . . . 1-25 .... 29-9 
 
 With the introduction of the ionic theory by Arrhenius 
 and van't Hoff in 1887 a more systematic investigation of the 
 anodic reactions taking place during the electrolysis of dilute 
 sulphuric acid was commenced. 
 
 It was shown that if the potential difference between two 
 platinum electrodes in dilute sulphuric acid be gradually 
 raised and the current intensity be plotted against the ap- 
 plied electromotive force a series of breaks occurs, which 
 breaks, on the ionic theory, correspond to different anodic 
 ionic discharges, the discharge of hydrogen being the only 
 cathodic reaction. Careful investigation has shown that the 
 ionic discharges associated with each break in sulphuric acid 
 and electrolyte are as follows : 
 
 P. D. Anodic Discharge. 
 
 1-08 0"->0 2 
 
 1-67 20H' -> 2 + H 2 
 
 1-95 S0 4 " -* H 2 S0 4 + 0, 
 
 2-60 HS0 4 '-(HS0 4 ) 2 
 
 2-83 30"_>O 3 
 
 
 
 In 1889 Nernst, by the introduction of the conception of 
 electrode solution pressure, pointed out the method of deter- 
 mining the influence of the anode potential on the discharge 
 
62 OZONE 
 
 of anions without having to take into account any cathodic 
 reactions. 
 
 If a platinum electrode be saturated with oxygen under a 
 definite pressure at a temperature of T, and immersed in a 
 sulphuric acid electrolyte, normal in respect to its hydrion 
 concentration, electrical equilibrium will finally be arrived at 
 between the oxygen molecules, atoms and ions in the elec- 
 trode and electrolyte, the electrode becoming positively 
 charged relatively to the solution by the discharge of nega- 
 tively charged oxygen ions, 
 
 + ve 
 
 0" 
 
 and a condition of equilibrium will obtain when the potential 
 difference between solution and electrode becomes sufficiently 
 great to prevent the discharge of any more negative ions. 
 
 If V be the electrode-electrolyte potential difference, 
 fju0 2 and /jiO" the molecular chemical potentials of the oxygen 
 gas in the electrode and of oxygen ions in the solution, then 
 if we imagine the transfer of a quantity of electricity 8e from 
 electrode to solution, the electrical work will be equal to 
 - VSe, the change in molecular chemical potential per mol. 
 will be yit0 2 - 2//.0", therefore, the work done on the transfer 
 of this quantity of electricity is equal to 
 
 - pO" 
 
 where e is the charge associated with one gram ion of a 
 monovalent element. 
 
THE ELECTEOLYTIC PEEPAEATION OF OZONE 63 
 
 If the conditions of reversible equilibrium obtain, then 
 
 Now yit0 2 = T((/>0 2 + E log 7r0 2 ) for a dilute solution, where 
 <p is independent of the concentrations, it being merely a 
 function of the temperature, 
 similarly pO" = T(c/>0" + E log CO"). 
 
 ET . CO" 
 
 V + - r log 
 
 Further, since in aqueous solutions 
 
 which gives an expression for the variation of the oxygen 
 electrode potential, with alteration in the hydrion concentra- 
 tion of the solution and the pressure of the oxygen gas. 
 
 The value of V is approximately - 1*35 volts, whence 
 the value of the cathode potential for a hydrogen electrode in 
 normal hydrion solution under a pressure of one atmosphere 
 is + 1-08 - 1-35 = - 0-27 volts. 
 
 If an oxygen electrode be set up against an ozone elec- 
 trode, the difference in potential between the two electrodes 
 can be calculated in a similar manner and found equal to : 
 
 Luther and Inglis ("Zeit. f. Physik. Chem.," 43, 203, 1903) 
 first attempted to obtain an approximate value for V/ by 
 
64 OZONE 
 
 measurement of the potential difference between an oxygen 
 and an ozone charged platinum electrode immersed in dilute 
 sulphuric or nitric acid. They obtained the value : 
 
 V; = - 0-736 volts. 
 
 We have noted that approximately the same value, viz. 
 - 0-83 volts, can be obtained by calculation from the Nernst 
 heat theorem. Subsequent investigators have found con- 
 siderably lower values : Nernst (" Zeit. Elektrochem.," 9, 89, 
 1903) obtained the value V = - 0'57, and Fischer and 
 Brauner("Ber.,"34, 3631, 1906) the values - 0'64 to - 0'46 
 volts. It would appear from the experiments of these latter 
 observers on the thermal equilibrium, that the lower value, 
 viz. - 0*50 volts is probably more correct. It is possible 
 that the higher values obtained by the earlier experimenters 
 were occasioned by the presence of oxozone 4 in the ozone 
 round the electrode, and a reinvestigation of this electrodic 
 reaction is clearly eminently desirable. 
 
 From the above calculation it is evident that an extremely 
 high anode potential is required to remove the formation of 
 ozone, whilst, in order to ensure its stability when produced, 
 both electrode and electrolyte must be kept cold. Fischer, 
 Massenez, and Bendixsohn ("Zeit. Anorg. Chem.," 52, 202, 
 1907; 11,229, 1907; 61, 13,153, 1909), realising these im- 
 portant factors, improved upon McLeod's results by adopting 
 the artifice of water-cooled electrodes, in addition to the sup- 
 plementary cooling of the electrolyte. 
 
 In their earlier experiments, a small platinum tube 6 mm. 
 long was sealed to two terminal glass tubes and served as 
 anode ; the tube itself was covered with glass with only a 
 thin strip, 0'4 mm. wide, exposed to the electrolyte. 
 
THE ELECTEOLYTIC PEEPAEATION OF OZONE 65 
 
 Cold water was circulated through this anode, and the 
 electrolyte was kept cool by immersion in ice water. 
 
 Utilising an anodic current density of 58 amperes per sq. 
 cm., and a sulphuric acid concentration of density 1*075 to 
 1*10, a yield of 17 per cent, ozone by weight (11 '3 per cent, 
 by volume) was obtained in the anodic oxygen. 
 
 A glass-covered, rhomboidal, platinum tube was then 
 substituted for the cylindrical one, and one edge, 0*1 mm. in 
 width, was exposed by grinding away the glass. The length 
 of the tube was 11*5 mm., and it was maintained at - 14 C. 
 by circulation of a solution of cold calcium chloride. A yield 
 of 28 per cent, ozone by weight (19 per cent, by volume) was 
 thus obtained. By embedding platinum foil in glass, and 
 exposing one edge only, 0*01 mm. wide, to the electrolyte, 
 slightly lower yields were obtained, viz. 23 per cent, of ozone 
 by weight. 
 
 In confirmation of McLeod's results, the optimum 
 concentration of sulphuric acid lay between = 1*075 and 
 1*10. 
 
 They noted that the quantity of ozone produced per kw. 
 hr. rose steadily with continued use of the platinum, which 
 became quite bright and burnished by the gas evolution in 
 course of time. 
 
 A yield of 7*1 gms. per kw. hr. was obtained at a potential 
 difference of 7*5 volts, and an anodic current density of 80 
 amperes per sq. cm. If we calculated the theoretical pro- 
 duction of ozone per kw. hr. from its heat of formation, i.e. 
 34,000 calories, the yield of 7*1 gms. per kw. hr. indicates an 
 
 7*1 
 
 electrical efficiency of only -^T^: = 0*6 per cent. 
 
66 OZONE 
 
 F. Fischer assumed that the primary discharge of ozone 
 occurs according to the equation 
 
 30" - 3 + 60, 
 
 which ozone is partly decomposed by the catalytic action of 
 the anode surface. 
 
 Other, but less efficacious methods have been suggested 
 from time to time for raising the anode discharge potential, 
 and thus increasing the yield of ozone. Donovan and Gard- 
 ner utilised a saturated solution of potassium permanganate 
 in from 5 to 10 per cent, of sulphuric acid, and obtained 
 relatively high concentrations of ozone ; chromic acid can 
 likewise be substituted for the permanganate. 
 
 St. Edme adopted the somewhat ingenious method of 
 obtaining a high anode current density, by employing 
 moistened crystals of phosphoric acid, or caustic potash or 
 soda, as the electrolyte. 
 
 In this way, the electrolyte was given sufficient conduc- 
 tivity for passage of the current, yet at the same time only 
 point contact between the moistened crystals and the anode 
 was ensured. 
 
 Archibald and von Wartenberg (" Zeit. Elektrochem.," 17, 
 812, 1911) pointed out that the low yields of ozone accom- 
 panying the electrolytic decomposition of dilute sulphuric 
 acid were probably occasioned by the high degree of anodic 
 polarisation that was produced when operating at high cur- 
 rent densities. In agreement with Fischer they considered 
 that the primary formation of ozone takes place according 
 to the equation 
 
 30" -> 3 + 60, 
 
THE ELECTBOLYTIC PEEPAEATION OF OZONE 67 
 
 but that the subsequent catalytic decomposition of ozone at 
 the electrode surface 
 
 20 3 -> 30 2 
 
 was not the most important factor. It was suggested that 
 the ozone thus formed is further oxidised at the anode 
 
 3 + 0" - 20 2 + 20, 
 
 consequently, if the anodic polarisation could be diminished 
 without alteration of the anodic current density, increased 
 yields of ozone could be obtained, since the secondary oxida- 
 tion would be diminished. 
 
 A series of experiments were carried out in which an 
 alternating current was super-imposed on the direct current 
 flowing through the cell ; this method of reducing the elec- 
 trode polarisation having been utilised in the Wohwill pro- 
 cess for the electrolytic parting of gold and silver, and in the 
 electrolytic preparation of hydrogen peroxide. As electrodes, 
 short platinum wires or platinum capillaries cooled with 
 water were utilised, as electrolyte sulphuric acid of varying 
 density, whilst a direct and alternating current of variable 
 periodicity was applied simultaneously to the cell. 
 
 It was established that the optimum acid density varies 
 with the area of the electrode and not only with the current 
 density, more concentrated electrolytes being desirable for 
 big electrodes as indicated by the following figures : 
 
 Area of Electrode Optimum Acid 
 
 in 53. Cm. Density. 
 
 0-041 1-34 
 
 0-333 1-478 
 
 The yield of ozone was also affected by the periodicity of the 
 
68 
 
 OZONE 
 
 alternating current, especially with currents of low frequency ; 
 above 20 periods per second the effect was not so marked. 
 
 Direct 
 Current. 
 
 Alternating 
 Current. 
 
 Periods per 
 Second. 
 
 Per Cent. O 3 . 
 
 
 Amperes. 
 
 
 
 0-61 
 
 0-50 
 
 11 
 
 3-04 
 
 0-61 
 
 0-50 
 
 17 
 
 4-10 
 
 0-61 
 
 0-50 
 
 25 
 
 4.77 
 
 It was noted that the applied potential difference necessary 
 for the passage of the current rose until the ratio 
 
 alternating current exceeded ^ then idl gunk ^ 
 direct current 
 
 this ratio became equal to 6, which was found to be an 
 optimum. The potential difference was found to sink with 
 increasing current density. 
 
 Their optimum yield was obtained under the following 
 conditions : 
 
 Electrode area : - 0*333 sq. cm. periodicity 18 v. per sec. 
 
 Temperature : 10 C. 
 
 Acid density : 1*478. 
 
 Direct 
 Current. 
 
 Alternating 
 Current. 
 
 A.C. 
 
 Anode 
 Current 
 Density. 
 
 Per Cent. O 3 
 by Volume 
 Calculated on the 
 Direct Current. 
 
 D.C. 
 
 Amperes. 
 
 0-25 
 
 1-50 
 
 6 
 
 0-75 
 
 37 
 
 The most important result from a technical point of view 
 was the effect of the alternating current on the potential 
 
THE ELECTROLYTIC PREPARATION OP OZONE 69 
 
 difference necessary to effect the passage of this current ; in 
 the above case only 2'75 volts being necessary with an anode 
 potential of 0'71 volt as opposed to 7'5 volts required by F. 
 Fischer for direct current. We can calculate from the above 
 data the production of ozone per kw. hr. as follows: 
 96,540 coulombs or 26*8 ampere hrs. liberate 1 gm. equiv- 
 alent, or 11 '2 litres of oxygen gas. Under the conditions of 
 operation, however, the liberated gas contains 37 per cent, 
 of ozone which would result from the condensation of 55'5 
 per cent. (37 per cent. + J 37 per cent.) of the oxygen, which 
 weighs 4'40 gms. Hence 26'8 ampere hrs. liberated 4'40 
 gms. of ozone. The potential difference which has to be ap- 
 plied to the cell to effect this liberation is 2' 75 volts, thus 
 4*4 gms. of ozone are produced by the expenditure of energy 
 equal to 26*8 x 2'75 or 73'7 watt-hrs., representing an output 
 of 59 gms. per kw. hr., or over eight times the yield. This 
 yield approximates to those obtained by the method of the 
 silent discharge, and it would appear possible, if larger 
 electrodes, and a cooled electrolyte were employed, to develop 
 this method of producing ozonised oxygen both for strong 
 and weak gas concentrations for the purposes of technical 
 production. 
 
CHAPTER VI. 
 
 PRODUCTION BY ULTRA-VIOLET RADIATION. 
 
 IN 1900, Ph. Lenard ("Ann. der Physik," i, 480, 1900), utilis- 
 ing a quartz mercury vapour lamp as a source of energy, 
 showed that ultra-violet light of extremely short wave length 
 was an effective agent for ozonising oxygen. Both Lenard 
 and E. Goldstein ("Ber.," 36, 3042, 1913) showed that 
 ultra-violet light in the Schumann portion of the spectrum 
 within the spectral region X = 120 ^ to X = 180 /z/-i exerted 
 the maximum activity in this respect; Goldstein (loc. cit.) 
 actually obtaining pure liquid ozone by means of a quartz 
 vacuum tube. Regener (" Ann. der Physik," 20, 1033, 1906), 
 who reinvestigated the matter, noticed the interesting fact 
 that although light of wave length X = 120 pn to 180 ^ was 
 a powerful ozonising agent, yet light still in the ultra-violet 
 portion of the spectrum of wave length X 230 /JL/J. to X = 290 fip 
 (especially X = 257 ft//,) exerted an equally effective catalytic 
 decomposing effect. Ozone is thus formed by light of short 
 wave length and decomposed again by light of slightly longer 
 wave length. According to E. Warburg the ozonisation 
 effected by ultra-violet light likewise increases steadily with 
 the pressure of the gas ("Deut. Phys. Ges. Vehr., 17, 10, 
 
 184, 1915). 
 
 (70) 
 
PRODUCTION BY ULTRA-VIOLET RADIATION 71 
 
 EFFECT OF RADIATION OF SHORT WAVE LENGTH. 
 
 Since marked concentrations of ozone result when oxygen 
 is subjected to irradiation in light of this wave length, it 
 necessarily follows that the energy necessary for the forma- 
 tion of ozone from oxygen is derived from the light, the 
 process of ozone formation being a typical photo-chemical 
 synthesis. 
 
 According to Planck's quantum theory (" Vorlesungen 
 iiber die Warmestrahlung," M. Planck, Leipzig, 1906, pp. 100, 
 et seq.), radiant energy is discrete, and can only be emitted 
 by an oscillator or absorbed by a resonator in definite quanta. 1 
 The magnitude of the quantum bears a definite relationship 
 to the frequency, of the light e = hv, where e is the magnitude 
 of the quantum, v the light frequency, and h Planck's constant 
 equal to 6*85 x 10~ 27 erg seconds. 
 
 Quanta, or the units of energy, may be emitted or absorbed 
 in single units or in even multiples of that unit at a time. 
 Two hypotheses have been advanced to explain the directional 
 motion of the quanta since it is evidently rectilinear in 
 motion. A. Einstein (" Ann. der Physik," 17, 133, 1905) pos- 
 tulates an entity for the quantum in the form of a light cell 
 which moves uniformly in the direction in which its centres 
 of gravity is projected. Sir J. J. Thomson (" Proc. Phys. Soc.," 
 14, 540, 1908; "Phil. Mag.," 792, 1913) has advanced the 
 ingenious hypothesis which assumes that the light travels in 
 the wave form as postulated on the old hypothesis, bat that 
 these waves are confined to certain directions, being virtually 
 
 , of course, possible, and indeed more probable, to assume that radiant 
 energy appears discrete because matter is discrete, and that the radiation 
 itself is continuous. 
 
72 OZONE 
 
 kinks in Faraday tubes which project from the point source. 
 A beam of light is thus compared to a bundle of a number of 
 Faraday tubes, and light transmission is effected by trans- 
 mission of pulses naturally of definite magnitudes, and there- 
 fore in quanta along these tubes. 
 
 The elements when raised to a high temperature emit 
 light in the form of spectral series. In all elements two 
 distinct types of light emission can usually be observed, 
 namely, band spectra and line spectra. Various investigators, 
 notably H. Delandres (" O.K.," 100, 1256, 1885, et seq.) have 
 shown that the elementary band spectra can be divided into 
 groups related by the expression v = Bn 2 + /3, where @ and B 
 are constants, and n a series of integers, whilst in each group 
 the frequency of the bands i/ are also related by the ex- 
 pression : 
 
 VQ = A (m + a) 2 + d, 
 
 where A, a, d are constants, and m a series of integers. 
 Again, in the line spectra, J. J. Balmer (" Verh. d. Natur. ges. 
 Basel," 2, 648, 750, 1885; " Wied. Ann.," 25, 40, 1885), 
 C. Runge ("B.A. Reports," 576, 1888), F. Paschen ("Ann. 
 der Physik," 27, 537, 1908, 35, 860, 1911), and J. R. Rydberg 
 ("K Svenska, Vet. Akad. Handl.," 23, 155, 1890) have 
 shown similar relationships. 
 
 From these and other considerations (see J. J. Thomson, 
 "Proc. Roy. Soc.," 14, 540, 1908; "Phil. Mag," 19, 331, 
 1910); J. Stark ("Prinzipien der Atorndynamik," Leipzig, 
 1911) we deduce that a chemical element is not composed of 
 homogeneous atoms or molecules, but that each atom or 
 molecule is composed of at least two parts, one which gives 
 rise to a line spectrum, and the other to a band spectrum 
 
PKODUCTION BY ULTRA-VIOLET RADIATION 73 
 
 when excited. The light thus emitted is periodic in character, 
 being produced by some form of oscillation or oscillators, 
 each periodic movement corresponding to one series of bands, 
 or lines in the spectrum. 
 
 From other considerations, such as the electrical pro- 
 perties and radioactivity of certain elements, the composite 
 nature of the atom receives confirmation. 
 
 Sir J. J. Thomson, who first suggested this actual structure 
 for the atoms, although speculations on the electrical nature 
 of matter had long been made for the purpose of calculation, 
 assumed the existence of a relatively large positive nucleus 
 with the negative electrons (or corpuscules) distributed in it. 
 A small positive nucleus with the electrons rotating round it, 
 in fact a small planetary system, is now a common hypoth- 
 esis. It is at present uncertain whether the inverse square 
 law or some higher power such as the inverse fifth power 
 conditions the rotation of the electrons. Information is also 
 lacking whether the electrons rotate in big or small circles, 
 i.e. whether the plane of their rotation passes through the 
 centre of gravity of the atom or not, and it is also a matter 
 of speculation whether the electrons are point charges, or 
 consist of rings such as are found in the satellites of Saturn. 
 
 It can easily be shown (see F. A. Lindemann, " Verh. d. 
 Phys. Ges.," 13, 482, 1911) that the amplitude of the vibrat- 
 ing particle in the oscillator emitting light radiation is of the 
 order 10~ 9 to 10~ 10 cms., or from 10 to 100 times smaller 
 than the actual diameter of an atom ; we are therefore forced 
 to the conclusion that the oscillators, both for the band and 
 line spectra emissions, are to be found in the atom itself. 
 Much evidence has been adduced to show that the source of 
 
74 t OZONE 
 
 infra-red radiation is the atom, of the visible light the charged 
 atom, and of the ultra-violet light the electron, the band 
 spectra owing their origin to the oscillations caused by the 
 swing of a valency electron about the positive nucleus. 
 
 If we imagine a valency electron circulating in its orbit 
 with a definite and constant momentum, a definite amount 
 of energy E must be supplied to remove the electron from 
 the system. If the electron be nearly but not quite removed 
 from the sphere of action of the atom it will oscillate about 
 its mean position of rotation and emit light. The energy of 
 oscillation must, according to Planck's hypothesis be a mul- 
 tiple of quanta, or : 
 nhv where n is a whole number, h Planck's constant, and v 
 
 the light frequency. 
 
 hence E must be > nhv to cause deformation and light 
 emission ; the smallest value of n is unity, so to cause light 
 emission by deformation of the orbit of a valency electron 
 
 E = liv, 
 
 TT 
 
 or v must be less than . It necessarily follows that as E, 
 
 n 
 
 the energy of deformation, is decreased, the wave length of 
 the light will increase or be shifted towards the infra-red 
 portion of the spectrum. The band spectrum of an element 
 thus, according to this view, optically expresses the configur- 
 ation of the valency electrons in the molecular system. 
 
 The relationship between position of the band spectrum 
 and complexity of the molecule in which the valency electron 
 is oscillating, can be clearly shown in the case of oxygen. 
 
 Monatomic oxygen, viz. 0, has a band spectrum in the 
 region X = 245 /JL/JL to X = 333 ^ (W. Steubing, " Ann. der 
 
PKODUCTION BY ULTKA- VIOLET KADIATION 75 
 
 Physik," 33, 353, 1913) . In diatomic or molecular oxygen, 
 viz. 2 , the band spectrum is shifted towards the ultra-violet, 
 viz. \ = 120 fifji to X = 190 /-tyu,, since the energy E, required 
 to remove a valency electron from two positive nuclei, is 
 much greater than is required to remove one of the two 
 valency electrons from the single positive nucleus of the 
 atomic oxygen. In ozone, on the other hand, not only is it 
 evident from its endothermic character, but also from a visual 
 representation of three positive nuclei coupled by valency 
 electrons, that the energy required to remove a valency 
 electron from the ozone molecule will be less than from the 
 molecular form, i.e. the band spectrum will be between the 
 two former. In fact, a strong absorption is noted at X = 
 258 ij.fi (W. N. Hartley, "Chem. News," 42, 268, 1888). 
 
 The oscillator of the series spectrum, on the other hand 
 (J. Stark, " Die Elektrizitat in Gasen," Leipzig, p. 447, 1902), 
 is to be found in the positive ion resulting from the complete 
 removal of a valency electron from an atom or a molecule. 
 The notable experiments of Sir J. J. Thomson (" Phil. Mag.," 
 13, 561, 1907, 21, 275, 1911, et seq.) on cathode ray analysis 
 have indicated that such ions which have lost or gained a 
 valency electron possesses actual entities, and can be distin- 
 guished one from another by their difference in electrical 
 charge ; thus in the case of oxygen there have been isolated 
 the charged gas ions : 
 
 ++ + - 
 0, 0, 0, 0, 
 
 as well as oxygen molecules of various charges. 
 
 F. Horton ("Phil. Mag.," 22, 214, 1914) has identified as 
 carriers of positive electricity, giving positive band spectra in 
 
76 OZONE 
 
 oxygen, the following polymers of electric atomic weights 
 8, 16, 32, 48, 96. 
 
 N. Bohr ("Phil. Mag.," 26, 476, 1913) attributes to the 
 oxygen atom a nucleus carrying eight unit position charges 
 with eight electrons, of which only two appear removable by 
 methods at present available. 
 
 The Mechanism of Ozone Formation. 
 We have already noted that the oxygen molecule when 
 subjected to ultra-violet light radiation of the correct fre- 
 quency for resonance may absorb quanta of energy. Similar 
 conditions obtain for the iodine molecule in the infra-red 
 spectral range and we may regard the primary cleavage to 
 occur in a similar manner, viz. : 
 
 I 2 ; I + I 
 2 ^ + O. 
 
 Warburg (" Preuss. Akad. Wiss.," Berlin, 872, 1914) has 
 adopted this hypothesis to explain the mechanism of ozone 
 formation. He assumes that the atomic oxygen resulting 
 from the cleavage of the molecule secondarily reacts either 
 with atomic oxygen to reform molecular oxygen as indicated 
 by the reversibility of the above equation, a point clearly 
 emphasised by Nernst from thermal considerations (see p. 29), 
 or it may react with molecular oxygen to form ozone 
 
 O 2 + = O 8 . 
 
 Warburg's experiments, conducted under pressures of from 
 30 to 400 kgm. per sq. cm., yielded a photo-chemical efficiency 
 of 55 per cent, at 120 kgm. cm. 2 and 29 per cent, at 300 kgm. 
 cm. 2 , indicating the plausibility of the above hypothesis. 
 Weigert (" Zeit. Wiss. Photochem.," n, 381, 1912) obtained 
 
PEODUCTION BY ULTKA-VIOLET EADIATION 
 
 77 
 
 a photo-chemical efficiency of 46 '0 per cent, and a thermo- 
 dynamic efficiency of 27'7 per cent. 
 
 The reaction is primarily a molecular one and the energy 
 of formation of a gram. mol. of 3 (34,000 cal.) should have 
 its corresponding photo-chemical equivalent equal, as cal- 
 culated by Warburg, to Nhv where N is the number of 
 molecules per gram. mol. From the relationship 
 
 (cal.) ncm. 
 
 with a value of 34,000 calories for the heat of formation of 
 ozone we obtain X = 800 /JLJJL, as the critical ozonising wave 
 length, or we must assume that one quantum of X = 200 
 will form four ozone molecules. 
 
 Ozone formation may therefore occur without gas ionisa- 
 tion, a fact which was first demonstrated by Lenard and con- 
 firmed by Ludlam (" Phil. Mag.," 
 23, 757, 1912). 
 
 We can easily deduce from our 
 previous considerations on the 
 mechanism of photo-chemical 
 processes in the light of the 
 quantum theory that ionisation 
 of oxygen will be brought about 
 by light of shorter wave length 
 than that required to produce 
 atomic oxygen and hence ozone. 
 
 If A, B, represent the two 
 positive nuclei of oxygen atoms 
 in a neutral molecule and c, c, 
 one valency electron of each atom which has come within the 
 attraction (partially saturated) of the positive charge of the 
 
 FlG * 
 
78 OZONE 
 
 neighbouring atom ; the simplest line of cleavage is along a, 
 a', resulting in the formation of an oxygen molecule with one 
 bond as link O, which can then react to form ozone 
 300 -> 20 3 . 
 
 This requires the smallest amount of energy, and hence is 
 effected by light of the longest wave length (> X = 200 /^). 
 The simplest cleavage into neutral atoms takes place along 
 the line b, b', requiring more energy than is necessary to 
 effect partial unsaturation, and hence shorter wave length 
 light X = 200 fifi. Cleavage along the line d, d' necessitates 
 the removal of one electron from an oxygen atom and conse- 
 quent increase of energy or light of a still shorter wave length, 
 ca. X = 180 pp. The energy required to move a valency 
 electron which is partially attached to two atoms on to one 
 atom is of the order 1 x 10~ 12 ergs, to completely remove 
 the electron requires a considerably greater expenditure 
 of energy. A quantum of light energy in the visible or 
 ultra-violet portion of the spectrum is of the order of 
 >3 x 10~ 12 ergs; thus with very short wave length light 
 electron removal can easily be effected. 
 
 Under these conditions we obtain monatomic oxygen 
 ions 
 
 2 ^ 6 + 6. 
 
 Ozone formation may result from this ionisation according to 
 the following reaction : 
 
 20 2 + O + 6 = 20 3 . 
 
 Light of still shorter wave length will actually remove elec- 
 trons from the monatomic oxygen ion (Ca. X = 130 fifi) 
 
 + ++ 
 
 -> + 0, 
 
 which electron may attach itself to the charged - ve residue 
 
PEODUCTION BY ULTKA-VIOLET EADIATION 79 
 
 provided that it be projected from the original ion with suf- 
 ficient kinetic energy 
 
 6 + -> a 
 
 Or again, it may attach itself more easily than as above to a 
 neutral molecule 
 
 2 + -> 6 2 . 
 
 In this way we can imagine the formation of the various 
 charged ions actually observed during irradiation of oxygen 
 by ultra-violet light of short wave lengths within the range 
 X = 130 to 200 fifjt. 
 
 Construction of Apparatus. 
 
 (a) Source of Ultra-violet Light. We have already in- 
 dicated that for the production of ozone a source of ultra- 
 violet light rich in lines of the Schumann region (below 
 X = 200) and if possible free from light of longer wave length, 
 especially in the region X = 230 to 290 /*//-, which exerts a 
 strong catalytic activity in deozonisation. 
 
 A glance at the curves representing the distribution of 
 energy over the spectrum radiated from a black body at 
 various temperatures will suffice to indicate that " black 
 body " radiation is unsuitable as an efficient source of Schu- 
 mann light. In agreement with the theoretical calculations 
 of Wien and Planck the experimental observations of Lummer 
 and Pringsheim (" Ver. d. Deut. Phys. Gesell.," i, 23, 1899 ; 
 2, 163, 1900) have indicated that, with elevation of the 
 temperature of the radiator, the maximum energy emission 
 E w shifts from the longer to the shorter wave length portion 
 of the spectrum. Even, however, at sun temperature, 
 ca. 5500 C., which temperature can only be approached with 
 difficulty by the utilisation of carbon arcs under high gas 
 
80 OZONE 
 
 pressures, a black body radiator will have its E position at 
 about X = 500 /*//,, or in the green of the visible spectrum. 
 The fraction of the total energy emitted which will lie in the 
 Schumann region of the spectrum XI to 200 /JL/J, will be re- 
 markably small. 
 
 We must, therefore, reject black body radiators and fall 
 back on methods of obtaining selective emission and as such 
 light sources we may utilise the arc, spark, or vacuum tube 
 illumination of various elements. 
 
 Arc and Spark Light Sources. 
 
 Most metals exhibit strong Schumann and ultra-violet 
 light radiation when the arc or spark electric discharge is 
 made to pass between metallic electrodes. 
 
 We may argue from the electronic structure of the atom 
 that since the removal of a second electron from an atom 
 which has already lost one, necessitates the supply of a still 
 greater quantity of energy for its removal than the first, and 
 as this energy is supplied in quanta, the value of hv must rise 
 with each subsequent removal. Atoms which can loose many 
 electrons without loss of atomic identity will therefore radiate 
 light corresponding to high values of hv, i.e. of extremely 
 short wave length. 
 
 The elements which have been most widely used as 
 sources of ultra-violet light are those of aluminium, iron and 
 especially mercury, which can loose as many as eight electrons. 
 
 The aluminium ultra-violet spectrum has been investigated, 
 more particularly by Lenard (" Sitz. Heidelberg, Akad. Wiss. 
 Abh.," 31, 1910) and Morris Airy (" Man. Lit. Phil. Soc.," 
 XLIX, 1, 1905), and iron by Lyman (" Astrophys. Jour.," 
 38, 282, 1913). 
 
PRODUCTION BY ULTRA-VIOLET RADIATION 81 
 
 Much work has been accomplished on the mercury arc, 
 which is especially rich, both in green, violet, and ultra-violet 
 radiation, most conspicuous where mercury vapour lamps are 
 used as light sources. 
 
 The investigations of Tian (" O.K.," 155, 141, 1912) and 
 Lyman (" Astrophys. Jour.," 38, 282, 1913) have shown that 
 "the spectrum is dominated by the broad unsymmetrical 
 line X = 184 '96 /*//, ". The spark spectrum of mercury is 
 rich in lines, whilst the arc spectrum contains only a few. 
 Other lines in the same series predicted by Paschen have 
 likewise been observed at X = 14O2 /JL/JL and X = 126*9 pp. 
 When viewed through a short column of air the line X = 
 184*96 /xyu, is replaced by three groups of faint lines observed 
 by Steubing (" Ann. der Physik," 33, 573, 1910). lonisation 
 of the mercury atom by collision commences at X = 253*6 /*/*, 
 equivalent to a fall of potential of the colliding electron of 
 4*8 volts A. Lande (" Phys. Zeit.," 15, 793, 1914), J. Franck 
 and G. Hertz ("Deut. Phys. Ges.," 16, 407, 1914). 
 
 Vacuum Tube Discharge. 
 
 Of the elements investigated by means of the vacuum 
 tube discharge, the remarkable activity of mercury vapour in 
 the emission of light of short wave length has already been 
 discussed, two other substances also exhibit a marked selec- 
 tive emission in the ultra-violet region, namely hydrogen 
 and carbon, the latter usually introduced into the vacuum 
 tube in the form of one of its oxides, carbon monoxide or 
 dioxide. Lyman (" The Spectroscopy of the Extreme Vio- 
 let," Longmans, 1914) states that hydrogen surpasses all 
 other gases in the wealth and strength of lines in the Schu- 
 mann region. They extend at pressures of 1 to 5 mm. from 
 
 6 
 
82 OZONE 
 
 \ = 90 /x/u, to X = 167'5 /*//,, and this light forms one of the 
 most important of all three distinct spectra which the 
 element possesses. St. John ("Astrophys. Jour.," XXV, 
 p. 45, 1907) found hydrogen to emit 250 times as much 
 energy of short wave length as a mercury vapour lamp. 
 There appears to be a distinct gap in light emission within 
 the spectral region X = 167*5 /z/i to X = 243 '3 /*/*. 
 
 Delandres (" C.R," 106, 842, 1888) noted a great number 
 of bands in the ultra-violet spectrum with the rarified oxides 
 of carbon in the vacuum tube within the range X = 130 /*//, 
 and X = 210 /^. 
 
 When consideration is taken of the difficulties in the 
 operation of an arc lamp with iron or aluminium electrodes, 
 such as the automatic adjustment of the arc gap, the removal 
 of the oxides produced during combustion, if the arc be open, 
 or the volatilisation of the metals on to the walls, if the arc 
 be of the enclosed type, as well as the great thermal effects 
 produced by an arc lamp in continuous operation, which, as 
 we have seen, militates against a high yield of ozone, it will 
 be clear that the mercury vapour lamp operating at low volt- 
 ages and relatively high internal mercury vapour pressures, 
 or working at high voltages with only a few millimetres 
 pressure of vapour, is the most suitable source of ultra-violet 
 light which, up to the present time, has received systematic 
 investigation. 
 
 The utilisation of hydrogen vacuum discharge tubes, 
 however, may possibly receive more attention in the future, 
 since discoloration of the tube walls, so frequently noticed in 
 mercury lamps, would be greatly minimised (see also Lyman, 
 "Astrophys. Jour.," 27, 87, 1908). 
 
PEODUCTION BY ULTEA- VIOLET KADIATION 83 
 
 (b) Material for Lamp Construction. 
 The walls of the mercury vapour lamp must be trans- 
 parent to radiation of this extremely short wave length, in 
 order that ozonisation of the surrounding oxygen may be ef- 
 fected. Ordinary glasses are singularly opaque, thus boro- 
 silicate crown-glass, which is the most transparent of the 
 ordinary glasses, passes only 8 per cent, of light of wave 
 length X = 309 pp, (Kriiss, "Zeit. f. Instrumentkunde," 23, 
 197, 1903), and is opaque to light below X = 297 ^. Schott 
 of Jena's uviol glass was specially prepared by Zschimmer 
 (" Zeit. f. Instrumentkunde," 23, 360, 1903) for ultra-violet 
 transparency. With a thickness of 1 mm. fifty per cent, 
 transmission of light at X = 280 pp is effected, whilst a uviol 
 microscope cover-slip is still transparent to X = 248 /*/*. 
 Zschimmer further indicated ("Phys. Zeit.," 8, 611, 1907) 
 that pure boric anhydride and silica are very transparent 
 even below X = 200 ^ (and even below X = 185 /*/*), 
 but that the addition of certain salts lessens their transpa- 
 rency. Boric anhydride is slightly inferior to silica, its lower 
 limit of transparency, according to Lyman, being X = 170 
 pp. Fritsch ("Phys. Zeit.," 8, 518, 1907) gives the follow- 
 ing composition of a glass extremely transparent down to 
 
 X = 185 pp : 
 
 CaF 2 6 parts. 
 
 B 2 3 14 
 
 M. Luckiesh ("J. Franklin Inst.," 186, 111, 1918) claims 
 that a special cobalt-blue glass is more transparent than 
 ordinary glass to ultra-violet. With the exception of Fritsch's 
 borate glass, which does not appear to have received any 
 technical application as yet, extremely pure fused silica is the 
 most suitable material for lamp construction. 
 
84 OZONE 
 
 Hughes (" Photo Electricity," 1913, p. 137) has shown 
 that fused quartz is still transparent down to X = 145*0 fj.fi ; 
 a thickness of 0'3 mm. will transmit 24 per cent, of X = 184'9 
 fji/ju, 36 per cent, of X = 197 yu./-t, and 40 per cent, of X = 
 
 200'2 yLtyu,. 
 
 Mention may be made of the naturally occurring sub- 
 stances, which are even more transparent to ultra-violet 
 light than fused silica, viz. quartz, fluorite, and rock salt. 
 Quartz in very thin laminae is transparent down to X = 145 
 ftp, rock salt to X = 175 /t/*, and fluorite to X = 123 /A/A. 
 
 In the Quain apparatus, which is the only form of quartz 
 mercury vapour lamp ozoniser in technical use, the lamp 
 which is of the vacuum type and operated by a coil or mag- 
 neto, at a terminal potential difference of circa 7000 volts, 
 is inserted in a hollow aluminium tube through which the 
 air or oxygen, undergoing ozonisation by irradiation from 
 the lamp, is passed at a relatively low velocity. No litera- 
 ture has been published dealing with the problem of the in- 
 fluence of gas velocity on ozone concentration and ozone 
 production per minute, but the following considerations will 
 indicate that the optimum conditions will very likely be 
 formed when only a thin film of air passes over the lamp at 
 high velocity. Dry and dust-free air is relatively transparent 
 to light above X = 186 /A/A, but nearly opaque to light below 
 X = 178 /A/A. Lyman (" Astrophys. Jour.," 27, 87, 1908) 
 states that 1 mm. of air will cut off most of the light below 
 X = 185 /A/A, which, as we have seen, is the active light for the 
 production of ozone. Kreusler ("Ann. der Physik," 6, 418, 
 1901) gives the following figures for the absorption produced 
 by 20 '45 cms. of oxygen : 
 
PEODUCTION BY ULTRA-VIOLET RADIATION 85 
 
 Light of Per Cent. 
 
 Wave Length. Absorption. 
 
 186 32-5 
 
 193 6-2 
 
 200 negligible 
 
 whilst Schumann observed an air film only 4 mm. thick 
 (equal to '8 mm. of oxygen approximately) was sufficient to 
 render all lines below X = 178 /*//, extremely faint. With 0'5 
 mm. of air, light down to X = 168 /JL/J, would be transmitted, 
 and below 0'05 mm. in air thickness, the spectrum stretched 
 considerably below X = 160 /-t/4. It will thus be observed 
 that the ozonising action of ultra-violet light, in so far as it 
 is caused mainly by light of wave length smaller than X = 
 200 fifji is confined to but a millimetre thickness or so of air. 
 
 Lamp Efficiency. 
 
 Figures are not available as to the optimum conditions 
 for the production of ultra-violet light from mercury vapour 
 lamps. As is to be expected the ultra-violet light fraction 
 increases with increasing voltage (see A. Tian., " C.R.," 155, 
 141, 1912). 
 
 J. N. Pring ("Proc. Roy. Soc.," 96, 204, 1914) showed 
 that no oxides of nitrogen or hydrogen peroxide were formed 
 during operation and that the average ozone content of the 
 air in the neighbourhood of the lamp was O'Ol per cent, at 
 760 mm. and 0*0014 per cent, at 30 mm. air pressure. W. 
 Chlopin (" Zeit. Anorg. Chem.," 71, 2198, 1911), on the other 
 hand, detected the presence of both hydrogen peroxide, ozone 
 and nitrous anhydride by exposure for a few minutes of 
 ordinary moist air to the rays of a Westinghouse quartz 
 mercury vapour lamp. 
 
86 OZONE 
 
 The ultra-violet efficiency of the various types of mercury 
 vapour lamps on the market was examined by C. Fabry and 
 Buisson in 1911 ("C.E.," 153, 93, 1911), who obtained the 
 following results : 
 
 Percentage of Power Sup- 
 Lamp, plied, Radiated in Wave 
 
 Lengths below 320. 
 
 Westinghouse 6 
 
 A.E.G 4-7 
 
 0-85 
 
 Westinghouse (ii) . . . . .0-13 
 
 PRODUCTION BY IONIC COLLISION. 
 
 In the previous discussion we have noted that molecular 
 cleavage of oxygen into neutral atoms with or without sub- 
 sequent ionisation may be brought about by absorption of 
 light energy, provided that this latter is of the correct fre- 
 quency for absorption. 
 
 The production of ozone depends primarily on the simplest 
 
 cleavage, viz : 
 
 2 -> + 0, 
 
 with subsequent synthesis of ozone, whilst secondary ozone 
 formation probably results from the combination of charged 
 ions, e.g. : 
 
 6 2 + 6 = O 3 
 
 (see W, W. Strong, " J. Amer. Chem. Soc.," 50, 104, 1913) 
 The cleavage and ionisation of the oxygen molecule may 
 
 also be brought about by other means than by the absorption 
 
 of light quanta, such as by direct impact by a particles or 
 
 electrons. 
 
 Madame Curie noticed that radium salts were effective in 
 
 ozonising oxygen (" C.E," 183), a point at first disputed by 
 
PEODUCTION BY ULTRA-VIOLET RADIATION 87 
 
 Kamsay and Soddy, but finally confirmed by Giesel and 
 Nasini and Levi (" Atti. K. Accad. Lincei," 17, 46, 1908). S. C. 
 Lind (" J. Amer. Chem. Soc.," 47, 397, 1912) and O. Schoner 
 (" C.E.," 159, 423, 1914) showed that the a particles projected 
 from radium ozonised oxygen ; Lind showed, inter alia, that 
 the number of ozone molecules formed were equal to the 
 number of ions made by the a particles 
 
 2 -> + O 
 20 2 + 20 = 20 3 . 
 
 In many of his experiments a slight deficiency in ozone 
 formation was observed from that calculated, but under no 
 circumstances was more ozone than the theoretical obtained. 
 (See also W. Duane, "C.K.," 153, 336, 1911.) It may be 
 noted in passing that similiar results were obtained for 
 hydrogen by W. Duane and Wendt (" Phys. Kev.," 10, 110, 
 1917), the presence of H 3 being clearly demonstrated. F. 
 Kriiger ("Phys. Zeit.," 13, 1040, 1912) obtained ozone by the 
 ionising action of Lenard rays obtained by the projection of 
 cathode rays through an aluminium window and showed, as 
 indicated in the following tables, that more ozone was formed 
 per second than ions in oxygen, the number of molecules of 
 ozone formed approximately more closely to the ionisation 
 of nitrogen under similar conditions : 
 
 No. Ions Produced per Sec. in No. Hols. O s Produc&A per Sec. 
 
 2 x 10 14 . tf 2 x 10 14 . x 10 14 . 
 
 0-70 6-0 7'0 
 
 0-56 1-2 1-1 
 
 0-41 1-4 1-4 
 
 0-21 0-5 0-33 
 
 In the case of radium, practically all the ozone produced is 
 formed through the agency of the a particles, the and y 
 
88 OZONE 
 
 radiation producing but minor and secondary effects. The 
 energy associated with each group of rays is clearly demon- 
 strated from the following figures ("Phil. Mag.," 22, 567, 
 1907) : 
 
 Heating effect of 1 gm. radium = 110 gms. cal. per hr. 
 a rays = 103*5 
 ^ = 2-0 
 7 = 4'5 
 
 With Kontgen or X-rays, on the other hand, ionisation is not 
 so marked, since only about 1 atom in 12 12 is ionised by the 
 penetration of the rays into a substance. 
 
 For ionisation to be effected by collision, the molecule or 
 atom must be struck by the a particle or electron with suffi- 
 cient energy to discharge a valency from its normal orbit in 
 the atomic sphere. It will thus leave the atom with a certain 
 critical velocity which it would also acquire if it had been 
 acted on by a definite potential difference. We may there- 
 fore equate the loss in kinetic energy sustained by the im- 
 pinging a particle or electron as a result of collision with the 
 molecule and the energy of discharge of the electron. 
 
 If m be the mass of the impinging electron, V Q its incident 
 and v l final velocity, its loss in kinetic energy will be : 
 l/2m(t; 2 - ^i 2 ) whilst the discharged electron of charge e will 
 possess an energy Ve. 
 
 Hence l/2m(^ 2 - ^ 2 ) = Ve. 
 
 A discharged electron or a particle will thus continue its 
 passage through the gas, causing ionisation by collision on its 
 way until its velocity sinks to the value v z where 
 
 = v, 
 
PRODUCTION BY ULTRA-VIOLET RADIATION 89 
 
 the minimum velocity necessary to cause ionisation by colli- 
 sion. Below this velocity the electron will merely adhere to 
 a neutral molecule to form a negatively charged ion, the a 
 particle will loose its charge to a neutral molecule to form an 
 atom of helium and a negatively charged gas ion, provided that 
 they have not come in contact with the walls of the con- 
 taining vessel before their journey is completed. 
 In the case of ionisation by electrons the value of 
 
 = 1-77 x 10 7 , 
 m 
 
 /. V - 2-82 x 10 - 16 v z 
 
 or vcc 
 
 In the following table are given the electron velocities in 
 cms. per second and the potential difference in volts required 
 to bring them to rest : 
 
 V. v. 
 
 1 5-9 x 10 7 
 
 10 1-88 x 10 8 
 
 100 0-595 x 10 9 
 
 1000 1-88 x 10 9 
 
 10,000 5-95 x 10 9 
 
 100,000 18-8 x 10 9 
 
 200,000 27 x 10 9 
 
 For the minimum velocity required for an electron to 
 cause ionisation of an oxygen molecule by collision Franck 
 and Hertz ("Verh. d. Deut. Phys. Ges.," XV, 34, 1913) 
 found the value 1*80 x 10 8 cms. per second corresponding to 
 a fall of 9'0 volts. 
 
 This value is in extremely good agreement with that cal- 
 culated from the critical wave length requisite for ionisation 
 by absorption of ultra-violet light quanta. It is evident that 
 
90 OZONE 
 
 the requisite energy equal to ~Ve can be supplied by the kinetic 
 energy lost by an impinging electron, i.e. l/2wv 2 , or by the 
 absorption of a light quantum hv thus 
 
 Taking \ = 135 pp we obtain the value 9 '20 volts for the 
 value of V determined in this manner. A value of 8*6 volts 
 being obtained by Compton (" Phys. Rev.," 8, 412, 1916), by 
 calculation of the work necessary to remove a valency electron 
 from an atom possessing Bohr's hypothetical structure. 
 
 Quantitative agreement between the yield of ozone calcu- 
 lated and that actually obtained has, as has already been 
 mentioned, been shown to hold for the case of ozonisation by 
 a particle discharge by Lind. Cases of ionisation and ozon- 
 isation by electron emission have given more variable 
 results. This is in part due to the great velocity and rela- 
 tively small size of the electrons which can pass through a 
 vessel containing gas and come to rest on the walls without 
 having made a great number of collisions, thus the major 
 part of its kinetic energy is still retained when it emerges 
 from the gas and strikes the walls. Again, it appears that 
 every collision which an electron makes with a molecule of 
 oxygen, with sufficient energy to dissociate the molecule, is 
 not always effective in doing so. According to P. Kirkby 
 (" Proc. Roy. Soc.," 85, 151, 1911), only 50 per cent, of such 
 collisions are effective. The yield of ozone by electron col- 
 lision in oxygen, therefore, usually falls far short of the 
 theoretical quantity. 
 
CHAPTEE VII. 
 
 PRODUCTION BY MEANS OF THE SILENT ELECTRIC DISCHARGE. 
 
 THE formation of ozone by the action of the silent discharge 
 on air is the only process of ozone production which has 
 received considerable technical development and a great 
 number of ozonisers of various types and designs have been 
 incorporated in installations for the economic manufacture 
 of ozone and ozonised air. 
 
 It may be stated at the outset that we do not possess 
 sufficient information about the mechanism of the silent 
 discharge to put forward a satisfactory explanation as to the 
 modus operandi of a " Siemens tube," nor can it be said that 
 the design and construction of ozonisers is on a scientific 
 basis, since, with the exception of a few generalisations based 
 on experiment and a few suggestions based upon somewhat 
 unsatisfactory and frequently incomparable theories, ozonisers 
 have been built on the rule of thumb and hit-or-miss 
 principle. 
 
 The following considerations, however, will indicate in 
 some measure the intricacy of the problem : 
 
 If the potential difference between a point and a plate 
 separated by a few millimetres of air space from it be 
 gradually raised, the current potential difference curves be 
 plotted, and they will be found to possess certain character- 
 istic features for both direct and alternating currents. The 
 V, i characteristic curves for point plate and plate discharges 
 
 have been obtained with great accuracy by Toepler ("Drud. 
 
 (91) 
 
92 
 
 OZONE 
 
 Ann.," 7, 477, 1902) and Brion (" Zeit. Elektrochem.," 14, 245, 
 1906) for direct currents, and by Cramp and Hoyle (" Electro- 
 chem. Ind.," 7, 74, 1909) for alternating currents. 
 
 The following indicate the series of changes in the char- 
 acter of the discharge obtained by Toepler : 
 
 ~ ve point to 
 "*" ue point 
 
 Low Tens ion 
 Arc. 
 
 I 
 
 FIG. 5. 
 
 + ve point to 
 t/e point 
 
 L 
 
 PIG. 6. 
 
PKODUCTION BY SILENT ELECTEIC DISCHAEGE 93 
 
 If a very small potential difference be applied between a 
 negative point and positive plate, a small amount will flow, 
 the current being carried entirely by the negative ions 
 present in the gas ; the i, V curve will then follow Ohm's 
 law until the rate of removal of gas ions by the electric cur- 
 rent becomes equal to the rate of supply, when the so-called 
 saturation current is arrived at, which is independent of the 
 applied potential difference. 
 
 As the P.D. is gradually raised, negative electrons are 
 discharged from the point and participate in carrying off the 
 current. The area around the point now becomes luminous, 
 which luminosity extends towards the plate with increasing 
 P,D., the discharge becoming a typical brush accompanied 
 by a slight crackling noise. At this point positive gas ions, 
 produced at the anode by detachment of an electron from a 
 gas molecule through collision with an electron travelling at 
 high speed, will also augment the current capacity of the 
 system. 
 
 The resistance of the air circuit now falls quite rapidly 
 owing to increase in conductivity by collision between 
 electrons and the gas molecules and the brush discharge is 
 converted into a high tension arc discharge. During the 
 high tension arc discharge, the cathode gets extremely 
 hot owing to bombardment by positive ions and the 
 thermionic emission of electrons as well as particles of 
 vaporised electrode material charged positively commences, 
 resulting in a still greater increase in conductivity ; the high 
 tension arc is therefore not stable but is transformed into the 
 more usual low tension arc. For the production of ozone 
 the electrically stable part of the discharge only, viz. the 
 
94 OZONE 
 
 non-luminous, the glow, and the brush discharge, come under 
 consideration, since, as we have already had occasion to 
 observe, the high thermal effects associated with both the 
 high and low tension arc discharges are more than sufficient 
 to mask any electronic formation of ozone. 
 
 We have noted that the transformation of the silent dis- 
 charge into the high tension arc discharge occurs after the 
 whole inter-electrode space has been filled with the discharge 
 glow, which makes its first appearance in the so-called 
 corona light. The nature of this luminescence is not clearly 
 understood ; that it is a function of the composition of the gas 
 is shown by the experiments of E. Riesenfeld (" Zeit. Elektro- 
 chem.," 17, 725, 1911), who noted that the discharge is pink 
 in nitrogen, blue in hydrogen, white in chlorine, " like the 
 combustion of iron wire in oxygen," and greenish-blue in 
 oxygen. 
 
 Sir J. J. Thomson (" Conduction of Electricity through 
 Gases," 1906, pp. 478-512) has shown that at the moment 
 when both anode and cathode glow make their appearance 
 there is a very great increase in conductivity of the gas 
 space, and advanced the hypothesis that just prior to the 
 appearance of the glow discharge the atoms have acquired 
 internal energy by collision with electrons and by absorption 
 of soft Eontgen rays, generated by collisions of electrons with 
 other atoms, until it has nearly approached the critical value 
 at which the atom becomes unstable and luminous; that 
 ionisation precedes the luminous discharge was clearly 
 indicated by D. Mackenzie (" Phys. Rev.," 5, 294, 1915). 
 
 According to K Nesturch (" Phil. Mag.," 30, 244, 1915) 
 there always exists a definite ratio between the amount of 
 
PRODUCTION BY SILENT ELECTEIC DISCHARGE 95 
 
 radiation and the number of gas ions formed by such 
 collision. 
 
 Sir J. J. Thomson and E. Threlfall ("Proc. Eoy. Soc.," 
 40, 340, 1886) clearly showed that ozone formation in the 
 silent discharge tube was associated with the production of a 
 luminous glow, whilst a similar conclusion was arrived at by 
 E. Warburg ("Ann. der Physik," 17, 1, 1905), who advanced 
 the hypothesis that ozone is only produced by electrons with 
 sufficient kinetic energy to cause themselves to become 
 luminous. 
 
 The view that the corona and brush discharges are at 
 least in part due to ionisation by collision is supported by a 
 series of experiments which have been made on the corona 
 "pressure" phenomenon by S. P. Farnweld, J. Kunz, and 
 especially Townsend and E. Warner (" Phys. Kev.," 8, 285, 
 1916). It is evident that if the molecules break up into ions 
 as a result of ionic collision an increase of pressure should 
 result. In an enclosed gas space subjected to the brush dis- 
 charge, this pressure increase has actually been noted, and 
 when corrected for the unavoidable temperature increase the 
 following relationship was found to hold good : 
 
 V* = vp, 
 
 where V is the applied voltage, i the corona current, v the 
 volume of the gas subjected to the silent discharge. 
 
 In oxygen, however, there is scarcely any increase in 
 pressure, due to the formation of ozone ; in other words, the 
 ozone formation is strictly proportional to the ionisation. 
 
 It would therefore appear that only a relatively small 
 portion of the discharge is effective in the production of 
 ozone and that the optimum results are to be obtained when 
 
yb OZONE 
 
 the ozoniser is so operated that the luminosity of the silent 
 discharge glow is at a maximum. 
 
 Many conflicting statements have been published relative 
 to the yield of ozone per kilowatt hour obtainable in ozonisers ; 
 on analysis these are found to be due to the fact that many 
 investigators have ignored the primary consideration affecting 
 the production, of ozone by this means, viz. the relationship 
 between ozone production per kilowatt hour and the concen- 
 tration of the ozone. It is evident that if a definite volume 
 of air be subjected to the silent electric discharge, the ozone 
 concentration in that air will rise to a certain definite value, 
 C , the "limiting " concentration. When this concentration 
 
 is reached, the rate of formation of ozone ~ will be equal to 
 
 its rate of destruction by thermal, catalytic and other effects. 
 Thus, in the enclosed volume of air, the apparent ozone pro- 
 duction per kilowatt hour will be zero whilst the actual pro- 
 duction will be -~. As a first approximation it may be 
 taken that the rate of catalytic deozonisation is proportional 
 to the concentration of ozone or ^ = KC ; thus the energy 
 
 required to produce strong concentrations of ozone in a 
 stream of gas will be a great deal more than is necessary to 
 produce the same amount of ozone in a very dilute state. 
 
 As clearly pointed out by Allmand (" The Principles of 
 Applied Electro-Chemistry "), the duty of an ozoniser cannot 
 be obtained without a knowledge of the following data : 
 
 (1) The limiting yield, or the yield per ampere hour at 
 zero concentration. 
 
 (2) The maximum concentration of ozone obtainable. 
 
PEODUCTION BY SILENT ELECTEIC DISCHAEGE 97 
 
 (3) The rate of variation of the yield with the concentra- 
 tion. 
 
 INFLUENCE OF CUEEENT ON YIELD OF OZONE. 
 
 Owing to the difficulties inherent in the construction of 
 high-tension (4000, 10,000 volts) generators the current for 
 ozone installations is usually derived from static transformers. 
 
 The static transformer may operate either on alternating 
 current or on direct current with a suitable interrupter in the 
 primary circuit. 
 
 (1) Static Transformer with Alternating Current Generator. 
 
 If a conductor such as a piece of wire describe simple 
 harmonic motion in front of the pole of a magnet as is ob- 
 tained in the rotation of an armature between the poles of a 
 magnet, a current varying in intensity from moment to 
 moment will be induced in the conductor. 
 
 Both E.M.F. and current time curves will follow those of 
 the sine or cosine curve. 
 
 Time 
 
 PlG. 7. 
 
 The E.M.F. at any time t being given by the relationship 
 E = Eo cos 
 
 the current in a similar manner by i = i, x sin -j~-. 
 
 1 
 
98 OZONE 
 
 If the circuit were entirely non-inductive, the volt-ampere 
 curves would naturally be superimposed, since at any time 
 the current flowing, would, in accordance with Ohm's law, 
 be strictly proportional to the applied E.M.F. In actual 
 practice self-induction is always present, being defined as the 
 value of the integral 
 
 L Hcoaedldl' 
 
 where e&dl dl' in the current circuit. 
 
 The product ~Li may be termed the electrical momentum 
 acquired by the current in the circuit, and Ohm's law has to be 
 modified to include the rate of change of electrical momentum 
 as well as the instantaneous current 
 
 or 
 
 di E. E 
 
 FT + ^ = =F- COS 
 
 L" L" T* 
 
 The solution of this equation is given by 
 
 E cos 
 
 ^-rrt \ 
 
 -j- - a) 
 
 VB- + OOu 
 
 , 
 
 where a = tan~ 
 
 There is therefore a lag between the current and E.M.F. 
 curves, and the maximum value of the current never exceeds 
 
 TM 
 
 where the expression under the square root 
 
 the " impedance " takes the place of E. 
 
PRODUCTION BY SILENT ELECTRIC DISCHARGE 99 
 
 Several generalisations which have an important bearing 
 on ozoniser designs follow from these considerations. 
 
 Firstly, large currents cannot be obtained in systems of 
 high inductances, and with increasing values of the periodicity 
 
 f J the inductance term becomes the only one of significance 
 in the resistance of circuit. 
 
 For high frequencies ^L will be large, consequently 
 augmenting tan a, making a the angle of lag approximately 
 
 i j 7T 
 
 equal to ^. 
 
 A 
 
 n, 
 
 E sin 
 
 or when cos -^- = 0, sin -^ = 1, the E.M.F. will therefore 
 
 be at a maximum when the current is zero and vice versa. 
 The Wattage consumption ~E> i will be equal to 
 
 E 2 cos n=p cos (^ aj 
 
 ==- = o when a is ? 
 
 or 
 
 V E 2 cos a 
 ===== for small values of a = 
 
 For E = o or a there is therefore no energy consumption, 
 whilst for some intermediate value there is a maximum energy 
 consumption. 
 
100 
 
 OZONE 
 
 For 
 
 E 
 
 to be a minimum 
 
 K 2 must equal L 2 or K = 
 
 If a condenser of capacity C be placed in circuit with the 
 secondary system we can in a similar manner obtain the ex- 
 pression for the relationship between the varying potential 
 difference and the charge on the plates. 
 
 / STT* 
 
 E (cos -JJT - a 
 
 By suitable adjustment of the condenser, i.e. making 
 
 C = 
 
 > we can g e ^ larger amounts through the circuit for 
 
 a given applied potential difference than if the circuit were 
 closed by a wire. 
 
 According to Kabakjian (" Phys. Eev.," 31, 117, 1910) the 
 limiting yield of ozone increases with decreasing capacity, 
 
 A 
 
 PIG. 8. 
 
 whilst the efficiency of ozone production at a definite con- 
 centration increases with decreasing capacity. 
 
 So far we have assumed that the resistance E of the cir- 
 
PEODUCTION BY SILENT ELECTRIC blSCHXBGE' 101 
 
 cuit is non-variant, but as we have had occasion to observe 
 the conductivity of the air gap in the discharge apparatus 
 varies with the current. Large currents cause the gap to 
 become more conducting, permitting under a constant applied 
 E.M.F. still higher currents to pass, ending finally in spark 
 and arc discharges. The sinuous character of the curve will 
 thus be altered, the maxima a> a being increased for this 
 reason to higher values b, b. 
 
 (2) Direct Current with Interrupter. 
 Small coils with magnetic or larger induction apparatus, 
 with mercury or Wehnelt type of make and break on the 
 primary, also yield a periodic current which, however, no 
 longer possesses the sinuous character of the alternating 
 current machine, but consists of a number of periodic current 
 makes and breaks as is depicted in the following curves : 
 
 FIG. 9. 
 
 It will be noted that under both conditions of operation 
 there exists a great danger of the spark discharge taking 
 
102 
 
 OZONE 
 
 place at the point of optimum current flow, which practically 
 coincides with the period of maximum conductivity of the 
 gas. The spark discharge itself is oscillatory in character 
 having a period T = 2-Tr^/LG and will possess a curve of the 
 following form : 
 
 FIG. 10. 
 
 According to the investigations of Kabakjian (" Phys. 
 Eev.," 31, 117, 1910) the brush discharge itself may under 
 conditions of high rates of discharge assume the oscillatory 
 character of the spark discharge. 
 
 (a) Influence on Voltage on Ozone Yield. Chassy (" Etude 
 sur TOzone, C.K.V.," 135, 1902) claimed that there was for a 
 fixed air gap in a given ozoniser a practically linear relation- 
 ship between the ozone yield per ampere hour and the 
 potential difference between the electrodes. Later experi- 
 ments have shown that Chassy's conclusions were not entirely 
 correct. E. Warburg ("Ann. der Physik," 13, 464, 1904) 
 showed that provided that the potential difference applied 
 was sufficient to maintain a uniform glow at the point of the 
 air gap, the yield of ozone was practically independent of the 
 voltage as is shown by the following figures : 
 
PEODUCTION BY SILENT ELECTEIC DISCHAEGE 
 
 103 
 
 Current 
 I x 10- 6 . 
 
 57 
 
 57-5 . 
 57-2 . 
 
 Voltage of 
 Point. 
 
 . 4,200 
 . 9,880 
 . 11,700 
 
 Oms. Ozone 
 per Coulomb. 
 
 0375 
 0386 
 0387 
 
 A. W. Gray ("Ann. der Physik," 13, 477, 1904), utilising a 
 standard Siemens ozoniser, likewise found that the yield per 
 coulomb was constant and independent of the voltage pro- 
 vided that uniform illumination was maintained in the dis- 
 charge space. 
 
 Kabakjian (" Phys. Rev.," V, 31, 17, 1910) found that the 
 ozone output per coulomb rapidly rose with the voltage until 
 a potential difference of 2,700 volts with a 1 mm. air gap and 
 3,200 volts with a 2 mm. air gap was reached after which no 
 further increase was noted. At these voltages presumably 
 "saturation" of the field with the brush discharge was just 
 effected. 
 
 Influence of Current Density. 
 
 With a constant regime established in the working of an 
 ozoniser the quantity of ozone produced per coulomb is 
 practically constant, for a point discharge on the other hand, 
 the yield per coulomb varies with the current flowing, as is 
 shown from the following figures obtained by Warburg: 
 
 Positive Point. 
 
 Negative Point and 
 Positive Plate. 
 
 Negative Point and 
 Positive Cylinder. 
 
 Voltage. 
 
 Current 
 1 x 10- 6 . 
 
 Grammes 
 O 3 per 
 Coulomb. 
 
 Current 
 1 x 10- s . 
 
 Grammes 
 O 3 per 
 Coulomb. 
 
 Current 
 1 x 10 - 6 . 
 
 Grammes 
 O s per 
 Coulomb. 
 
 8,420 
 10,400 
 12,000 
 
 28-8 
 57-2 
 94-2 
 
 0-0172 
 0-0600 
 0-0630 
 
 17-4 
 25-1 
 57-2 
 
 0-0489 
 0-0459 
 0-0375 
 
 29-1 
 94-2 
 
 0-0431 
 0-0386 
 0-0370 
 
104 OZONE 
 
 The variation in these figures is attributed by Warburg 
 to the alteration in the volume of the corona glow in the dis- 
 charge space under the varying voltages. In a Siemens type 
 of ozoniser operating under the optimum conditions, he ob- 
 tained a value of 0*260 gms. per coulomb, a figure confirmed 
 by A. W. Gray (" Ann. der Physik," 13, 477, 1904). 
 
 These figures show some light on the mechanism of ozone 
 formation. Since, as can be shown by electrolytic decom- 
 position or by measurement of the charge of the electron, 
 together with the number of molecules in a gram molecule, 
 96,540 coulombs are associated with one equivalent of a 
 substance, it necessarily follows that if the ozone were pro- 
 duced by some form of electrolytic action in the gas space, a 
 limit is set to the quantity of ozone obtained per coulomb. 
 
 We may, of course, make various assumptions as to the 
 magnitude of the electronic transfer associated with the for- 
 mation of a molecule of ozone, but it will be evident that the 
 maximum yield for the minimum transfer will be obtained 
 on the assumption of the following possible sequence of 
 
 reactions : 
 
 > 
 
 o 2 + 20 -> 6 + 6 
 o 2 + 6 -> o 3 
 
 i.e. one molecule of ozone would be formed for each electron 
 transferred. Forty-eight gms. of ozone should therefore be 
 formed for a current consumption of 96,540 coulombs, or 
 O'OOOS gms. per coulomb. It is evident that the quantities 
 of ozone actually produced per coulomb exceed the amount 
 some 520 times, even though these conditions represent those 
 most favourable to ozone formation by electronic transfer. 
 
PEODUCTION BY SILENT ELECTEIC DISCHAEGE 105 
 
 We are forced to the conclusion, assuming the accuracy of 
 Warburg and Gray's experimental data, that most of the 
 ozone is of secondary origin and is produced by collision 
 between electrons both primary and secondary and gas 
 molecules. 
 
 Kriiger and Moeller ("Nernst Festschrift," 240, 1912) 
 have suggested that one electron may liberate, in the case of 
 the positive point discharge, seventeen secondary electrons, 
 and for a silent discharge in metallic tubes 287 secondary elec- 
 trons, which would necessitate velocities produced by applied 
 voltages of approximately 100 and 50,000 volts respectively. 
 
 Influence of Wave Form. 
 
 The hypothesis advanced from the previous considerations 
 that ozone formation is produced by inter-molecular and 
 electronic collision, and is not a phase of gaseous electrolysis 
 between the electrodes, is supported by a consideration of the 
 effect of the wave form on the ozone yield. A. Vosmaer 
 ("Ozone," Constable, 1916, p. 70) states, "a very peaked 
 wave form would cause a greater distance between regular 
 working tension and ordinary maximum tension and thus 
 facilitate the brush discharge. On the other hand a flattened 
 curve would give more available energy in the domain of 
 working and would give a better output of ozone . . . there 
 is not so much difference in wave form to be of any import- 
 ance." It is clear that this investigator does not consider 
 wave form of great importance and his views have been sup- 
 ported by many other observers. Those, however, who have 
 had occasion to make use of the oscillograph and thus have 
 been able to plot the wave form with accuracy have noticed 
 
106 OZONE 
 
 that the wave form does have an important bearing on the 
 yield of ozone. Amongst the more important investigations 
 published, may be mentioned those of 0. Frohlich (" Elek- 
 trotech. Zeitschrift," 12, 340, 1901). E. Warburg and Leit- 
 hauser ("Ann. der Physik," 4, 28, 17, 1909), and G. Lechner 
 (" Zeit. Elektrochem.," 17, 414, 1911 ; 21, 309, 1915). 
 
 The following is a brief summary of the more important 
 conclusions : 
 
 The ozone yield per coulomb rises at first rapidly with the 
 periodicity of the alternating current and thereafter more 
 slowly. Up to 500 periods have actually been employed in 
 technical installations. It is evident from a consideration of 
 the form of the sine curve that an increasing periodicity in- 
 creases the steepness of the curve, i.e. high periodicity ensures 
 
 a large value for the rate of alteration of the current flow -r,. 
 
 Qjt 
 
 A high periodicity likewise lowers the minimum potential 
 difference to produce a silent discharge across a fixed inter- 
 polar space (E. Eiesenfeld, " Nernst Festschrift," 374, 1912). 
 More ozone is produced per coulomb with a periodically 
 broken direct current than with an alternating current of the 
 same current strength and periodicity. A glance at the 
 typical wave forms for these two types of current flow will 
 
 suffice to indicate that in the former case the -r. values are 
 
 at 
 
 much larger than in the symmetric alternating currents. A 
 further advantage of the direct current is that for the same 
 effective potential difference a larger current can be passed 
 through the silent discharge tube with a consequent increase 
 in ozone production. 
 
PRODUCTION BY SILENT ELECTRIC DISCHARGE 107 
 
 Puschin and Kauchtschev (" J. Euss. Phys. Chem. Soc.," 
 46, 576, 1914) have likewise shown that the yield of ozone 
 increases with the frequency, but that the optimum frequency 
 was dependent on the applied voltage as indicated by the 
 following figures : - 
 
 Periodicity. Applied Voltage. 
 
 1240 6500 
 
 950 7000 
 
 660 8000 
 
 For a constant air-flow an increase in periodicity above 
 these limits decreases the output of ozone, whilst an increas- 
 ing air-flow displaces the maximum towards increasing 
 frequencies. 
 
 In a general way it is not difficult to offer an explanation 
 of the increase in yield of ozone per coulomb with rapid 
 alteration of the current flow or increasing tension of a 
 current impulse, if we regard the formation of ozone due to 
 electronic and molecular collisions. 
 
 At any given instant we may regard the current flow as 
 constant ; then proceeding from the negative to the positive 
 electrode there will be a stream of electrons which, in their 
 passage through the gas space will ionise part of the gas 
 therein. The effect of the electron stream on the oxygen 
 fraction of the gas is three-fold : 
 
 (a) A splitting of the molecule into two neutral atoms by 
 direct impact 
 
 2 = + 0. 
 
 As we have seen it is by this disruption that ozone is chiefly 
 formed in ultra-violet light. 
 
 (6) An ionisation of the molecule or atom by impact 
 
108 OZONE 
 
 o, -> 6 2 + e 
 
 O ->6 + Q 
 
 the positive ions so formed which may be atomic or consist 
 of molecule clusters (we have already indicated that evidence 
 for clusters up to 6 in complexity is at hand, this probably 
 represents the extreme upper limit, as large clusters easily 
 break down again). The atoms or molecules with one or two 
 positive charges naturally proceed in the reverse direction to 
 the stream of electrons and by impact and combination with 
 them neutralisation to atoms and molecular groups is once 
 more effected 
 
 6 2 +20 = 2 . 
 
 (c) An ionisation of the molecule or atom by impact and 
 adherence of the electron. 
 
 The electron having spent most of its kinetic energy with 
 which it left the electrode or dielectric, may, on contact with 
 a neutral atom or molecule, not possess sufficient energy to 
 detach a valency electron from its orbit of rotation and may 
 actually adhere to the system it strikes forming a negatively 
 charged atom or molecule 
 
 2 + = 0' 2 
 + = 0'. 
 
 Oppositely charged atoms and molecules may then react to 
 form ozone 
 
 6 2 + 0' = 0,. 
 
 If the current flux be subject to violent changes then the 
 stream density of both electrons and gas molecules will not 
 be constant, but will proceed by a series of irregular spasmodic 
 bursts of varying velocities, thus greatly enhancing the pos- 
 sibilities of collision. 
 
PRODUCTION BY SILENT ELECTRIC DISCHARGE 
 
 109 
 
 Influence of Gas Flow and Composition. 
 
 We have already referred to the fact that the output of an 
 ozoniser is governed by the concentration of the ozone formed 
 during the discharge, since for high concentrations the rate 
 of deozonisation is increased and the apparent yield of ozone 
 per kw. hr. decreased. In the curves on next page are 
 shown the relationships obtained between yield and concen- 
 tration by utilising a standard Siemens and Halske industrial 
 ozoniser (Erlwin, " Zeit. f. Sauerstoff and Stickstoff," e, 143, 
 1911). 
 
 Warburg and Leithauser ("Drud. Ann.," 28, 1, 1908) 
 made a very extensive series of experiments on both glass 
 and metallic ozonisers to determine the influence of the ozone 
 concentration derived upon the yield. Their results are 
 tabulated in the following columns : 
 
 
 
 
 
 
 
 Gms. per 
 
 Limiting 
 
 
 
 
 
 
 
 Ampere 
 
 
 
 
 
 
 
 
 Hour for 
 
 
 
 Type of 
 Ozoniser. 
 
 Distance 
 between 
 Electrodes 
 
 foltage. 
 
 Period- 
 icity. 
 
 Amperes. 
 
 Cose. 
 
 Concentra- 
 tions of 
 Gms. per 
 Cubic 
 
 Yield 
 
 Concen- 
 tration 
 
 
 in Cms. 
 
 
 
 
 
 Metre. 
 
 Gms. j Amp. 
 
 in Gms. 
 
 
 
 
 
 
 
 
 Hr. 
 
 per Cubic 
 
 
 - 
 
 
 
 
 
 10 
 
 4 
 
 
 
 Glass 
 
 0-51 
 
 8,050 
 
 50 
 
 0-182 
 
 0-185 
 
 38-3 
 
 41-9 
 
 45-5 
 
 3-5 
 
 
 1-40 
 
 10,080 
 
 50 
 
 0-102 
 
 0-314 
 
 
 
 
 
 
 
 
 
 i 
 
 1-40 
 
 16,900 
 
 50 
 
 0-193 
 
 0-243 
 
 52-3 
 
 55-1 
 
 56-8 
 
 59-2 
 
 
 3-72 
 
 17,500 
 
 50 
 
 0-160 
 
 0-415 
 
 51-1 
 
 56-1 
 
 60-1 
 
 31-2 
 
 M tal 
 
 2-26 
 
 10,800 
 
 50 
 
 0-182 
 
 0-431 
 
 72-2 
 
 78-4 
 
 82-6 
 
 40-5 
 
 
 4-66 
 
 13,900 
 
 50 
 
 0-169 
 
 0-450 
 
 53-3 
 
 62-4 
 
 68-4 
 
 20-2 
 
 
 2-26 
 
 9,480 
 
 100 
 
 0-308 
 
 0-451 
 
 75-7 
 
 81-4 
 
 88-2 
 
 51-6 
 
 
 4-66 
 
 12,300 
 
 100 
 
 0-280 
 
 0-447 
 
 54-0 
 
 63-0 
 
 69-0 
 
 16-8 
 
 
 2-26 
 
 9,340 
 
 510 
 
 1-58 
 
 0-537 
 
 57-1 
 
 66-0 
 
 71-9 
 
 18-3 
 
 
 4-66 
 
 12,100 
 
 510 
 
 1-19 
 
 0-704 
 
 33-0 
 
 58-0 
 
 74-7 
 
 11-4 
 
110 
 
 OZONE 
 
 10 15 30 25 30 
 Cubic Metres per Hour 
 
 35 
 
 45 
 
 Air How Rate 
 
 10 15 20 25 30 
 
 Cubic Metres per Hour 
 
 FlG. 11. 
 
PRODUCTION BY SILENT ELECTRIC DISCHARGE 111 
 
 It will be noticed that the conditions most favourable for 
 the economic production of ozone obtain with low concentra- 
 tions of ozone or relatively high flow rate of air. High air 
 flow rates likewise serve to keep the ozoniser cool, an import- 
 tant consideration since the catalytic decomposition of ozone 
 is considerably accelerated by high temperatures. 
 
 For technical operations the air-flow rates are accordingly 
 adjusted as to give the minimum concentration of ozone 
 which will prove effective for the process under consideration ; 
 for under these conditions, although more energy must be 
 expended for pumping air, yet a very considerable economy 
 is effected in the ozone production. 
 
 The concentrations of ozone and the yields obtainable per 
 kw. hr. are higher in oxygen than in air, but the employment 
 of oxygen instead of air does not prove to be economical in 
 practice, although concentrations up to 150 gms. per cubic 
 metre, or nearly three times the maximum concentration 
 attainable with air, can be carried out. 
 
 The yield, however, does not increase indefinitely with 
 the oxygen pressure, thus, H. von Wartenberg and L. Max 
 (" Zeit. f. Elektrochem.," 14, 879, 1913), operating with an 
 ozoniser constructed to withstand high pressures with an 
 interpolar space of from 2 to 6 mm. and a current of 1 milli- 
 ampere at 23,000 volts and 50 periods, obtained the maximum 
 ozone concentration and ozone yield per watt-second at 0'5 to 
 1 atmosphere. Pressures up to 5 atmospheres were reached 
 during the course of their experiments. 
 
 It will be noted that 60 gms. of 3 in air, and 180 gms. 
 of ozone in oxygen per kw. hr. represent the best results 
 yet obtained with ozonisers operating under the optimum 
 
112 OZONE 
 
 conditions. Taking 34,000 calories as the heat of formation 
 of ozone, this represents a theoretical yield of 1'2 kgms. per 
 kw. hr., or industrial ozonisers have an efficiency of only 
 5 per cent, in air or 15 per cent, in oxygen. 
 
 Air suitable for ozonisation should be free from dust, 
 which favours the passage of sparks, and from certain gaseous 
 impurities such as oxides of nitrogen, chlorine, and more 
 especially water vapour. All three gases appear to exert a 
 distinct inhibiting effect on the formation of ozone, in addi- 
 tion to a deozonising action, which is especially marked in 
 the case of chlorine and nitrogen dioxide. The function of 
 these gases as catalytic deozonisers will be referred to later. 
 The inhibiting action, which is most marked in the case of 
 water vapour, has been attributed to the formation of mist, 
 which serve as nuclei for the condensation of the gas ions. 
 Under these conditions of condensation, the velocity of the 
 ions is naturally reduced and their power of ionising or 
 breaking down a molecule into atoms is correspondingly 
 lowered. Undried air at ordinary temperatures having a 
 water vapour pressure of ca. 7mm. H 2 has a limiting yield 
 of ozone which is only 60 to 70 per cent, of the air when 
 dry ; in the presence of moisture likewise the formation of 
 oxides of nitrogen, due to the thermal effects of sparking as 
 well as the possible interaction of ozone with some form of 
 active nitrogen produced in the spark discharge, is usually 
 occasioned. T. Lowry (" J.C.S.," 101, 1152, 1912), in an in- 
 teresting research on the effect of the silent and spark dis- 
 charges on air, showed that in dry air oxides of nitrogen were 
 not formed under the experimental conditions by passage 
 through the ozoniser or the spark discharge gaps. When 
 
PEODUCTION BY SILENT ELECTEIC DISCHAEGE 113 
 
 passed, however, through both in series, or when the air 
 currents subjected to each discharge were mixed, oxides of 
 nitrogen were produced. 
 
 Lowry came to the conclusion that in the spark discharge 
 an active variety of nitrogen was formed which was easily 
 oxidised by ozone. 
 
 DIELECTEIC MATEEIAL. 
 
 We have already noted that the yield of ozone per kw. 
 hr. at a definite concentration increases with the increase in 
 capacity of the ozoniser, but that the limiting yield decreases. 
 In addition it must be remembered that extreme variation in 
 the size of the air gap or interpolar free space is not permis- 
 sible, since too small a gap will permit the passage of sparks 
 and possible arcing with minute variations in the applied 
 voltage, whilst with a large interpolar free space the luminous 
 discharge, on which the formation of ozone appears to be 
 largely dependent, will not fill or " saturate " the field. It is 
 evident that the use of dielectric material other than air, by 
 which alterations in the capacity and interpolar distances 
 can readily be made, offers the designers of ozonisers a very 
 considerable latitude in these factors. 
 
 For the purposes of calculation we may take a simple 
 plate form of ozoniser and consider the effect of inserting a 
 plate of dielectric material in the air space between the two 
 metallic electrodes. 
 
 "T~ 
 i 
 
 FIG. 12. 
 8 
 
114 OZONE 
 
 If the two plates are charged with a quantity of electricity 
 of surface density <r, the attraction, at a point P situate in one 
 plate, by the other plate which is separated from it by the 
 interpolar distance (a) is 
 
 rrdro- cos 6 f^TrB tan fl<rQ sec 2 Odd cos 
 
 _ _ 
 
 " a 2 sec 2 B " , a 2 sec 2 
 
 = I - 27TO- sin 6 = 27Tcr. 
 
 J ir 
 
 -jT 
 
 At a point between the two plates the attraction due to each 
 plate is 2-7TO-, thus, with a positive charge on one plate and a 
 negative charge on the other, F = 47rcr. 
 
 If the potential difference be V, and the total charge Q, 
 
 then = 47T0- = -~z, A being the area of each plate, or the 
 d> A. 
 
 capacity =| = A-, 
 
 if we neglect the irregular distribution of the stream lines 
 near the edges of the plates. 
 
 On the insertion of a piece of plate glass of thickness b 
 
 between the plates, the equivalent air thickness is ^, where 
 
 H 
 
 K is the specific inductive capacity of the glass, hence the 
 
 O A 
 
 new capacity will be augmented to ~ = -- - -- j- , and 
 
 the interpolar distance of air space reduced to a - b. By 
 this means we have augmented the capacity and decreased 
 the interpolar distance of the ozoniser, and thus increased its 
 efficiency ; at the same time the tendency to sparking and 
 rupture of the gap between the electrodes is diminished, since 
 the mechanical force per unit area is likewise reduced, and 
 
PEODUCTION BY SILENT ELECTEIC DISCHAEGE 115 
 
 the possibility of the flow of currents of high densities natur- 
 ally excluded. 
 
 The following are the approximate values of the specific 
 inductive capacities of the more common dielectric materials, 
 dry air being taken as unity : 
 
 Material. K. 
 
 Paraffin wax 2-3 
 
 Eosin 2-6 
 
 Ebonite 3-2 
 
 Sulphur 3-8 
 
 Glass 6 to 7 
 
 The choice of dielectric material is naturally limited, since 
 it has to withstand both high temperatures and the destruc- 
 tive oxidising action of the ozone. Amongst those which 
 have been suggested may be mentioned : shellac, mica, 
 quartz, glass, and artificial insulators formed by condensation 
 of phenol and formaldehyde, the so-called Baekelites ; glass, 
 however, is the only material which has received extended 
 technical application. 
 
 The effect of insertion of a solid dielectric in the inter- 
 polar space is, however, not so simple as indicated by the 
 above considerations, since like other materials, not only are 
 they imperfect insulators, but many of them possess the in- 
 teresting property of acquiring residual charges. We may 
 regard the dielectric medium to consist of a number of con- 
 ducting particles embedded in an insulating material, the 
 fraction being smaller in the case of the more perfect insula- 
 tors. If this fraction be denoted by u, then the specific in- 
 ductive capacity can be calculated approximately from the 
 relationship : 
 
116 OZONE 
 
 1 + 
 
 K 
 
 1 - u 
 
 u 
 
 u being determined from the molecular specific volume -%. 
 
 When a strip of dielectric material is charged up to a 
 high potential, after discharge it will be found to acquire a 
 small charge on standing, which is often sufficiently great to 
 raise the potential of the dielectric up to 300 volts. This 
 property of acquiring a residual charge is only possessed by 
 those bodies which possess the property of exhibiting elastic 
 after-effects, it is never shown by simple substances, but only 
 by mixtures such as the glasses; thus xylene and paraffin 
 oil alone do not show this effect, but on mixing the two, the 
 residual charge is apparent. One of the constituents must 
 also possess a certain amount of electrical conductivity. 
 
 It is interesting to note that Kiesenfeld (" Zeit. Elektro- 
 chem.," 725, 1911) failed to obtain a brush discharge with 
 pure quartz glass, although such discharges are readily ob- 
 tained with all forms of glass which may contain quite large 
 percentages of silica, attributable to the slight electrical con- 
 ductivity of the glass, and its possession of a residual charge. 
 
 Two other important properties of dielectric materials 
 must be briefly alluded to, namely, the alteration in conduct- 
 ivity with elevation of the temperature, and the modification 
 which the dielectric undergoes when subjected to electrical 
 stress. 
 
 It is well known that the conductivity of glass at even 
 slightly elevated temperatures rapidly increases. Mond and 
 Langer and Huber have actually used solid glasses as 
 electrolytes at temperatures between 200 and 500 C. Local 
 
PRODUCTION BY SILENT ELECTEIC DISCHARGE 117 
 
 overheating at one point in the dielectric, due to a slight ir- 
 regularity in the current flow, will thus cause an increase of 
 conductivity at this point, with a corresponding augmentation 
 of the current. Fusion and finally perforation of the glass 
 results. It is for this reason that porcelains, which possess 
 temperature coefficients even higher than those of the glasses, 
 are unsuitable for dielectric material in ozonisers. 
 
 Kerr noted that the optical properties of dielectrics were 
 considerably modified by the application of electric stresses. 
 These modifications are influenced by the period of time for 
 which the stress has been applied, thus Fleming (see " Amer. 
 SuppL," 45, 1912) showed that the conductivity of a dielec- 
 tric varied with the frequency of the applied alternating 
 current, and Lipp (" Hochspanning Technik ") found that 
 the applied voltages necessary for perforation of thin sheets 
 of dielectric material varied with the period to which the 
 dielectric material had been subjected to the electrical stress. 
 The perforating potentials for the usual dielectric materials 
 are approximately as follows : 
 
 Material Perforating Potentials 
 
 in Kilovolts per Cm. 
 
 Mica 600 to 750 
 
 Micanite 400 500 
 
 Porcelain 100 
 
 Glass 75 300 
 
 Lead glass 1000 
 
 Air (" Amaduzzi N. Cimenta," 3, 51, 1912) 
 
 7000 volts per 1-5 cm. 
 97,000 13-5 
 100,000 14 
 
 The potential difference necessary for discharge between 
 two conductors in air varies with the size and shape of the 
 
118 OZONE 
 
 conductor, point discharge taking place much more readily 
 than discharge across plane surfaces. The following figures 
 (Abraham and Villard, "Physical Constants," 1910) indicate 
 the potential difference required to cause a 30 mm. spark to 
 strike between two spherical electrodes of varying radius : 
 
 Eadius in Mm. Potential Difference. 
 
 a (plane) 82,700 
 
 300 85,100 
 
 100 84,400 
 
 50 90,000 
 
 (point) 30,500 
 
 The relationship between the potential difference and the 
 striking distance, is also not a simple one, as indicated from 
 the experimentally derived figures for spherical electrodes 
 1 cm. in radius 
 
 Distance in Cms. Kilovolts. 
 
 0-06 27 
 
 O'lO 41 
 
 0-40 13-0 
 
 0-50 15-6 
 
 For very small interpolar distances, say 1 to 50 pp, the volt- 
 age necessary is independent of the distance and equal to 
 about 350 volts (E. Williams, "Phys. Chem.," 31, 216, 1910). 
 0. Hoveda ("Phys. Bev.," 34, 25, 1912) gives the follow- 
 ing relationship for point to plane discharges : 
 
 c+ 
 
 where a, b, c, are constants, and D is the interpolar distance. 
 The use of minute points corrugated or roughened metal 
 electrodes in industrial ozonisers is very frequent, since, as 
 we have seen, the presence of points facilitates the electrical 
 discharge, 
 
PBODUCTION BY SILENT ELECTEIC DISCHARGE 119 
 
 L. Decombe (" Jour, de Physique," 2, 181, 1912) has at- 
 tempted to calculate the energy dissipated in a condenser 
 when connected to an alternating current source ; he shows 
 that the energy absorbed, i.e. V8g (where V is the applied 
 E.M.F. and q the charge on the condenser), can be expressed 
 in the form : 
 
 where m is the polarisation and K a constant, this is equi- 
 valent to the dissipated energy : 
 
 or the dissipated energy is proportional to the square of the 
 polarisation current and independent of the periodicity. 
 V. Ehrlich and R Euss (" Zeit. Elektrochem.," rp, 330, 1913), 
 as a result of an investigation on the measurement of the 
 electrical quantities in a Siemens ozone tube, showed that 
 the ionisation or polarisation current and applied potential 
 were always in phase, and that the potential difference 
 across the gas gap was a direct measure of the energy. 
 
 Chassy (" Jour, de Physique," 2, 876, 1912), on the other 
 hand, showed that the energy absorbed per second by a gas 
 under alternating fields, varied as the charge Q and not as 
 the square of the charge, as in metallic conductors. 
 
 An increase in conductivity of the solid dielectric is also 
 to be expected from its exposure to the ultra-violet light 
 generated by the brush discharge in the interpolar air gap. 
 A. Goldmann ("Ann. der Physik," 36, 3584, 1911) has 
 shown that solid dielectrics exhibit both an increase in con- 
 ductivity and a negative discharge of electrons similar to the 
 Hallwachs effect in metals when subjected to ultra-violet 
 
120 OZONE 
 
 irradiation ; the conductivity of sulphur is said to increase 
 1500 times when thus illuminated. 
 
 Types of Industrial Ozonisers. 
 
 Industrial ozonisers may be grouped into two distinct 
 types : those in which the silent discharge passes across the 
 air gap without the interposition of any solid dielectric, and 
 those in which one or both of the electrodes are protected by 
 some suitable dielectric material, usually glass. 
 
 Non-dielectric Ozonisers. 
 
 Several attempts have been made to produce an ozoniser 
 without any dielectric and although large units on various 
 systems have been constructed from time to time, their 
 efficiency has usually been extremely low and at the present 
 time all industrial ozonisers contain dielectric material. 
 
 Schneller, Wisse and Sleen, in 1894, were the first to 
 construct large ozonisers without a dielectric. One dis- 
 charging surface consisted of a sheet of platinum gauze to 
 provide a great number of small points some 30 mm. from 
 the other surface formed of perforated metal sheet, the air 
 current being forced through the perforations in the sheet. 
 The two electrodes were cylindrical in shape and were 
 mounted in glass tubes. The operating voltage was at first 
 15,000, which was subsequently raised to 50,000. To avoid 
 the formation of an arc discharge, which in the absence of 
 any dielectric between the electrodes would be attended with 
 disastrous effects, a high resistance was inserted in series 
 with the ozoniser. 
 
 These investigators found much difficulty in the con- 
 struction of a resistance suitable for the purpose. It was 
 
PEODUCTION BY SILENT ELECTKIC DISCHAKGE 121 
 
 necessary to obtain a suitable resistance of 1'5 megohms, 
 capable of carrying O'Ol ampere ; moist unglazed porcelain 
 and glass tubes containing 80 per cent, glycerine were found 
 most suitable. Vosmaer (" Ozone," p. 94) showed that 
 sparking and arcing could not be avoided even with this 
 artifice, and that the external resistance served to increase 
 the capacity of the circuit rather than the resistance. 
 
 Slate was found to be more suitable than either glycerine 
 or moist unglazed porcelain, which rapidly lost its humidity 
 and suffered an increase in resistance. 
 
 Patin's ozoniser followed a similar construction ; re- 
 frigeration of the air prior to ozonisation was employed to 
 increase the yield, and a number of small metallic perforated 
 prisms enclosed in a single unit comprised the electrodes in 
 lieu of the perforated plates and gauzes in Schneller's ap- 
 paratus. 
 
 Various improvements in design of this type of ozoniser 
 were introduced by Tindal in 1894 and more especially by 
 De Frise in 1904. De Frise's plant was actually employed 
 for a short period in the sterilisation of water on a large 
 scale at the Saint-Maur Water Works at Paris, but the 
 dielectric Siemens-Halske ozoniser was subsequently in- 
 stalled and adopted as proving itself more economical in 
 operation. 
 
 Tindal employed Schneller's method of augmenting the 
 capacity of the system by the insertion of liquid resistances 
 in series with the ozoniser. The ozoniser consisted essenti- 
 ally of a system of compartments separated by perforated 
 metal plates, containing alternately a set of electrodes and a 
 water-cooling device. 
 
122 
 
 OZONE 
 
 The perforated metal plates attached to the cooling tubes 
 served as one set of electrodes of the ozoniser and fine gauze 
 as the other set. 
 
 Fio. 13. 
 
 The water was maintained in active circulation to aug- 
 ment the cooling. 
 
 De Frise likewise adopted Tindal's arrangement of alter- 
 nating ozonisation and refrigeration, but adopted a different 
 arrangement for the distribution of the electrodes. 
 
 These consisted essentially of a series of crescent-shaped 
 discs mounted in parallel each with its own liquid inductive 
 capacity in a water-cooled metallic trough which served as 
 the other electrode. 
 
 FIG. 14, 
 
PEODUCTION BY SILENT ELECTEIC DISCHAEGE 
 
 123 
 
 Each disc was furnished with a number of minute points to 
 facilitate the discharge. 
 
 These ozonisers were in successful operation with voltages 
 up to 10,000, although an applied voltage of from 15,000 to 
 20,000 was usually employed. 
 
 It possessed distinct advantages over Tindal's apparatus 
 in that the distance between the electrodes could be reduced 
 very considerably without risk of arcing, thus increasing the 
 efficiency of the apparatus. 
 
 Various types of mechanical ozonisers operating without 
 the interposition of any dielectric have been constructed from 
 time to time, the most successful being that of Otto. 
 
 FIG. 15. 
 
 The frequent occurrence of arcing in non-dielectric 
 ozonisers as well as the necessity for obtaining a very small 
 polar distance between the electrodes for efficient working, 
 led Otto to construct an ozoniser in which by the rotation of 
 one electrode the polar distance was always varying, so that 
 if for any accidental cause an arc should be formed at one 
 
124 OZONE 
 
 point, it would speedily be broken again by the subsequent 
 increase in the arc gap. A diagram of an improved form of 
 Otto ozoniser is shown above. One electrode, which is fixed, 
 consists essentially of an aluminium disc, studded with a 
 great number of small points, the rotor being a metallic disc 
 segmented with insulating material. Up to 80,000 volts have 
 been employed on these machines. 
 
 None of these ozonisers have proved sufficiently economi- 
 cal or reliable for industrial operations in which, at the 
 present time, ozonisers containing one or more dielectrics 
 are practically universally employed. 
 
 Ozonisers Containing a Dielectric. 
 
 Tubular Ozonisers. The present construction of service- 
 able apparatus for the production of ozone by means of the 
 silent discharge has developed from the simplest forms of 
 ozone tubes constructed by von Siemens, in 1857, in Germany, 
 by Brodie in England, and Berth elot in France. 
 
 Siemens' first ozoniser consisted essentially of two coaxial 
 glass tubes, the outer coated externally and the inner intern- 
 ally with tin-foil, the air being passed through the annular 
 space. Brodie substituted water as electrode material in the 
 place of tin-foil, and Berthelot used sulphuric acid. 
 
 These ozonisers thus contained two dielectric plates cover- 
 ing each electrode, and in practice gave serviceable and 
 uniform results. 
 
 Brodie's and Berthelot's system gives somewhat better 
 results than that of Siemens', since it is a simple matter in 
 these forms to arrange for efficient cooling of the electrodes 
 and interpolar space. When oxygen is employed instead of 
 
PKODUCTION BY SILENT ELECTEIC DISCHAEGE 125 
 
 air, a 10 per cent, ozone concentration can easily be obtained 
 at room temperature and over 20 per cent, at -25 C. 
 
 Yet another type of apparatus was introduced by Dr. 
 Oudin and Andreoli in 1893. As one electrode a simple 
 form of glass vacuum tube was used, electrical connection 
 being furnished by means of a sealed-in platinum wire, 
 which in Andreoli's apparatus ran through the whole length 
 of the tube. The second electrode consisted of a series of 
 equally spaced indented steel rings, surrounding the vacuum 
 tube, or a copper spiral coiled in the form of a helix round 
 the tube ; the whole unit being inserted in a glass tube 
 through which the current of air was passed. Gaiffe and 
 E. Chatelain at a later date modified Oudin and Andreoli's 
 apparatus by substituting a second annular vacuum tube in 
 place of the steel or copper electrode by the former investi- 
 gators. 
 
 Small types of Oudin and Andreoli's machines were at 
 one time popular for medical work but have not been 
 developed for industrial purposes. 
 
 The Siemens type of ozoniser was developed by Froh- 
 lich and Erlwein, of Siemens and Halske, to its present form, 
 which has proved to be eminently suitable for industrial 
 purposes. 
 
 The tubular form of the earlier form of apparatus was 
 preserved, but various modifications were made in the dis- 
 position of the electrodes. 
 
 For the internal tin-foil coated glass tube, a cylindrical 
 aluminium tube was substituted, maintained in position in 
 the outer tube (of glass), and separated from it by T5 mm. 
 by means of three ebonite spring plungers. As the other 
 
126 
 
 OZONE 
 
 electrode, Brodie's idea of using water was adopted and pro- 
 vision was made for circulation to ensure cooling, the cast- 
 iron frame containing the ozoniser and water cooling being 
 carefully earthed. The smallest technical unit contains two 
 ozone tubes, the largest eight in the same water-cooling 
 frame, which is provided with a glass inspection plate. 
 The operating voltage varies between 4000 and 10,000 
 volts. 
 
 Ozone 
 
 \tuminium Cylinder 
 G/ass Tube 
 
 FIG. 16. 
 
 -<. 
 
 We have already discussed the results obtained by 
 Erlwein with this type of apparatus and will refer to the 
 technical applications in a subsequent section. 
 
 A similar type of apparatus has been developed by the 
 General Electric Company, but the somewhat cheaper and 
 equally efficient enamelled iron has been substituted for the 
 aluminium. 
 
 The Westinghouse Company in the Gerard ozoniser sub- 
 
PRODUCTION BY SILENT ELECTRIC DISCHARGE 
 
 127 
 
 stitute oil for water as the external electrode, and utilise, as 
 in the early Brodie tube, a double dielectric system. 
 
 According to the investigations of Vosmaer, the economi- 
 cal production of high concentrations of ozone is realisable 
 in this type of apparatus. 
 
 Various other forms of tubular ozonisers have been the 
 subject of patent literature, a few of which, such as the 
 
 
 M Air 
 
 
 
 g Ozonised 
 
 
 
 \ 
 
 Jf 
 
 
 
 
 
 
 Oil 
 
 Oft 
 
 
 J. 
 
 
 
 
 k 
 
 /V- 
 
 y 
 / 
 
 
 0/v 
 
 Metal 
 
 FIG. 17. 
 
 Elworthy, Yarnold and Gaiffe, have been sporadically de- 
 veloped for a short time by small companies, only to sink 
 again into oblivion. In Europe the only representation of 
 this type of ozoniser which may be said to have established 
 its footing industrially is that of Siemens and Halske. 
 
 Plate Type Ozonisers. In the tubular type of ozoniser, 
 owing to the disposition of the electrodes, it would appear 
 
128 OZONE 
 
 that the use of cooling water is essential in order to keep 
 the interpolar space and the inner electrode relatively cool, 
 which, as we have seen, is one of the most important factors 
 in the economic production of ozone. Since the thermal 
 radiation obtainable from two parallel plates is much greater 
 than in the tubular form, where the radiation is practically 
 confined to the interpolar area and to the external surface of 
 one electrode, the provision of cooling water is not essential 
 for efficient operation. Although better results are obtain- 
 able in well-designed ozonisers, where water cooling is used 
 in addition to the cooling effected by the air passage, yet 
 with air-cooled plate ozonisers of the proper design and with 
 the enhanced capital and running costs entailed in water- 
 cooling devices, the cost of ozone production by either 
 method is approximately equal. Nevertheless, in those cases, 
 usually exceptional, where high ozone concentrations or low 
 air current velocities are required, provision for water would 
 appear desirable. Modern plate form ozonisers have con- 
 sequently developed on these two distinct lines, those in 
 which air cooling only is utilised and those in which supple- 
 mentary water cooling is made use of. 
 
 Air-cooled Plate Ozonisers. The earlier ozonisers of 
 this type, such as those of Villon and Genin of Prepoignot, 
 and an experimental one of Otto's, were not a success, since it 
 was found impossible to prevent a very considerable rise in 
 temperature during continuous operation, resulting in a 
 serious loss of efficiency. 
 
 The first technical ozoniser which showed promise was 
 that of Andreoli, and possessed the great merits of simplicity 
 of construction and uniformity in operation. 
 
PRODUCTION BY SILENT ELECTRIC DISCHARGE 129 
 
 Andreoli's ozoniser consisted of a series of serrated alu- 
 minium plates separated from each other by a sheet of glass, 
 ca. 2 mm. thick, and an air gap. Each plate had an area of 
 3O5 x 30 '5 cms., and possessed 17,760 points formed by in- 
 dentation of the serrations. These units were mounted in a 
 wooden box ; provision was made for possible expansion and 
 contraction, and various artifices were devised to ensure the 
 plates being inserted quite parallel with each other. 
 
 With an eight-plate ozoniser, operating with a voltage as 
 low as 3,300 volts, an energy consumption of 550 watts could 
 easily be maintained, producing ca. 60 to 100 gms. of ozone 
 per kw. hr., naturally at very low concentrations. In his 
 later models Andreoli likewise introduced water cooling, and 
 could consequently elevate to applied voltage from 3,300 to 
 10,000, without any undue rise in temperature. 
 
 Experimental ozonisers of similar construction have been 
 designed by Vohr and Vosmaer but do not appear to have 
 been developed for industrial purposes. The only air-cooled 
 type plate ozoniser which appears to have outgrown the ex- 
 perimental state, and to be actually employed in the various 
 industries, is that of the Ozonair Company. 
 
 This ozoniser is extremely simple in construction and 
 efficient in operation. The electrodic discharge through the 
 glass dielectric plates and air gap is facilitated by employing 
 flat sheets of metallic aluminium alloy gauze as electrode ma- 
 terial ; this possesses the dual advantage of an even distribution 
 of points over the whole electrode area, and of being practi- 
 cally resistant to corrosion or tarnishing. Cooling is effected 
 by the air current and concentrations up to 3, and, for 
 
 short periods, even 4 gms. of ozone per cubic metre may be 
 
 9 
 
130 
 
 OZONE 
 
 obtained under conditions of actual operation without undue 
 elevation of the temperature. The operating voltage is usu- 
 ally 5000 at periodicities varying from 25 to 100. Tests have 
 shown that at a concentration of 2 gms. per cubic metre the 
 output may exceed 100 gms. per kw. hr. 
 
 Aluminium 
 
 Glass. 
 
 FIG. 18. 
 
 Water-Cooled Plate Ozonisers. This type of ozoniser 
 was developed chiefly by the early investigations of Otto, 
 and subsequently by those of Marmier and Abraham ; as a 
 result the Otto Marmier Abraham Ozoniser was constructed, 
 which has found a by no means insignificant number of in- 
 dustrial applications, especially in France. 
 
 The unit consists of a pair of hollow r disc electrodes with 
 perfectly plain faces which are protected by plates of the 
 dielectric, in this case glass, some 2 mm. thick. The inter- 
 polar air space through which the air flows in a radial 
 direction varies from 1 '3 to 1'8 mm. in width. Each electrode 
 
PEODTJCTION BY SILENT ELECTEIC DISCHAEGE 
 
 131 
 
 is cooled by running water, a broken fall providing against 
 short circuiting through the earth. 
 
 Air 
 
 FIG. 19. 
 
 The outer casing is made of earthenware, the air being 
 forced in at one end and emerging after ozonisation at the 
 other. 
 
 The ozoniser operates successfully at a voltage of 30,000 
 to 40,000 volts, although voltages much lower than this can 
 naturally be employed, 12,000 being quite normal. 
 
132 OZONE 
 
 Otto introduced for these machines a simple form of 
 electric safety valve consisting of two horn-shaped electrodes, 
 separated by a variable air gap and placed in series with the 
 ozoniser. 
 
 An accidental rise in the operating voltage would cause a 
 discharge to take place across the air gap when suitably ad- 
 justed, and thus obviate any break-down in the dielectric 
 plates of the ozoniser. 
 
 Labille suggested the use of mica as a dielectric in place 
 of glass plate in ozonisers of this type, but in practice, owing 
 to the disintegration of this laminated material, unfavourable 
 results were obtained. 
 
CHAPTEK VIII. 
 
 THE CATALYTIC DECOMPOSITION OF OZONE. 
 
 SINCE the quantity of ozone in equilibrium with atmospheric 
 oxygen at ordinary temperatures is, as we have seen, almost 
 vanishingly small, it follows that in ozonised oxygen or 
 ozonised air conditions of unstable equilibrium obtain, and 
 the apparent stability of the ozone is due to the fact that the 
 equilibrium is " frozen," or the rate of decomposition of ozone 
 at these temperatures is almost negligible. The rate of de- 
 composition of the ozone in excess of the minute amount 
 present at the true equilibrium can be accelerated in a variety 
 of ways, such as by the addition of catalytic materials, either 
 solid, liquid, or gaseous, by photo-chemical action, or by 
 purely thermal methods, by slightly elevating the temperature. 
 
 The rate of decomposition of ozone has been the subject 
 of many investigations. Shenstone in 1897 considered dry 
 ozone to be extremely unstable, and to undergo decomposi- 
 tion with extreme rapidity. At a later period, H. E. Arm- 
 strong showed that, in the absence of oxides of nitrogen, the 
 rate of decomposition was sensibly lessened. 
 
 Jahn (" Zeit. Anorg. Chem.," 48, 260, 1906), at Nernst's 
 suggestion, conducted a series of experiments on the rate 
 of decomposition with a view to elucidating the mechan- 
 ism of decomposition. The ozone molecule may undergo 
 
 (133) 
 
134 OZONE 
 
 decomposition in a variety of ways, as indicated in the fol- 
 lowing equations : 
 
 (i) 20 3 = 30 2 
 
 (ii) 3 = 2 + 
 
 (iii) O 3 + = 20 2 . 
 
 Jahn's experimental results appeared to indicate that the rate 
 of decomposition could be formulated in the expression 
 
 dc(0 3 ] -frCCO,) 2 
 dt C(O 2 )' 
 
 He consequently argued that the decomposition followed the 
 course indicated by equations (ii) and (iii) on the assumption 
 that the first reaction was rapid and reversible 
 
 3 ^ 2 + 0, 
 
 and the second slow and irreversible 
 3 + - 20 2 . 
 
 Perman and Greaves ("Proc. Eoy. Soc.," 4, 807, 353, 1908) 
 likewise stated that the rate of decomposition was inversely 
 proportional to the concentration of the oxygen 
 
 K 
 
 = 
 dt = C(0 2 )' 
 
 Chapman and Clark (" Trans. Chem. Soc.," 93, 1638, 1908) 
 and Chapman and Jones (" Trans.," 254, 2463, 1910) subjected 
 the whole matter to an exhaustive examination ; they showed 
 that Jahn's and Perman' s interpretation was not correct and 
 that very serious errors due to the catalytic effect of the sur- 
 face of the glass vessel vitiated the accuracy of their results. 
 These investigators found that the rate of decomposition was 
 proportional to the square of the concentration of the ozone 
 and independent of the oxygen pressure ; 
 
THE CATALYTIC DECOMPOSITION OF OZONE 135 
 
 ^ = KC(0 3 ) 2 . 
 
 E. Weigert (" Zeit. Phys. Chem.," 80, 78, 1912), however, 
 obtained values corresponding to an order of decomposition 
 somewhat exceeding two in the dark. 
 
 The rate of decomposition is greatly accelerated by rise 
 in temperature, being almost instantaneous at 270 C. and 
 quite rapid at 100 C. This fact, namely, that the equilibrium 
 is not to be regarded as " frozen " until room temperatures 
 are arrived at is, as we have seen, the factor militating against 
 the successful thermal production of ozone in contradistinc- 
 tion to nitric oxide where the equilibrium is practically 
 " frozen " at 600 C. 
 
 The decomposition may likewise be accelerated by solid 
 catalytic agents such as the following : 
 
 Ag, Cu, Co, Ni, Cr 2 3 , Pb 3 4 , V 2 5 , Mn0 2 , Ti0 2 , Th0 2 , 
 
 Ce0 2 , U 3 8 , W 2 3 , BaO, CaO, Hg, Ni, Pt, Pd, V, 
 and powdered glass. 
 
 The catalytic activity of platinum black is extremely 
 great. Mulder and v. d. Meulen (" Eec. Trans. Chem. Pays.- 
 Bas.," i, 167) and Warburg (" Berl. Akad., Ber.," i, 1900, 1176, 
 1901) noted the rapidity of decomposition of ozone rich gas 
 when passed over cold or warm platinum black, whilst Elster 
 and Geitel (" Ann. der Physik," 2, 39, 321, 1890) and Well and 
 Kopp (" Jahresber. Chem.," 270, 1889,322, 1890) observed 
 the formation of ozone by passing oxygen over hot platinum, 
 indicating the reversibility of the catalytic activity of the 
 platinum. 
 
 In many cases the activity of these catalytic materials 
 
136 OZONE 
 
 can be attributed to the formation and subsequent decom- 
 position of an unstable oxide or peroxide, e.g. 
 
 2Ag + 3 -> Ag 2 + O 2 - Ag. 
 
 Manchot ("Ber.," 39, 1510, 1906; 40, 2891, 1907; and 42, 
 3948, 1908), as a result of a series of experiments, obtained 
 some interesting results with silver as a catalytic agent. He 
 noted that pure silver was relatively stable in the presence 
 of ozone, it being necessary to warm the metal up to 24 C. 
 before decomposition of the ozone and formation of silver 
 oxide occurred. 
 
 A trace of iron, usually obtained from the emery powder 
 employed for cleaning the silver, serves as an excellent 
 catalyst promoter, silver containing but a minute trace of 
 iron reacts already at normal temperatures, and less than 
 O'Ol per cent, of ozone can be detected by this means. Man- 
 chot states that silver thus prepared is even more sensitive 
 than alcoholic tetramethyl base paper and that the ozone 
 present in hot flames can easily be detected. 
 
 J. Strutt ("Proc. Koy. Soc.," 87, 302, 1912) examined 
 Chapman and Jones' results from a statistical point of view. 
 He showed that in the case of catalytic decomposition at a 
 gas-solid surface, the rate of decomposition depends on the 
 number of collisions of a gas molecule with the surface neces- 
 sary to effect the rupture of the molecule or reaction with 
 the surface. 
 
 Calculating the number of collisions necessary to effect 
 the decomposition of ozone in the presence of metallic silver 
 from the minimum area of silver necessary to effect such 
 change, he showed that only 1*6 collisions of an ozone mole- 
 cule with the silver was necessary, or practically every col- 
 
THE CATALYTIC DECOMPOSITION OF OZONE 137 
 
 lision was effective. In the absence of anj' surface, i.e. in free 
 space, Strutt concluded that at 100 C. two molecules of ozone 
 
 must collide 6 x 10 11 times before a favourable collision re- 
 sulted. 
 
 The effect of various gases on the rate of decomposition 
 of ozone was likewise investigated by Chapman and Jones, 
 who showed that oxygen, nitrogen and carbon dioxide had 
 no effect, whilst nitrogen dioxide and chlorine accelerated 
 the rate of decomposition. The influence of water vapour 
 was not very marked. Shenstone (" Trans. Chem. Soc.," 71, 
 47, 1897) stated that water vapour did not retard the forma- 
 tion of ozone ; but, as pointed out by Armstrong, Shenstone 
 probably included any nitrogen dioxide formed at the same 
 time in his ozone estimations. Warburg and Leithauser 
 ("Ann. der Physik," IV, 20, 757, 1906) found that the for- 
 mation of ozone was retarded by water vapour whilst the rate 
 of decomposition was not affected. The specific influence of 
 water vapour in accelerating the decomposition of ozone was, 
 however, noted by Warburg (" Sitzungs K Akad. Wiss.," 
 Berlin, 644, 1913) in the course of his investigations on the 
 photo-catalytic decomposition of ozone. 
 
 It may be concluded from these experiments, as well as 
 from the somewhat scanty data collected on the rate of 
 decomposition of ozone in solutions, that water vapour exerts 
 a slight yet distinct catalytic action. This is only to be ex- 
 pected if it can be assumed that ozone is slightly acidic, since 
 when passed into strong alkalis it forms the somewhat un- 
 stable coloured ozonates MH0 4 . In this case (see Chapman 
 and Jones, "Trans. Chem. Soc.," 208, 1811, 1911) an equili- 
 brium is probably set up represented by the equation 
 
138 OZONE 
 
 20H' + 20 3 2 2H0 4 ' 
 2H0 4 ' -> H 2 + 2 + 20. 
 
 If the rate of decomposition in the presence of moisture is 
 thus accelerated by the HO/ ion then the observed rate of 
 decomposition will be 
 
 2^ = K(0 3 ) + K'C(0 3 ) 2 (OH) 2 , 
 
 where C(OH) is very small and K' relatively small compared 
 toK. 
 
 Eoth (" Monatsheft," 34, 665, 1913) investigated the de- 
 composition of ozone in acid solutions ; he showed that the 
 rate of decomposition in strong acids was nearly bimolecular 
 and in very weak acids practically monomolecular, the rate 
 for any acid strength being determined by the equation 
 
 = KC(0 3 ) 2 + K'C(0 3 ), 
 
 by suitable choice of the values from K and K', the velocity 
 coefficients. 
 
 The catalytic activity of light in decomposing ozone was 
 first studied by Eegener (" Ann. der Physik," 20, 1033, 1906), 
 who showed that light of a certain wave length in the ultra- 
 violet portion of the spectrum, viz. in the region 230 to 290 fifM, 
 exerted a very marked deozonising action, and we have 
 already discussed the interesting fact that light of shorter 
 wave length exerts an ozonising action. 
 
 The reaction of kinetics of this reaction was investigated 
 by von Bahr (" Ann. der Physik," 4, 33, 598, 1910) and especi- 
 ally by Weigert (" Zeit. Phys. Chem.,' J 80, 78, 1912), who 
 showed that with complete absorption of light the reaction 
 velocity of decomposition was proportional to the ozone con- 
 
THE CATALYTIC DECOMPOSITION OF OZONE 139 
 
 centrations. On the other hand, under conditions of homo- 
 geneous illumination, when the emergent and entrant beams 
 are equally intense, conditions obtaining approximately in 
 very thin gas films, the reaction velocity was found to be pro- 
 portional to the square of the ozone concentration. From 
 Weigert's data it may be calculated that in his experiments, 
 approximately 100 molecules of ozone were decomposed per 
 quantum of light energy absorbed. We have already noted 
 that the magnitude of the quantum hv necessary to effect 
 any given photo-chemical action increases as the required 
 energy increases, and consequently the photo-chemical activity 
 of light is greatest as we approach the extreme ultra-violet. 
 In the case of the formation of ozone from oxygen we have 
 already discussed the effect of quanta of various magnitudes 
 on both the oxygen atom and molecule; we would expect 
 that the magnitude of the quantum necessary to detach an 
 oxygen atom from an ozone molecule would be very small, on 
 account of its instability ; and again that the period of natural 
 vibration of an ozone molecule, which determines the absorb- 
 tive power for light of a definite wave length, would be larger 
 than that for the smaller oxygen molecule or atom. Both 
 these expectations are fulfilled in the experimental results 
 since deozonising light has a longer wave length than that 
 effective in ozonisation. Photo-chemical equivalence, how- 
 ever, is not obtained as in the case of ozonisation. The 
 absorption of a quantum of light energy by the already ex- 
 tremely unstable ozone molecule causes it to explode with 
 considerable violence, and, as we have seen, the energy liber- 
 ated during the explosion is able to cause the primary and 
 secondary decomposition of over 100 other molecules before 
 
140 OZONE 
 
 the energy is dissipated into the surrounding medium. M. 
 Saltmarsh (" Proc. Phys. Soc.," 27, 357, 1915) regards the 
 ultra-violet light in deozonisation as the source of nuclei in 
 the ozonised oxygen which serve as centres of decomposition. 
 Similar results were obtained by E. Warburg (" Berlin Akad. 
 Sitzungsber.," 2, 644, 1913). He obtained values for the 
 specific photo-chemical activity by filling a little quartz cell 
 with ozonised oxygen and exposing it to radiation for a 
 definite length of time. 
 
 The specific photo-chemical activity </> was obtained from 
 
 the ratio 
 
 m - m, 
 *" ~^> 
 
 where E is the energy absorbed from the light m , and m s the 
 ozone concentration before and after irradiation. 
 
 The rate of deozonisation was calculated from the relation- 
 
 ship 
 
 dc(0 B ) 
 
 J = Light energy in gm. calories per second. 
 A = Fraction of light energy absorbed = am. 
 m = Concentration of ozone in cell. 
 
 g = A cell constant correcting for the spontaneous decomposi 
 tion of the ozone in the cell. 
 
 fyy\ 
 
 Hence log = t log (</>AJ -f <?), 
 
 m s 
 
 and in the absence of radiation 
 
 He obtained the following values for the specific photo- 
 chemical activity with light of wave length X = 253 ^ for 
 
THE CATALYTIC DECOMPOSITION OF OZONE 141 
 
 Ozone in <f> x 10 5 . 
 
 Oxygen . 0-253 
 
 Nitrogen 0-975 
 
 Helium 1'520 
 
 Hautefeuille and Chappuis (" C.R," 91, 762, 1880) were the 
 first to notice the inhibiting effect of chlorine on the forma- 
 tion of ozone, which result was confirmed by Shenstone and 
 Evans (" J.C.S.," 73, 246, 1898). The work of Bodenstein 
 and others on the hydrogen chlorine combination leads one to 
 conclude that chlorine is an optical sensitiser for the decom- 
 position of ozone. F. Weigert (" Zeit. Elektrochem.," 14, 
 591, 1908) clearly showed that ozone containing small 
 quantities of chlorine was rapidly decomposed by blue and 
 violet light, whilst pure ozone is, as we have seen, only sensi- 
 tive to ultra-violet radiation. The rate of decomposition was 
 found to be proportional to the intensity of the light and 
 independent of the ozone concentration. It would appear 
 that the molecules of chlorine absorb the smaller quanta of 
 longer wave length and are consequently endowed with an 
 excess of kinetic energy equal in magnitude to the quantum 
 absorbed. By subsequent collision with an ozone molecule 
 this increment of energy is transferred, causing a rupture of 
 the ozone molecule into oxygen. The chlorine molecule thus 
 serves as a conveyor of energy and makes the medium sensi- 
 tive to this particular radiation frequency. 
 
CHAPTEK IX. 
 
 INDUSTRIAL USES OF OZONE. 
 
 APPLICATION TO HYGIENIC PURPOSES. 
 
 The Sterilisation of Water by Ozone. The earliest experi- 
 ments on the use of ozone as a germicide for the sterilisation 
 of water were made by De Meritens in France (1886). 
 
 He showed that even dilute ozonised air would effect the 
 sterilisation of polluted water, provided that intimate contact 
 between gas and liquid was effected. A few years later the 
 subject was reinvestigated by Frohlich(" Elektrochem. Zeit.," 
 344, 1891), of the firm of Siemens and Halske, who erected 
 a semi-technical experimental plant at Martinikenfeld. 
 Ohmuller and Prall (" Arbeit. Kais. Gesund," 229, 1892), at 
 the request of the German Government, investigated the 
 process in great detail and as a result showed that ozone 
 energetically attacked bacteria in water from which any 
 excess of inert organic matter had been previously removed. 
 Sufficient evidence as to its practical utility was thus at 
 hand to warrant a closer examination as to its suitability for 
 municipal work. 
 
 Chemically, ozone is the ideal agent for purification, since 
 it leaves behind it nothing foreign in the treated water, with 
 the exception of oxygen, which assists in the normal aeration 
 
 and greatly augments the palatability of the water. 
 
 (142) 
 
INDUSTRIAL USES OF OZONE 143 
 
 For technical purposes, however, the important factors of 
 reliability and cost are all important. 
 
 As a result of Frohlich's experiments and the satisfactory 
 report of Ohmuller and Prall, the firm of Siemens and Halske 
 developed their process of sterilisation with ozone, and large 
 plants were installed at Wiesbaden and Paderborn and at a 
 later date at St. Maur, Paris, and Petrograd. 
 
 Contemporary with these developments, Tindal, Schneller 
 and Van Sleen installed an ozone sterilisation plant at Oud- 
 shoorn in Holland which was subjected to a detailed in- 
 vestigation by Van Ermengem on behalf of the Belgian 
 Board of Agriculture, and by Drs. Ogier, Koux and Kepin 
 (" Eev. Gen. des Sciences," 596, 1896) for the municipality 
 of Paris, with the result that an ozone installation on Tindal 
 and de Frise's system was installed at St. Maur, Paris. At 
 a later date a combination of the Tindal-de Frise and 
 Siemens-Halske systems was utilised by the Parisian muni- 
 cipal authorities. 
 
 In 1898, Abraham and Marmier erected an installation at 
 Lille, and in 1904 an ozone plant on Otto's system was sup- 
 plied to the municipality of Nice. 
 
 Small installations were likewise erected on Vosmaer's 
 system in Holland and at Philadelphia, U.S.A. 
 
 As a result of these developments at the present time 
 there are in operation over fifty plants for the industrial 
 sterilisation of water by this means. 
 
 We have already given a brief description of the various 
 types of ozonisers developed for the purpose of the economical 
 production of ozonised air ; the various installations naturally 
 differed in their methods of ensuring contact of the water 
 
144 
 
 OZONE 
 
 with the ozonised air and in the preliminary treatment of 
 the water. 
 
 Systems of Ensuring Intimate Contact Between the Ozone 
 and Water. In the earlier Siemens and Halske installations 
 at Wiesbaden and Paderborn, ozonised air containing from 
 2-J to 3 grammes of ozone per cubic metre passed upwards 
 
 Water 
 
 FIG. 21. 
 
 through a tower filled with broken flint and met a descend- 
 ing current of water which had passed through roughing 
 filters. 
 
 Fairly satisfactory bacterial reduction was effected by 
 this method, the count ranging from 9 to 2 organisms per 
 c.c. after ozonisation, and these organisms were of the 
 resistant and harmless B. subtilis types. Economically, 
 
INDUSTEIAL USES OF OZONE 145 
 
 however, these contact towers left much to be desired, the 
 cost being about 2'75d. per 1000 gallons of water treated. 
 
 The fundamental point to be considered in the design of 
 a contact tower, viz. the intimate contact of every drop of 
 water with the ozonised air, was clearly not realised in this 
 elementary type of tower. It is evident that every flint is 
 covered with a thin film of water into which the ozone can 
 only penetrate by diffusion. Now the rate of diffusion of the 
 ozone into the liquid film is proportional to its partial pres- 
 sure in the gas above the film and approximately inversely 
 proportional to the square root of its density. Thus, although 
 it is not a difficult matter to oxygenate water in this type of 
 tower by means of air, since the partial pressure of oxygen 
 in air is about 20 per cent, of one atmosphere, yet in the 
 case of ozonised air containing 2'4 gms. per cubic metre 
 the partial pressure of ozone is only 1/1000 of an atmosphere 
 and its relative rate of diffusion only ^/f f, or three-quarters 
 that of oxygen. As a consequence, to ensure sterility, high 
 concentrations of ozone had to be employed and much ozone 
 was lost at the top of the tower. 
 
 Trouble was also caused, especially at Wiesbaden, by 
 clogging of the contact tower with precipitated ferric 
 hydroxide. The water contained small quantities of ferrous 
 salts up to 0' 5 parts per million, which were precipitated on 
 oxidation as ferric hydroxide in the interstices of the quartz 
 packing. 
 
 Vosmaer states that by proper design of this type of 
 tower, relatively long and slender in proportion, sterilisation 
 can be accomplished by uzing an ozone concentration of 
 
 only 1 gramme per cubic metre, and that complete removal 
 
 10 
 
146 
 
 OZONE 
 
 of the ozone is effected by such a tower when gas and liquid 
 flow rates are properly adjusted ; thus a tower 1 foot in 
 diameter and 33 feet high was shown capable of treating 
 10,000 gallons an hour, one 3 feet diameter at Philadelphia 
 dealt with 50,000 gallons per hour. 
 
 In the first plants of Tindal and de Frise both ozonised 
 air and raw water entered by separate pipes at the base and 
 flowed in the same direction through the flint packed tower, 
 which was divided up into segments by means of perforated 
 metal or celluloid plates. 
 
 FIG. 22. 
 
 A considerable advance on this practice was made by 
 Otto .by the introduction of an emulsifier for ensuring inti- 
 mate mixture of ozonised air and water at the base of the 
 tower. The emulsifier is constructed on the lines of a Korting 
 injector or simple water vacuum pump, the water supplied 
 under pressure drawing in the ozonised air under vacuum. 
 Not only was a very intimate mixture of the ozone and the 
 water effected by this means but by utilising the vacuum 
 produced by the water stream the necessity for pumping 
 ozonised air, which requires pumps of special construction, is 
 avoided. Otto's installation at Nice included a sectional 
 contact tower of cement using shingle as packing, and 
 
INDUSTKIAL USES OF OZONE 
 
 147 
 
 operated by means of these emulsifiers. The plant is 
 capable of dealing with 5,000,000 gallons per day. 
 
 In the Ozonair System, which has been successfully 
 developed for installations of capacity from 1000 to 2000 
 gallons per hour, the ozonised air is drawn into an injector 
 by the pressure of the water which is of the order of about 
 two atmospheres. It is thoroughly emulsified by the injector 
 and passes into a sterilising tank which is so proportioned as 
 to give a contact of about 80 minutes. The concentration of 
 ozone used is from 2 to 3 gms. per cubic metre produced in 
 the standard type of ozonisers (see p. 129). 
 
 The greatest economy in the utilisation of ozone was 
 effected by de Frise, who introduced a cyclic system at the 
 St. Maur waterworks, where it operated with great success. 
 
 A diagrammatic sketch of De Frise's system is shown 
 below. 
 
 FIG. 23. 
 
 The ozone compressor draws the ozonised air from the 
 ozonisers and forces it into the plate sterilising tower ; from 
 the top of the steriliser the air returns to the ozonisers, 
 
148 OZONE 
 
 passing on its way through a separator and a dryer in which 
 it comes in contact with calcium chloride, which is cooled in 
 hot weather. A small suction valve on the inlet of the 
 separator admits fresh air to make up for the loss through 
 absorption by the sterilised water. The sterilising tower is 
 constructed on the sectional system, being built up of 
 enamelled iron sections 20 inches long and 3 feet in diameter, 
 each section (there being 30 sections in the tower) being 
 separated from the others by perforated plates. 
 
 With a ratio of ozonised air to water of two to five, 
 88 per cent, removal of the ozone is effected in the tower 
 system, the residue being returned after drying to the ozon- 
 isers. 
 
 Pre-treatment of the Water. Since inert organic matter 
 is readily oxidised by ozone to carbon dioxide and water, in 
 order to exercise every possible economy in the utilisation of 
 ozone, it is necessary to subject polluted waters to some form 
 of preliminary purification. 
 
 The earlier installations at Wiesbaden and Paderborn 
 were equipped with simple roughing filters with small pebbles 
 as filtration medium, through which the water passed at 
 relatively high speeds. In the later plants, where it was 
 realised that the cost of removing organic matter by oxida- 
 tion with ozone was higher than by the usual methods of 
 sedimentation and filtration, more attention was paid to the 
 pre-treatment of the water, and the ozone was utilised merely 
 to remove the last traces of organic matter and ensure steri- 
 lising of the water, conditions unobtainable by any method 
 of sedimentation and filtration except at prohibitive costs. 
 
 In the Otto installation at Nice sand filtration alone is 
 
INDUSTRIAL USES OF OZONE 149 
 
 employed as pre- treatment. At Ginnekin, in Holland, the 
 Mark water is subjected to gravel filtration, followed by a 
 sand filter. At Oudshoorn the Rhine water is passed through 
 sedimentation basins and sand filters prior to sterilisation. 
 At St. Maur, Paris, sedimentation basins in conjunction with 
 gravel and sand filters are utilised for the Marne water, 
 whilst at Petrograd the Neva water, which is subjected to 
 very serious pollution and liable to contain pathogenic or- 
 ganisms, is passed through a battery of mechanical filters 
 using alumina as coagulant prior to the ozone treatment. 
 
 .For the sterilisation of upland waters which are subjected 
 to sporadic or seasonal pollution such pre-treatment is gener- 
 ally unnecessary, especially in those cases where adequate 
 storage is provided by means of impounding reservoirs which 
 act as sedimentation basins of large capacity. Exceptions, 
 however, are to be found when the upland water is derived 
 from a peaty or recently flooded watershed, where, during 
 the summer months, the quantity of oxidisable organic matter 
 in the water suffers a considerable augmentation and neces- 
 sitates a correspondingly increased dosage of ozone. 
 
 For the treatment of river supplies which are already 
 polluted, and with the growth of the riparian population are 
 becoming increasingly contaminated in organic matter, some 
 form of pre-treatinent is necessary, and it must be emphasised 
 that the utility of ozone in the present stages of its industrial 
 manufacture is most marked when it is employed as a finisher 
 or sterilising agent for waters which are already good enough 
 but not safe enough for public supplies, rather than when at- 
 tempts are made to purify highly contaminated streams. 
 
150 
 
 OZONE 
 
 EESULTS EFFECTED BY OZONE TEEATMENT. 
 
 The change in appearance of many waters after ozonisa- 
 tion is frequently very striking, attributable to the removal 
 of traces of coloured organic compounds and to the aeration 
 and oxygenation effected during the process. Marshy or 
 polluted river waters, which, although clear after effective 
 sand nitration, are brown or greenish-brown in colour, and 
 generally flat and insipid, after ozonisation they become 
 bright and sparkling, generally colourless or acquiring a faint 
 bluish tint. 
 
 The physical changes are accompanied by chemical 
 changes no less marked and characteristic of the process, 
 thus the oxygen consumed figure determined by the potas- 
 sium permanganate under standard conditions (usually at 
 60 C. in acid solution) is usually reduced from 30 to 70 per 
 cent., as is indicated by the following figures : 
 
 Installation. 
 
 Oxygen Consumed Figure. 
 Parts per 100,000. 
 
 Observers. 
 
 Before 
 Treatment. 
 
 After 
 Treatment. 
 
 Nice .... 
 Seine .... 
 St. Maur . 
 
 Philadelphia 
 
 1-4 
 0-43 
 0-124 
 
 0-3 
 0-21 
 0-060 
 (mean of 14) 
 40 per cent, 
 reduction 
 
 Buisine & Bouriez. 
 Van der Sleen. 
 Rideal. 
 
 Vosmaer. 
 
 The quantity of organic matter oxidised by the ozone as in- 
 dicated by the difference in the oxygen consumed figures 
 naturally varies with the three factors, the nature of the 
 organic matter, the concentration of the ozone, and the time 
 
INDUSTEIAL USES OF OZONE 151 
 
 of contact between the ozone and the water. Thus S. Rideal 
 (" J. Roy. S. Inst.," 30, 32, 1909) found at St. Maur the 
 following distribution of the ozone at two different gas con- 
 centrations : 
 
 Ozone in Gins. Per Cubic Metre, in 
 
 Entering air 1'68 2-65 
 
 Entering water 0-670 1-071 
 
 Escaping into the air 0'137 0*291 
 
 Escaping in solution in the water . . . 0'048 0-162 
 
 Used in oxidising organic matter . . . 0-494 0-618 
 
 Gins. Per Cub. Metre. 
 Oxygen consumed for the water untreated 
 
 (from acid KMn0 4 at 60 C.) 1'73 1-12 
 Oxygen consumed for the water treated 
 
 (from acid KMn0 4 at 60 C.) T25 0-61 
 
 Decrease 0-48 0-51 
 
 = 28 % = 45 % 
 
 During these experiments it was noticed that the dissolved 
 ozone left in the water after it left the ozonising column dis- 
 appeared after a few hours, and it seemed probable that 
 oxidation of the more resistant organic matter was still pro- 
 ceeding. This was confirmed by the determination of the 
 oxygen consumed figures from time to time. 
 
 Two typical tests giving the following results : 
 
 Oxygen Consumed Gms. 
 
 Per Cubic Metre. 
 Filtered water 1-73 1-12 
 
 After ozonisation 1-25 0-61 
 
 One hour later 1*15 0-58 
 
 Two hours later 0-90 0-51 
 
 A corresponding decrease in the free and albuminoid 
 ammonia, takes place simultaneously with the reduction in 
 
152 OZONE 
 
 the oxygen consumed figures, as is shown from the investiga- 
 tions of Van der Sleen on the Seine water : 
 
 Before Treatment. After. 
 
 Free ammonia . 0-271 0-136 
 
 Albuminoid ammonia . . . 0-536 0-189 
 
 As in the case of the organic matter in the water, a 
 marked reduction in the number of bacteria takes place. In 
 common with other germicides ozone is selective in its action 
 on micro-organisms, certain organisms being more resistant 
 than others to its action. It so happens that pathogenic 
 organisms which are already enfeebled by their unsuitable 
 environment are easily destroyed, whilst the non-pathogenic 
 sporing organisms, such as B. subtilis, are frequently found in 
 ozonised waters. This point was fully investigated by Otto 
 in his experiments on the treatment of the Seine and Vannel 
 waters. 
 
 He showed that all pathogenic organisms and the follow- 
 ing common bacteria which he isolated were rapidly removed : 
 these included B. fluorescens liquefaciens, B. coli, B. termo, 
 B. proteus vulgaris, B. prodigiosus and Aspergillus niger. 
 More resistant were B. subtilis, B. luteus and Penicillium 
 glaucum, which only succumbed to prolonged treatment with 
 ozone. 
 
 Camlette showed that Lille installation effected an average 
 reduction in the bacterial pollution of the water from 28,000 
 per c.c. in the untreated water to 10 c.c. in the treated water ; 
 these latter were found to consist entirely of the B. subtilis 
 type. These conclusions were confirmed by S. Eideal in his 
 investigations on the Siemens-de Frise plant at St. Maur. 
 In the filtered water, prior to ozonisatjon, 3B, coli was usually 
 
INDUSTEIAL USES OF OZONE 153 
 
 present in 40 c.c. of water and occasionally in 20 c.c., whilst 
 it was never found in the treated water. The 20 c.c. count 
 was reduced from an average of 74 to 11 per c.c., which were 
 found to be spore-bearing organisms of B. subtilis type. 
 Whilst purification can be effected by the consumption of 
 0'53 gms. of ozone per cubic metre of water, the ozone 
 consumption per cubic metre of water naturally varies with 
 the quality of the water to be treated, the economic limit 
 appears to be in the neighbourhood of 2 gms. ozone per cubic 
 metre of water. If, by preliminary testing with permanganate, 
 it is found that more than this quantity would be required 
 then some form of preliminary treatment is essential for 
 economic operation. 
 
 The Purification of Air. 
 
 Several types of apparatus have been designed for the 
 the purification of air in confined spaces, such as theatres, 
 lecture halls, slaughter houses, tanneries and breweries, whilst 
 extended use of ozonised air has been made upon the Central 
 London Railway systems and public lavatories. 
 
 Although the immunity of the motor drivers on the 
 London Tubes during the recent epidemics of influenza has 
 been somewhat remarkable it is more probable that this must 
 be attributed to the uniformity of temperature and to the 
 circulation of the air rather than to any specific action of the 
 ozone in the air. 
 
 The evidence for the specific utility of ozone as a means 
 of purifying air is somewhat conflicting. 
 
 Dewar and McKendrick ("Pogg. Ann.," 152, 329, 1874) 
 showed that by the inhalation of strongly ozonised air the 
 frequency of pulsation of the heart is lowered very consider- 
 
154 OZONE 
 
 ably, the blood temperature sinks from 3 to 5, and post 
 mortem examination showed that the blood had become 
 venous in appearance. Thenard (" O.K.," 82, 157, 1876) and 
 Biny ("Med. C. Bl.," 20, 721, 1882) confirmed these observa- 
 tions of Dewar's. Schultz ("Arch. f. Exper. Path. u. Plan.," 
 29, 365, 1892) records several cases of chronic poisoning by 
 ozone. Jordan and Carlson x (" J. Amer. Med. Assoc.," 61, 1007, 
 1913) confirmed the deodorant action of ozone on air but 
 showed that long before the concentrations reached those 
 necessary for germicidal action injury was caused to the res- 
 piratory tract. 
 
 The lowest concentrations of ozone in air which can 
 exert a definite disinfecting action (Schultz, " Zeit. f. Hyg.," 
 75, 1890 ; and de Christmas, " Ann. de 1'inst. Pasteur," 7, 689, 
 1893) appears to be in the neighbourhood of 13'53 mgm. per 
 litre. With such concentrations sterilisation can usually be 
 effected in air, but the presence of large quantities of moisture 
 lowers its germicidal activity. According to Labbe and 
 Oudin ("C.K.," 113, 141, 1891) the highest concentration 
 which may be inhaled without deleterious effects is approxi- 
 mately O'll to 0'12 mgm. per litre. 
 
 They state that beneficial results obtain by the inhalation 
 of ozonised air of this concentration, a marked increase in the 
 oxyhaemoglobin contact of the blood taking place after an 
 interval of from ten to fifteen minutes. 
 
 It is therefore evident that there is no question of 
 germicidal activity in ozonised air of concentrations suitable 
 for respiration. As a powerful oxidant it doubtless removes 
 
 1 Leonard Hill and Flack (" Proc. Roy. Soc.," B 84, 405, 1911) state that a 
 concentration of 1 in 10& of ozone produces irritation of the respiratory tract, 
 
INDUSTEIAL USES OF OZONE 155 
 
 small traces of hydrogen sulphide and other impurities in air, 
 whilst the unpleasant smells associated with crowded places 
 are amenable to treatment with oxidising agents such as 
 ozone. It is somewhat remarkable that most odoriferous sub- 
 stances contain unsaturated valencies and as such would 
 naturally be attacked by means of ozone. The odour of ozone 
 in itself when very dilute is by no means unpleasant and thus 
 provides a counter irritant to the olfactory organs, a matter 
 of psychological importance if of no physiological significance. 
 
 The use of ozone even in dilutions of O'll to 0'12 mgm. 
 per litre has frequently been condemned on account of its 
 supposed physiological activity, in many cases erroneously, 
 since, unless the usual precautions such as drying the air and 
 the avoidance of sparking are taken in the preparation of the 
 ozonised air, oxides of nitrogen are liable to be formed which 
 have an extremely high physiological activity even in extreme 
 dilutions. Schwarg and Munchmeyer ("Zeit. f. Infekt. 
 Krankh," 75, 81, 1913) investigated in great detail the de- 
 odorising action of small concentrations of ozone in air ; they 
 observed that hydrogen sulphide was rapidly oxidised, sulphur 
 dioxide slowly converted to sulphuric anhydride, whilst the 
 mercaptans, skatol, and indol were oxidised to somewhat 
 pleasant smelling substances. 
 
 Carbon monoxide was but slowly oxidised whilst ammonia 
 was not affected. Franklin (" IV. Int. Congress on School 
 Hygiene," 1918) and Eiesenfeld and Egidius (" Zeit. Anorg. 
 Chem.," 85, 217, 1914) confirmed these observers' results on 
 the action of ozone on hydrogen sulphide. This reaction 
 was studied in detail by Biesenfeld, and it was found that 
 oxidation takes place through a series of intermediate com- 
 pounds according to the following scheme ; 
 
156 OZONE 
 
 hyposulphate, sulphite 
 Sulphide -> thiosulphate * 
 
 ^polythionates. 
 
 -> dithionates - sulphates. 
 
 Although the germicidal activity of ozone in concentrations 
 which exert no injurious action on the respiratory organs is 
 practically negligible, yet as has been shown by Dibden some 
 sterilisation is effected by ozonisation of air, since a marked re- 
 duction is obtained in the bacterial count of the air which has 
 actually passed through the ozoniser. This air, which has been 
 relatively strongly ozonised and subjected to the ultra-violet 
 radiation in the ozoniser, is practically sterile, and a consequent 
 improvement in the bacterial pollution of air which has been 
 admixed with this purer air was naturally expected and in 
 fact obtained. The largest system of air purification by 
 means of ozone is that of the Central London Kailway which 
 is equipped with air screens, washers, and ozonisers of the 
 Ozonair type. The total air supply treated for this system 
 being of the order of eighty million cubic feet per daj^. The 
 average concentration of ozone in the Tube air is of the order 
 of one part in two millions (1*2 mgm. per cubic metre) which, 
 under special circumstances, it is stated that it is increased to 
 five parts per million (12 mgm. per cubic metre). 
 
 The installation on the Central London Kailway was so 
 successful in operation that the system of ventilation has been 
 extended to practically every other London Tube. 
 
 Surgical and Therapeutic Uses of Ozone. 
 
 During the period of the war small portable ozonisers 
 have been in use in many of the military hospitals of the 
 nations for the treatment of wounds. Major Stokes, of the 
 
INDUSTRIAL USES OF OZONE 157 
 
 Queen Alexandra Military Hospital ("Lancet," 1918), de- 
 scribes the following method of wound sterilisation which 
 appears to have given excellent results. Wounds and sinuses 
 are washed twice daily with boiled water and a dressing of 
 oxy gauze is applied. The ozone (ozonised air) is applied on 
 the wounded surface or on to the cavities and sinuses twice 
 daily for a maximum time of fifteen minutes or until the sur- 
 face becomes glazed. At first ozone causes an increase of the 
 discharge of pus, which is gradually replaced by clear serum. 
 The serum at a still later stage becomes coloured reddish or 
 purplish. It was found that ozonised air applied in this way 
 
 was a strong stimulant, and increased^ the flow of blood to the 
 
 **/-s/jy! ^c&oferM^'v 
 affected part, that it was a germicide ana exerted great powers 
 
 in the formation of oxyhaemoglobin. A record of seventy- 
 nine cases so treated is given, in which the period of treat- 
 ment varied from a few days to three months and was in 
 practically all cases completely successful. 
 
 Curie (" Practitioner," 864, 1912) describes the application 
 of ozonised air to the liberation of iodine in the lungs for the 
 treatment of phthisis. Potassium iodide is introduced into 
 the lungs by ionisation, and the iodine is subsequently liber- 
 ated by the inhalation of ozonised air. For disinfection of 
 the intestinal canal in cases of enteritis and dysenteries, 
 Lessing (" Lancet," Nov., 1913) records the improvement 
 obtained by washing out the intestinal canal with ozonised 
 water. The treatment of ulcers and pyorrhoea of the teeth has 
 been successfully accomplished with ozonised air. Ozonised 
 medicaments and ointments, such as vaseline, have been stated 
 to possess a superior curative value to those not so treated. 
 Ozone appears to be slightly soluble in these semi-fluid 
 
158 OZONE 
 
 hydrocarbons, and would naturally give the medicaments a 
 germicidal activity. Information is not available as to the 
 extent of the solubility of ozone in vaseline, fats, and lards, 
 or how long the substance retains any germicidal activity due 
 to dissolved ozone or to the presence of unstable ozonides. 
 As has already been mentioned, the action of dilute ozonised 
 air in stimulating the production of oxyhaemoglobin has been 
 successfully utilised for the treatment of cases of anaemia, 
 whilst its application in cases where there is a shortage of 
 oxygen absorbed in the system, such as in asthma or heart 
 weakness, is not without benefit. 
 
 Applications for Bleaching Purposes. 
 Houzeau (" C.K.," 75, 349, 1872) and Boillot (" C.E.," 80, 
 1187, 1875) showed that dilute ozonised air possessed, in 
 common with chlorine and bromine, the property of selectively 
 oxidising the colouring matters present in various natural 
 substances. Ozone appears to be even more efficacious than 
 either chlorine or bromine, since not only is the danger of 
 forming coloured substitution chloro- or bromo-derivatives 
 avoided, but the oxidising power of ozone greatly exceeds that 
 of chlorine or bromine. For this reason only very dilute 
 concentrations of ozone may be utilised for bleaching purposes, 
 and many unsuccessful results are directly attributable to the 
 employment of air relatively highly ozonised. 
 
 Fibres. 
 
 Linen and cotton goods are slowly attacked by ozone 
 (" Kolb. Bull. Soc. Ind. Mulhouse," 38, 94, 1868), and accord- 
 ing to Witz("Bull. Soc. Eouen," u, 198, 1883), in the 
 presence of moisture, oxycellulose is formed. The subject 
 
INDUSTRIAL USES OF OZONE 159 
 
 was reinvestigated by Cunningham and Doree (" J.C.S.," 
 103, 1347, 1912) employing high concentrations of ozonised 
 oxygen (20 to 25 gms. per cubic metre). They showed that 
 an oxycellulose and cellulose peroxide accompanied by a de- 
 struction of the fibre were produced with the liberation of 
 carbon dioxide. 
 
 Unbleached samples of cotton became white in from one to 
 two hours, but when dried the fibre was found tendered and 
 dusty. Jute fibres were likewise rapidly bleached, but became 
 acid and tender when subjected to a similar treatment for a 
 few hours. The cellulose peroxide affected a photographic 
 plate, and possessed oxidising properties such as the liberation 
 of iodine from potassium iodide. In the presence of water 
 hydrogen peroxide was formed. In general the action of 
 ozone on cellulose closely resembles that of ammonium per- 
 sulphate ("Ditz. Chem. Zeit," 31, 833, 1907 ; " J. Prakt. Chem.," 
 78, 343, 1908), and with more dilute concentrations of ozonised 
 air bleaching without subsequent tendering might be ob- 
 tained. 
 
 The application of ozonised air to the conditioning of tex- 
 tile materials is stated to be entirely successful ; exceedingly 
 dilute concentrations of ozonised air are employed, and under 
 suitable conditions of humidity the period of conditioning 
 can be reduced from several months to a few days. 
 
 Oils, Fats, and Waxes. 
 
 The applications of the use of ozone in the oil, fat, and 
 wax industries are very considerable, the reagent being useful 
 for a great diversity of purposes. Amongst the more im- 
 portant of these may be mentioned : 
 
160 
 
 OZONE 
 
 1. Kemoval of odour, flavour, and colour from oils and 
 fats intended for edible purposes. 
 
 2. Bleaching and refining of oils and fats for use in the 
 soap industry. 
 
 A. Air Cleaner. 
 
 B. Electrically driven Blower. 
 
 C. Air Delivery Pipe. 
 
 D. Air Cooling Machine. 
 K. Electric Motor. 
 
 E. Cold Air Delivery Pipe. 
 
 F. Ozone Generators 
 
 G. Transformer. 
 H. Valve. 
 
 I. Ozone Pipe. 
 
 J. Ozone Injectors. 
 
 L. Switchboard. 
 
 FIG. 24. 
 
 3. Bleaching of oils for paint and varnish-making. 
 
 4. Bleaching and refining of waxes for use in the manu- 
 facture of candles, polishes, and ointments. 
 
 In the above sketch is given a diagrammatic arrange- 
 
INDUSTKIAL USES OF OZONE 161 
 
 ment of a bleaching plant on the Ozonair system. The 
 temperatures at which selective oxidation of the colouring 
 matter or objectionable odoriferous substances in the oil, fat, 
 or wax commences, the concentration of ozone, and the period 
 of action naturally vary with the nature of the substance 
 treated and the degree of refining required. Generally a fairly 
 dilute concentration of ozone at not too elevated temperatures 
 passed through the material for a relatively long time gives 
 the best results. In certain cases it is found advantageous 
 to add small quantities of catalytic materials, such as salts 
 of manganese, vanadium, or cerium, to accelerate the process 
 of oxidation. 
 
 The application of ozonised air to the bleaching and 
 deodorisation of oils, fats, and waxes is well exemplified by 
 the following list (see pp. 162, 163), in which a summary is 
 given of the effect of dilute ozonised air on various com- 
 mercial products. 
 
 Application to the Paint and Varnish Industry. 
 
 As has already been observed all the industrial oils and 
 waxes readily undergo partial or complete deodorisation by 
 fractional oxidation of the coloured chlorophyllic constituents 
 utilising ozonised air as oxidising agent. For the production 
 of clear and transparent varnishes this is a matter of some 
 technical importance, and the use of ozone for this purpose has 
 been frequently suggested. 
 
 It would appear that the use of ozonised air as a substitute 
 for siccatives in the preparation of drying oils is already 
 out of the experimental state. It has long been known that 
 the drying of oils is a process of slow oxidation and poly- 
 merisation (see Lippert, " Zeit. Angew. Chem.," u, 412, 1895; 
 
 11 
 
162 
 
 OZONE 
 
 Oil, Fat, 
 or Wax. 
 
 Bleaches. 
 
 Deodorises. 
 
 Remarks. 
 
 Arachide or 
 ground nut 
 
 Yes 
 
 Yes 
 
 Bleaches easily, but requires subsequent 
 treatment for deodorisation and im- 
 provement of flavour. 
 
 Bone fat 
 
 Yes 
 
 Yes 
 
 Bleaches easily, at the same time the 
 odour is much improved, in fact rank 
 or ill-smelling fats are readily con- 
 verted into perfectly sweet substances. 
 
 Borneo 
 Tallow 
 
 Coconut oil 
 
 Yes 
 Yes 
 
 No 
 
 bleaches readily, but requires subse- 
 quent treatment for deodorisation 
 and improvement of flavour. 
 
 1 Bleaches readily, but requires subse- 
 quent treatment for deodorisation 
 and improvement of flavour. 
 
 Coconut acid 
 oil 
 
 Yes 
 
 Yes 
 
 Yields readily to treatment for bleach- 
 ing and deodorisation. 
 
 Cod oil 
 
 Yes 
 
 Yes 
 
 
 Cotton seed 
 
 Yes 
 
 No 
 
 bleaches fairly easily according to 
 grade or quality, but requires subse- 
 quent treatment for deodorisation 
 and improvement of flavour. 
 
 Japan wax 
 tallow 
 
 Yes 
 
 Yes 
 
 Bleaches well according to grade or 
 quality. 
 
 Lard oil 
 
 Yes 
 
 No 
 
 Bleaches very easily. 
 
 Mowrah 
 
 Yes 
 
 
 
 Bleaches readily, but requires subse- 
 quent treatment for deodorisation 
 and improvement of flavour. 
 
 Neatsfoot oil 
 
 Yes 
 
 Yes 
 
 Bleaches readily. 
 
 Oleine 
 
 Yes 
 
 Yes 
 
 Bleaches easily. At the same time the 
 odour is much improved, but the 
 characteristic odour still remains. 
 
 Olive 
 
 Yes 
 
 Yes 
 
 Bleaches very easily, the darkest 
 varieties, such as that known as 
 "sulphur," are bleached to the or- 
 dinary olive colour. 
 
 Palm 
 
 Yes 
 
 No 
 
 Bleaches very easily in most cases. 
 Even " unbleachable " varieties such 
 as Congo and saltpond are much im- 
 proved in colour. 
 
 Sesame 
 
 Yes 
 
 No 
 
 Bleaches fairly easily according to 
 grade or quality, but requires subse- 
 quent treatment for deodorisation 
 and improvement of flavour. 
 
INDUSTEIAL USES OF OZONE 
 
 163 
 
 Oil, Fat, 
 or Wax. 
 
 Bleaches. 
 
 Deodorises. 
 
 Remarks. 
 
 Soya 
 
 Yes 
 
 Yes 
 
 Bleaches easily. At the same time 
 
 
 
 
 the odour is much improved, but the 
 
 
 
 
 characteristic odour still remains. 
 
 Stearine 
 
 Yes 
 
 Yes 
 
 Bleaches easily. At the same time 
 
 
 
 
 the odour is much improved, but the 
 
 
 
 
 characteristic odour still remains. 
 
 Tallow 
 
 Yes 
 
 Yes 
 
 Bleaches easily. Odour is much im- 
 
 
 
 
 proved. 
 
 Turnip seed 
 
 Yes 
 
 Yes 
 
 1 Bleaches fairly easily according to 
 
 
 
 
 grade or quality, but requires subse- 
 quent treatment for deodorisation 
 
 
 
 
 and improvement of flavour. 
 
 Beeswax 
 
 Yes 
 
 
 
 There are at least twenty well-known 
 
 
 
 
 commercial varieties all of which 
 
 
 
 
 possess different characteristics, and 
 
 
 
 
 respond to the treatment in varying 
 
 
 
 
 degree. Some are bleached entirely 
 
 
 
 
 by ozone, others only partially, so 
 
 
 
 
 requiring subsequent sun treatment 
 
 
 
 
 in the ordinary way. Contrary to 
 
 
 
 
 what might be presumed, the dark 
 
 
 
 
 varieties, such as " Cuba" or "West 
 
 
 
 
 Indian," give the best, whereas the 
 
 
 
 
 lighter varieties may give the least 
 
 
 
 
 satisfactory results. 
 
 Cardililla 
 
 No 
 
 
 
 
 
 Carnauba 
 
 No 
 
 
 
 
 
 Lanoline 
 
 Yes 
 
 - 
 
 The bleaching is effected very readily 
 
 
 
 
 after removal of all the free fatty 
 
 
 
 
 acids. The degree of colour or bleach 
 
 
 
 
 obtained depends very much on the 
 
 
 
 
 quality of the original wax and the 
 
 
 
 
 care with which it has been collected. 
 
 Montana 
 
 No 
 
 
 
 
 
 l ln all these cases, if deodorisation only is desired, this can be obtained 
 in certain cases by using a low concentration of ozone and without any sub- 
 sequent treatment, but when bleaching and deodorisation are both required, 
 subsequent treatment is necessary. 
 
 It has been stated that the actual cost of bleaching and deodorising oils 
 and fats for edible purposes or for soap-making, by means of ozonised air, 
 does not exceed five shillings per ton, a figure which compares quite favour- 
 ably with the alternative methods of treatment with superheated steam, fuller's 
 earth or charcoal filtration. 
 
164 OZONE 
 
 Weger, " Chem. Eev. Fett. Harz. Ind.," 4, 301, 1899). A. 
 Genthe ("Zeit. Angew. Chem.," 19, 2087, 1906), who investi- 
 gated the process in detail, showed that the action was auto- 
 catalytic in character, or that the rate of oxidation of the oil 
 after a time when a quantity x had already been oxidised was 
 
 given by the equation -^ = K(& - x)(b + x), where a and b 
 
 were the initial concentrations of linseed oil and catalyst 
 originally present. Further experiments (see also Engler 
 and Weiszberg, " Chem. Zeit.," 27, 1196, 1903) showed that 
 the catalytic material naturally formed was some form of 
 unstable peroxide which accelerated the oxidation of the oil 
 by air. 
 
 This unstable peroxide could be supplemented or replaced 
 by other peroxides of a similar character, such as those ob- 
 tained on the exposure of turpentine to the air or even by 
 the agitation of an ether air mixture. These peroxides are 
 destroyed by boiling the oil but can be regenerated by aera- 
 tion. (For a consideration of the composition of these 
 peroxides formed in the drying of linseed oil see Orloff (" J. 
 Kuss. Phys. Chem. Soc.," 42, 658, 1910); Fahrion ("Zeit. 
 Angew. Chem.," 23, 723, 1910); Salway and Kipping 
 ("J.C.S.," 95, 166, 1909), and others.) 
 
 The addition of siccatives, such as salts of lead, manganese 
 zinc, frequently with the addition of certain promoters as 
 cobalt, vanadium, cerium, and uranium is now common 
 practice. The salts are either those of weak acids such as 
 the borates, or of soluble organic acids and oleates, linoleates, 
 or resinates (see Ingle, "J.C.S. Ind.," 454, 1917). The 
 siccatives are pseudo-catalytic in behaviour, and serve to 
 
INDUSTKIAL USES OF OZONE 165 
 
 stabilise or assist in the formation of the auto-catalytic per- 
 oxide. 
 
 Identical results are obtained with the use of ozonised 
 air, the oils can be easily thickened and the process conducted 
 at much lower temperatures. Linseed, Chinese wood, poppy 
 seed, rape, and similar oils, rapidly thicken at comparatively 
 low temperatures (upwards of 35 C.), and at the same time 
 their colour is much improved by selective oxidation of the 
 coloured constituents. Bleaching usually proceeds anterior 
 to the thickening process, consequently an improvment in 
 colour may be obtained without drying the oil. 
 
 Linseed oil may readily be thickened to a syrup or to a 
 jelly for the manufacture of linoleum ; Chinese wood oil 
 likewise rapidly undergoes oxidation, whilst poppy-seed and 
 rape oils thicken less readily than linseed oil. The utilisation 
 of ozonised air in the oxidation of linseed oils has naturally 
 been extended from the simple preparation of drying oil to 
 the drying of the oil in its various technical applications, 
 such as linoleum manufacture, the preparation of waterproof 
 materials, fish netting, and other similar manufactures. 
 
 Ozone in the Fine Chemical Industries. 
 
 The oxidation of organic substances by means of ozone has 
 been the subject of numerous investigators. Carbon monoxide 
 is readily oxidised to carbon dioxide in the presence of moisture 
 (Clausman, " C.E.," 150, 1332, 1910), but only slowly when 
 the gases are dry (Kemsen, " Ber.," 8, 1414, 1875). Alde- 
 hydes are readily oxidised to alcohols ; iodobenzene to iodoso 
 benzene whilst the saturated hydrocarbons themselves are 
 readily attacked at low temperatures. Thus, methane is 
 
converted below 100 C. into a mixture of methyl alcohol 
 formaldehyde and formic acid. Drugmaii (" J.C.S.," 89, 1614, 
 1906) has shown that gradual hydroxylation of one carbon 
 atom takes place, the corresponding alcohol is first formed 
 which is then oxidised rapidly to the more stable aldehyde 
 acid. Subsequently slow oxidation to the acid proceeds : 
 
 /OH 
 
 C . CH 3 -> C . CH 2 OH -=> C . CH<; -> C . HOH 
 
 X OH 
 
 ->C - OH 
 
 \>OH -> COOH 
 \OH 
 
 Harries ("Ann.," 374, 288, 1910) suggests that oxidation 
 proceeds through the intermediary formation of unstable 
 peroxides, e.g., C . CH 2 OH - C . CH 2 = 0, and not through 
 a series of unstable hydroxylations as postulated by Drugman. 
 
 Ozone is finding many applications as an oxidising agent 
 in the fine chemical industries. It has been successfully 
 employed for the preparation of several synthetic perfumes, 
 such as the methyl ether of pyrocatechaldehyde (vanillin), 
 piperonal (heliotrope), and anisic aldehyde ; the manufacture 
 of vanillin being accomplished on a very large scale in France, 
 America, and in this country. 
 
 Vanillin is prepared from eugenol according to the follow- 
 ing scheme : 
 
 C 3 H 5 CH = CH . CH 3 CHO 
 
 CH 3 CHO 
 COCH 
 
 Eugenol. Isoeugenol. Vanillin. 
 
INDUSTRIAL tJSES OF OZONE 167 
 
 the isoeugenol being oxidised by ozone to vanillin and 
 acetaldehyde by rupture of the double bond. 
 
 According to Trillat the process is conducted as follows : 
 Eugenol is converted into isoeugenol by treatment with 
 caustic potash and amyl alcohol, from which solution it is 
 liberated by sulphuric acid. 
 
 About 25 litres of isoeugenol are dissolved in acetic acid 
 and subjected to a current of ozonised air (2 to 2'5 gms. 
 per cubic metre) for a period of six hours at a low temperature 
 (ca. 2 C.) in an enamel-lined vessel fitted with a tall recti- 
 fication column. When the oxidation to the aldehyde is 
 completed the acetic acid is removed by distillation, ether and 
 sodium bisulphate are added and the solution warmed to 30 C. 
 The bisulphate aldehyde compound is washed with ether 
 decomposed with sulphuric acid, and the vanillin finally ex- 
 tracted with ether. 
 
 In a similar manner heliotropin is prepared from safrol 
 according to the equations 
 
 CK, - CH = CH 2 CH = CH.CH 3 
 
 CHO 
 
 L 2 
 
 Safrol. Isosafrol. Piperonal (heliotropin). 
 
 The safrol is converted into isosafrol by heating with an 
 alcoholic solution of caustic potash, from which it is subse- 
 quently extracted by means of ether. For treatment with 
 ozone (as in the case of isoeugenol) it is dissolved in acetic 
 acid, from which the heliotropin is recovered in like manner. 
 
168 OZONE 
 
 Anisaldehyde can be prepared in a similar manner by the 
 oxidation of anethol 
 
 CH = CH . CH, CH 9 . CH = GEL CHO 
 
 OCH 3 OCH 3 OCH 3 
 
 Attempts have also been made to utilise ozone for the 
 oxidation of aniline to aniline black and the leuco bases of 
 various dyes, such as indigo, to the coloured dye-stuffs, but do 
 not appear to have received any extensive technical applica- 
 tion. 
 
 For analytical purposes in organic chemistry ozone merits 
 some attention, since by the preparation of ozonides and their 
 subsequent decomposition the structure of various complex 
 compounds containing ethylene linkages has been elucidated. 
 
 The investigations of Harries (" Ber.," 37, 839, 842, 2708, 
 3431, 1904, et seq.) and his co-workers, have been the most 
 remarkable in this direction. Unsaturated compounds are 
 ruptured at the double bond and converted into aldehydes 
 and ketones. 
 
 In the absence of water, however, a direct addition of 
 ozone to the ' double bond occurs with the formation of 
 ozonides 
 
 >C : C< - >C . C< 
 
 I I 
 
 o o 
 
 which on the subsequent addition of water undergo decom- 
 position to ketones and hydrogen peroxide 
 
INDUSTRIAL USES OF OZONE 169 
 
 >c . c< 
 
 I I + H 2 ->>C = + O = C<+ H 2 2 . 
 O 
 
 V 
 
 The ozonides are colourless, viscid, oily substances, highly 
 explosive and possessing a penetrating odour. 
 
 In common with ozone they affect a photographic plate, 
 attributable to the chemioluminescence produced on oxida- 
 tion of organic matter by means of ozone (see pp. 8, 159). 
 
 The ozonides behave like powerful oxidising agents them- 
 selves, akin to the peroxides in chemical behaviour in that 
 they bleach indigo, liberate iodine from potassium iodide and 
 react with potassium permanganate. 
 
 Benzene triozonide or ozobenzene, isolated by Renard, is 
 a relatively stable substance, easily produced by passing 
 ozonised air into dry benzene, from which it is precipitated 
 as a gelatinous amorphous product. It explodes somewhat 
 violently on the addition of warm water. 
 
 Oleic acid, either when, dissolved in acetic acid (Harries 
 and Thieme, " Ber.," 39, 28, 44, 1906), or when treated with 
 ozonised air without a solvent (Molinari and Sonicini, " Ber.," 
 39, 27, 34, 1906), forms a normal ozonide. In chloroform 
 four atoms of oxygen are taken up to form an ozonide per- 
 oxide. 
 
 By analysis of the products of decomposition the position 
 of the unsaturated linkage in oleic acid was established be- 
 tween the atoms C 9 and C 10 , thus giving oleic acid the structure 
 CH 3 (CH 2 ) 7 CH = CH(CH 2 ) 7 COOH. 
 
 If strong concentrations of ozone be employed for the 
 preparation of ozonides, oxozonides are said to be formed at 
 
170 OZONE 
 
 the same time (Harries, " Ber.," 43, 936, 1912, et seq.\ thus 
 s. butylene yields the following substances on ozonisation : 
 
 (1) CH 3 . CH . CH . CH 3 (2) f CH 3 . CH . CH . CH 3 ] 
 
 \X \ \ / \ 
 
 3 I 3 J 2 
 
 Normal ozonide. The dimeric ozonide. 
 
 (3) CH 3 . CH . CH . CH 3 
 and \ / 
 
 4 
 
 Oxozonide. 
 
 and a dimeric oxozonide 
 
 (4) fCH 8 . CH . CH . CH S } 
 
 { v }, 
 
 cyclo pentene exhibits a similar behaviour in that two ozonides 
 and two oxozonides are formed on oxidation (Harries, " Ann.," 
 4, 4101, 1915). 
 
 Harries (loc. cit.) attributed the formation of oxozonides 
 to the presence of oxozone in the ozonised oxygen which he 
 postulated to be present by analysis, employing the iodide 
 method of estimation. The existence of oxozone in ozonised 
 air or oxygen has, however, not been confirmed (see p. 184) 
 and some other structural formula must, consequently, be 
 adopted for the oxozonides. 
 
 The ozonisation of rubber was first attempted unsuccess- 
 fully by Wright in 1897 ("Bull. Soc. Chem.," 18, 438, 1897), 
 and was subsequently investigated in great detail by Harries 
 and his co-workers, Langheld and Haefmer, in the hope of 
 elucidating the complex structure of the isoprene polymer. 
 
 Harries (" Ber.," 37, 2708, 1904) ozonises rubber by the 
 following procedure : Ozonised air, washed with caustic soda 
 
INDUSTRIAL tfSES 
 
 OZOlSfiJ 
 
 171 
 
 and sulphuric acid to remove the oxozone, containing from 
 6 to 12 per cent, of ozone, is passed for ten hours into a 1 per 
 cent, solution of purified rubber in chloroform. The end of 
 the reaction is ascertained by the decolorisation of bromine. 
 
 The ozonide is obtained by evaporation at 20 C. in vacuo, 
 subsequently reprecipitating from ethyl acetate by petroleum 
 ether in the form of a thick oil solidifying to a vitreous mass. 
 
 Rubber ozonide is soluble in ethyl acetate, benzene and 
 alcohol is explosive and like ozone it acts on a photographic 
 plate. 
 
 Analysis gives the average composition C 49 per cent., 
 H 2 6'9 per cent., and the molecular weight about 526, corre- 
 sponding to the compound 
 (C 10 H 16 6 ) 2 (C 51*72 per cent., H 2 6'70 per cent., m.w. = 464). 
 
 Adopting Harries' structural formula for isoprene rubber 
 
 fCH 3 C CH 2 CH 2 CH 1 
 
 \ S S 
 
 [He CH 2 CH 2 C.CH 3 L 
 
 the rubber ozonide would possess the following structure : 
 CH, 
 
 H 
 
 C CH 2 CH 2 C 
 
 ^0 
 
 \ 
 
 
 
 CH 2 CH 2 C 
 
 I / 
 
 H CH, 
 
 Eubber ozonide suffers decomposition on boiling with 
 water, forming levulinic aldehyde and levulinic aldehyde 
 peroxide, as indicated by the following equation : 
 
172 OZONE 
 
 CH 3 
 /O C CH 9 CH CH 
 
 X 
 
 CH 2 CH 2 C C 
 
 I I 
 
 H CH 3 
 
 CH 3 
 
 0= C CH 2 CH 2 CH= 
 
 -> II II 
 
 0= ==0 
 
 + 
 CH 3 .CO.CH 2 CH 2 .CHO 
 
 On continued boiling of the aqueous solution the levulinic 
 aldehyde peroxide undergoes autoxidation to levulinic acid, 
 and the a and ft lactones of this acid 
 
 CH 3 
 
 0=C CH 2 CH 2 CH-0 
 
 II [I -> CH 3 . CO . CH 2 CHjjCOOH 
 
 .Q Levulinic acid. 
 
 CH 3 C : CH CH 2 
 
 -* I I 
 
 O- CO 
 
 Levulinic lactone. 
 
 Gottlob (" Zeit. f. Anal. Chem.," 20, 2213, 1907) investigated 
 the action of ozone on many varieties of African rubbers, 
 especially those from Uganda and the Upper and Lower 
 Congo areas ; he obtained the following mean values for the 
 decomposition products : 
 
 Per Cent. 
 
 Yield. 
 Levulinic acid 49-8 
 
 aldehyde 25*8 
 
 ,, ,, peroxide 3'8 
 
 Resin 5-0 
 
INDUSTEIAL USES OF OZONE 173 
 
 Paulsen ("Le Caoutchouc et la Guttapercha," 7, 4177, 1913) 
 has shown that the ozonides of various resins, such as sandarac 
 and dammar, are precipitated by carbon tetrachloride, a 
 property which Dubosc and Luttringer (" Eubber, Its Pro- 
 duction, Chemistry and Synthesis," Griffin, 1918) has applied 
 to the estimation of rubber resins. Molecular weight deter- 
 mination has shown that the natural rubber molecule is 
 exceedingly complex, whilst the ozonide consists of but two 
 molecules of ozonised dimethyl cyclo octadiene. It there- 
 fore follows that ozone exerts a depolymerising action on the 
 rubber molecule. 
 
 Attempts to prepare dimethyl octadiene itself by reduction 
 of the ozonide have, however, proved fruitless. 
 
 Cyclo octadiene, the simplest of the cyclo octane deriva- 
 tions, has, however, been isolated by Willstatter, and on 
 ozonisation and hydrolysis this yields succinic dialdehyde, 
 whilst on polymerisation it yields a product very similar to 
 natural rubber 
 
 CH CH 2 CH 2 CH n 2 . CHO . CH 2 . CH 2 CHO 
 
 ' 
 
 CH CH 2 CH 2 CH M 
 
 On this evidence, supported by his previous work on the 
 nitrosites and tetrabromide of rubber, Harries adopted the 
 somewhat unusual eight-ring structure as the unit in the 
 polymer of natural rubber. Although Harries' views have 
 received wide acceptation, yet this theory is contested by 
 several workers in the field; notably by Pickles, who ad- 
 vances various arguments why rubber should be represented 
 as an open-chain polymer : 
 
174 OZONE 
 
 JOB. 
 
 I = C . CH 2 . CH 2 CH -J. 
 
 the ozonide of which would naturally possess the following 
 structure : 
 
 CH 3 
 
 I /O 
 
 C CH 2 CH 2 CH X 
 
 A few applications of the use of ozonised air have been 
 made in preparative inorganic chemistry ; thus the oxidation 
 of manganates to permanganates, chlorates to perchlorates, 
 ferrous chloride to ferric chloride are reactions which pro- 
 ceed smoothly and rapidly with the aid of ozone. The 
 preparation of permanganates by means of ozone is said to 
 possess advantages over the usual chemical methods of 
 manufacture. 
 
 APPLICATION TO BEEWING AND FOOD PRESERVATION. 
 
 Ozonised air has found increasing application in the 
 brewery, not only to prevent the ingress of adventitious 
 micro-organisms during the process of fermentation and 
 cooling of the wort, but also the refrigerating and bottling 
 the beer. By enclosure of the fermenting tuns and the 
 cascade coolers in a suitable air shaft high concentrations of 
 ozone may be used, which ensures the sterility of the air in 
 contact with the liquid. Minor applications are found in the 
 treatment of filtering material, the cleansing of clarifying 
 chips and the sterilisation of bottles and casks. 
 
 According to V. Vetter (" Zeit. f. Brauerie," Feb., 1911) 
 all utensils, with the exception of rubber goods, which are 
 
INDUSTEIAL USES OF OZONE 175 
 
 rapidly attacked by ozone can be cleaned after washing with 
 water by subjecting these to an air current containing 0'5 gm. 
 ozone per cubic metre for half an hour. 
 
 Will and Wiensiger and V. Vetter (loc. cit.) have shown 
 that yeast has a higher power of resisting ozone than other 
 organisms met with in brewing. On this observation pro- 
 cesses of selective sterilisation of the fermenting liquid have 
 been devised in which by the aeration with ozonised air for a 
 suitable period of time all organisms such as sarcina, with 
 the exception of the yeast cells, are destroyed. It was found 
 that aeration with ozonised air containing 3 gms. per cubic 
 metre at the rate of 12 cubic metres per hour per kilogram of 
 pressed yeast, for a period of from fifteen to twenty minutes, 
 ensured the production of a normal and energetic fermenta- 
 tion. It was further claimed that the flavour of the beer 
 was unchanged, its keeping qualities improved, together with 
 its power of resisting infection on storage. In forcing tray 
 experiments beer from untreated yeast turned at the end of 
 thirty-two days, whereas that from ozonised yeast remained 
 good for eighty-six days, showing only a slight haze at the 
 end of this period. Similar applications of ozone in the 
 other fermentative industries, such as the manufacture of 
 wines, cider, perry, alcohol and vinegar, have been frequently 
 proposed. From time to time proposals have been advanced 
 to accelerate the normal ageing of wines and especially spirits 
 by treatment with ozonised air. According to De la Coux 
 (" L'Ozone," p. 378) the process of ageing is virtually one of 
 slow oxidation by means of atmospheric oxygen. Not only 
 is a small fraction of the alcohol oxidised direct to acetic acid, 
 as indicated by an increase in the quantity of ethyl acetate 
 
176 OZONE 
 
 in the spirit, but the bouquet is in part due to the formation 
 of acetal produced by interaction of aldehyde and alcohol in 
 the presence of ozonised oxygen. Resinous matter, which in 
 the normal process of ageing is precipitated from the wine or 
 spirit, is, it is said, also removed by treatment with ozone 
 (together with fusel oils from whisky). De la Coux likewise 
 records an improvement in colour. Such a process of 
 artificial ageing, first proposed by Pasteur, has been at- 
 tempted on a small industrial scale by numerous investi- 
 gators, notably Villon, Broyer, and Petit, and others. The 
 type of plant employed follows closely on those adopted for 
 the sterilisation of water, either plate towers or spray 
 systems being utilised to ensure intimate contact between 
 the ozonised air and the liquid. For the artificial ageing of 
 wines it is said that 20 to 40 litres of oxygen should be 
 utilised per hectolitre of liquid by the continued passage of 
 cooled ozonised air for a suitable length of time. The wine 
 so treated, it is claimed, will become fully mature in from 
 two to three months. 
 
 For spirits some 50 litres should be utilised per hectolitre. 
 Three days after treatment the spirit is clarified by precipita- 
 tion with magnesia or filtration through suitable clarifying 
 agents and reozonised, until another 50 litres of oxygen are 
 absorbed. The process is repeated three or four times and 
 the spirit finally stood for a few months. Villon claims that 
 twenty year old cognac may thus be prepared in less than 
 six months. 
 
 Although these claims are distinctly interesting, and any 
 method of rapidly maturing wines and spirits would possess 
 great economic advantages, yet it must be confessed that 
 
INDUSTRIAL USES OF OZONE 177 
 
 apart from the fact that no process of industrial ageing on 
 these lines appears to be in actual operation, ageing is prob- 
 ably not entirely a process of oxidation but results from a 
 great number of chemical reactions produced from enzyme 
 activity taking place but slowly in the wine. Probably not 
 the least important are the proteoclastic ferments effecting 
 the gradual hydrolysis of the small quantities of protein sub- 
 stances present in the liquid. 
 
 A useful field for the application of ozonised air is to be 
 found in the preservation of food, especially in connection 
 with refrigeration. 
 
 For the prolonged storage of fresh meat it is necessary to 
 maintain it at a low temperature, in order to lower the rate 
 of hydrolysis both proteoclastic and lipoclastic produced in 
 the meat by the naturally occurring enzyme. At normal 
 temperatures meat can only be preserved for a few days 
 without its quality being seriously affected by such changes 
 occurring. During storage and transit not only are the sub- 
 stances subjected to internal attack by the natural enzymes 
 but frequently external sources of contamination are to be 
 found, especially flies and air-borne micro-organisms. 
 
 Thus maintaining perishable foodstuffs at a low tempera- 
 ture and in a sterile atmosphere ideal conditions for preserva- 
 tion obtain. Several large refrigerating warehouses and ship 
 holds have been equipped with ozonisers on the air-circulating 
 systems. 
 
 We have already alluded to the enhanced efficiency of 
 ozonisers at low temperatures, hence the conjunction of an 
 ozoniser in a refrigerating system is a particularly economical 
 
 installation. 
 
 12 
 
178 OZONE 
 
 Useful applications for ozone are likewise to be found in 
 the drying of copra, which, when subjected to the ordinary 
 sun-drying process, is liable to acquire an exceedingly offensive 
 smell. The conditioning of air in flour mills by means of 
 ozonised air is said to be attended with a possible increase 
 in the mill capacity of 30 per cent, and the practical elimina- 
 tion of the flour moth. 
 
CHAPTEE X. 
 
 METHODS OF DETECTION AND ANALYSIS. 
 
 THE presence of one part of ozone in a million of air can be 
 detected by means of its characteristic odour, which, however, 
 is liable to be confused with that of dilute chlorine or nitrogen 
 peroxide. 
 
 In common with other powerful oxidising agents, it will 
 readily liberate iodine from the usual starch iodide papers, 
 colouring them a brilliant blue, a method employed by Schon- 
 bein, "Wolffhugel and Van Bastelaer. 
 
 Houzeau indicated that by a simple modification of the 
 test paper ozone could be distinguished from acid oxidants, 
 such as chlorine or nitrogen peroxide. A strip of filter paper 
 is impregnated with a solution of neutral potassium iodide 
 and one half is then treated with starch and the other with 
 an alkali indicator such as phenolphtalein or rosolic acid. 
 Ozone is sharply distinguished from chlorine and nitrogen 
 peroxide by liberating both iodine and alkali from neutral 
 potassium iodide 
 
 3 -+ 2KI + H 2 = 2 + 2KOH + I 2 . 
 
 The paper is not entirely diagnostic for ozone, since carbon 
 dioxide will slowly liberate iodine from neutral potassium 
 iodide solutions and will not turn the paper distinctly acid. 
 Hydrogen peroxide will give the same indications as ozone 
 
 itself. Iodine is also set free by photolysis (Loew, " Zeit. f. 
 
 (179) 
 
180 OZONE 
 
 Chein.," 5, 625, 1869), and the papers should be guarded from 
 direct sunlight. Guiacum test papers are turned blue by 
 ozone in common with other oxidants. 
 
 Cazeneuve has shown that m.-phenylene diamine test 
 papers are sensitive to oxidants and that ozone can be dis- 
 tinguished from hydrogen peroxide, since the former gives a 
 brown coloration and the latter an intense blue. 
 
 Arnold and Mentzel (" Ber.," 35, 1324, 2902, 1902), as a 
 result of a series of experiments, showed that benzidine and 
 dimethyl p.-phenylene diamine, or better the tetramethyl 
 derivative (the sensitivity to oxidising agents increases with 
 the number of methyl groups inserted), were extremely 
 sensitive and at the same time gave a ready means of dis- 
 tinguishing between the different oxidising substances, as 
 indicated by the following table : 
 
 Colour shown by : 
 
 Oxidant. Benzidine. Tetramethyl 
 
 Base. 
 
 Ozone Brown. Violet. 
 
 Nitrous acid .... Blue. Straw yellow. 
 
 Halogens .... Blue then red. Deep blue. 
 
 Hydrogen peroxide . . nil. nil. 
 
 Mention may be made of the following test papers which 
 indicate the presence of ozone : Silver foil is coloured black 
 by the formation of the somewhat unstable silver oxide in 
 the presence of ozone ; hydrogen sulphide will, of course, 
 form a sulphide coloration somewhat similar to that of 
 oxide ; lead acetate paper is sufficiently diagnostic of hydro- 
 gen sulphide ; whilst lead sulphide paper is bleached by ozone 
 and hydrogen peroxide owing to conversion into lead sulphate. 
 
METHODS OF DETECTION AND ANALYSIS 181 
 
 Manchot (" Ber.," 39, 3570, 40, 289, 1907, 42, 3948, 1908) 
 notes that silver is extremely sensitive to the presence of a 
 little metallic iron as catalytic agent, a coloration is easily 
 produced by O'Ol per cent, ozone, and he claims this to be 
 more sensitive than the tetramethyl base paper. 
 
 Manganous sulphate impregnated filter papers turn brown 
 in the presence of ozone, due to the formation of Mn 2 O 3 . 
 Manganous oxide may, of course, be formed if any alkali be 
 present in the gas, e.g. ammonia, and this in turn will undergo 
 atmospheric oxidation to the coloured manganese oxide, 
 especially in the presence of light (Danhary, "J.C.S.," 5, 
 1, 1867). Thallous oxide is converted into the brown thallic 
 oxide T1 2 3 by the action of ozone. Nitrous acid is without 
 effect, since the nitrite and nitrate of thallium are not coloured. 
 Halogens and hydrogen sulphide, however, produce a brown 
 coloration, the former due to oxidation and the latter due to 
 conversion into a coloured sulphide. Carbonic acid present 
 in the gas to be detected causes a considerable decrease in 
 sensitivity of thallous oxide paper owing to conversion to the 
 somewhat insoluble carbonate. 
 
 Various investigators, notably Poe'y and Berigny, have 
 used these test papers in the form of long strips in a suitable 
 recorder mechanism for the continuous detection of ozone in 
 gases. By means of a simple clockwork escapement a small 
 piece of a ribbon of impregnated paper is exposed to the gas 
 stream for a short period, and by noting the time and coloration 
 of the paper the presence of ozone in the gas at any time 
 during the period of operation can easily be detected. 
 
 Some attempts have been made to convert these so-called 
 cronozoscopes into cronozometers for giving some idea as to 
 
182 OZONE 
 
 the quantity of ozone in the gas at different intervals of time. 
 These experiments have usually been directed along one of 
 the following lines : either the time of exposure is increased 
 until the test slip becomes sufficiently coloured to be indis- 
 tinguishable from a standard colour, when the amount of 
 ozone present is naturally inversely proportional to the time ; 
 or a series of standard colours are made up and each test slip 
 is exposed for a definite and constant time interval. It would 
 appear that the former method gave more accurate results. 
 It is evident that human control is necessary for this type of 
 cronozometer, but a simple mechanical mechanism could 
 doubtless be fitted to make the machine automatic and not 
 merely semi-automatic in action. Thus, for example, the 
 difference in reflecting powers of various shades of thallium 
 oxide T1 2 3 could be made to actuate a system of balanced, 
 thermocouples. 
 
 METHODS OF ESTIMATION. 
 
 (a) Iodide Method. Bunsen's method of estimating ozone 
 by the liberation of an equivalent of iodide from a neutral 
 solution of potassium iodide according to the equation 
 
 (1) 2KI + O 3 + H 2 = 2KOH + I 2 + 2 , 
 followed by back titration after acidification with sodium 
 thiosulphate or sodium hydrogen sulphite, using starch as 
 indicator, is liable to give unsatisfactory results owing to the 
 further oxidation of the liberated iodine into iodite, iodate, 
 and periodate (Garzarolli, " Thurnlackh. Monatsh.," 22, 
 455, 1901). With a large excess of potassium iodide and in 
 a slightly acid solution, however, the sensible error due to the 
 formation of iodate and periodate is not large. According to 
 
METHODS OF DETECTION AND ANALYSIS 183 
 
 Ladenburg, hydrogen peroxide is formed under these condi- 
 tions (" Ber.," V., 34, 1187, 1901) 
 
 (2) 40 3 + 10HI = 5I 2 + 4H 2 + 30 2 + H 2 2 , 
 
 and although a slight loss may result, due to interaction of 
 the ozone and the hydrogen peroxide thus formed 
 
 H 2 2 + 3 = H 2 + 20 2 , 
 
 the results are usually somewhat higher than those determined 
 by physical methods, which we will shortly refer to. Laden- 
 burg ("Ber.," 36, 115, 1903) obtained excellent results by 
 performing the estimation in the reverse manner, viz. passing 
 ozone through standardised sodium hydrogen sulphite and 
 back titrating with standard iodine solution. 
 
 Ingles (" J.C.S.," 98, 1010, 1903) showed that the acid 
 iodide method invariably gave high results, neutral solutions 
 yielding more accurate determinations. He found that 
 neutral potassium bromide solutions gave discordant results. 
 
 Houzeau's modification of the iodine method is extremely 
 accurate. A consideration of the equation (1) will indicate 
 that for every equivalent of ozone two equivalents of alkali 
 are liberated, and consequently the increase in alkalinity of 
 the solution gives a measure of the amount of ozone. A very 
 dilute sulphuric acid solution of potassium iodide is usually 
 employed, followed after absorption of the ozone by back 
 titration with standard alkali, using litmus or congo-red as 
 an indicator. The liberated iodine can be determined in the 
 neutral solution by means of sodium thiosulphate, thus giving 
 a check on the former figure. 
 
 E. H. Eiesenfeld and F. Bencher (" Zeit. Anorg. Chem.," 
 98, 167, 1916) investigated the effect of the addition of acid 
 
184 OZONE 
 
 to neutral potassium iodide solutions for the estimation of 
 ozone in great detail. They showed that in all cases, although 
 the main reaction proceeded according to the equation 
 
 3 + KI -> I + O 2J 
 a side reaction took place simultaneously 
 
 3 + SKI -> 31 
 
 The side reaction was found to be uninfluenced by the ozone 
 concentration in the gas, but greatly favoured by low tempera- 
 tures and relatively strong acid solutions. Any values be- 
 tween 1 and 3 atoms of iodine per molecule of ozone could 
 be obtained by altering these conditions, a value of 2*7 being 
 readily obtainable. 
 
 They suggest that the intermediary ions 10', I0 3 ', I0 4 ' play 
 a part in the reaction, and that in all solutions containing 
 ozone and potassium iodide an equilibrium between the fol- 
 lowing ions is invariably obtained, K, OH', I', 10', I0 3 ' and 
 I0 4 '. The production of oxozone (0 4 ), in the silent electric 
 discharge was suspected by Harries (" Zeit. Elektrochem.," 17, 
 629, 1911) as a result of analyses of the ozonised air by the 
 iodide method. It appears more than probable that the 
 above side reaction fully accounts for Harries' results. 
 
 N 
 
 Vosmaer (" Ozone," Constables, 1916) employs ^r sul- 
 phuric acid, and finds that no appreciable loss of accuracy 
 results by using acid of this strength. 
 
 (b) Arsenious Oxide Method. Thenard's method of esti- 
 mating ozone, similar to that employed for evaluating bleach- 
 ing powder or permanganate solutions, is based upon the 
 oxidation of arsenious acid to arsenic acid by these oxidising 
 agents, according to the equation 
 
METHODS OF DETECTION AND ANALYSIS 185 
 
 3As(OH) 3 + 3 + 3H 2 = 3As(OH) 5 . 
 
 A dilute solution of potassium arsenite is prepared by the 
 solution of arsenious oxide in potassium bicarbonate, and after 
 
 N 
 
 filtration is standardised by ^ n n iodine solution, using starch 
 
 iUUU 
 
 as indicator. 
 
 The ozonised air is metered after absorption of the ozone 
 by the potassium iodide or arsenious acid solutions (1 wash 
 bottle is sufficient for gas flow rates up to 10 litres per hour); 
 for higher flow rates a greater number must be employed. 
 De la Coux (" L'Ozone," p. 530) states that five 1-litre wash 
 bottles are ample up to 500 litres per hour. 
 
 Ozone concentrations are usually expressed in grammes 
 per cubic metre of air. It will be noted that in the case of 
 arsenious acid absorption the equivalent volume of oxygen 
 is not returned to the gas, as is the case in absorption by 
 means of potassium iodide, and consequently in this case the 
 gas suffers a diminution in volume when passing through the 
 absorbers; this correction, however, is but a small one. (4*8 
 gms. 3 per cubic metre would give a diminution of but 
 0'224 per cent.) 
 
 (c) David (" O.K.," 164, 430, 1917) suggests the use of 
 
 N 
 
 z^. ferrous ammonium sulphate solution, slightly acidified 
 lUU 
 
 with sulphuric acid, as absorbent ; back titration is accom- 
 plished with potassium permanganate. It is stated that the 
 solution is unaffected by air at this concentration. 
 
 (d) Physical Methods. Otto has made use of a barograph 
 for obtaining a continuous record of the density of a stream 
 of ozonised air ; it is claimed that the apparatus is extremely 
 
186 OZONE 
 
 sensitive owing to the great difference in densities between 
 ozone and oxygen. The same investigator also devised a 
 dilatometer for the estimation of ozone based upon the in- 
 crease of volume which ozone undergoes when subject to 
 thermal decomposition. A 500 c.c. flask, terminating in a 
 graduated neck, is filled at atmospheric pressure and at a given 
 temperature with the ozonised oxygen. The flask is inverted 
 and the graduated neck is immersed in a mercury bath, and 
 the ozone is then decomposed by heat. Boiling amyl benzoate 
 (b.p. 261 C.) has been found to be a suitable substance for 
 this purpose. The increase in volume after cooling to the 
 original temperature and readjusting to the original pressure 
 is noted and thence the ozone content of the gas can be cal- 
 culated from the equation 
 
 20 3 = 30 2 . 
 
 Otto further devised an optical method based on the principle 
 of the tintometer, more recently applied by Lovibond to similar 
 purposes. A series of coloured cobalt-blue glasses serve as 
 standards of comparison with a tube of definite length of 
 ozonised air or oxygen under standard conditions of tempera- 
 ture and pressure. 
 
 F. Kriiger and M. Moeller (" Physik. Zeit.," 13, 7J9, 1912) 
 have suggested the measurement of ozone concentrations by 
 the absorption of ultra-violet light. The maximum absorp- 
 tion of ultra-violet light by ozone is found in the region X = 
 200 to 300 /JL/JL, especially at X = 254 pp. According to Beer's 
 law the absorption coefficient may be expressed in the form 
 I = I o e- ked or log I = log I _ ked , where I is the initial inten- 
 sity of the ultra-violet light, I the intensity after absorp- 
 tion by a layer of ozonised oxygen d cms. thick containing 
 
METHODS OF DETECTION AND ANALYSIS 187 
 
 e gms. per cubic metre, and K the absorption coefficient for 
 ozone ; thus, under constant illumination in a tube of con- 
 stant length, the ozone concentration is proportional to the 
 logarithm of the intensity of the emergent ultra-violet light, 
 determined by means of a potassium photo-electric cell. 
 
 None of these methods of physical analysis have, however, 
 received technical application, reliance having usually been 
 placed on some modification of the iodide volumetric method. 
 
 The analysis of gases containing ozone and other oxidising 
 agents has been the subject of investigation by Tommasi, as 
 early as 1879 (" Chem. News.," 29, 289, 1874). 
 
 The gases containing ozone and chlorine or nitrous acid 
 are passed into a normal solution of potassium ferrocyanide, 
 and the total oxidising power determined by the conversion 
 to potassium ferricyanide effected. Another portion of the 
 gas is then passed over hot platinum black, or through a hot 
 tube containing manganese dioxide, when the ozone is de- 
 stroyed. The potassium ferrocyanide conversion is then de- 
 termined, and from the difference in the two estimates the 
 ozone content of the gas can be determined. 
 
 Analysis of mixtures containing ozone and the oxides of 
 nitrogen may also be accomplished by passing the gases 
 into liquid air, when the oxides of nitrogen are solidified and 
 may be separately determined. 
 
 Hauser and Herzfeld ("Ber.," 45, 3575, 1912) cite an in- 
 teresting method for the analysis of small quantities of 
 methane, which, it would appear, would also be applicable to 
 the estimation of ozone. They note that methane is quanti- 
 tatively oxidised at ordinary temperatures to formaldehyde 
 by ozone according to the equation 
 
188 OZONE 
 
 CH 4 + 20 3 = HCHO + H 2 + 2 . 
 
 It has been suggested that this method of ozone estimation 
 might be applied to the detection of electrical leaks and corona 
 discharge in tunnels through which insulated high-tension 
 electric cables are led (e.g. the Lotschberg Simplon Tunnel), 
 as the amount of ozone resulting from the ionisation of the 
 air in the tunnel would give an approximate idea as to the 
 magnitude of the electrical leak. 
 
 An instrument for the detection and estimation of dissolved 
 ozone in water was devised some years ago by U. Evans 
 and the author (" An Electro-chemical Indicator for Oxidising 
 Agents," " The Analyst," August, 1913). This consists essen- 
 tially of a small cell formed by a platinum rod surrounded 
 by a copper tube. The water containing the ozone flows 
 through the annular space between the platinum and the 
 copper at a good rate, and forms the electrolyte of the cell, 
 whilst the platinum rod and the copper tube forming the 
 electrodes of the cell are connected to a microammeter or 
 thread recorder. In the absence of any oxidising agents in 
 the water, the small cell rapidly becomes polarised, the current 
 flowing through the microammeter sinks to zero, and the 
 platinum becomes charged with hydrogen corresponding to 
 the electrolytic solution pressure of the copper in the water. 
 Most potable waters contain a quantity of dissolved salts to 
 make the internal resistance of the cell sufficiently low for 
 practical operation. On the addition of any oxidising agent 
 to the water, the cell is partly depolarised by the removal of 
 hydrogen from the platinum electrode, and the rate of re- 
 moval of hydrogen by the ozone is measured on the micro- 
 ammeter, 
 
METHODS OF DETECTION AND ANALYSIS 189 
 
 Since 96,540 coulombs are associated with 1 gm. equivalent 
 or 8 gms. of ozone, assuming the electrode reaction 
 
 60 3 = 30", 
 
 this quantity passing through the cell per second would 
 generate a maximum current of 96,540 amperes. If the 
 liquid flow rate were 1 c.c. per second, a fairly normal rate for 
 the instrument, one part of ozone in 10,000,000 of water would 
 correspond to a passage of 10~ 4 mgm. of ozone through the 
 cell per second, equivalent to a possible current of 12 x 10~ 4 
 amperes. It is, however, evident that all the ozone cannot 
 act as a depolariser, since half of it at least is wasted at the 
 other electrode, and for convenience of operation the cell and 
 flow rates are not so proportioned as to effect complete re- 
 duction of the ozone in the water flowing through the cell. 
 In actual operation the recorded current is about 25 per cent, 
 of the theoretical maximum. The instrument is remarkably 
 sensitive, easily estimating or recording one part of ozone in 
 10,000,000 of water or '00001 per cent. 
 
NAME INDEX. 
 
 ABRAHAM, 130, 143. 
 Allmand, 96. 
 Andrdoli, 125, 128, 129. 
 Andrews, 2, 58. 
 Archibald, 66. 
 Armstrong, 41, 132, 137. 
 Arnold, 50, 180. 
 Arny, 19. 
 Arrhenius, 61. 
 
 BACH, 42. 
 
 Baeyer, 14. 
 
 Balmer, 72. 
 
 Barus, 40. 
 
 Baumann, 40. 
 
 Baumert, 2, 57. 
 
 Becquerel, 2, 11. 
 
 Beger, 9. 
 
 Bencher, 183. 
 
 Bendixsohn, 64. 
 
 Berigny, 21, 22, 181. 
 
 Berthelot, 7, 59, 124. 
 
 Besson, 26. 
 
 Bineau, 16. 
 
 Binz, 154. 
 
 Birkeland, 24. 
 
 Blanc, 40. 
 
 Bloch, 11, 40. 
 
 Bodenstein, 141. 
 
 Bohr, 76, 90. 
 
 Boillot, 158. 
 
 Bonyssy, 20. 
 
 Bottger, 49. 
 
 Brauner, 47, 51, 64. 
 
 Brion, 92. 
 
 Erode, 51. 
 
 Brodie, 28, 38, 124, 126, 127. 
 
 Brooksbank, 10. 
 
 Broyer, 176. 
 
 Brunck, 33. 
 
 Buisson, 17, 18, 23, 86. 
 
 Bunsen, 182. 
 
 CAMLETTB, 152. 
 Carius, 7, 59. 
 Carlson, 154. 
 Cazeneuve, 180. 
 Chapman, 48, 134, 137. 
 
 Chappuis, 6, 9, 141. 
 Chassy, 102, 119. 
 Chatelain, 125. 
 Chlopin, 85. 
 Chree, 26. 
 Clark, 134. 
 Clausius, 29, 38. 
 Clements, 50, 51, 52. 
 Cloez, 16. 
 Compton, 90. 
 Cramp, 92. 
 Croze, 10. 
 
 Cruickshank, 1, 57. 
 Cunningham, 159. 
 Curie, 86. 
 Curie, 157. 
 
 DAVID, 185. 
 
 De Christmas, 154. 
 
 De"combe, 119. 
 
 De Frise, 121, 122, 143, 146, 147. 
 
 De la Coux, 6, 31, 32, 58, 175, 176, 
 
 185. 
 
 Delandres, 72, 82. 
 De Marignac, 57. 
 De Meritens, 142. 
 De Varigny, 19. 
 Dewar, 6, 8, 48, 49, 153, 154. 
 Dibden, 156. 
 Dixon, 41. 
 Donovan, 66. 
 Dore"e, 159. 
 Drugman, 166. 
 Duane, 87. 
 Dubosc, 173. 
 
 EGIDIUS, 155. 
 Ehrlich, 119. 
 Einstein, 71. 
 Elster, 39, 135. 
 Elworthy, 127. 
 Engler, 19, 39, 41, 42. 
 Erlwein, 109, 125, 126. 
 Evans, 141, 188. 
 
 FABRY, 17, 18, 23, 86. 
 Fanrweld, 95. 
 Finck, 53. 
 
 (191) 
 
192 
 
 OZONE 
 
 Fischer, 8, 47, 51, 52, 54, 55, 56, 64, 
 
 66, 69. 
 
 Fleming, 117. 
 Fowler, 10, 17, 23. 
 Franck, 81, 89. 
 Franklin, 155. 
 Freny, 2. 
 Friend, 37. 
 Fritsch, 83. 
 
 Frohlich, 13, 106, 125, 142. 
 Fulhame, 41. 
 
 GAIFFE, 127. 
 Gardner, 66. 
 Geitel, 39, 135. 
 Genin, 128. 
 Genthe, 36, 37, 164. 
 Gerard, 126. 
 Giesel, 87. 
 Goekel, 40. 
 Goldmann, 119. 
 Goldstein, 6, 10, 70. 
 Gottlob, 172. 
 Graham, 5. 
 Gray, 103, 104, 105. 
 Greaves, 134. 
 
 HABEB, 116. 
 
 Haeffner, 170. 
 
 Hallwachs, 119. 
 
 Harries, 15, 166, 168, 170, 171, 173, 184. 
 
 Hartley, 16, 18, 23. 
 
 Hatcher, 19. 
 
 Hauser, 187. 
 
 Hautefeuille, 6, 50, 51, 141. 
 
 Hayhurst, 20. 
 
 Hazura, 37. 
 
 Henriet, 20. 
 
 Hertz, 81, 89. 
 
 Herzfeld, 187. 
 
 Hollman, 7. 
 
 Holmes, 20. 
 
 Hoppe, 40. 
 
 Horton, 11, 75. 
 
 Houzeau, 11, 20, 22, 23, 30, 37, 158, 
 
 179, 183. 
 Hoveda, 118. 
 Hoyle, 92. 
 Hughes, 84. 
 
 ILOSVAY, 49. 
 Ingles, 183. 
 Inglis, 47, 63. 
 
 JAHN, 7, 132, 134. 
 Jellinek, 53. 
 Jones, 48, 134, 137. 
 Jordan, 154. 
 Jorrisen, 38. 
 
 KABAKJIAN, 100, 102, 103. 
 
 Kauchtschev, 107. 
 
 Kerr, 117. 
 
 Kirkby, 90. 
 
 Kissling, 37. 
 
 Kopp, 135. 
 
 Kreusler, 84. 
 
 Kron, 17. 
 
 Kriiger, 105, 186. 
 
 Kunz, 95. 
 
 LABBE, 154. 
 
 Labille, 132. 
 
 Ladenburg, 5, 6, 7, 9, 10, 18, 183. 
 
 Lande, 81. 
 
 Langer, 126. 
 
 Langheld, 170. 
 
 Lechner, 106. 
 
 Lehmann, 9, 10, 18. 
 
 Leithauser, 106, 109, 137. 
 
 Lenard, 10, 27, 70, 80. 
 
 Leroux, 50. 
 
 Lessing, 157. 
 
 Levi, 87. 
 
 Lind, 87. 
 
 Lipp, 117. 
 
 Lippert, 36. 
 
 Liveing, 43. 
 
 Loew, 49. 
 
 Lovibond, 186. 
 
 Lowry, 112, 113. 
 
 Luckiesh, 83. 
 
 Ludlam, 77. 
 
 Lummer, 79. 
 
 Luther, 47, 63. 
 
 Luttringer, 173. 
 
 Lyman, 80, 81, 83, 84. 
 
 MACKENZIE, 94. 
 McKendrick, 153. 
 McLellan, 26. 
 McLeod, 7, 60, 61, 65. 
 Mailfert, 7. 
 Malaquin, 31. 
 Manchot, 14, 136, 181. 
 Marie Davy, 22. 
 Marmier, 130, 143. 
 Marx, 51, 54. 
 Massenez, 64. 
 Meidinger, 57. 
 Meissner, 49. 
 Mentzel, 50, 180. 
 Mercanton, 25. 
 Meyer, 17, 23. 
 Moeller, 105, 186. 
 Moissan, 31. 
 Mond, 116. 
 Morris Airy, 80. 
 Moufgang, 8. 
 
NAME INDEX 
 
 193 
 
 Mulder, 135. 
 Miinchmeyer, 155. 
 
 NASINI, 21, 87. 
 
 Nernst, 29, 44, 47, 50, 53, 55, 61 
 
 76, 133. 
 Nesturch, 94. 
 
 ODDO, 27. 
 
 Ogier, 143. 
 
 Ohmiiller, 142, 143. 
 
 Olzweski, 6. 
 
 Ostwald, 42. 
 
 Otto, 4, 8, 123, 124, 128, 130, 132, 
 
 146-152, 185, 186. 
 Oudin, 125, 154. 
 Ozonair, 129, 147, 161. 
 
 PASCHEN, 72, 81. 
 Patin, 121. 
 Paulsen, 173. 
 Perman, 134. 
 Petit, 176. 
 Peyron, 22. 
 Pickles, 173. 
 Pincus, 49. 
 Planck, 71, 79. 
 Poey, 181. 
 Pollitzer, 45. 
 Prall, 142, 143. 
 Pre"poignot, 128. 
 Priestly, 22. 
 Pring, 20, 21, 85. 
 Pringsheim, 79. 
 Przibram, 40. 
 Puschin, 107. 
 
 QOAIN, 84. 
 
 EAMMBLSBEBQ, 32. 
 
 Ramsay, 87. 
 
 Rao, 21. 
 
 Regener, 10, 70, 138. 
 
 Reicher, 38. 
 
 Remsen, 7. 
 
 Renard, 169. 
 
 Repin, 143. 
 
 Richarz, 59. 
 
 Rideal, 151. 
 
 Riesenfeld, 12, 94, 116, 155, 183. 
 
 Rossi, 50. 
 
 Roth, 138. 
 
 Rothmund, 8, 12. 
 
 Roux, 143. 
 
 Runge, 72. 
 
 Russ, 119. 
 
 Rydberg, 72. 
 
 , 64, 
 
 143, 
 
 SALTMARSH, 140. 
 
 St. Edme, 66. 
 
 St. John, 82. 
 
 Schneller, 120, 121, 143. 
 
 Schonbein, 1, 16, 22, 29, 35, 39, 49, 57, 
 
 179. 
 
 Schone, 7, 9, 19, 58. 
 Schoner, 87. 
 Schott, 83. 
 Schultz, 154. 
 Schumann, 85. 
 Schuster, 10. 
 Schwarg, 155. 
 Segler, 40. 
 
 Shenstone, 133, 137, 141. 
 Siemens, 124, 125. 
 Simpson, 26. 
 Sleen, 120. 
 Soddy, 87. 
 Soret, 2, 3, 4, 5, 56. 
 Stark, 9, 10. 
 Steubing, 10, 11, 81. 
 Stokes, 157. 
 Strong, 86. 
 
 Strutt, 9, 17, 18, 23, 136, 137. 
 Struve, 49. 
 Swann, 26. 
 
 TAIT, 2. 
 
 Than, 49. 
 
 Thenard, 154, 184. 
 
 Thierry, 19. 
 
 Thomson, 11, 71, 73, 75, 94, 95. 
 
 Thornton, 25. 
 
 Threlfall, 95. 
 
 Tian, 81. 
 
 Toepler, 91, 92. 
 
 Tommasi, 187. 
 
 Townsend, 95. 
 
 Traube, 41. 
 
 Trillat, 167. 
 
 Troost, 7, 50, 51. 
 
 Tropsch, 8. 
 
 USHER, 21, 22. 
 
 VAN BASTELAER, 177. 
 
 Van de Meulen, 7, 135. 
 
 Van der Willigen, 49. 
 
 Van Ermengen, 143. 
 
 Van Marum, 1. 
 
 Van Sleen, 145, 152. 
 
 Van't Hoff, 29, 38, 39, 42, 61. 
 
 Vetter, 174, 175. 
 
 Villiger, 11. 
 
 Villon, 128, 176. 
 
 Vohr, 129. 
 
 Von Bahr, 138. 
 
 13 
 
194 
 
 OZONE 
 
 Von Wartenberg, 66, 111. 
 Vosmaer, 105, 121, 127, 129, 143, 145, 
 184. 
 
 WARBURG, 70, 76, 77, 95, 102, 103, 
 
 104, 105, 106, 109, 135, 137, 140. 
 Warner, 45. 
 Weger, 36. 
 
 Weigert, 76, 135, 138, 139, 141. 
 Well, 135. 
 Wendt, 87. 
 Wien, 79. 
 Wiensiger, 175. 
 Wild, 19, 41, 42. 
 
 Will, 175. 
 Williamson, 2, 56. 
 Willstatter, 173. 
 Wisse, 120. 
 Witz, 158. 
 Wohlwill, 67. 
 Wolff, 51. 
 Wolff hiigel, 179. 
 Wright, 170. 
 
 YARNOLD, 127. 
 
 ZENGHILIB, 49. 
 Zschimmer, 83. 
 
SUBJECT INDEX. 
 
 ACCEPTORS, 37, 42. 
 
 Active nitrogen, 113. 
 
 oxygen, 35, 37. 
 
 Air conditioning, 179. 
 
 Air-cooled ozonisers, 128. 
 
 Air purification, 13, 153-156. 
 
 Allotropes, 2, 10, 29. 
 
 Alternating currents, 67-69. 
 
 Altitude, effect of, on O 3 content, 17, 
 
 19, 20, 26. 
 Aluminium arc, 82. 
 
 oxidation, 49. 
 Ammonia, 151. 
 Ammonium ozonate, 14. 
 Analysis of, 180-187. 
 
 uses in, 168. 
 Anethol, 168. 
 Aniline black, 168. 
 Anisaldehyde, 166, 168. 
 Anode material, 58. 
 Anodic polarisation, 67. 
 Antozone, 28, 29, 39. 
 Antozonides, 2, 28, 39. 
 a particles, 87, 88, 90. 
 
 Arc discharge, 51, 80, 82, 93, 94, 120. 
 Arsenious oxide, 184. 
 Atmospheric ionisation, 24-27. 
 
 ozone, 17-27. 
 Atomic diameters, 73. 
 
 hydrogen, 28, 87. 
 
 oxygen, 29, 56. 
 
 structure, 72-76. 
 Aurora, 25. 
 Autocatalysis, 36, 37, 164. 
 
 Autoxidation, 1, 23, 29, 34, 35, 37, 38, 
 
 39. 
 
 BACTERIA, 142. 
 
 Baekelite, 115. 
 
 Band spectra, 9, 72, 73, 90. 
 
 Barium peroxide, 30, 31. 
 
 Barograph, 185. 
 
 Benzaldehyde, 35. 
 
 Benzene ozonide, 169. 
 
 Benzidine, 180. 
 
 derivatives, 180. 
 
 Black body radiators, 79, 80. 
 
 Boiling-point, 7. 
 
 Boric acid, 83. 
 Boro-silicate glass, 83. 
 Brewing, 175. 
 Brush discharge, 93, 95. 
 
 CALCIUM peroxide, 14. 
 Capacity influence, 100, 113. 
 Carbon arc, 79, 81. 
 
 monoxide oxidation, 165. 
 Catalytic autoxidation, 161, 164. 
 
 decomposition, 12, 30, 58, 67, 96, 
 
 134, 135, 136, 137, 181. 
 
 preparation, 49. 
 Cathode rays, 75. 
 Cellulose, action on, 158. 
 Chemical properties, 11. 
 Chemioluminescence, 169. 
 Chinese wood oil, 165. 
 Chlorates decomposition, 33. 
 Chlorine, action of, 112, 137, 141, 179, 
 
 180. 
 
 discharge, 94. 
 Chlorophyll, 22, 161. 
 Chromic acid, 38. 
 Cinnamon oil, 37. 
 Citronella, 35. 
 Cobalt-blue glass, 83. 
 Cognac ageing, 175. 
 Cold electrolysis, 57-69. 
 
 storage, 13, 177. 
 Colloidal platinum, 51. 
 Colour, 6, 8, 17, 51, 150. 
 Combustion, formation in, 35, 49, 51. 
 Condensers, 100. 
 
 Conditioning of textiles, 159. 
 Contact towers, 144. 
 Copra drying, 178. 
 Corona, 25, 95, 188. 
 
 pressure, 95. 
 Corrosion, 35, 129. 
 Cronozometers, 181. 
 Cronozoscopes, 181. 
 
 Current density influence, 59, 103. 
 
 DAMMAR, 173. 
 Decomposition, 134-139. 
 Density, 4, 5. 
 Deodorant action, 13, 154. 
 
 (195) 
 
196 
 
 OZONE 
 
 Depolymerisation, 173. 
 
 Detection, 180-187. 
 
 Dielectric ozonisers, 124. 
 
 Dielectrics, 113-125, 132. 
 
 Diffusion, 5, 145. 
 
 Drying oils, 13, 36, 37. 
 
 Dust, influence of, 17, 24, 84, 112. 
 
 ELECTRIC safety valve, 132. 
 Electrode materials, 58. 
 
 potentials, 47, 61, 63, 64. 
 Electrolysis, 31. 
 Electrolytic preparation, 1, 57. 
 Electron velocities, 25, 89, 105. 
 Emulsifiers, 146. 
 
 Endothermic actions, 7, 43, 44, 75. 
 Enzymes, 177. 
 
 Equilibrium, 29, 46, 48, 49. 
 Essential oils, 23. 
 Estimation, chemical, 48, 179. 
 
 electrical, 188. 
 
 photo-chemical, 187. 
 Ethylene linkages, 13, 15, 168. 
 Explosive properties, 6, 52, 169. 
 
 FAT bleaching, 13, 162, 163. 
 
 deodorising, 13, 162, 163. 
 Ferric chloride preparation, 174. 
 Fibres, action on, 158. 
 
 Field saturation in corona, 103. 
 
 Filtration, 144, 148. 
 
 Flour mills, 178. 
 
 Fluorine, 31. 
 
 Fluorite, 84. 
 
 Food preservation, 177. 
 
 Formaldehyde, 22, 166. 
 
 Formic acid, 166. 
 
 Frauenhofer lines, 17, 23. 
 
 Frozen equilibria, 49, 133, 135. 
 
 7 BAYS, 40. 
 Gas velocities, 109. 
 Germicidal action, 13, 142. 
 Glass, 83, 115, 117. 
 Guiacum test, 180. 
 
 HEAT of formation, 7, 77. 
 
 theorem, 45. 
 Heliotropin, 14, 167. 
 Hide preservation, 13. 
 Historical, 1. 
 
 Hydrogen peroxide, 2, 12, 22, 38, 41, 
 50-56, 59, 84, 180. 
 
 sulphide, 155. 
 
 vacuum, 82. 
 
 IMPEDANCE, 98. 
 Indigo, 35, 168. 
 Indol, 155. 
 
 Induction, 98. 
 
 Infra-red radiation, 9, 10, 74, 76. 
 
 Intermediate compounds, 41. 
 
 Interpolar distances, 114, 115. 
 
 Iodide estimation, 19-21, 182, 184. 
 
 Ionic theory, 61. 
 
 lonisation, 24, 25, 26, 77, 86, 88, 90. 
 
 Ionising potentials, 25, 81, 89. 
 
 Iron arc, 80. 
 
 Isatin, 35. 
 
 Isoeugenol, 167. 
 
 Isosafrol, 167. 
 
 JUTE, action on, 159. 
 KERR effect, 117. 
 
 LAVENDER, 23. 
 Leucobase oxidation, 163. 
 Levulinic acid, 171. 
 aldehyde, 172. 
 Limiting yield, 96. 
 Line spectra, 9, 72, 73-90. 
 Linoleum, 13, 165. 
 Linseed oil, 165. 
 Lithium ozonate, 14. 
 Liquefaction, 6. 
 
 MAGNETIC susceptibility, 11. 
 Manganese dioxide, 33, 58. 
 Manganous salts, action on, 12, 19, 181. 
 Matter, structure of, 73. 
 Mechanism of formation, 76. 
 Medicinal uses, 156. 
 Mercaptans, 155. 
 Mercury arc, 6, 70, 81. 
 M.-phenylene diamine, 180. 
 Methane oxidation, 165, 187. 
 Methyl alcohol, 166. 
 Mica, 115, 132. 
 Monatomic oxygen, 29, 56, 74, 78. 
 
 NASCENT state, 40. 
 
 Natural sources, 16. 
 
 Nitrogen peroxide, action of, 112. 
 
 detection of, 180. 
 
 formation of, 52-56. 
 
 Nitrosites, 173. 
 Non-dielectric ozonisers, 120. 
 
 ODOUR, 5, 6. 
 
 Oleic acid, 169. 
 
 Optical sensitation, 141. 
 
 Organic combustion, 12, 13, 142, 150, 
 
 152. 
 
 Oscillatory sparks, 102. 
 Oscillators, molecular, 73, 75. 
 Oscillograph, 105. 
 
SUBJECT INDEX 
 
 197 
 
 Oxides of metals, 33, 34. 
 
 nitrogen, 14, 16, 21, 25, 50, 54, 
 
 85, 112, 133. 
 
 Oxidising powers, 11, 169. 
 Oxozone, 15, 173, 174, 184. 
 Oxozonides, 15, 170. 
 Oxycellulose, 158. 
 Oxyhsemoglobin, 154, 158. 
 Ozobenzene, 169. 
 Ozonates, 14, 137. 
 Ozonic acid, 14. 
 Ozonide peroxides, 169. 
 Ozonides, 2, 14, 15, 41, 168-171. 
 Ozonised water, 21. 
 
 waxes, 157. 
 Ozonisers, 120-132. 
 
 PAINTS, 161. 
 
 Palladium, 35. 
 
 Pathogenic organisms, 149, 152. 
 
 Perchlorates, 13, 174. 
 
 Periodic acid, 32. 
 
 Periodicity, 68, 99, 106, 130. 
 
 Permanganates, 13, 30, 32, 66, 150, 
 
 153, 169, 174. 
 
 Peroxides, 30, 31, 33, 37, 42, 136, 164. 
 Persulphates, 31, 59. 
 Phosphorescence, 8, 9. 
 Phosphoric acid electrolytes, 57, 66. 
 Phosphorus, 28, 38, 39, 40, 42. 
 Photo-chemical activity, 140. 
 
 decomposition, 23, 70, 138. 
 
 efficiency, 76-77. 
 
 production, 23, 70, 77, 138-141. 
 Photo-electric cells, 187. 
 
 effect, 119. 
 
 Photographic action, 159, 169, 171. 
 Photolysis, 180. 
 Physiological activity, 154. 
 Piperonal, 166. 
 Plate ozonisers, 127. 
 Platinum electrodes, 58-67, 120. 
 Point discharge, 118, 123. 
 Polymerisation, 28, 76, 161. 
 Poppy-seed oil, 165. 
 Positive carriers, 11, 40, 93. 
 Potassium chlorate, 33. 
 
 iodide, 2, 4, 11, 20, 157. 
 Potential electrode, 47, 58, 63. 
 Pressure, effect of, 46, 70, 76, 111, 
 
 134. 
 
 Pre-treatment of waters, 148. 
 Promoters, 164. 
 Pseudocatalysis, 164. 
 Pyrocatechaldehyde, 166. 
 
 QUANTA, 71, 72, 77, 89, 139, 141. 
 Quarte, 70, 84, 115, 116, 140. 
 
 RADIUM, 86, 88. 
 
 Rape oil, 165. 
 
 Rate of decomposition, 24, 49, 63, 133, 
 
 138, 137, 177. 
 Reaction rates, 12, 30, 53. 
 Refrigeration, 122, 177. 
 Residual charge, 115. 
 Rock salt, 84. 
 Rontgen rays, 88, 94. 
 Rotating electrode, 124. 
 Roughing filters, 144, 148. 
 Rubber analysis, 15, 173, 175. 
 Rusting of metals, 35. 
 
 SAPROL, 167. 
 
 Sandal-wood oil, 23. 
 
 Sandarac, 173. 
 
 Saturation current, 93, 113. 
 
 Schumann rays, 80. 
 
 Seasonal variations, 20, 21. 
 
 Secondary production, 26, 86, 105. 
 
 Selective oxidation, 13, 161, 165. 
 
 Self-induction, 98. 
 
 Shellac, 115. 
 
 Siccatives, 161, 164. 
 
 Silent discharge, 25, 91, 95. 
 
 Silica, 83. 
 
 Silver, action of, 3, 33, 49, 136, 181. 
 
 Skatol, 155. 
 
 Solubility, 7, 8. 
 
 Spark discharge, 1, 51, 102, 118. 
 
 Specific heat, 45. 
 
 inductive capacity, 114, 115. 
 Spectra, 9, 10, 72, 73-90. 
 Spirit ageing, 179. 
 Spontaneous ionisation, 27. 
 Stability, 48. 
 
 Starch iodide tests, 16, 23, 180. 
 Static transformers, 97. 
 Storms, 21. 
 
 Sulphates, action on, 12. 
 Sulphuric acid electrolytes, 57. 
 
 TANNERIES, 153. 
 
 Temperature, influence of, 46-52, 58. 
 
 Tetrabase papers, 136, 180, 181. 
 
 Tetrabromides, 173. 
 
 Textile conditioning, 159. 
 
 Thallous salts, action on, 181. 
 
 Therapeutic uses, 156. 
 
 Thermal decomposition, 32, 111, 134. 
 
 equilibrium, 22, 45. 
 
 production, 44, 50, 51-56. 
 Thermionic emission, 93. 
 Thiosulphates, action on, 12. 
 Thunderbolts, 25. 
 Tintometer, 186. 
 
 Tubular ozonisers, 124. 
 Turpentine, 3, 25, 37. 
 
198 
 
 OZONE 
 
 ULTRA-VIOLET lamps, 70, 85. 
 
 light, 17, 23, 26, 70, 76, 79, 86, 107. 
 
 estimation by, 186. 
 
 Uviol glass, 83. 
 
 VANILLIN, 14, 166, 167. 
 Varnishes, 161. 
 Vaseline, ozonised, 157. 
 Velocity of decomposition, 49. 
 Voltage, influence of, 102. 
 perforating, 117, 118. 
 
 WATER-COOLED ozonisers, 130. 
 Waterfalls, ionisation in, 27. 
 Water, influence of, 58, 112, 137. 
 
 sterilisation, 142-153. 
 Wave form, influence of, 105. 
 Waxes, bleaching of, 13, 162, 163. 
 Wine ageing, 175. 
 
 Wound treatment, 13, 157. 
 
 YIELDS, chemical, 30, 31, 33. 
 
 electrolytic, 65, 69. 
 
 silent discharge, 111. 
 
 ABERDEEN: THE UNIVERSITY PRESS 
 
THIS BOOK IS DUE ON THE LAST DATE 
 STAMPED BELOW 
 
 AN INITIAL FINE OF 25 CENTS 
 
 WILL BE ASSESSED FOR FAILURE TO RETURN 
 THIS BOOK ON THE DATE DUE. THE PENALTY 
 WILL INCREASE TO 5O CENTS ON THE FOURTH 
 DAY AND TO $1.OO ON THE SEVENTH DAY 
 OVERDUE. 
 
 JAN 2 
 
 
 FEB M5 1&n 
 
 
 
 j 
 
 ILL 3 
 
 JUN 15ia3& 
 
 AUG 29 1947 
 
 
 
 
 APK 1 4 1948 
 
 AU8 2* 
 
 
 LIBRARY i; 
 
 'J 
 
 r8 * D 
 
 r ri D Xr i ^^^PJ^ 
 
 LAM -1 f\ 
 
 
 1951 
 
 JAN 19 
 
 1942 
 
 lOJan'SSMHf 
 
 *" 4 to,. 
 
 ^H:C-D ODF 
 
 
 
 JAW q igro 
 
 
 tf iC/OO 
 
 AUG 20 
 
 VJ4i: 
 
 
 
 
 
 NOV 6 
 
 1943 
 
 ^^^ 
 
 i 1/5 ^ <f 
 
 1945 
 
 s ^ 
 
 AUG Jt 
 
 J I Ot/ 
 
 . il? -. 
 ;ro LD- o5> 
 
 JAN 30 
 
 1947 
 
 ~OCT2" 
 
 2 % 652^rm 
 
 LD 2 1-10 C 
 
YU 18383 
 
 , 
 
 UNIVERSITY OF CALIFORNIA LIBRARY