SMOKELESS POWDER, 
 NITRO-CELLULOSE, 
 
 AND 
 
 THEORY OF THE CELLULOSE 
 MOLECULE 
 
 BY 
 
 JOHN B. BERNADOU, 
 
 Commander, United States Navy 
 
 FIRST EDITION 
 , FIRST THOUSAND 
 
 NEW YORK 
 
 JOHN WILEY & SONS 
 
 LONDON: CHAPMAN & HALL, LIMITED 
 1908 
 
^ 
 
 Copyright, 1901, 
 
 BY 
 JOHN B. BERNADOU 
 
 
 ROBERT DRUMMOND, PRINTER, NEW YORK 
 
PREFACE 
 
 FOR purposes of comparative study, the writer has 
 brought together in the present volume a series of 
 papers, by various investigators, upon the composition 
 of cellulose and the properties of explosives prepared 
 therefrom. He has supplemented these with an ac- 
 count of experiments made by himself; and from the 
 whole has drawn certain conclusions as to the possible 
 ultimate chemical composition of cellulose and the 
 nitro-celluloses. 
 
 While the general development of war-material 
 from the mechanical and metallurgical standpoints 
 the production of ordnance and armor is so largely 
 identified with progress in the useful arts in the 
 United States, yet, until very recently, but little has 
 been accomplished in our country in the way of im- 
 provement in explosives. Within the last few years, 
 however, a particular form of smokeless powder has 
 largely supplanted the old black and brown powders 
 for military uses; and the last decade of the past 
 century has witnessed the virtual abandonment of 
 <*, propellant that has held its place in war, with 
 
 361224 m 
 
IV PREFACE 
 
 comparatively little modification, for four hundred 
 years. 
 
 This new smokeless powder, which is adapted for 
 use in arms of all calibres, is prepared from a particu- 
 lar type of colloid nitro-cellulose. Such an extension 
 of the employment of this latter body from its original 
 use for detonating purposes, to its new use as a pro- 
 gressive explosive, has attracted general attention, and 
 led to a more careful and extended study of the nitro- 
 celluloses in general. It is with the view of further 
 extending such study and of possibly preparing the 
 way for the introduction of future improvements in 
 progressive explosives that this book has been pre- 
 pared. 
 
 Before presenting it, the author wishes to express 
 his thanks to certain eminent scientists for the privi- 
 lege that they have courteously afforded him of 
 making his own translations of certain portions of their 
 works upon explosives: to Professor D. Mendeleef, 
 of Russia, for his paper entitled " Pyrocollodion 
 Smokeless Powders"; to M. Vieille, of the French 
 Service des Poudres et Salpetres, for his article upon 
 the nitration of cotton ; and to M. Bruley, of the same 
 service, for a similar paper. Thanks are due also to 
 Messrs. Longmans, Green & Co. for the privilege 
 kindly extended of making certain extracts from 
 Messrs. Cross and Bevan's valuable work, ' ' Cellulose, 
 published by them. 
 
 Finally, the author wishes to express his indebted- 
 ness to Dr. Alfred I. Colin, of New York, for the care 
 he has bestowed upon the reading of the proof of 
 
PREFA CE V 
 
 this book at a. time when the writer's absence from 
 the United States on active service afloat prevented 
 his giving the matter the personal care and attention 
 it otherwise would have had; and for his prepara- 
 tion of a comprehensive index. 
 
 U. S. S. " Dixie/' April 24, 1901. 
 
AUTHOR'S NOTE, 1908 
 
 SEVEN years have elapsed since the publication of the 
 first edition of the present work. With the exception 
 of minor modifications of processes previously established, 
 the interval has witnessed no radical changes in the 
 methods of manufacture of or the composition of smoke- 
 less powder. The substitution of centrifugals for pots 
 in the nitrating process to promote rapidity of nitration 
 a step of doubtful expediency and the introduction 
 into the solvent of a small amount of an ether less volatile 
 than the ethylic ether, to prevent the too rapid "drying 
 out" of the powder, .represent the most important changes 
 made. 
 
 Several years ago Mr. George W. Patterson, the able 
 chemist in charge of the laboratories at the U. S. Govern- 
 ment powder works at Indian Head, Maryland, re- 
 marked to me that he did not believe that "gun-cotton" 
 was soluble in ethylic ether, even at as low a temperature 
 as that produced by liquid air; and he proposed that we 
 conduct a second series of experiments, employing as 
 the solvent absolute ether dehydrated over sodium. 
 Results showed that the "gun-cotton" was not soluble 
 in the modified solvent and at the low temperature stated; 
 whereas, it was soluble in the Squibb 's ether that I had 
 
 vii 
 
viii PREFACE 
 
 previously employed and which was marked "Squibb's 
 Ether, C. P., for Anaesthesia." This Squibb's ether was 
 subsequently found to contain traces of ethyl alcohol. 
 
 I had but one morning at my disposal in which to make 
 these experiments, and so could make no effort to deter- 
 mine whether or not ethylic ether treated with sodium 
 contained traces of some other impurity that acted to 
 render it incapable of dissolving the gun-cotton. The 
 presence in the ethlyic ether of exceedingly small quanti 
 ties of ethyl alcohol on the one hand, or of sodium on 
 the other, may have developed katalytic tendencies of 
 diametrically opposite character, the one tending to 
 promote, the other to prevent, the solution of the "gun- 
 cotton" under the conditions stated. 
 
 But the fact has been brought out clearly that, starting 
 from the 2-1 ethylic-ether ethyl- alcohol compound 
 solvent, the greater the diminution of the amount of 
 alcohol present the lower the temperature required to 
 effect the solution of the "gun-cotton." As this is in 
 accordance with experimental results already cited in 
 Chapter IV, upon which the development of the theory 
 of the nitro-cellulose molecule is based, no change will 
 be made in that part of the present work. Attention is 
 called, however, to the fact that the use of ethylic ether 
 alone as a solvent, in connection with insoluble nitro- 
 cellulose of highest nitration, probably represents a limit 
 at which solution can only be obtained through the 
 employment of some temperature not far from absolute 
 zero. 
 
 ROME, Italy, April 28, 1907. 
 
TABLE OF CONTENTS 
 
 CHAPTER I 
 
 PAGE 
 
 Origin 1 
 
 Nomenclature 4 
 
 Definitions 5 
 
 CHAPTER II 
 
 EARLIER VIEWS AS TO NITRO-CELLULOSE COMPOSITION AND 
 CONSTITUTION . 8 
 
 CHAPTER HI 
 
 THE CONCEPTION OF PROGRESSION IN RELATION TO COMPO- 
 SITION AND CONSTITUTION 21 
 
 CHAPTER IV 
 
 SOLUTIONS OP NITRO-CELLULOSE. THEORY OF THE CELLU- 
 LOSE MOLECULE.. . 38 
 
 APPENDIX I 
 
 RESEARCHES UPON THE NITRATION OF COTTON. BY M. 
 
 VIEILLE 81 
 
 I. First Method of Nitration 81 
 
 II. Second Method of Nitration 87 
 
 III. Higher Limit of Degree of Nitration 91 
 
 IV. Cellulose Incompletely Nitrated 92 
 
 V. Resume ,nd Conclusions 94 
 
 ix 
 
X TABLE OF CONTENTS 
 
 APPENDIX II 
 
 PACE 
 
 PYROCOLLODIOW SMOKELESS POWDERS. BY PROFESSOR D. 
 
 MENDELEEF 97 
 
 APPENDIX III 
 
 THE NITRATION OF COTTON. BY M. BRULEY 127 
 
 I. Method of Nitration and Mode of Representing Results. . 130 
 
 II. Results Obtained by M. Vieille 131 
 
 III. Experiments in Nitration (First Series) 134 
 
 IV. Experiments in Nitration ( Second Series) 140 
 
 V. Experiments in Nitration (Third Series) 141 
 
 VI. Various Experiments 144 
 
 VII. Resume and Conclusions 152 
 
 APPENDIX IV 
 
 THE DEVELOPMENT OF SMOKELESS POWDER. BY COMMANDER 
 
 JOHN B. BERNADOU, U. S. N 1 56 
 
 INDEX . . 195 
 
SMOKELESS POWDER 
 
 CHAPTER I 
 ORIGIN 
 
 THE discovery that cellulose, by treatment with 
 nitric acid, is converted into a highly inflammable or 
 explosive body was made during the first half of the 
 nineteenth century. The action of nitric acid on 
 starch was investigated to some extent in 1833 by 
 Braconnot, who found that a very rapidly burning 
 material was produced, and which he named xyloidine. 
 Pelouze further investigated this substance in 1838, 
 and also studied similar bodies prepared from paper, 
 linen, etc., which he held to be identical with the one 
 from starch. 
 
 In 1846 Schonbein discovered that cellulose in the 
 form of cotton, when immersed in nitric acid, freed 
 from the acid excess, and dried, was converted into a 
 highly explosive compound. This latter substance, 
 subsequently known as gun-cotton, constitutes the 
 base of modern smokeless powders.* 
 
 * The name " gun-cotton," as originally employed, was generic 
 for all varieties of the more highly explosive nitro-celluloses pre- 
 pared from cotton. 
 
2 SAfOR'ELESS PO WDER 
 
 On its first appearance gun-cotton was hailed by 
 ordnance experts as a smokeless substitute for gun- 
 powder, and extended series of experiments were con- 
 ducted to prove its adaptability to the requirements of 
 war. The combined efforts of chemist, powder-maker, 
 and artillerist failed, however, to secure such a control 
 of its combustion as would enable it to be safely em- 
 ployed, in charges of a given weight, to develop a uni- 
 form bore-pressure with standard projectiles in guns of 
 a given calibre. After the occurrence of a number of 
 detonations, unlocked for and more or less disastrous, 
 its use was abandoned. Nevertheless, the study of 
 the composition, properties, and methods of production 
 of gun-cotton and kindred bodies continued to be sys- 
 tematically prosecuted, the theory of nitro-substitution 
 was evolved, and the knowledge of the applicability of 
 the new materials to military requirements extended. 
 But more than forty years were to elapse from the 
 time of the first experiments before a thoroughly re- 
 liable smokeless powder, free from all tendency to 
 detonation, producing minimum erosion, of good 
 keeping qualities, and of maximum propulsive effect, 
 was to be obtained. 
 
 To trace the various steps of progress in the deter- 
 mination of the properties of gun-cotton is exceed- 
 ingly difficult. It is primarily difficult for the reason 
 that the base-material under investigation possesses 
 an organic structure ; and because the resultant nitrated 
 product varies in chemical composition according to 
 strength and temperature of the acids employed in 
 its preparation, their water content, duration of re- 
 
ORIGIN 3 
 
 action, and temperature at which the reaction is con- 
 ducted. It is difficult secondarily for the reason that 
 the ultimate product of nitration is either resolvable, 
 by treatment with certain solvents, into different sub- 
 components, or else is capable of direct preparation in 
 a form soluble in one or another solvent; and because 
 certain industries, sciences, and professions the man- 
 ufacture of explosives, of fabrics, photography, sur- 
 gery have required the development of special forms 
 of material and demanded special lines of research. 
 
 The existence to-day of a very large number c5f 
 unclassified names to denote the different varieties of 
 nitro-cellulose, serves as an illustration of the necessa- 
 rily involved and complex nature of such investigations 
 as have been made for the purpose of determining its 
 composition, and of the incompleteness of these inves- 
 tigations. The confusion in nomenclature that has 
 arisen is attributable to the fact that the formula for 
 cellulose, a.nd therefore for the cellulose nitrates, not 
 being established, investigators in different countries 
 started in independently to determine simultaneously 
 both the properties of these compounds and their 
 chemical constitutions. Matters became further in- 
 volved through the efforts to translate the accounts of 
 the work of one investigator into the language of 
 another. 
 
 With the view of avoiding confusion in the future it 
 may be well to anticipate somewhat here, to indicate 
 the origin of the various names applied to forms of 
 nitro-cellulcse, to define and classify them, and to 
 indicate such of them as will be employed hereafter in 
 
4 SMOKELESS POWDER 
 
 the body of the present work to denote those distinct 
 varieties that possess significance from the standpoint 
 of explosive effect. 
 
 NOMENCLATURE 
 
 Names may be divided into classes, according to 
 origin, as follows : 
 
 I. Those devised by discoverers to denote new 
 chemical compounds. Thus, xyloidin, from uAo=, 
 wood; pyroxylin, from Ttvp, fire, and <?t>Ao, wood. 
 
 2. Those implying uses to which special forms of 
 the material are applied, as gun-cotton, collodion-cot- 
 ton. 
 
 3. Those originating in references to the theory 
 of nitro-substitution ; thus, nitro-cellulose, mono-, di-, 
 tri-, tetra-, penta-, etc., nitro-cellulose. 
 
 4. Names based upon the physical characteristics 
 of the^material, as "soluble nitro-cellulose," "insol- 
 uble nitro-cellulose," " friable cottons" (Vieille). 
 
 5. Names obviously incorrect as to application and 
 meaning, as " soluble cotton " for soluble nitro-cellu- 
 lose; ** insoluble cotton " for insoluble nitro-cellulose. 
 
 6. Generic names for groups of varieties between 
 which differences in chemical composition were recog- 
 nized as existing; thus, nitro-celluloses as opposed to 
 nitro-hydrocelluloses. 
 
 7. Names referring to chemical composition as 
 modified by extent and character of nitro-substitution, 
 as " soluble nitro-cellulose of high nitration," of ' low 
 nitration." 
 
ORIGIN 5 
 
 DEFINITIONS 
 
 Cellulose. The cell-wall or envelope of plant-tissues, 
 to which the name cellulose has been applied as to a 
 chemical individual. Unless the contrary is stated, 
 the term cellulose, as employed in the present work, 
 refers to pure cotton, unbleached and unspun, freed by 
 mechanical treatment (ginning, picking, boiling, etc.) 
 from wood, dirt, greases, resins, and foreign matter in 
 general ; either in the natural state, or as waste product 
 from industrial processes. 
 
 Nitration. The displacement of a number of atoms 
 of replaceable hydrogen in cellulose, and the substitu- 
 tion therefor of the univalent radicle nitryl (NO 2 ). 
 The term " nitration " is also employed to indicate the 
 percentage of nitrogen in a given nitro-cellulose. 
 
 Nitro-cellulose. Products resulting from the treat- 
 ment of cellulose with strong nitric acid under the 
 condition that they retain the cellular structure of the 
 original cotton. 
 
 Nitro-cellulose of higJi nitration. Those forms of 
 nitro-cellulose in which a relatively large number of 
 the replaceable hydrogen atoms are replaced by nitryl. 
 
 Nitro-cellulose of mean nitration. Those forms of 
 nitro-cellulose in which a mean number of the replace- 
 able hydrogen atoms are replaced by nitryl. 
 
 Nitro-cellulose of low nitration. Those forms of 
 nitro-cellulose in which a relatively small number of 
 replaceable hydrogen atoms are replaced by nitryl. 
 
 Insoluble nitro-cellulose. Those forms of nitro-cel- 
 lulose of high nitration insoluble at ordinary atmos- 
 
SMOKELESS POWDER 
 
 pheric temperatures in a mixture of two parts by 
 weight of ethyl ether and one part by weight of ethyl 
 alcohol. 
 
 Soluble nitro-cellulose. Those forms of nitro- cellu- 
 lose of low or mean nitration soluble at ordinary at- 
 mospheric temperature in a mixture of two parts by 
 weight of ethyl ether and one part by weight of ethyl 
 alcohol. 
 
 Hydrocellulose. The product obtained by exposing 
 cotton to the action of hydrochloric-acid fumes ; or by 
 immersing it in hydrochloric, dilute sulphuric, or very 
 dilute nitric acid ; a white pulverulent mass which, 
 under the microscope, is seen to consist of fragments of 
 the original fibre of modified cellular form. 
 
 Nitro-hydrocellulose. Products resulting from the 
 treatment of hydrocellulose with strong nitric acid 
 under the condition that the resultant product retains 
 the cellular structure originally possessed by the hy- 
 drocellulose. 
 
 Nitro-hydrocellulose of high nitration. Those forms 
 of nitro-hydrocellulose in which a relatively large 
 number of the replaceable hydrogen atoms are replaced 
 by nitryl. 
 
 Nitro-hydrocellulose of mean nitration. Those 
 forms of nitro-hydrocellulose in which a mean number 
 of the replaceable hydrogen atoms are replaced by 
 nitryl. 
 
 Nitro-hydrocellulose of low nitration. Those forms 
 of nitro-hydrocellulose in which a relatively small 
 number of replaceable hydrogen atoms are replaced by 
 nitryl. 
 
ORIGIN 7 
 
 Insoluble nitro-hydrocellulose.- Those forms of nitro- 
 
 hydrocellulose of high nitration insoluble at ordinary 
 
 atmospheric temperatures in a mixture of two parts 
 
 -by weight of ethyl ether and one part by weight of 
 
 ethyl alcohol. 
 
 Soluble nitro-hydrocellulose. Those forms of nitro- 
 hydrocellulose of low or mean nitration soluble at 
 ordinary atmospheric temperatures in a mixture of 
 two parts by weight of ethyl ether and one part by 
 weight of ethyl alcohol. 
 
 Gun-cotton. The military name for those forms of 
 highly explosive nitro-celluloses employed in war, and 
 which are generally mixtures of a large quantity of in- 
 soluble with a small quantity of soluble nitro-cellulose 
 and a very small quantity of unnitrated cotton. 
 
 Pyrocellulose. Soluble nitro-cellulose of high uni- 
 form nitration, possessing a sufficient content of oxy- 
 gen to convert its carbon into carbonic oxide and its 
 hydrogen into aqueous vapor. 
 
 These terms, as defined above, will be employed so 
 far as possible in the body of this treatise. The follow- 
 ing list of synonyms is supplied for reference : 
 
 TABLE I 
 
 Name Synonym 
 
 Nitro-cellulose, Pyroxyline or pyroxylin. 
 
 Insoluble nitro-cellulose, Insoluble gun-cotton ; insolu- 
 
 ble cotton. 
 
 Soluble nitro-cellulose, Soluble gun-cotton; soluble 
 
 cotton; collodion-cotton; col- 
 lodion-pyroxylin. 
 
 Soluble nitro-cellulose of low Friable cotton, 
 nitration. 
 
CHAPTER II 
 
 EARLIER VIEWS AS TO NITROCELLULOSE COM- 
 POSITION AND CONSTITUTION 
 
 EXISTING knowledge of the composition and consti- 
 tution of nitro-cellulose still remains in a state of great 
 confusion. A number of papers, scattered through- 
 out the literature of experimental chemistry, and 
 which throw valuable light upon the subject, have been 
 published by independent investigators, while from 
 time to time more elaborate articles summarizing the 
 results of early workers have appeared. At least one 
 excellent treatise upon cellulose has been published, * 
 but it devotes only a few pages to the consideration 
 of the nitro-celluloses. The reason for the existing 
 confusion is that as yet no definite chemical structure 
 has been determined for nitro-celluloses ; they are re- 
 garded as nitro-substitution compounds, as ethers, or 
 else their composition is considered as still doubtful. 
 
 If, however, the various original experimental re- 
 sults and the conclusions that have been drawn there- 
 from be examined in sequence, it will be observed that 
 there does exist a tendency to account for the compo- 
 sition of these bodies on a definite hypothesis. There- 
 
 *" Cellulose," Cross and Bevan. London: Longmans, Green 
 & Co., 1895. 
 
 8 
 
EARLIER VIEWS AS TO NITRO-CELLULOSE 9 
 
 fore the writer, in his endeavor to throw some light 
 upon ultimate cellulose- and nitro-cellulose composi- 
 tion, will take up for consideration, first of all, in 
 natural order of succession, results obtained by a 
 number of students, each of whom, in his work, has 
 supplied material from which subsequent investigators 
 have drawn important conclusions. 
 
 Almost from the beginning of the study of the body 
 the existence of more than one form was recognized. 
 Domonte and Menard's discovery that pyroxylin of 
 low nitration was soluble in ether-alcohol, while the 
 more highly nitrated variety remained insoluble there- 
 in, furnished a positive differentiation of nitro-cellu- 
 loses into two groups; Bechamp showed that there 
 existed nitro-celluloses of different nitrations soluble 
 in ether-alcohol ; and the only way of explaining the 
 existence of such differences, in accordance with 
 chemical theory, was by assuming that the substances 
 of varying degrees of nitration obtained consisted of 
 mixtures of various quantities of different, definite, 
 chemical .compounds. 
 
 The numerous attempts to explain both the mode 
 of formation of nitro-cellulose as well as its constitu- 
 tion, and to reconcile results from its analyses, are well 
 illustrated by the formulae selected by early investi- 
 gators to represent its varieties. Of these formulas, 
 numerous series exist, extending from the original 
 formula of Schonbein, through the later formulated, 
 so-called mono-, di-, and trinitro-celluloses, to the 
 series of six nitrates of 'Eder. At an early date the 
 belief became established, based in all probability upon 
 
10 SMOKELESS POWDER 
 
 analogies furnished by the nitro-glycerins and nitro- 
 benzols, that nitro-celluloses were mixtures of nitro- 
 substitution products, which, assuming the composition 
 of cellulose as C 6 H 10 O B , were formulated as the tri-, di-, 
 and mononitrates, with compositions C 6 H 7 O 6 (NO 2 ) 8 , 
 C 6 H 8 5 (N0 2 ) 2 , and C 8 H,O 6 (NO 2 ), respectively. Of 
 these, trinitro-cellulose was considered as identical 
 with the insoluble variety of high nitration, and di- 
 nitro-cellulose with that soluble in ether-alcohol ; the 
 existence of the mononitro-cellulose was predicated. 
 Certain reactions due to these bodies led to a change 
 of views concerning their chemical structure. Bchamp 
 found that nitro-celluloses dissolved in ether-alcohol 
 surrendered nitric acid upon addition of potash or 
 ammonia, with the resultant formation of nitro-cellu- 
 loses of lower nitration ; he claimed for them the 
 composition of ethers.* 
 
 Other reactions of the nitro-celluloses tend to sup- 
 plement this view of their composition; e.g., ferric 
 chloride and potassium and ammonium sulphydrates 
 
 * An ether is one of a class of organic bodies divided into two 
 groups : (i) simple ethers, consisting of two basic hydrocarbon 
 radicles united by oxygen, and corresponding in constitution to 
 the metallic oxides, as CH 3 OCH 3 , methyl ether, or methyl oxide, 
 analogous to AgOAg, silver oxide ; (2) compound ethers, consist- 
 ing of one or more basic or alcohol radicles and one or more acid 
 or hydrocarbon radicles united by oxygen, and corresponding to 
 the salts of the metals, as CH 3 COOC 2 H 6 , ethyl acetate or acetic 
 ether, corresponding to CH 3 COONa, sodium acetate. Thus, if a 
 cellulose be represented as possessing a composition CeHioOs, 
 and be written as a tribasic alcohol, C 8 H 7 O 2 (OH) 3 , then, on sub- 
 stituting nitryl, NO a , for the replaceable hydrogen, we obtain 
 C6H 7 O 2 O 3 (NO 2 ) 3 , a compound nitric ether. 
 
EARLIER VIEWS AS TO NITRO-CELLULOSE II 
 
 (Hadow, von Pettenkofer, cite.d by Guttmann) occasion 
 the recovery of their cellulose, while the liberated 
 nitric acid oxidizes the iron of the ferric chloride and 
 transforms the sulphydrates into nitrates. 
 
 In 1878 Dr. J. M. Eder conducted a series of in- 
 vestigations into the character and composition of the 
 nitro celluloses which has had much influence upon 
 subsequent formation of thought in relation to the 
 character and composition of these bodies; and from 
 the results of his labors he was led to conclude that 
 there existed no less than six distinct varieties of them, 
 three of which, the hexa-, penta-, and di-, he was able 
 to isolate ; two of them, the tetra- and tri-, he obtained 
 in admixture; the mononitro-cellulose, however, he 
 was unable to prepare. Doubling the coefficients of 
 cellulose to avoid fractional coefficients in the deriva- 
 tives, he formulated his series of nitrates as follows : 
 TABLE II 
 
 Name Composition 
 
 Cellulose hexanitrate Ci a Hi4O 4 (NO3)e 
 
 Cellulose pentanitrate CuHisOaCNOsJs 
 
 Cellulose tetranitrate C, a H 16 O 6 (NO3)4 
 
 Cellulose trinitrate Ci 2 Hi 7 O 7 (NO3)3 
 
 Cellulose dinitrate Ci2Hi 8 Oe(NO 3 )a 
 
 Cellulose mononitrate 
 
 Dr. Eder's work constituted a purely scientific re- 
 search into the chemical properties of a group of 
 bodies. Subsequent investigations are characterized 
 by the partial subordination of their scientific aims to 
 the demands of the useful arts. These investigations 
 are of two kinds, and refer to the technical uses of the 
 soluble and insoluble varieties respectively. Those 
 
12 SMOKELESS POWDER 
 
 upon the soluble nitro-celluloses relate, for example, to 
 photography, to questions of transparency and uni- 
 formity of thickness of film ; those upon insoluble nitro- 
 celluloses, to the art of war and to the attainment of 
 the highest explosive effect consistent with the main- 
 tenance of stability. The general purpose of the 
 present work is the study of the physical and chemical 
 properties of nitro-celluloses and nitro-cellulose col- 
 loids, their methods of preparation, and their explosive 
 qualities; and to the prosecution of this study investi- 
 gations of the second class have, until recently, 
 afforded the more direct aid. Conditions of research 
 have, however, been recently modified. For, whereas 
 the explosives prepared from cotton were formerly 
 insoluble nitro-celluloses (gun-cottons) susceptible of 
 detonation as well as of combustion, and employed in 
 mines and torpedoes, yet in recent times the field of 
 research has been extended to the soluble varieties in 
 the effort to secure a suitable base material for the 
 preparation of anon-detonating, progressively-burning, 
 smokeless powder. 
 
 The following account of Eder's nitrates (taken 
 from Cross and Bevan, " Cellulose " *) may be quoted 
 here : 
 
 11 Hexanitrate, C ia H M O 4 (NO 3 ) 6 , gun-cotton. In 
 the formation of this body, nitric acid of 1.5 sp. gr. 
 and sulphuric acid of 1.84 sp. gr. are mixed in vary- 
 ing proportions, about 3 of nitric to I of sulphuric ; 
 sometimes this proportion is reversed, and cotton 
 
 * By permission of Longmans, Green & Co. 
 
EARLIER VIEWS AS TO NITRO-CELLULOSE 13 
 
 immersed in this at a temperature not exceeding 
 10 C. for 24 hours; IOO parts of cellulose yield 
 about 175 of cellulose nitrate. The hexanitrate so 
 prepared is insoluble in alcohol, ether, or mixtures 
 of both, in glacial acetic acid or methyl alcohol. 
 Acetone dissolves it very slowly. This is the most 
 explosive gun-cotton. It ignites at i6o-i7o C. 
 According to Eder the mixtures of nitre and sul- 
 phuric acid do not give this nitrate. Ordinary gun- 
 cotton may contain as much as 12 per cent, of nitrates 
 soluble in ether-alcohol. The hexanitrate seems to 
 be the only one quite insoluble in ether-alcohol." 
 
 " Pentanitrate, C 12 Hj 6 O & (NO,) 6 . This composition 
 has been very commonly ascribed to gun-cotton. 
 It is difficult, if not impossible, to prepare it in a state 
 of purity by the direct action of acid on cellulose. 
 The best method is the one devised by Eder, making 
 use of the^ property discovered by De Vrij, that gun- 
 cotton (hexanitrate) dissolves in nitric acid at about 
 80 or 90 C., and is precipitated, as the penta- 
 nitrate, by concentrated sulphuric acid after cooling 
 to o C. ; after mixing with a large volume of water, 
 and washing the precipitate with water and then with 
 alcohol, it is dissolved in ether-alcohol and again pre- 
 cipitated with water, when it is obtained pure. This 
 nitrate is insoluble in alcohol, but dissolves readily 
 in ether-alcohol, and slightly in acetic acid. Strong 
 potassa solution converts this nitrate into the di- 
 nitrate, C ia H 18 O 8 (NO 3 ) 8 ." 
 
 "The tetra- and trinitrates (collodion-pyroxylin) 
 are generally formed together when cellulose is 
 
14 SMOKELESS POWDER 
 
 treated with a more dilute nitric acid, and at a higher 
 temperature, and for a much shorter time (13-20 
 minutes), than in the formation of the hexanitrate. 
 It is not possible to separate them, as they are sol- 
 uble to the same extent in ether-alcohol, acetic ether, 
 acetic acid, or wood-spirit. On treatment with con- 
 centrated nitric and sulphuric acids, both the tri- and 
 tetranitrates are converted into pentanitrate and hexa- 
 nitrate. Potassa and ammonia convert them into 
 dinitrate." 
 
 "Cellulose dinitrate, C 18 H 18 O 8 (NO,) 3 , is formed 
 by the action of alkalies on the other nitrates, and 
 also by the action of hot dilute nitric acid on cellu- 
 lose. The dinitrate is very soluble in alcohol-ether, 
 acetic ether, and in absolute alcohol. Further action 
 of alkalies on the dinitrate results in a complete 
 decomposition of the molecule, some organic acids 
 and tarry matters being formed." 
 
 The next after Eder to increase our knowledge of 
 the character and relationships of the members of the 
 nitro-cellulose series was M. Vieille (Comptes Rendus, 
 95, 132), the eminent French savant, whose efforts (in 
 collaboration with M. Sarrau) were crowned with 
 success in the production and establishment of 
 the manufacture of a successful smokeless powder in 
 France. M. Vieille's researches were published in 
 part in the French official journal of explosives, the 
 Memorial des Poudres et Salpetres,* a translation of 
 
 * " Recherches sur la Nitrification du Colon," by M. Vieille, 
 Ingenieur des Poudres et Salpetres, Vol. II., 1889; Paris : 
 Gauthiers-Villars et Fils. 
 
EARLIER VIEWS AS TO NITRO-CELLULOSE I 5 
 
 the article referred to constituting Appendix I of the 
 present work. 
 
 Vieille's method consisted in selecting acid mix- 
 tures of specified standard strengths, nitrating under 
 regularly controlled conditions as to mass, time, and 
 temperature, determining the nitration of the resul- 
 tant products, and plotting upon diagrams the resul- 
 tant nitrations as referred to the strengths of the acid 
 mixtures used to produce them. In this manner 
 he was enabled to recognize a tendency of nitration 
 towards groupings or discontinuities, rather than 
 towards progressively increasing nitration, advancing 
 in measure with increase in strength of acids; and he 
 interpreted this tendency as indicating the existence 
 of a number of definite cellulose nitrates. He sum- 
 marizes the results of his work as follows : 
 
 "In order to account completely for the different 
 changes (groupings, discontinuities) by the production 
 of nitro-products corresponding to definite formulae, the 
 equivalent of nitro-cellulose must be quadrupled. Nitro- 
 celluloses corresponding to such formulae agree with 
 the theoretical yields of nitrogen dioxide per gram of 
 material indicated in the following table, and which 
 correspond either to the discontinuities to which we 
 have alluded, or else to a change of physical properties. " 
 
 While Dr. Eder formulated six varieties of nitrated 
 material, doubling the coefficients of cellulose by 
 writing it C, a H 20 O 10 , instead of C 6 H 10 O 6 , M. Vieille 
 formulates no less than eight compounds, to express 
 which he quadruples coefficients, writing cellulose as 
 C a4 H 40 O ao . He thus obtains : 
 
16 
 
 SMOKELESS POWDER 
 
 O co d O co 
 ON r>. vo "t d 
 
 g^ 
 
 : endecan 
 
 : decanitr 
 
 : enneani 
 
 octonitn 
 
 heptanit 
 
 hexanitr 
 
 pentanit 
 
 'c 
 
 5 
 
 
 
 en 
 
 
 en 
 O 
 
 en 
 O 
 
 en 
 
 O 
 
 3S 
 
 
 
 o 
 
 t/5 
 
 g 
 
 3 
 
 3 
 
 3 
 
 | 
 
 p 
 
 3 
 
 3 
 
 
 
 
 
 
 
 
 i) 
 
 UU UO UUU U 
 
 oo oo ooo o 
 
 o ^^ ^o^ ^o ^S ~& *~ ^~o 
 
 qq qq qqq q 
 
 EE EE EEE E 
 
 UU UU UUU U 
 
EARLIER VIEWS AS TO NITRO-CELLULOSE I/ 
 
 To illustrate the accordance of theory with prac- 
 tice, the content of nitrogen (dioxide), as it should 
 exist according to theory, is compared in each case 
 with that actually determined from practical experi- 
 ment. From the above there would appear to be 
 two varieties of gun-cotton insoluble in ether-alcohol, 
 two varieties of soluble nitro-celluloses, and a number 
 of varieties of nitro-celluloses of lower nitration. M. 
 Vieille's paper was published shortly after the official 
 announcement by the French Government of the de- 
 velopment and successful establishment of the manu- 
 facture of an efficient smokeless powder, these results 
 being the declared fruits of M. Vieille's researches, 
 and it is therefore authoritative. 
 
 The points to which attention are especially called 
 in M. Vieille's work are the quadrupling of the expo- 
 nents of cellulose, and the formulation of as many as 
 eight varieties of its nitro-derivatives. 
 
 The abandonment of the old types of smoke-form- 
 ing powders that had been in use for hundreds of 
 years, and the substitution therefor, by the French 
 government, of a new and efficient smokeless powder, 
 was a step that naturally attracted the attention of all 
 civilized powers. The composition of the French 
 powders and their methods of manufacture remained 
 carefully guarded secrets. Efforts to achieve similar 
 results were at once inaugurated in other countries, 
 and the period of inactivity following the abandon- 
 ment of efforts to employ gun-cotton as a propellant 
 gave way to one of marked activity in all that related 
 to the study of explosives. Shortly after the an- 
 
1 8 SMOKELESS POWDER 
 
 nouncement of the results obtained in France, the 
 Russian government commissioned Professor D. Men- 
 dele'ef, a chemist who had already achieved world- 
 wide reputation as the expounder of the Periodic 
 Law of existence of the elements, to conduct a series 
 of researches with a view to the production of an 
 efficient smokeless powder for Russia. As the result 
 of his labors Professor Mendele"ef succeeded in de- 
 veloping a powder called by him " pyrocollodion," 
 which proved satisfactory, and which was adopted 
 in Russia for use in arms of all calibres. As in the 
 case of its predecessor in France, the composition and 
 method of manufacture of pyrocollodion remain care- 
 fully-guarded national secrets. Outlines of ballistic 
 results have been published, however, and it is by 
 these that the next light is thrown upon the composi- 
 tion of nitro-celluloses. 
 
 The scope of Professor Mendeleef's work, and the 
 character of the analytical methods followed by him, 
 may be ascertained from an examination of a paper 
 published by him, my translation of which constitutes 
 Appendix II of the present work. In relation to 
 the structure of the nitro-celluloses he states: 
 
 " In all aldehydes, beginning with the formic and 
 the acetic, a 'tendency towards polymerization is to be 
 noted, due, doubtless, to the property of aldehydes 
 of entering into various combinations (with H 2 O, 
 NaHSO 3 , etc.); whence the composition C e H ]0 O B , 
 containing an aldehyde grouping, should also possess 
 this property, so far as relates thereto. We may 
 therefore safely assume that the molecular composi- 
 
EARLIER VIEWS AS TO NITRO-CELLULOSE IQ 
 
 tion of cellulose, judging from its properties, is poly- 
 merized, i.e., it is of the form C 6M H 10M O 6M , where n is 
 probably very great." 
 
 It will be seen from the foregoing that in their ef- 
 forts to explain the atomic constitution of nitro-cellu- 
 lose, investigators have invariably been led to increase 
 the common multiple of the elements entering into 
 the composition of cellulose, until finally Mendele"ef 
 states as his opinion that this multiple may be very 
 great through reason of the presence of certain ten- 
 dencies towards polymerization. It is interesting to 
 observe how different chemists have been led to es- 
 tablish conclusions as to the existence of compounds 
 corresponding to the formulae they have written. 
 Eder dissolves out of gun-cotton its soluble content, 
 and as the result of the analysis of the remaining por- 
 tion writes its formula as double tri-, or hexanitro- 
 cellulose, C 12 H, 4 (NO 2 ) 8 O 10 ; he isolates from cotton 
 treated with weaker acids, a compound corresponding 
 in content of nitrogen to, what a body of constitution 
 C la H 16 (NO 2 ) B O 10 , a pentanitro-cellulose, should give, 
 and obtains through the agency of solution a nitro- 
 cellulose precipitate satisfying the desired conditions. 
 He also obtains, by employing acids of still further 
 reduced strengths, a body of variable composition 
 which he regards as a mixture of tetra- and trinitro- 
 celluloses, but does not succeed in isolating the two 
 distinct bodies from the mixture. Vieille employs 
 mixtures of acids increasing in strength progressively; 
 continues the process of nitration through a period of 
 time sufficiently long to insure the total conversion of 
 
2O SMOKELESS POWDER 
 
 the cellulose into nitro-cellulose (as shown by the em- 
 ployment of a solution of iodine in potassium iodide 
 as an indicator of free cellulose) ; refers to coordinate 
 axes the results of each observation, employing 
 strengths of acids as abscissae, and numbers of cubic 
 centimetres of nitric oxide evolved as ordinates. 
 Upon comparing results he notes a tendency towards 
 the existence of progressive steps in nitration; i.e., 
 for an acid varying somewhat above and below a cer- 
 tain point in strength there appears to be formed a 
 definite product of nitration. Taking into account the 
 number of such steps observed and their distances 
 apart, he formulates the series of compounds enumer- 
 ated in a preceding paragraph. 
 
CHAPTER III 
 
 THE CONCEPTION OF PROGRESSION IN RELA- 
 TION TO COMPOSITION AND CONSTITUTION 
 
 CELLULOSE is distinguishable from other materials 
 employed for purposes of nitration by the possession 
 of a continuous, complex, and definite cell struc- 
 ture. The plant producing it has developed by 
 growth from the protoplasmic state, through influences 
 of soil, atmosphere, and sunlight; after its death the 
 cellulose tissues remain, the skeleton of the once liv- 
 ing organism, possessing a structure bestowed by suc- 
 cessive growth processes and not by any definite and 
 general chemical change occurring at any stated time. 
 It is this remaining tissue, an organization of cells 
 infinite in their variety and form, that constitutes the 
 base material employed for nitration. We may pro- 
 ceed to nitrate it in various ways ; we may first par- 
 tially destroy or disintegrate the cell structure (con- 
 vert it into hydrocellulose) ; the nitration may be con- 
 ducted at various temperatures and continued for 
 different lengths of time with the employment of vari- 
 ous acid mixtures of different strengths, and with the 
 use of different relative quantities of cotton and acids. 
 If all of these governing conditions be taken account 
 of, and if proper allowance be made for their effect, 
 
 21 
 
22 SMOKELESS POWDER 
 
 we may predicate for the nitrated product an exact 
 composition and exact qualities ; if any of them be 
 overlooked, we lose control of physical character and 
 chemical composition of the final product. 
 
 The method of accounting for the composition of 
 nitro-cellulose by assuming it to consist of a mixture 
 of the different members of a graded series of cellulose 
 nitrates themselves definite chemical compounds 
 was a rational procedure in accordance with established 
 chemical usage; yet there existed grave difficulties in 
 the way of maintaining such views. In the first 
 place, the members of the series could not be sepa- 
 rated from one another. Eder was unable to separate 
 cellulose tetra- and trinitrates. The impossibility of 
 resolving a given nitro-cellulose into definite quantities 
 of the various compounds formulated by Vieille, Eder, 
 or their predecessors, is generally recognized by every 
 investigator to-day. In the second place, the concep- 
 tion of a definite period of time being required to ef- 
 fect nitration is inseparably connected with the forma- 
 tion of these bodies. Thus properly nitrated gun- 
 cotton consists of a large quantity of insoluble with a 
 small quantity of soluble and a very small quantity of 
 unnitrated cotton ; but if nitration be arrested before 
 completion, there results a mixture of different 
 quantities of nitrated and unnitrated cotton. 
 
 That the nitration of cellulose is a gradual progres- 
 sive process, advancing from incipiency through lapse 
 of time towards completion, was the next theory ad- 
 vanced. This change in treatment of the problem, 
 which virtually led to the abandonment of the old 
 
PROGRESSION 2$ 
 
 simple formulation advocated, is discussed with great 
 clearness by M. Bruley, a French Government chem- 
 ist, in a paper entitled " Sur la Fabrication des Cotons 
 Nitres," published in the Memorial des Poudres et 
 Salpetres (Vol. VIII, 1895-6), my translation of which 
 constitutes Apendix III of the present work. 
 
 For purposes of elucidation, Bruley's work will be 
 considered comparatively, in relation to what has been 
 accomplished by others. The first investigators 
 treated cellulose with nitric acid direct, and obtained 
 a product to which they endeavored to ascribe a for- 
 mula; their successors demonstrated the existence of 
 more than one variety of nitro-cellulose, and formulated 
 a series of compounds. The various acid mixtures 
 employed in these researches may be divided into two 
 classes: those composed of nitric acid and water taken 
 in various proportions, and those containing sulphuric 
 acid in addition to nitric acid and water. It was un- 
 derstood how the absorptive action of the sulphuric 
 acid rendered possible the formation of cellulose ni- 
 trates of higher nitration than those that could be ob- 
 tained by the employment of concentrated nitric acid 
 alone. But mixtures of concentrated nitric and sul- 
 phuric acids necessarily contain water, anhydrous sul- 
 phuric acid being a solid and anhydrous nitric acid 
 being impossible to prepare; while with the more 
 highly diluted mixture of the two acids cellulose 
 nitrates identical with those formed with nitric acid 
 and water alone could be formed ; therefore the widest 
 range of nitration resulted from the employment of 
 
SMOKELESS POWDER 
 
 mixtures containing nitric and sulphuric acids and 
 water. 
 
 Vieille had already studied the nitro-products 
 formed by the use of mixtures containing various pro- 
 portionate quantities of nitric and sulphuric acids of 
 definite strengths. Bruley's investigations covered the 
 broader field resulting from the employment of mix- 
 tures of the three elements, in which the quantity pres- 
 ent of each element varied in all practicable proportions 
 in reference to the quantities of the other two. The 
 scope of the latter work may be graphically indicated 
 as follows : 
 
 15. 
 
 -nQ 
 
PROGRESSION 25 
 
 Let OX and OY be coordinate axes. On OY lay 
 off OY' t which divide into one hundred equal parts; 
 and let OY" represent the number of parts of nitric 
 acid to one ' hundred parts of sulphuric acid. Then 
 Y"X'", parallel to OX, will be the locus of all points 
 corresponding to mixtures in which the amounts of 
 nitric and sulphuric acids present bear the definite 
 
 OY" 
 
 ratio -Qyj to each other. On OX lay off OX' and 
 
 divide it into one hundred equal parts, and let OX" 
 represent the number of parts of water to one hundred 
 parts of sulphuric acid. Then X"Y"', parallel to OY, 
 will be the locus of all points corresponding to mix- 
 tures in which the quantities of water and sulphuric acid 
 
 OX" 
 present bear the definite ratio , to one another. 
 
 The point P, in which the lines Y"X'" and X"Y" 
 intersect, corresponds to a mixture containing definite 
 quantities of nitric acid, sulphuric acid, and water. 
 The area OY' QX f is a rectangle, every point of which 
 corresponds to some one combination of the three 
 elements. M. Bruley explores this area by choosing 
 a number of approximately equidistant points dis- 
 tributed over its surface, preparing the acid mixtures 
 corresponding to them, and determining the nitration 
 and solubility of the nitro-celluloses prepared from 
 these mixtures. Proceeding in this manner, and join- 
 ing points of equal nitration, he maps out a series of 
 parallel or nearly parallel curves, between which are 
 included areas of equal or similar solubility. The fol- 
 
26 
 
 SMOKELESS POWDER 
 
 lowing diagram, taken from M. Bruley's work, illus- 
 trates the method : 
 Y 
 
 55- 
 
 u-40 
 O 
 
 
 
 DC 
 
 <35- 
 
 030 
 
 o 
 
 DC 
 
 h 
 
 z20 
 
 LL 
 
 O 
 
 Z 
 
 015 
 
 H 
 DC 
 
 2 
 010 
 
 ou 
 
 XVI XI 
 
 5 10 15 20 25 30 35 40 45 
 
 PROPORTION OF WATER TO 100 PARTS OF SULPHURIC ACID 
 The Roman numeral is the serial number of experiment; the 
 upper Arabic numeral, the nitration in cubic centimetres of NO 2 , 
 the second, the solubility; and the third, when given, the vis- 
 cosity. 
 
 The nearer the points corresponding to nitrations 
 lie to the line Ef, which is the locus of mixtures con- 
 
PROGRESSION 2? 
 
 taining the least possible quantities of water, the 
 higher the nitrations. The area is divided into belts 
 corresponding to mixtures forming insoluble nitro- 
 celluloses, " intermediate " nitro-celluloses, collodions 
 of higher and of lower nitration, and * * friable " cottons. 
 The published investigations do not consider mixtures 
 containing more than 55 parts of nitric to 100 parts of 
 sulphuric acid, M. Bruley stating that mixtures con- 
 taining more than this relative proportion of nitric 
 acid are " not practicable commercially from their in- 
 creased cost " ; nor those containing less than 15 parts 
 of nitric to 100 parts of the sulphuric acid present; 
 in these nitration proceeds with exceeding slowness. 
 Similarly, mixtures containing less than 35 per cent, 
 of water, compared with the quantity of sulphuric acid 
 present, embrace all those capable of producing nitra- 
 tions higher than those of "friable" cottons; while 
 the lower limit of water is fixed by the strength of 
 the strongest acids commercially obtainable. 
 
 Besides nitration and solubility, a new characteristic 
 of the nitrated product is here introduced, viscosity, 
 as determined by the rate of flow, in drops per min- 
 ute, through a standard orifice, of a standard solution 
 of the collodion under examination. M. Bruley 
 states that increased temperature of nitration, as well 
 as continued pulping and washing in warm water, 
 all have the effect of diminishing the viscosity of the 
 collodions formed from nitro-cellulose ; but he does 
 not discuss the practical bearing of this fact on the 
 manufacture of colloids. 
 
 M. Bruley also discusses briefly relations of time 
 
28 SMOKELESS POWDER 
 
 and temperature to nitration. The duration of the 
 reaction is found to be the more prolonged the smaller 
 the quantity of nitric acid present in the nitrating mix- 
 ture; an immersion of two hours is found to suffice, 
 in general, for the production of the soluble nitro- 
 celluloses ; but as much as eight to ten hours are re- 
 quired to complete the nitration of gun-cotton. A 
 comparison is made of nitrations conducted at three 
 different temperatures, employing the same acids and 
 cotton ; and the conclusion is reached that, in the 
 case of soluble nitro-celluloses, increase of tempera- 
 ture during dipping and reaction increases ultimate 
 nitration and solubility; while for gun-cotton, though 
 the effect of temperature upon solubility is less dis- 
 tinctly marked, yet high temperatures have the effect 
 of increasing solubilities. 
 
 During the years 1895 and 1896, I conducted series 
 of experiments at the Naval Torpedo Station, at New- 
 port, Rhode Island, with the view to the production 
 of a nitro-cellulose base suitable for conversion, by 
 direct colloidization, into an efficient smokeless pow- 
 der.* The problem presented itself in the form of an 
 attempt to transform well-known types of nitro-cellu- 
 loses, with or without the addition of solid, non- 
 colloidable ingredients, and by the use of standard sol- 
 vents, into colloid powders; and it ultimately resolved 
 into an effort to overcome certain ballistic inconven- 
 iences, the existence of which was not recognized when 
 the experiments were begun. In the endeavor to over- 
 
 * I have continued experimenting in this field, at intervals, 
 ever since this time. J. B. B. 
 
PROGRESSION 2$ 
 
 come these inconveniences, it became apparent that 
 there was little hope of preparing from the old mate- 
 rials a powder capable of fulfilling service require- 
 ments ; and the results, wholly negative, of elaborate 
 and exhaustive series of experiments with the old 
 materials, forced me to the conclusion that, if a suit- 
 able powder were developed, it could only be through 
 the discovery of a new form of nitro-cellulose capable 
 of colloidization, and which would possess physical 
 and chemical properties not pertaining to any hitherto 
 existing known form of this substance. 
 
 As the result of my labors I succeeded in develop- 
 ing such a new form of nitro-cellulose, which I was 
 able to convert directly, by colloidization, without 
 addition of other ingredient, into a colloid smokeless 
 powder, suitable for arms of all calibres. With this 
 powder I conducted extended series of experiments, 
 testing and establishing its keeping qualities, and stand- 
 ardizing weights of charge and dimensions of grain for 
 the different calibres of guns. Its manufacture was 
 established on a commercial scale at the Torpedo Sta- 
 tion, at Newport, Rhode Island, where the represen- 
 tatives of the various private powder-manufacturing 
 establishments were instructed in the method of prep- 
 aration of the new material.* Subsequently these firms 
 established plants of their own on a scale far larger 
 
 * At this time (1894-1897) the Naval Torpedo Station, at New- 
 port, Rhode Island, was under the command .of Captain George 
 A. Converse, U. S. Navy, to whose ability as an administrator, 
 as well as mechanical skill as a specialist, the successful devel- 
 opment of smokeless powder is largely attributable. 
 
30 SMOKELESS POWDER 
 
 than that undertaken at Newport; and finally an ap- 
 propriation was secured for the Navy, sufficient for the 
 erection of a plant large enough to enable this branch 
 of the service to manufacture an appreciable share of 
 the powder consumed afloat. 
 
 The steps leading up to the development of the 
 new nitro-cellulose are cited here in relation to the 
 light they throw upon ultimate cellulose and nitro- 
 cellulose structure. Taken in order of approximate 
 sequence they may be outlined as follows: 
 
 I. Powders were prepared by colloiding insoluble, 
 soluble, and mixtures of insoluble and soluble nitro- 
 celluloses in acetone (both varieties of nitro-cellulose 
 readily colloid in this solvent), forming the resultant 
 material into strips of different thicknesses, firing 
 charges of different weights of each thickness from 
 standard guns, and tabulating resultant velocities and 
 pressures. Besides the pure colloids thus obtained 
 there were experimented with other colloids having in- 
 corporated into them various percentages of barium and 
 potassium nitrates, both deposited from solution and 
 employed in the dry pulverulent state. It was observed 
 that the addition of the nitrates led to the develop- 
 ment of smoke, but increased the value of V/P, the 
 ratio of the maximum pressure to the initial velocity 
 realized. 
 
 2. A large quantity of gun-cotton, then stored at 
 the Torpedo Station, was available for conversion into 
 powder. It was found that different manufactured 
 lots of this material produced powders differing widely 
 in ballistic properties. To determine the cause of 
 
PROGRESSION 31 
 
 these differences, I took blocks of each gun-cotton in 
 store (some fifty odd lots), washed, dried, and sifted 
 each sample, the washing to remove the sodium car- 
 bonate, colloided each with the same acetone, formed 
 the resultant masses into small rectangular grains of 
 the same dimensions and dried them. I then placed 
 one- gram weights of each small lot upon slips of glass 
 and ignited them in the open. The ashes remaining 
 after combustion differed very greatly among them- 
 selves, some assuming the form of a fine white resi- 
 due, others of a sooty mass of unconsumed carbon. 
 Upon comparing the ash residues with the results of 
 the complete chemical analyses of the original cottons 
 subsequently made, I found that, for the cases where 
 carbon was deposited, the gun-cottons from which the 
 powder had been made contained an abnormally large 
 quantity of free unnitrated cotton, the existence of 
 which had not hitherto been suspected. 
 
 3. Those lots of gun-cotton containing large quanti- 
 ties of free unnitrated cotton being eliminated, experi- 
 ments with powders prepared from the remaining lots 
 were continued. Ballistic differences were still found to 
 obtain, although less marked than those primarily ob- 
 served. Upon comparing the nitration of the original 
 gun cottons, which differed among themselves con- 
 siderably, ranging between N = 13.4 and N = 12.4, 
 with the ballistic results from the powders prepared from 
 them, it was found that the brusquer powders, which 
 developed high pressures as corresponding to compar- 
 atively low velocities, were prepared from gun-cottons 
 of highest nitration. It occurred to me, therefore, 
 
32 SMOKELESS POWDER 
 
 that uniformity in ballistic conditions might be pre- 
 served through the maintenance of mean nitration, 
 i.e., by blending different lots of nitro-celluloses to- 
 gether, so that the mean nitrogen contents of all 
 blends should be maintained a constant ; at the same 
 time providing that dimensions of powder-grains and 
 weights of charge remained equal. The truth of this 
 theoretical assumption was abundantly sustained by 
 the results of practical experiment, and the problem 
 of the attainment of ballistic uniformity, using gun- 
 cotton and the old forms of soluble nitro-celluloses as 
 basis materials for the manufacture of colloid powder, 
 was solved. Bearing in mind the actual nitrations of 
 the various lots of gun-cottons on hand, I chose a 
 mean of 12.75 per cent. N as a standard of nitration, 
 and blended the successive lots to this figure, adding 
 soluble nitro-cellulose when it became necessary to re- 
 duce lots of especially high nitration down to the stan- 
 dard. 
 
 4. I observed, as the result of experiment, that the 
 velocity corresponding to a given bore-pressure could 
 be increased by incorporating oxygen-carriers, such as 
 barium and potassium nitrates, in certain percentages 
 into colloid powders ; and to such an extent that for 
 a pressure of fifteen tons a gain of about 300 foot- 
 seconds could be realized. I attributed this difference 
 in ballistic effect to an accelerative action due to the 
 oxidation of the products of combustion of the nitro- 
 cellulose after their evolution in the bore of the gun 
 and before leaving the gun, by the gaseous oxides of 
 
PROGRESSION 33 
 
 nitrogen evolved from the metallic nitrates.* As the 
 bodies to which accelerative effect was attributable 
 were free, inert particles distributed throughout the 
 substance of the colloid, I argued that free, uncol- 
 loided gun-cotton distributed throughout an ether- 
 alcohol colloid of soluble nitro-cellulose would produce 
 the same effect. Experiments with free gun-cotton as 
 an " accelerator" proved failures, however, practically 
 no ballistic difference being observed between the ac- 
 tion of the ether-alcohol colloid of the soluble, contain- 
 ing free insoluble nitro-cellulose and certain other col- 
 loids of the same in which both ingredients were present 
 in the colloid state. The fact was that both the sol- 
 uble and insoluble forms of nitro-cellulose were, in 
 both cases, consumed away at practically the same rate. 
 5. The effort to prepare powders from colloided 
 soluble nitro-celluloses into which uncolloided insoluble 
 nitro-cellulose was incorporated as an accelerator, led 
 to extended series of experiments with ether-alcohol 
 colloids. These proved at first difficult to manufacture, 
 but, once prepared, were found manifestly superior to 
 the acetone colloids, as they were extremely tough 
 and well capable of resisting disintegration in the gun 
 at the instant of firing, in which regard the brittle 
 acetone colloids had not proved altogether satisfac- 
 tory. Actual experiments in firing had also shown 
 that the greater the quantity of uncolloided (insoluble) 
 
 *See Appendix IV, which is a reprint of my paper entitled 
 " Development of Smokeless Powder," published in the Proceed- 
 ings of the U. S. Naval Inst., in which the subject of acceleration 
 is treated at some length. 
 
34 SMOKELESS POWDER 
 
 nitrocellulose present in an ether-alcohol colloid 
 powder, the brusquer the powder proved, developing 
 a higher bore-pressure as corresponding to the muzzle 
 velocity realized. But, in order to maintain a standard 
 nitration, explosive strength, it was necessary to 
 have present always a considerable quantity of the un- 
 colloided (insoluble) form of nitro-cellulose. As this 
 was objectionable for the above reason, I therefore 
 began experimenting, to see how far the nitration of 
 the soluble nitro-cellulose could be raised, with the 
 view of minimizing the amount of insoluble nitro- 
 cellulose that would be required 
 
 6. I also made investigations in another line, and 
 conducted firing trials with a series of powders pre- 
 pared from nitro-hydrocelluloses. A note upon the 
 constitution of this body may be pertinent here. 
 
 When cotton is exposed to the action of hydro- 
 chloric acid, in the form of gas or of concentrated 
 aqueous solution, it undergoes a change of form and 
 composition, being converted into a material known as 
 hydrocellulose, the composition of which has been 
 variously regarded. Viewed as a substitution product, 
 it is formulated as a cellulose hydrate (Girard, cited by 
 Cross and Bevan) ; considered physically, it appears to 
 consist of an aggregation of fragments of the cells 
 from which the original cellulose was b,uilt up. The 
 fibres seem attacked along planes where their sub- 
 stance is most readily susceptible of decomposition by 
 the acid and are separated from one another; the ac- 
 tion of the acid appears to be a cutting of cell-joints 
 over planes of weakness, whereby the fibres are di- 
 
PROGRESSION 35 
 
 vided into small lengths which are clearly visible un- 
 der the microscope. (There would thus seem to be 
 some portions of the cell more readily susceptible to 
 attack than other portions.) If the process be incom- 
 plete, partial separation only occurs, long unattacked 
 fibres remaining in quantity. 
 
 Samples of nitro-hydrocellulose, wholly soluble in 
 ether-alcohol, of nitration as high as 12.6 to 12.7 
 per cent., were obtained by nitrating hydrocellulose 
 in the acid mixtures I was employing to prepare 
 directly mixtures of insoluble and soluble nitro-cellu- 
 lose from cotton. This led to extended series of trials 
 of powders, of both the accelerated and the unaccel- 
 erated types, containing different quantities of the two 
 varieties of nitro-hydrocelluloses in various propor- 
 tions, with a view of investigating their stability and 
 their ballistic properties. 
 
 On account of their extreme brittleness, these col- 
 loid powders proved too brusque, and experiments 
 with them were abandoned. They also failed in cer- 
 tain cases to develop normal stability. 
 
 7. The observation of certain remarkable phenom- 
 ena connected with the effect of decrease in tempera- 
 ture in prpmoting the solution and colloidization of 
 certain forms of nitro-cellulose in certain solvents, 
 to which reference is made in detail in a subsequent 
 portion of this work, led me to take up the consider- 
 ation of temperature, as an important factor in the con- 
 trol of the character and degree of nitration. I con- 
 ducted extended series of experiments in which cotton 
 was nitrated in various acid mixtures at temperatures 
 
36 SMOKELESS POWDER 
 
 ranging between o C. and 80 C. The result of these 
 experiments was the development of pyrocellulose, a 
 form of soluble nitro-cellulose of high nitration, which, 
 for a given weight of its substance, converted into 
 colloid grains of standard dimensions and dried, de- 
 veloped, when fired from the standard gun, the high- 
 est muzzle velocity, as compared with a given limiting 
 bore pressure. 
 
 The results of the observations in this regard may 
 be briefly summarized as follows : 
 
 The physical and chemical characteristics of prod- 
 ucts of nitration are subject to radical modifications, 
 through variation of temperature at which the reac- 
 tion is conducted. Increase of temperature has the 
 effect of raising the nitration of both the soluble and 
 the insoluble varieties, and would appear also to in- 
 crease the percentage of the soluble component in a 
 blend of the soluble and insoluble varieties. The heat 
 employed may be derived from two sources: (i) that 
 evolved during the exothermic reaction of the nitra- 
 tion of cellulose; (2) that which is supplied from ex- 
 ternal sources to additionally raise the temperature of 
 the nitrating mass. If the temperature of the nitrat- 
 ing mass be raised above a certain point, the structure 
 of the cellulose is attacked, and the nitro-celluloses 
 cease to form, the material dissolving in the acid, with 
 the resultant formation of nitro-substitution bodies of 
 other genera, such as the nitro-saccharoses; or else, 
 direct decomposition of the nitrated body ensues, 
 with evolution of copious fumes of nitric-oxide gas. 
 The chemical and physical phenomena attending 
 
PROGRESSION 37 
 
 both the decomposition and formation of nitro-cellu- 
 loses are largely controlled by temperature. Thus, 
 the explosive force of gun-cotton is greatly reduced 
 by freezing the latter. Gun-cotton saturated with 
 liquid air is not only not an explosive, but is practically 
 a non-combustible ; while non-nitrated cotton under 
 similar conditions is a violent explosive.* 
 
 Again, as will be shown hereafter, the extent of solu- 
 bility of certain nitro-celluloses in certain solvents is 
 a function of the temperature at which the solution is 
 undertaken ; so that the relative quantities of the 
 soluble and insoluble constituents in a mixture of 
 nitro-celluloses formed at one operation may depend 
 upon the temperature at which the separation of the 
 sub constituents is effected. 
 
 * The result of experiments made by me in New York in Octo- 
 ber, 1899. 
 
CHAPTER IV 
 
 SOLUTIONS OF NITRO-CELLULOSE. THEORY OF 
 THE CELLULOSE MOLECULE 
 
 CELLULOSE is found to possess, after nitration, a re- 
 markable property that unnitrated cellulose does not 
 possess: it dissolves freely in a number of liquids, in 
 which it is not soluble in the unnitrated state. Of these, 
 the solvents ethyl ether, ethyl alcohol, ether-alcohol, 
 and acetone are of special interest, both from their 
 connection with the manufacture of smokeless powder, 
 and from the physical and chemical bearings of their 
 methods of effecting solutions. From the solutions 
 the* nitro-cellulose forms with them it may be precipi- 
 tated by the addition of an excess of water or other 
 liquid in which it is not soluble, in a flocculent form ; 
 and ultimately, after drying, forms a pulverulent mass. 
 It is to be specially remarked that the nitro-cellulose 
 cannot be recovered in its original cellular state ; the 
 process of solution has destroyed its organic structure, 
 which may not be recovered or recreated. 
 
 The process of effecting the solution of the nitro- 
 cellulose may be regarded as preliminary to the forma- 
 tion of the colloid. All forms of nitro-cellulose dis- 
 solve freely in an excess of those solvents, in contact 
 with smaller quantities of which they form colloids 
 
 38 
 
SOLUTIONS OF NITRO-CELLULOSE 39 
 
 directly. If the quantity of solvent be reduced below 
 a certain point in proportion to. the quantity of nitro- 
 cellulose employed, colloidization ensues without pre- 
 vious liquefaction; if the solvent be sufficient in 
 quantity to effect liquefaction, evaporation of excess 
 of solvent must precede colloidization ; if the one state 
 can be produced, the possibility of forming the other 
 may be predicated. 
 
 The line between solution and colloidization is not 
 to be drawn sharply ; the two states of matter merging 
 into one another. There are, however, two distinct 
 sets of progressive steps or series of physical changes 
 to be observed, through one or the other of which 
 these bodies pass in their transformation from the 
 liquid to the solid state, and which may be expressed 
 with reference to their progression as follows : 
 
 First series: (a) liquid; (b) jelly ; (<:) elastic mass; 
 (d) tough colloid. 
 
 Second series : (a) liquid ; (b) slime; (r) plastic mass ; 
 (d) brittle colloid. 
 
 To Series I belong most ether-alcohol colloids; to 
 Series II, most acetone colloids. 
 
 The purpose of the present chapter is to describe 
 nitro-cellulose solutions, the methods of forming them, 
 and their characteristics apart from consideration of 
 their colloidal evaporated residues; and to present in 
 relation thereto certain theoretical considerations 
 throwing light upon the chemical constitution of cel- 
 lulose and its nitrates. The subjects of colloids, their 
 properties and methods of formation, will be treated 
 subsequently. 
 
4O SMOKELESS POWDER 
 
 A discussion of the solubility of nitro-celluloses may 
 best be prefaced by an account of a remarkable prop- 
 erty of soluble nitro-celluloses in general, especially 
 characteristic of pyrocellulose and of those forms of 
 nitro-hydrocellulose soluble in ether-alcohol; viz., 
 the direct solubility of these bodies in the solvent ether * 
 when subjected to the influence of cold. 
 
 In August, 1896, I conducted a series of experi- 
 ments with soluble nitro-hydrocellulose of very high 
 nitration, to determine its adaptability for conversion 
 into smokeless powder; this material, colloided in 
 ether-alcohol and dried, had given promise of value as 
 a progressive explosive. 
 
 Upon a very warm Sunday afternoon, I visited a 
 dry-house in which a small tray of soluble nitro-hydro- 
 cellulose, of high nitration (N 12.4-]-), was exposed 
 to a moderate drying temperature. Removing about 
 one gram of the amorphous gray powder from the dry- 
 ing-tray, I introduced it into a test-tube, which I 
 tightly closed with a rubber stopper to prevent the 
 absorption of moisture. I then visited the chemical 
 laboratory and partly filled the tube with what I sup- 
 posed to be ether-alcohol. To my disappointment 
 the precipitate did not dissolve. Upon agitation, it 
 diffused itself throughout the liquid, but rapidly settled 
 to the bottom, in its original pulverulent form, when 
 brought to rest. I at first attributed this behavior 
 to excessive nitration, believing the material to con- 
 
 *When " ether " and " alcohol " are referred to without quali- 
 fication, they are intended to designate ethyl ether and ethyl 
 alcohol. 
 
SOLUTIONS OF N 17 RO-CELLULOSE 4! 
 
 sist of insoluble instead of soluble nitro cellulose. Hav- 
 ing observed, however, during the previous winter that 
 certain of the insoluble nitro-hydrocelluloses appeared 
 to exhibit a tendency to enter into solution upon ex- 
 posure to cold (a portion of the precipitate appearing 
 to rise in the tube, like a jelly, under the influence 
 of cold), it occurred to me to try the effect of a salt- 
 and-ice freezing mixture upon the tube and its con- 
 tents. I therefore mixed a small quantity of salt and 
 ice in a beaker, into which I introduced the tube in a 
 central vertical position. To my great satisfaction, 
 the whole of the precipitate went rapidly into solution, 
 forming a yellowish-brown, mobile fluid. 
 
 While considering the phenomenon that I had wit- 
 nessed, I remained seated at my desk, holding the 
 test-tube enclosed within the palm of my hand. Sud- 
 denly I noticed that, although I had shifted the tube 
 into the inverted position (cork downwards), yet the 
 contents, masked by my hand, had not observed the 
 law of liquid flow, for the lower uncovered end of the 
 test-tube, which extended downwards, was empty. 
 Carefully opening my hand, I was surprised to find 
 that the contents of the tube had condensed into a 
 dense jelly, which remained fixed in the upper part of 
 the tube. I re-introduced the tube into the freezing 
 mixture ; its contents liquefied as before, being trans- 
 formed into a liquid as mobile as maple-syrup ; re- 
 moval from the source of cold and immersion in water 
 heated to about 100 F. (38 C.), or holding in the 
 palm of the hand, sufficed to cause the liquid to con- 
 geal. 
 
42 SMOKELESS POWDER 
 
 The phenomenon above described related to the 
 behavior of soluble nitro-hydrocellulose in the pres- 
 ence of an excess of solvent (ether) when contained in 
 a sealed vessel which is exposed successively to differ- 
 ent temperatures; removal of the, cork with subsequent 
 evaporation of the solvent results in the formation of 
 a solid colloid residue, non-liquefiable upon variation 
 of temperature. 
 
 After producing the results above described, I next 
 endeavored to duplicate them. To my surprise, the 
 soluble nitro-cellulose, which I supposed the same as 
 that I employed the day before, promptly went into 
 solution in ether-alcohol at a temperature of about 
 70 F. Investigation led to the discovery of an un- 
 fortunate interchange of trays in the dry-house, which 
 rendered it impossible to repeat with certainty the 
 experiment cited. There was, therefore, but one 
 mode of procedure left to select sample lots from all 
 the trays, one of which must represent the lot origin- 
 ally taken, and experiment with them in the presence 
 of ether-alcohol. Still I was unable to obtain the 
 original results. It next occurred to me that I might 
 have employed by mistake some solvent other than 
 ether-alcohol, and as the result of a day's experiment- 
 ing, I found that phenomena identical with those 
 originally developed could be obtained with soluble 
 nitro-hydrocelluloses using ethyl ether as a solvent. 
 
 Experimenting subsequently, 1899-1900, to ascer- 
 tain whether the above phenomena were peculiar to 
 nitro-hydrocelluloses, or whether they were character- 
 istic of all soluble nitro-celluloses, I found: 
 
SOLUTIONS OF NITRO-CELLULOSE 43 
 
 I. That pyrocellulose could be readily dissolved 
 in ether upon application of cold. 
 
 2. That all soluble nitro-celluloses acted similarly 
 in presence of an excess of ether, but that some were 
 more readily disintegrated by ether upon application of 
 cold than others. 
 
 3. That if the necessary degree of cold were de- 
 veloped, soluble nitro-celluloses were not only soluble 
 to any desired extent in ether, but that they could be 
 colloided directly therein, without recourse to lique- 
 faction. 
 
 4. That the addition of a few drops of alcohol in 
 difficult cases appeared to be equivalent to a lowering 
 of temperature, i.e., it rendered a given soluble nitro- 
 cellulose more soluble in ether, in the presence of a 
 given degree of cold, than it otherwise would have 
 been. 
 
 The problem of nitro-cellulose solution may be ap- 
 propriately prefaced with the query : Why are certain 
 forms of soluble nitro-cellulose readily soluble at ordi- 
 nary atmospheric temperatures in a compound of two 
 parts by weight of ethyl ether with one part by weight 
 of ethyl alcohol ; whereas, the said forms of nitro-cel- 
 lulose are not soluble to any appreciable extent, at 
 ordinary atmospheric temperatures, in an excess of 
 either ethyl ether or ethyl alcohol, when either is 
 employed alone as a solvent? The point emphasized 
 is, why should the single material, soluble nitro-cellu- 
 lose, prove more soluble in a mixture of two solvents 
 than in either solvent separately? 
 
 If one-tenth of a gram of soluble nitro-cellulose be 
 
44 SMOKELESS POWDER 
 
 placed in a test-tube and covered with, say, 25 c.c. of 
 ethyl alcohol, and the test-tube be corked and then 
 violently agitated, it will be found that solution will 
 not ensue, but that the soluble nitro-cellulose will (if 
 pulped) gradually settle to the bottom after the tube 
 is brought to rest, or else remain suspended in an 
 undissolved state in the liquid. Similarly, the same 
 quantity of nitro-cellulose will remain undissolved in, 
 say, twice the same quantity of ether under similar 
 treatment. If, however, the contents of the two tubes 
 be combined, the soluble nitro-cellulose will promptly 
 go into solution in the mixture of the two solvents. 
 
 Solubility in ether and solubility in alcohol must be 
 touched upon before proceeding to the question of 
 solubility in the compound solvent. It is known that 
 certain nitro-celluloses of low nitration are soluble in 
 ethyl alcohol. To this Vieille and Mendeleef attest, 
 and the latter recommends that this fact be taken 
 advantage of to remove traces of nitro-celluloses of 
 low nitration from those of higher nitration, the 
 higher soluble nitro-celluloses not being soluble in 
 alcohol, at least not at ordinary temperatures. 
 
 If it happened that the soluble nitro-cellulose dis- 
 solved with the same ease in warm alcohol that it does 
 in cold ether, the action of the compound solvent 
 could be accounted for on simple physical grounds. 
 A mixture of the two solvents might so balance differ- 
 ences of temperature as to effect solution at some tem- 
 perature that represented a mean proportional to the 
 percentage of each solvent in the mixture. The at- 
 tempt was made, therefore, to dissolve pyrocellulose 
 
SOLUTIONS OF NITRO-CELLULOSE 45 
 
 in alcohol (95 -per cent.) heated to near its boiling 
 point, but proved unsuccessful. On the contrary, 
 the tendency was rather away from than toward solu- 
 tion. This theory was, therefore, untenable. 
 
 Having alluded briefly to the effect upon soluble 
 nitro-cellulose of the individual solvents, (i) ethyl 
 alcohol, and (2) ethyl ether, the solubility in the com- 
 pound ether-alcohol solvent may next be considered. 
 
 In their work, "Cellulose" (London: Longmans, 
 Green & Co.), Cross and Bevan, referring to cellulose 
 and its hydration, state (p. n): 
 
 " According to modern views on the subject of solu- 
 tion generally, and the solution of colloids in particu- 
 lar, the lines drawn by the older investigators of these 
 phenomena are of arbitrary value ; gelatinization being 
 expressed as a continuous series of hydrations between 
 the extreme conditions of solid on the one side and 
 aqueous solution on the other." 
 
 In their physical forms the solutions of nitro-cellu- 
 loses in ether and ether-alcohol present striking analo- 
 gies to the hydrated and gelatinized forms of cellulose 
 itself. If, as stated, the gelatinized or hydrated form 
 may be regarded as a continuous series of hydrations 
 of cellulose, then the colloids can be regarded as what 
 may be termed a continuous series of etherizations or 
 alcoholizations of nitro-cellulose. In this manner we 
 may establish an analogy in point of behavior between 
 cellulose in the presence of water and nitro-cellulose in 
 the presence of ether, alcohol, and ether-alcohol. 
 
 Referring again to Cross and Bevan, "Cellulose," 
 we find (pp. 4 and 5): 
 
46 SMOKELESS POWDER 
 
 "All vegetable structures in the air-dry condition 
 retain a certain proportion of water, or hygroscopic 
 moisture, as it is termed, which is readily driven off at 
 1 00 (C.), but reabsorbed on exposure to the atmos- 
 phere. under ordinary atmospheric conditions." 
 
 " The phenomenon is definitely related to the pres- 
 ence of OH groups in the cellulose molecule, for in 
 proportion as these are suppressed by combination 
 (with negative radicles to form the cellulose esters) the 
 products exhibit decreasing attractions for atmospheric 
 moisture. It is to be noted that some of these syn- 
 thetical derivatives are formed with only slight modifi- 
 cations of the external or visible structure of the cellu- 
 lose, of which, therefore, the phenomenon in question 
 is again shown to be independent." 
 
 In parallel to this we may state: 
 
 The series of the cellulose esters known as the nitro- 
 celluloses exhibit at ordinary atmospheric tempera- 
 tures, with ether and alcohol, effects similar to those 
 of cellulose with water in what relates to, (i) hygro- 
 copic moisture, and (2) gelatinization. 
 
 In relation to the effect of alkalies on concentrated 
 solutions and this is of primary importance in connec- 
 tion with our subject Cross and Bevan state (p. 23) : 
 
 " Cold solutions of the alkaline hydrates of a certain 
 concentration exert a remarkable effect upon the cellu- 
 loses. Solution of sodium hydrate, at strengths ex 
 cecding 10 per cent. Na a O, when brought into contact 
 with the cotton fibre, at the ordinary temperature, in- 
 stantly changes its structural features, i.e., from a flat- 
 tened riband, with a large central canal, produces a 
 
SOLUTIONS OF NITRO-CELLULOSE 47 
 
 thick cylinder with the canal more or less obliterated. 
 These effects in the mass, e.g., in cotton cloth, are 
 seen in a considerable shrinkage of length and width, 
 with corresponding thickening, the fabric becoming 
 translucent at the same time. The results are due to 
 a definite reaction between the cellulose and the alka- 
 line hydrates, in the molecular ratio C ia H ao O 10 : 2NaOH, 
 accompanied by combination with water (hydration). 
 The compound of the cellulose and alkali which is 
 formed is decomposed on washing with water, the 
 alkali being recovered unchanged, the cellulose ap- 
 pearing in a modified form, viz., as the hydrate 
 C J2 H 20 O J0 .H a O. By treatment with alcohol, on the 
 other hand, one half erf the alkali is removed in solu- 
 tion, the reacting groups remaining associated in the 
 ratio C ia H ao O 10 : NaOH. The reaction is known as 
 that of mercerization, after the name of Mercer, by 
 whom it was discovered and exhaustively investigated. 
 Although, however, it aroused a good deal of attention 
 at the time of its discovery, it remained for thirty years 
 an isolated observation, i.e., practically undeveloped. 
 Recently, however, the alkali cellulose has been made 
 the starting point of two series of synthetical deriva- 
 tives of cellulose, which must be briefly described." 
 
 " From the points established by Mercer in connec- 
 tion with this reaction, the following may be further 
 noted : 
 
 "At ordinary temperatures a lye of 1.225-1.275 
 sp. gr. effects 'mercerization' in a few minutes; 
 weaker liquors produce the result on longer exposure, 
 the duration of exposure necessary being inversely 
 
48 SMOKELESS POWDER 
 
 as the concentration. Reduction of temperature pro- 
 duces, within certain limits, the same effect as in- 
 creased concentration. The addition of zinc oxide 
 (hydrate) to the alkaline lye also increases its activity. 
 Caustic-soda solution of i.ioo sp. gr., which has only 
 a feeble 'mercerizing' action, is rendered active by 
 the addition of the oxide in the molecular proportion 
 Zn(OH) a : 4 NaOH." 
 
 Two points in the above merit special attention in 
 connection with our subject. They are: 
 
 I. The removal of one-half of the alkali on treat- 
 ment with alcohol, the reacting groups remaining as- 
 sociated in the ratio C 13 H 20 O 10 : NaOH. 
 
 2. The acceleration, on exposure to a lye of 
 1.225-1.275 sp. gr., of the process of " mercerization " 
 by reduction of temperature. Here is presented an 
 analogy to the increased solubility of nitro-cellulose in 
 ether and ether-alcohol upon application of cold. 
 
 In referring to the production of cellulose thio- 
 carbonates and to the quantitative regeneration of 
 cellulose from solution as thiocarbonate, Cross and 
 Bevan state (pp. 29-31): 
 
 4< The occurrence of this reaction, under what may 
 be regarded as the normal conditions, proves the 
 presence in cellulose of OH groups of distinctly alco- 
 holic function. The product is especially interesting 
 as the first instance of the synthesis of a soluble cellu- 
 lose derivative i.e., soluble in water by a reaction 
 characteristic of the alcohols generally. The actual 
 dissolution of the cellulose under this reaction we can- 
 not attempt to explain, so long as our views of the 
 
SOLUTIONS OF NITRO-CELLULOSE , 49 
 
 general phenomena of solution are only hypotheses. 
 There is this feature, however, common to all pro- 
 cesses hitherto described, for producing an aqueous 
 solution of cellulose (i.e., a cellulose derivative), viz., 
 that the solvent has a saline character. It appears, 
 in fact, that cellulose yields only under the simulta- 
 neous strain of acid and basic groups, and therefore 
 we may assume that the OH groups in cellulose are 
 of similarly opposite function. In the case of the 
 zinc-chloride solvents there cannot be any other de- 
 termining cause, and the soluble products may be re- 
 garded as analogous to the double salts. The reten- 
 tion of the zinc oxide by the cellulose, when precipi- 
 tated by water, is an additional evidence of the pres- 
 ence of acidic OH groups; and conversely, the much 
 more rapid action of the zinc chloride in presence of 
 hydrochloric acid indicates the basicity of the mole- 
 cule, i.e., of certain of its OH groups. On the other 
 hand, in both the cuprammonium and thiocarbonate 
 processes there may be a disturbance of the oxygen 
 equilibrium of the molecule ; and although there is no 
 evidence that the cellulose regenerated from these 
 solutions respectively is oxidized in the one case or 
 deoxidized in the other, it is quite possible that tem- 
 porary migration of oxygen or hydrogen might be 
 determined, and contribute to the hydration and ulti- 
 mate solution of the cellulose. But, apart from hy- 
 potheses, we may lay stress on the fact that these pro- 
 cesses have the common feature of attacking the cellu- 
 lose in the two directions corresponding with those of 
 electrolytic strain ; and it is on many grounds prob- 
 
50 SMOKELESS POWDER 
 
 able that the connection will prove casual and not 
 merely incidental." * 
 
 It is the feature of double attack upon the cellulose 
 that suggests the cause of the increased solubility of 
 the nitro-cellulose in the compound ether-alcohol sol- 
 vent, as compared with its relative solubility in the 
 ether and alcohol separately. In the case of the 
 " mercerized " cellulose we observe that a removal of 
 one-half the alkali is effected by the alcohol treat- 
 ment. The reaction of the alcohol here relates to the 
 basic OH groups, the alkali appearing to be retained 
 as a base in the groups of acid reaction. A lower 
 temperature facilitates the " mercerization " process; 
 this may be interpreted into the statement that raising 
 the temperature of the molcule beyond a certain point 
 retards the process. So, in the case of the increased 
 solubility of soluble nitro-cellulose in ethyl alcohol at 
 low temperatures, the employment of artifical cold 
 accelerates the process of solution. We may here as- 
 sume that the material, from its dual character, is 
 under an electrolytic strain, from which it is removed 
 by application of cold (abstraction of heat) ; the alco- 
 
 * Referring to the character of the regenerated cellulose in 
 comparison with the original material, Cross and Bevan state, 
 inter al. : 
 
 " (i) Its hygroscopic moisture, or water of condition, is some 3 to 
 4 per cent, higher, viz., from 9 to 10.5 per cent. 
 
 " (2) Empirical composition. The mean results of analysis show 
 c = 43-3 P e r cent., H = 6.4 per cent., which are expressed by the 
 empirical formula 4CHi O6.H a O." 
 
 It will be observed that the regenerated material is an amor- 
 phous hydrocellulose. 
 
SOLUTIONS OF NITRO-CELLULOSE 5 1 
 
 hol then attacks and dissolves the basic OH groups; 
 the structure of the cellulose is destroyed ; the acid 
 groups yield subsequently and enter into solution, to 
 effect which there may occur, as suggested by Cross 
 and Bevan for cellulose, a temporary transfer, within 
 the molecule of nitro-cellulose, of hydrogen and oxy- 
 gen ; and the whole substance finally yields to alco- 
 holization in a manner analogous to that by which 
 cellulose itself yields to hydration. 
 
 Similarly, for the increased solubility of soluble 
 nitro-cellulose in ethyl ether at low temperatures, the 
 employment of artificial cold accelerates the process 
 of solution. The ether dissolves the acid OH groups 
 (in contradistinction to the action of the alcohol, 
 which attacks the basic OH groups); the structure of 
 the cellulose is destroyed; the basic groups' yield sub- 
 sequently, to effect which there occurs a transfer of 
 hydrogen and oxygen within the molecule in a direc- 
 tion opposite to that in which it occurs in the previ- 
 ous case, and the whole substance finally yields to 
 etherization in a manner similar to that by which cel- 
 lulose itself yields to hydration. 
 
 As bearing upon the theory it may be remarked that 
 the two solvents, ethyl alcohol and ethyl ether, differ 
 by H-O-H, since ether, (C,H 6 ),O, + H 2 O = alcohol, 
 2C,H 6 O. 
 
 In the case of the mixed ether-alcohol solvent we 
 subject the cellulose in the one operation to the double 
 tendency to disintegration. The necessity for the 
 application of cold (or the absorption of heat) to ef- 
 fect the necessary arrangement of the H-O-H groups, 
 
52 SMOKELESS POWDER 
 
 as they may be styled, i.e., to strain them into basic 
 and acid relations, is no longer required, for the double 
 attack in two directions, corresponding to those of 
 electrolytic strain, is provided for by the parallel double 
 composition of the ether-alcohol solvent. 
 
 There is reason for supposing that the above-de- 
 scribed reaction is not limited in its occurrence to 
 forms of soluble nitro-cellulose alone, but that it 
 obtains equally for the insoluble varieties. If the 
 theory be correct, then insoluble nitro-cellulose should 
 be soluble in ether. Macnab has shown that insoluble 
 nitro-cellulose dissolves in ether-alcohol at a very low 
 temperature.* 
 
 In the fall of 1899, the time that liquid air first be- 
 came readily obtainable in quantity in New York, I 
 decided to check Macnab's experiment by employing 
 this material as a refrigerant; also (i) to experiment 
 with a view of ascertaining whether insoluble nitro- 
 cellulose of very high nitration in its unpulped fibrous 
 form, the form in which it might be supposed to 
 oppose a maximum resistance to the disintegrating 
 action of solvents, would not actually go into solu- 
 tion in ethyl ether alone under influence of extreme 
 cold; and (2) to determine how far nitro-cellulose is 
 soluble in absolute ethyl alcohol. The ether em- 
 ployed in these experiments was Squibbs', C.P., for 
 
 * See experiments cited by Guttmann in his article " Manufac- 
 ture of Explosives," published in the Journal of the Society of 
 Chemical Industry, London, June 30, 1894 ; also p. 404 appendix 
 to his recently published work, " Manufacture of Explosives," 
 London: Macmillan. 
 
SOLUTIONS OF N2TRO-CELLULOSE 53 
 
 anesthesia; the insoluble nitro-cellulose was gun- 
 cotton of high nitration (N = 13.4-!-) and purity, in 
 the unpulped state. 
 
 The solvent employed in the alcohol experiments 
 was absolute ethyl alcohol ; the nitro-celluloses were 
 
 (1) soluble nitro-hydrocellulose of nitration, N 11.4$; 
 
 (2) soluble nitro-hydrocellulose of N 12.6$; (3) pulped 
 pyrocellulose, N 12.42$; and (4) unpulped gun-cotton 
 of high nitration, N 13.4$. 
 
 Experiment I. I placed about one-tenth gram of 
 gun-cotton in a test-tube, poured over it 25 c.c. of 2 : 1 
 ether-alcohol, tightly closed the tube with a rubber 
 stopper and immersed it in a vessel containing about 
 750 c.c. of liquid air. When the contents of the tube 
 became sufficiently cooled, the gun-cotton went read- 
 ily into solution, forming a clear mobile liquid of a 
 honey-yellow color. I found that the gun-cotton re- 
 mained in the colloid form after the removal of the 
 tube from the liquid and the subsequent heating of its 
 contents to the temperature of the atmosphere. On 
 removing the tube from the liquid-air bath, uncorking 
 and allowing the contents to evaporate, the residue 
 formed a tough, homogeneous amber-colored film. 
 
 Experiment II. Having succeeded readily in col- 
 loiding insoluble nitro-cellulose in ether-alcohol, I next 
 decided to ascertain whether it were possible to col- 
 loid it in ether alone. I placed one-tenth gram of the 
 gun-cotton in a test-tube, poured over it about 25 c.c. 
 of ether, tightly closed the tube and immersed it in 
 the same quantity of liquid air as before. 
 
 Upon sufficiently reducing the temperature, the 
 
54 SMOKELESS POWDER 
 
 contents of the tube became solid and the ether froze 
 into a snow-white mass. On removing the tube from 
 the cold and allowing the ether to melt, I found that 
 the gun-cotton had disintegrated and gone into solu- 
 tion, forming a mobile slightly clouded liquid with a 
 yellowish tinge. The gun-cotton remained in solution 
 after the withdrawal of the tube from the liquid air 
 and the subsequent heating of its contents to the 
 temperature of the atmosphere. On removing the 
 tube from the liquid-air bath, uncorking and allowing 
 its contents to evaporate, the residue formed a tough, 
 homogeneous, slightly yellowish colloid. 
 
 Experiment III. About one-tenth gram of each 
 of the above-mentioned series of nitr'o-celluloses was 
 each placed in a separate test-tube and each had 
 poured over it about 25 c.c. of absolute ethyl alcohol. 
 The tubes thus partly filled were closed tightly with 
 rubber stoppers, immersed in liquid air (for about 
 two-thirds of their lengths), and allowed to remain 
 therein for five minutes.* 
 
 The contents of each tube froze in about half a 
 minute and the rest of the time the liquid air was 
 acting upon the frozen contents. On removing the 
 tube containing the gun-cotton from the bath and 
 allowing its contents to melt, the gun-cotton was, to 
 
 * The liquid air for these experiments was very kindly fur- 
 nished me by Messrs. Vandivert and Gardenhire, of 32 Broadway, 
 New York, pioneers in the development of liquid air in the United 
 States. I am also indebted to Mr. M. Burger, president of the 
 company, for his kindness; and especially to Mr. O Ostergren 
 of New York, one of the original experimenters and developers 
 of the commercial manufacture of liquid air. 
 
SOLUTIONS OF NITRO-CELLULOSE 55 
 
 all appearances, unaltered. This showed that, whether 
 or not the fibre of the cotton had been attacked, it 
 was not affected to the same extent by the ethyl 
 alcohol as it was by the ethyl ether, which, as pre- 
 viously shown, destroyed the fibre of the gun-cotton, 
 causing it to go into solution. 
 
 The sample of pyrocellulose and the two small lots of 
 nitro-hydrocellulose in the other tubes were found to 
 have gone into solution in the ethyl alcohol, through 
 the action of the cold, the former affording a jelly- 
 like straw-colored colloid ; the two latter, the usual 
 brownish-yellow colloids. 
 
 We may now call attention in parallel to the follow- 
 ing remarkable reactions ; 
 
 I. That nitro-celluloses dissolve in ethyl ether 
 under the influence of intense cold. 
 
 2. That nitro-celluloses, with the apparent excep- 
 tion of the highly nitrated insoluble variety (this will 
 be further experimented with later, in the presence of 
 extreme cold) dissolve in ethyl alcohol under the in- 
 fluence of intense cold. 
 
 3. That the so-called soluble forms of nitro-cellu- 
 lose of mean and low nitrations go into solution in a 
 mixture of ethyl ether and ethyl alcohol at ordinary 
 atmospheric temperatures. 
 
 In connection with the above may be mentioned : 
 
 (a) That the various forms of nitro-celluloses seem 
 more readily soluble in ethyl ether than in ethyl alcohol. 
 
 (b That a mixture of two volumes of ethyl ether 
 to one of ethyl alcohol seems to extract at ordinary 
 atmospheric temperatures the greatest quantity of 
 
56 SMOKELESS POWDER 
 
 soluble constituents from a mixture of the two forms 
 of nitro-cellulose (soluble and insoluble as formed by 
 one dipping of cotton.) 
 
 It now remains to consider the theory of compo- 
 sition of cellulose and nitro-cellulose from the chemical 
 standpoint. In this connection it is necessary, first of 
 all, to emphasize the fact of the similarity in compo- 
 sition of the cellulose and the nitro-cellulose molecule. 
 The latter is formed from the former by the substitu- 
 tion, for a certain number of replaceable hydrogen 
 atoms, of the same number of equivalents of nitryl 
 NO,. The effect of the introduction of nitryl into the 
 substance of the cellulose is an apparent weakening 
 of the stability of the latter, rendering it susceptible 
 of decomposition in a number of additional ways, and 
 making actual tendencies to decomposition, traces of 
 which exist in the original material, but which are held 
 in control by stronger counteracting tendencies derived 
 from other sources. 
 
 For the sake of simplicity the formulae presented in 
 what follows will be those of cellulose. It is to be 
 understood that whenever questions arise referring 
 directly to nitro-cellulose, such as solubility in ether 
 or ether-alcohol, the cellulose molecule as presented is 
 to be considered as tacitly representing nitro-cellulose, 
 without actual expression of the substitution of NO, 
 in the replaceable hydrogen atoms. 
 
 Bearing the above in mind, the following additional 
 facts may be taken advantage of in formulating a 
 theory of the composition of the cellulose and the 
 nitro-cellulose molecule : 
 
SOLUTIONS OF N1TRO-CELLULOSE 57 
 
 I. The two solvents of soluble nitro-cellulose, ethyl 
 ether and ethyl alcohol, differ in composition by H O 
 H. Like H 3 O, they may be written (C 3 H 6 ) 3 O and 
 C 3 H.O or C 3 H 6 .O.H. 
 
 2. Cellulose, C 13 H 20 O 10 , is transformed by treatment 
 with alkali in aqueous solution into C 13 H 30 O 10 .2NaOH. 
 
 3. By subsequent treatment with water it is con- 
 verted into cellulose hydrate, C 13 H ac .O 10 .H 3 O. 
 
 4. By treating C 12 H aj) O 10 .2NaOH with alcohol, 
 C 3 H 6 O, the former is converted into the mercerized 
 form, C 13 H ao O 10 .NaOH, with half the alkali removed. 
 
 To account for the occurrence of reactions (2) and 
 (3), the composition of the cellulose molecule is 
 regarded as doubled, i.e., raised from C,H 10 O 6 to 
 C 12 H 30 O J0 . (4) is of especial importance and interest, as 
 it exhibits the action of alcohol upon non-nitrated cel- 
 lulose; establishing a basis for the assumption that the 
 action of the double ether-alcohol solvent upon nitro- 
 cellulose is based upon the original composition of 
 cellulose itself. 
 
 Assuming a duality of composition as indicated (i) 
 by its basic and acid reactions; (2) by its greater solu- 
 bility in the compound solvent; (3) by the H-O-H 
 difference of the components of the original compound 
 solvent ; and (4) by the creation and absorption of the 
 H-O-H groups in the formation of cellulose hy- 
 drates and alkaline compounds, we may proceed as 
 follows : 
 
 If alcohol is capable of effecting the solution of a 
 nitro-cellulose, it must effect the solution of both its 
 components or resolve itself into two sub-components, 
 
58 SMOKELESS POWDER 
 
 each reacting on one sub-component of the nitro-cel- 
 lulose, in order that it may effect the solution of the 
 whole. Under the latter assumption, we may con- 
 sider alcohol, C,H 6 O, as having the composition 
 
 H H 
 
 H H 
 
 which may be divided into 
 
 H H 
 
 i _ n i 
 
 and >C 
 
 . >H H/l 
 
 H H 
 
 Similarly, we may regard ethyl ether, (C a H 6 ),O, as 
 having the composition 
 
 H H 
 
 I I 
 
 H C O C H 
 
 I I 
 
 H C HH C H 
 
 I I 
 
 H H 
 
 capable of dividing into 
 
 H H 
 
 H C O C-H 
 
 | and | 
 
 H C H H C-H 
 
 I I 
 
 H H 
 
SOLUTIONS OF NITRO-CELLULOSE 59 
 
 If such a tendency as illustrated above exists, it 
 will become a reality through the disintegration of 
 equivalents which, in combination, correspond to a 
 molecule of water, H a O.* The removal of one-half of 
 the alkali from cellulose in the mercerization process 
 is effected by the action of the alcohol upon the 
 basic groups therein ; the molecule of cellulose should, 
 therefore, be represented so as to permit a resolution 
 of the original material into acid and basic sub-com- 
 ponents through the disintegration of the water- 
 groups. Under this assumption, we may write cellu- 
 lose as 
 
 H H 
 
 I I 
 
 H C O H H O C H 
 
 I H KK J 
 
 C O H H O C 
 
 H C O 
 
 * It will be observed that the constituents of water, H-O-H, 
 form the central linking group in each of the expressions 
 
 H H 
 
 H H || 
 
 H C O C H 
 : C and 
 
 I TT /- TI [T / TT 
 
 rl \^ rl rl l^ rl 
 
 H 
 
 \ 
 
 H 
 
60 SMOKELESS POWDER 
 
 which would resolve into 
 
 H H 
 
 I I 
 
 H C O H H O C H 
 
 I /H | 
 
 c< .o=c 
 
 I X H I 
 
 C O H H O C 
 
 Form I. || || 
 
 C O H H O C 
 
 I H, | 
 
 c=o >c 
 
 I H/ | 
 
 H C O H H O C H 
 
 I I 
 
 H H 
 
 or 
 
 H H 
 
 I I 
 
 H C O H H O C H 
 
 I/H I 
 
 c< o=c 
 
 I X H I 
 
 C O H H O C 
 Form II. || || 
 
 C O H H O C 
 H | 
 
 c o=c 
 
 H C O H H O C H 
 
 The molecule as above written contains double cen- 
 tral carbon bonds, which fact permits it, on its enter- 
 ing into combination, to be written as 
 
SOLUTIONS OF NITRO-CELLULOSE 
 
 H H 
 
 H C O H H O C H 
 
 6i 
 
 | ^H H- | 
 
 _C_O H H O C 
 
 C O H H O C 
 
 I ^^ H \ I 
 
 C< >C 
 
 I -O" I 
 
 H C O H H O C H 
 
 A 
 
 H 
 
 Without radical modification, it may be expressed 
 
 H C O H H O C H 
 
 I -H B> J 
 
 H C O H H O C H 
 
 H C O-H H O C H 
 
 I H H\ I 
 
 C< ^C 
 
 ^O/^ 
 H C O H H O C H 
 
 as corresponding to C,,H, O 10 . 
 
62 SMOKELESS POWDER 
 
 Halving it, we will obtain 
 
 H H 
 
 H O C 
 
 C O H 
 
 I H H | 
 C O H H O C 
 
 corresponding to C,H 10 O 6 .* 
 
 This latter form, which is the simplest expression 
 for cellulose, represents, not the molecule, but the type 
 unit of cellulose, as it enters into combination, 
 through its four free single carbon bonds, either with 
 other similar units, by polymerization, or with other 
 substances by chemical combination. 
 
 * Written singly, in a similar form, but with closed bonds, the 
 single molecule, CH 10 O 6 , may be expressed as 
 
 H 
 
 H H 
 
 Compare also with Cross and Bevan, " Cellulose," p. 38, where, 
 in reference to cellulose acetates, it is stated if the above formula 
 (alluding to a formula cited) be established by further and ex- 
 haustive investigation, the cellulose unit must be C8H 8 O.(OH) 4 . 
 
SOLUTION'S OF NITKO-CELLULOSE 03 
 
 The multiplicity of the cellulose derivatives and the 
 generally recognized tendency towards polymerization 
 already alluded to suggest the further amplification 
 of the molecule by inter-combination of its units 
 through connection of their carbon bonds into poly- 
 meric forms, the type of which may be expressed as 
 follows : 
 
 Such a method of representation possesses a special 
 interest, as it prepares the way for other considerations. 
 The actual number of phases for the polymerized 
 molecule thus expressed may vary from 2 to any 
 desired number. In the above diagram a 5-phase 
 form is presented for purposes of illustration. The 
 simple molecule with its four free carbon bonds as 
 
6 4 
 
 SMOKELESS POWDER 
 
 presented in each of the five sectors combines with 
 the simple molecule in each adjacent sector on either 
 hand. The simplest polymerized form that, under our 
 theory, can stand alone, is C J2 H 20 O 10 , composed of 
 two * 'sectors," the carbon bonds in each of which 
 unite with those of the other. 
 Thus we have 
 
 H 
 
 H 
 
 correcting the form previously written, 
 
 The 5-phase molecule, or its polymer, corresponds 
 to pyrocellulose. 
 
 It is evident that under such an assumption the 
 molecule may possess an infinity of phases. There is 
 no limit to their number. On this assumption, and it 
 
SOLUTIONS OF N1TRO-CELLULOSE 65 
 
 seems to me on this assumption only, may we account 
 for definite chemical composition of the cellular form 
 in the plant structure. For we may regard the cell 
 as built up from an aggregate of molecules of identical 
 composition but of progressively-varying numerical 
 phase. The cell may begin with molecules of low 
 phase and end with molecules of high phase, or con- 
 versely. Molecules of progressively-varying phase- 
 magnitude may be deposited in turn from protoplas, 
 mic matter in particular forms of different density, the 
 successive evolutions providing for the infinity of cel- 
 lular structures appreciable to the human eye, which, 
 in their successive deposition, build up the fibrous sub- 
 stance of the cell. 
 
 It is to be remembered that the graphic represen- 
 tation of molecular arrangement upon a plane surface, 
 space of two dimensions, cannot be regarded as 
 more than a conventional device illustrating an ar- 
 rangement that exists in nature in space of three 
 dimensions. Nevertheless, the conventional ring- 
 formed combination of elemental particles shown in 
 the polyphase molecule strongly suggests the vortex- 
 ring theory of the composition of matter (as applica- 
 ble to the molecule). For whatever the atom may be, 
 the molecule need not be limited in composition to 
 the simplest collection of the lowest possible number of 
 atoms capable of entering into combination, but may 
 be built up from a very great number of the elemental 
 particles taken together in their proper ratios.* Such 
 
 * On this hypothesis the molecular weight of cellulose would be 
 represented by an average. 
 
66 SMOKELESS POWDER 
 
 a molecule would increase in amplitude according to 
 the number of elemental particles entering into its 
 composition ; and the thought therefore suggests it- 
 self, that progressive variation in the amplitude of the 
 molecular ring is a characteristic of organic life. Or, 
 conversely, we may state that we may seek for the 
 beginnings of organic life, at least of plant-life, in 
 the polymerization of the carbohydrates. 
 
 Briefly summing up, the conditions governing the 
 formulation of cellulose may be stated as follows: 
 
 I. The capability of the expression of the molecu- 
 lar formula as ^(C 8 H 10 O B ) or C 8W H 10W O 6M justifies the as- 
 sumption that the molecule may be represented as 
 composed of n number of similar atomic aggregations 
 of the form C 6 H 10 O 6 , these aggregations conjointly 
 forming the molecule. 
 
 2. Under the assumption that Eder's nitrates repre- 
 sent limits of nitration in the sense defined by Vieille, 
 n cannot possess a value less than 2 ; that is, C, a H 20 O 10 
 represents the lowest expression for the molecule. 
 
 The fact that 2 : i ether-alcohol dissolves certain 
 forms of nitro-cellulose at ordinary atmospheric tem- 
 peratures with greater ease than any other compound 
 solvent containing ether-alcohol in other than the 2 : I 
 proportion, tends to show that all the OH groups in 
 which nitro-substitution takes place are not similarly 
 placed within the molecule. 
 
 4. The fact that, at very low temperatures, either 
 ether or alcohol singly will dissolve soluble nitro-cellu- 
 lose, whereas at ordinary atmospheric temperatures 
 a mixture of the two solvents is required to effect 
 
SOLUTIONS OF NITRO-CELLULOSE f 
 
 solution, implies that, under influence of cold (absence 
 of heat), both ether and alcohol, on the one hand, 
 nitro-cellulose on the other, may undergo a certain 
 atomic rearrangement within the molecule. 
 
 5. If symmetry of arrangement exists in the sub- 
 components of the molecular section C 6 H, O b per- 
 mitting the representation of atomic aggregations 
 which, according to the influences to which they are 
 exposed, may exhibit acid actions on the one hand, 
 basic reactions on the other, it would be by groupings 
 of the carbon and hydrogen atoms, which are even 
 in number, around the oxygen atoms, which are odd 
 in number. 
 
 6. There is good reason for the assumption (basis of 
 theory of nitro-substitution) that the molecules of both 
 cellulose and nitro-cellulose are of similar structure ; 
 that there is no general rearrangement of the atoms of 
 the cellulose in the process of nitration, but that nitra- 
 tion is accomplished through the substitution of nitryl 
 for replacable hydrogen in certain hydroxyl groups, the 
 said groups retaining, after nitration, the position in 
 the molecule that they held before nitration, as already 
 stated. It is upon this assumption that we represent 
 the molecule of cellulose as typical, both of cellulose 
 and nitro-cellulose, and do not represent the substitu- 
 tion of the nitryl in the molcule, unless actually refer- 
 ring to some specific property of the nitro-cellulose 
 distinguishing it from the original cellulose. 
 
 7. Under the assumption (i)of the basic and acid, 
 positive and negative, distribution of the atoms within 
 the molecule, based upon the conception of the con- 
 
68 SMOKELESS POWDER 
 
 stitution of the molecule as a double salt, and (2) of 
 the action of the compound ether-alcohol solvent upon 
 nitro-cellulose at ordinary atmospheric temperatures, 
 and of the action of both the ether and alcohol sep- 
 arately upon nitro-cellulose at low temperatures, 
 we may regard (a) the ethyl ether, (ft) the ethyl 
 alcohol (which possesses the composition but not the 
 atomic arrangement of an ether [methyl]), and (c) 
 the nitro-cellulose, which possesses the characteristic 
 properties of an ether, as all splitting up into dual 
 sub-groups. The manner in which ethyl ether and 
 ethyl alcohol could so split up symmetrically is easily 
 shown ; regarding nitro-cellulose as an ether, we may 
 seek to split it into sub-groups within the molecule 
 in a similar manner. 
 
 The three type forms may then be presented as fol- 
 lows: 
 
 Ethyl alcohol Ethyl ether 
 
 C,H 6 (C,H 6 ) 3 
 
 (Under strain as methyl ether) 
 
 H H 
 
 H H || 
 
 | /H H v | H C O C H 
 
 C< >C | ! 
 
 I \O/ I H C HH C H 
 
 H H || 
 
 H H ' 
 
 Nitro-cellulose 
 (As expressed by its type, cellulose) 
 
 H C O H H O C H 
 
 H C O H H O C H 
 
SOLUTIONS OF NITRO-CELLULOSE 
 
 That all nitre-cellulose's are soluble in the compound 
 ether-alcohol solvent, notwithstanding the wide differ- 
 ences in temperatures at which solution is effected, 
 leads me to the conclusion that all nitro-cellulose 
 molecules are of similar constitution and organization. 
 That they are of a dual composition is evinced by 
 their more ready solubility in the compound solvent ; 
 again the fact that they are all soluble in the single 
 solvent ethyl ether leads me to the conclusion that 
 the two halves of the dual molecule are, with respect 
 to each other, similar, or of similar inverted forms. 
 Therefore, on page 60, Form I is the correct diagram 
 of the separation of the factors, rather than Form II. 
 
 We have considered briefly in the preceding the 
 composition of the molecule in relation to mercer- 
 ization, polymerization, nitration, and colloidization. 
 It remains to dwell somewhat more fully upon the re- 
 actions of nitration and hydration, and to show how 
 they are reflected in the modified structure of the 
 molecule. 
 
 The type cellulose has been written : 
 
 By changing the relative position of atoms of hydro- 
 
SMOKELESS POWDER 
 
 gen, and without altering the constitution of the 
 molecule, it may be written : 
 
 The polymerized molecule is composed of sections of 
 the type : 
 
 m i 
 
 From an examination of the unit of polymerization it 
 
 H C O H H O C H 
 
 H C O H H O C H 
 
 will be seen that the atoms of carbon are connected (i), 
 by single bonds with each other; (2), with hydrogen 
 bonds; (3), with hydroxyl bonds; and (4), with what 
 
 may be termed " water bonds," of the form C" 
 
 H 
 
 Each carbon atom is connected with two other car- 
 bon atoms, one on either side. Every third carbon 
 
SOLUTIONS OF N ITRO-CELLULO^E fl 
 
 atom is connected with the di-valent <( water " radicie 
 (A). The remaining two of each set of three carbon 
 atoms are each connected with one atom of hydrogen 
 and one hydroxyl radicle (B). The symmetrical order 
 of arrangement of the carbons is BAB BAB 
 BAB. 
 
 The free carbon bonds of each unit represent poly- 
 merization. 
 
 The hydroxyl radicles represent nitration, in the 
 mean and higher stages. 
 
 The " water" radicles represent mercerization, and 
 nitration in the lower stage. 
 
 The open oxygen bond in the " water" radicle rep- 
 resents colloidization by combination of the half mole- 
 cule of the nitro-cellulose with the half molecule of 
 the solvent. 
 
 Nitration. The process of nitration may here be 
 taken up for more extended consideration. The fibrous 
 cellulose reflects in its chemical behavior a character- 
 istic of plant life, namely, the possibility of altering 
 growth conditions through changes of temperature. 
 The reactions into which it enters may be varied 
 through the variation of the temperature at which the 
 said reaction takes place. These changes in rate and 
 character of reaction as due to temperature are espe- 
 cially characteristic of nitration, mercerization and 
 colloidization ; and doubtless, during the original 
 growth processes of the plant, affect polymerization. 
 The effect of temperature upon nitration is reflected in 
 the resultant nitrated product in two ways: (i) in 
 raising the degree of nitration in proportion to the 
 
72 SMOKELESS POWDER 
 
 increase in temperature (within certain limits) ; and (2) 
 in increasing and modifying solubilities corresponding 
 to given nitrations. 
 
 Until very recently, our actual knowledge of the 
 nitration of cellulose may be said to have been con- 
 fined to the following facts: 
 
 I. That there was a lower limit of obtainable nitra- 
 tion, somewhere between that represented by the 
 compounds C ]3 H 14 O 4 (OH) 6 (NO 9 ), for which N = 3.80, 
 and C 13 H 14 O 4 (OH) 4 (NO,),, for which N = 6.76. 
 
 2. That there exists an upper limit of attainable 
 nitration, somewhere near that of the hexanitrate of 
 Eder, C n H 14 O 4 (NO 3 ) B , for which N = 14.14. 
 
 3. That nitration is progressive between these 
 limits ; and that when determined at a given tempera- 
 ture, it advances gradually and progressively from the 
 lower towards the higher limit, while under certain 
 conditions it may be made, similarly, to decrease. 
 
 4. That somewhere between the attainable limits of 
 nitration a point is reached (and the point is not con- 
 stant so far as relates to degree of nitration) where the 
 nitrated product ceases to be soluble at a given mean 
 atmospheric temperature in a mixture of ethyl ether 
 and ethyl alcohol. This fact was utilized to separate 
 the so-called "soluble" nitrates of celluloses of low 
 nitration from the "insoluble" nitrates of high nitra- 
 tion. 
 
 Let us consider the formula for the double-type 
 molecule in connection with the six nitrates for- 
 mulated by Eder, and let us assume that there has 
 been isolated from the substance of the cellulose for 
 
SOLUTIONS OF NITRO-CELLULOSE 73 
 
 purposes of experimentation a homogeneous body 
 
 Negative 
 Acid 
 
 H-C 0-H H-^O-C H 
 
 <H=5--nH> 
 
 H-C 0-H H-O-C H. 
 H-C 0-H H-O-C H 
 
 H-C-O-H H-O-C H 
 
 Positive 
 Basic 
 
 composed of molecules of the above 2-phase form. 
 Then, if there are six nitrates, we could form them 
 successively by substituting nitryl for the replaceable 
 hydrogen in the " water" and "hydroxyl" radicles of 
 the cellulose on one side of the molecule. Substituting 
 thus in the above diagram we would have, for the 
 highest nitrate 
 
 Negative 
 Acid 
 
 H-C O N0 2 H-O-C H 
 
 ( r^ r NoT" "~TP = ' ( r 
 
 H-C O N0 2 H-O-C tt 
 H-C-0 NO 2 H-O-C H 
 
 H-6- O NO 2 H-O C H 
 
 Positive 
 Basic 
 
 On account of the ready solubility of the lower 
 nitrates in ethyl alcohol at ordinary temperatures the 
 maximum efficiency of the 2 : I ether-alcohol, and the 
 solubility of the highest nitrates in ethyl ether, we 
 assume that the replaceable hydrogen atoms in the 
 water radicles are replaced by nitryl first, and that 
 subsequently the nitro-substitution continues through 
 the displacement of the hydrogen in the hydroxyl 
 radicles. If we regard alcohol as connected with the 
 
74 
 
 SMOKELESS POWDER 
 
 positive (basic) side of the molecule, and ether with 
 the negative (acid) side, we may represent nitration, 
 for the higher nitrates (say, the hexanitrate), as follows: 
 
 Negative 
 Acid 
 
 H-C O N0 2 H-O-C H 
 
 H-C O NO 2 "H-O-C H 
 H-C-0 NO 2 H-O-C H 
 
 9<S_ -^^c 
 
 H-C-0 N0 2 H-O-C-H 
 
 Positive 
 Basic 
 
 Under such conditions, we may formulate the 
 nitrates of cellulose for the 2-phase molecule as fol- 
 lows : 
 
 TABLE V 
 
 Cellulose hexanitrate, C ia H 14 (NO 2 ) 6 O 10 . 
 
 Cellulose pentanitrate, C 12 O 1B (NO 3 ) 6 O 10 . 
 
 Cellulose tetranitrate, C ia O ]6 (NO 3 ) 4 O i0 . 
 
 Cellulose trinitrate, C ai H 17 (NO 2 ) 3 O 10 . 
 
 Cellulose dinitrate, C 11 H 18 (NO 9 ) 1 O 10 . 
 
 Cellulose mononitrate, C 12 H 19 (NO 3 )O J0 . 
 
 If we turn from the consideration of the theoretical 
 2-phase cellulose above represented to that of actual 
 cellulose, we will find that the latter may be resolved 
 into cellulose nitrates, not in six, but in a very great 
 number of ways. Every molecule may be differently 
 affected in the process of nitro-substitution. The 
 character and composition of the resultant product may 
 vary with strength of acid, its relative quantity in pro- 
 portion to the cotton taken (this determines the rate 
 of absorption of water formed during the reaction), 
 
SOLUTIONS OF N1TRO-CELLULOSE 75 
 
 the temperature at which the reaction takes place, and 
 the duration of immersion in the acid bath. The 
 cotton may be nitrated on the surface of the fibre or 
 nitrated wholly throughout ; it may represent a mix- 
 ture of soluble and insoluble nitrates or it may possess 
 a uniform degree of nitration ; and it may exhibit 
 variations (for the same chemical composition) in its 
 solubilities in standard solvents ; for the successive 
 polyphase molecules are differently constituted and 
 differently placed, and doubtless afford progressively- 
 varying resistances to the effect of the acid bath. 
 . It will be remembered that the theoretical limiting 
 of nitration (14. 14 per cent, nitrogen) was not obtained 
 by Eder. Probably the ultimate substitution of nitryl 
 for every atom of replaceable hydrogen is a limit that 
 may not be attained in practice. Doubtless some of 
 the molecules do represent complete replacement, 
 while others are considerably removed from complete 
 transformation. It will be shown hereafter that by 
 colloidization the molecular amplitude of the nitro- 
 cellulose is reduced to a mean, but such is not the 
 case for the fibrous material. 
 
 Hydration, When, in the attempt to nitrate cellu- 
 lose, the strength of the acid mixture is reduced below 
 a certain figure, a remarkable phenomenon may be ob- 
 served instead of a nitration there may be produced a 
 hydration. We may, therefore, assume that from the 
 same point of weakness in the molecule there may de- 
 velop either a nitro-substitution (nitration), or else an 
 atomic rearrangement accompanied by an absorption of 
 water (hydration). It has already been stated that ni- 
 
? SMOKELESS POWDER 
 
 tration starts, in accordance with our theory, in nitro-sub- 
 stitution in the water radicles, producing nitro-celluloses 
 soluble in the alcohol (basic) solution. Thehydration, 
 then, should start in the rearrangement of the parts of 
 the molecule adjacent to these water radicles. The 
 quantity of water that is to be absorbed or, rather, 
 taken up into the substance of the molecule, may, as 
 shown by Cross and Bevan in their references to mer- 
 cerization, approach the limit C 12 H 20 O 10 .2H 2 O ; corre- 
 sponding to C 12 H 20 O 10 .2NaOH ; while the definite hy- 
 drate, C 12 H 20 O 10 . K 2 O, containing half as much water, 
 and corresponding to C 12 H 20 O 10 .NaOH, is readily iso*- 
 lated. We have, therefore, as definite limits, the in- 
 corporation into the molecule of cellulose, C ]2 H 20 O 10 , 
 of water in the two ratios, 2H 2 O and H 2 O. 
 
 1 / \ I 
 Now the radicle C<f /C plus H 2 O becomes 
 
 | \H H/ | 
 
 I I 
 
 H C O H H O C H ; and as the water radicle 
 
 I I 
 
 occurs twice in the double molecule C ia H ao O 10 ,we have 
 
 for the full transformation, 
 
 H C O H H O C H 
 
 H C O H H O C H 
 
 I I + H,0 = 
 
 H C O H H O C H 
 
 H C O H H O C H 
 
and 
 
 SOLUTIONS OF NITRO-CELLULOSE 77 
 
 H C O H H O C H 
 
 I I 
 
 H C O H H O C H 
 
 I I 
 
 H C O H H O C H 
 
 H C O H H O C H 
 
 H C O H H O C H 
 
 H C O H H O C H 
 
 H C O H H O C H 
 
 I I + 2H a O = 
 
 H C O H H O C H 
 
 I 
 
 H C O H H O C H 
 
 I I 
 
 H C O H H O C H 
 
 I I 
 
 H C O H H O C H 
 
 I I 
 
 H C O H H O C H 
 
 I I . 
 
 H C O H H O C H 
 
 I I 
 
 H C O H H O C H 
 
 H C O^H H O C H 
 
 I ' I 
 
? SMOKELESS POWDER 
 
 In considering these forms of modification of the 
 molecular structure it should be remembered that the 
 molecules of the original fibre are characterized by 
 their variations in amplitude. And we know, actually, 
 that hydration represents a breaking down of the cell- 
 fibre. This breaking down is also illustrated in the 
 two diagrams of the modified molecules just presented. 
 In the former it will be observed that but one of the 
 two water-bonds remains; in the latter, that both 
 water radicles have disappeared and that the two 
 halves of the molecule appear wholly separated. 
 
 The effect of hydration in breaking down the fibrous 
 structure of the cotton and ultimately putting it into 
 solution may therefore be explained by the following 
 steps : 
 
 i . The fibre exists as an aggregation of cells of vary- 
 ing amplitude. The solvent at first attacks those cells 
 that are either the weakest or most exposed, and trans 
 
 \ / \\ 
 forms in them some of the water radicles, C<^ s&t 
 
 \^HH/ \ 
 
 I I 
 
 into the hydroxyl forms, H C O H H O C H. 
 
 I I 
 
 The result is a partial transformation into C ia H ao O lu '- 
 
 H,0 = (Q.H.A,). 
 
 2. As hydration proceeds, the attack takes place 
 throughout all molecules irrespective of amplitude un- 
 til in each molecule one-half of the water radicles are 
 transformed into hydroxyl forms, and a material, yet 
 fibrous, built up of modified cells of varying ampli- 
 tude in each of which one-half of the water radicles 
 
SOLUTIONS OF NIT RO-CELLU LOSE 79 
 
 are transformed into hydroxyl forms, results. This is 
 the true hydrocellulose, C^H^Oi^HjO, or, more prop- 
 erly, C,,H,,O,,. 
 
 3. As hydration proceeds, transformation advances 
 towards the total conversion of all the water radicles 
 into hydroxyl radicles. This is accompanied by dis- 
 integration of the fibre and a tendency to enter into 
 solution. The formula shows that with total trans- 
 formation of the water radicles into hydroxyl forms 
 the molecule splits up into two halves, between which 
 there is no chemical union ; therefore actual chemical 
 combination of the two halves ceases, and there can be 
 no chemical connection between them except such as 
 may be represented by electrolytic strain. 
 
 4. As solution becomes actually effected, the fourth 
 and last, and perhaps the greatest, change in the series 
 occurs. By their intimate contact and admixture the 
 dissolved molecules, freed from their organic form of 
 aggregation, are reduced to a common amplitude.. 
 They hereafter constitute amorphous cellulose, and 
 may be represented as an aggregate of the form 
 C 6 H 10 O b . It is probable, however, that the number 
 of atomic particles in each molecule still remains ex- 
 ceedingly great, as progressive nitration still appears 
 to occur for this material. 
 
 5. The transformation of the whole substance of the 
 cellulose into the form corresponding to C 12 H, O 10 .2H a O 
 represents the dividing line between the chemical and 
 the physical aspects of the absorption of water into 
 the substance of the molecule. Any greater absorption 
 of water than that corresponding to C^H^O^^HjO 
 
80 SMOKELESS POWDER 
 
 pertains to simple solution; any less absorption, to 
 chemical change, affecting the structure of the cellulose 
 cell. 
 
 6. The hydrocellulose obtained by the usual pro- 
 cesses represents, simultaneously, a combination of 
 all the above-described processes. Part of the cellu- 
 lose remains wholly unattacked, and, if allowed to 
 exist, renders the whole mass what the workmen style 
 ''woolly"; part is converted into C la H 2 ,O n , a true 
 hydrocellulose, which, inasmuch as its fibre still ex- 
 ists, combines to exercise the same effect upon the 
 physical constitution of the mass as the unnitrated 
 portion; part is precipitated as wholly hydrated cellu- 
 lose of the form C ia H 20 O 10 .2H 2 O, and part goes into so- 
 lution. 
 
 7. Once the fibrous structure is lost by solution, 
 and the molecules reduced to a common amplitude, 
 the organic constitution is gone and may not be re- 
 covered in any way. The cellulose thereafter remains 
 as an amorphous body. This condition finds a parallel 
 in the ultimate state of colloided nitro-cellulose, when 
 the last traces of solvent are totally expelled there- 
 from, and there results a pulverulent amorphous mass. 
 
APPENDIX I 
 
 RESEARCHES UPON THE NITRATION OF COTTON 
 
 By M. VIEILLE 
 
 VERY different formulae have been suggested to 
 represent the composition of the mtro-products de- 
 rived from celluloses, and particularly the composi- 
 tion of products of maximum and minimum nitration. 
 These products were, moreover, obtained by processes 
 differing at the same time both as to temperature of 
 reaction, concentration of acids, and the nature of the 
 sulpho-nitric mixtures employed. Therefore the re- 
 sults were not susceptible of any general interpreta- 
 tion. 
 
 We have thought it well to take up this study 
 again in a methodical manner, and to investigate the 
 influence of different methods of nitration and of 
 temperature upon resultant products. 
 
 I. FIRST METHOD OF NITRATION* 
 
 Conditions of dipping. We first decided to deter- 
 mine the law in accordance with which the degree of 
 
 * In translating this paper I have converted the chemical for- 
 mulae employed therein into those of the system employed in the 
 United States (O = 16). 
 
 81 
 
82 SMOKELESS POWDER 
 
 nitration and physical qualities vary under well-de- 
 fined conditions of dipping, namely, dipping in pure 
 nitric acid of various degrees of concentration and at 
 a temperature of 11 C. 
 
 The nitration was effected by immersing wadded 
 cotton in from 100 to 150 times its weight of acid; 
 all elevation of temperature was thus obviated and 
 the strength of acidity of the bath could be regarded 
 as constant during the whole of the dipping. 
 
 The operations were conducted in large-mouthed, 
 stoppered flasks of about 500 c.c. capacity and cooled 
 externally by a current of water. The flasks con- 
 tained about 250 c.c. of nitric acid. Three grams of 
 well-carded cotton were introduced into the upper part 
 of the vessel which was then corked and shaken four 
 or five times, whereby the cotton was equally dis- 
 tributed throughout the interior, so that each fibre 
 might be considered as surrounded by 150 times its 
 weight of acid. 
 
 Percentage of nitration. The content of nitrogen 
 in the nitrated product was determined by M. Schloes- 
 sing's method for estimating nitrogen in nitrates. In 
 order to apply this method to the determination of 
 nitrogen in explosives, it had to undergo some modi- 
 fications in detail, which show at the same time the de- 
 gree of exactness realized in the conditions we have 
 adopted. The following table presents a resume of 
 the reults of our experiments: 
 
? NITRATION OF COTTON 
 
 
 B-S 
 
 V 
 
 S 
 
 co co o ' 
 
 3 p J; 
 
 " 
 
 2 i 
 
 
 O to 
 
 
 <u 
 
 O ^C 0-> Q 
 
 U "*^ 
 
 *rt O 
 
 
 O ,. 
 CO .^ 
 
 CU 
 
 c -c 
 
 O *- O (/j 
 
 ^J 4) c 
 
 C w 
 rt ) 
 
 
 
 
 03 *- 
 
 "^ ^ v*-i iri 
 
 *~^ O ^ 
 
 
 
 *- 43 
 
 c 
 
 s* 
 
 -n 
 
 y T3 2* 
 
 j-5 
 
 
 *"^ cJD 
 
 '" 
 
 g W>_ 
 
 O G 'p tj 
 
 52 o <-> 
 
 3 "o 
 
 
 C *" 
 
 C 
 
 S o 
 
 <S o-2 {^ 
 
 T3 rt u 
 
 rt 
 
 
 O ^"o 
 
 rt Tl 
 
 rt r^ *O 
 
 aj o "S 0> 
 
 C v- <t3 
 
 ., Q 
 
 
 
 CJ 
 
 
 
 rt 4> 4> 
 
 
 G 
 <U 
 
 > 
 U ^JJ 
 
 ll 
 
 Q 10 (d 
 
 oj 3 .\ 
 
 2 rC 10 W 
 
 
 D C 
 
 1 
 
 'o 
 
 
 
 cx 
 
 C/) 
 
 O t) ' 
 
 3-S 
 G 
 
 U* 
 
 a| 
 
 co "Tj *O 
 
 ifiil 
 
 ^ ^ 03 g 
 
 ?l * 
 
 .2^ w 
 <i3 
 >, o 3 
 
 -=S 
 
 ill 
 
 Character o: 
 
 he nitrated product res 
 pletely soluble in aceti 
 uble in pure ether or < 
 
 cd co 
 C o 
 
 i| 
 
 o ^ 
 
 <U J3 
 
 11 
 
 he nitrated product ha 
 cotton. It becomes g' 
 tion of acetic ether an 
 
 otton dissolves in the i 
 liquid precipitable by 
 obtained swells up thr 
 ether and becomes gel 
 ing. Ether-alcohol pi 
 
 jj! 
 
 CO JJ 
 
 "^ 43 
 
 2^ / 
 
 O, w v- 
 
 esidue becomes more 
 strongly blackened by 
 tion. Nitration, insig 
 
 
 H 
 
 C^ 
 
 H 
 
 U 
 
 H 
 
 & 
 
 i|J' 
 
 
 
 
 M CO O I^ 
 
 
 
 S-olJ 
 
 
 
 4 r^. co 
 
 Oco co 
 
 o'vo" 
 
 4- co Tj- co 
 
 co CN in co 
 
 CN (N y O 
 
 i 
 
 Q ^ *_i O 
 
 
 
 
 f~~ 
 
 
 
 d 
 
 q 
 
 q q 
 
 qq 
 
 q q 
 
 OO o 
 
 c< c< CT 
 
 q 
 
 (U 
 
 ffi: 
 
 K ; K 
 
 EH 
 
 K : ffi 
 
 HH H-( - ^ 
 
 ffi- 
 
 .C 
 
 
 co m 
 
 CO 
 
 ^* O 
 
 O O co 
 
 co 
 
 SH 
 
 . 
 
 vO CO 
 
 !** t-i 
 
 10 
 
 in r**- O 
 
 o 
 
 
 
 M 
 
 i-H (^ 
 
 ci CN* 
 
 w cJ 
 
 CN CN CO 
 
 CO 
 
 c 
 
 
 
 
 
 
 
 o 
 
 i 
 
 ~h ~r 
 
 i ~h 
 
 I *T~ 
 
 -] (- -f- 
 
 -J- 
 
 1 
 
 0. 
 
 6. 6 
 
 00 
 
 6 , 6 
 
 66, 6 
 
 o\ 
 
 e 
 
 o 
 U 
 
 I 
 
 ffi K 
 
 XX 
 
 ffi K 
 
 Effi'ffi 
 
 *" 
 
 "o ^(J 
 
 CN r^ 
 
 O CN O 
 
 CO CO 
 
 co co 
 
 r^ r^ vo 
 
 co O m o 
 
 N 
 
 SK II 
 
 in rj- 
 
 T* . 
 
 *TT 
 
 >* ^ ^f 
 
 ^t rf ^- TT 
 
 1" t 
 
 Q H 
 
 
 
 
 
 
 
-84 SMOKELESS POWDER 
 
 The number of cubic centimetres indicated in the 
 third column corresponds, for each degree of concen- 
 tration of the acid taken, to the maximum nitration. 
 
 Determination of tJie maximum nitration. This 
 maximum was determined in each case by the analy- 
 sis of specimens exposed to times of dipping of in- 
 creasing lengths. The limit is, moreover, very clearly 
 indicated by employing a solution of iodine in iodide 
 of potassium, which produces a black or greenish dis- 
 coloration of nitro-products containing traces of non- 
 attacked cotton. We may state that, beyond the 
 point where the discoloration ceases to be produced, 
 prolongation of dipping does not increase the degree 
 of nitration. 
 
 Thus a specimen dipped in acid of density 1.488 
 gave at the end of 24 hours of dipping 161 c.c., and 
 was discolored by iodine. At the end of 70 hours it 
 gave 165.7 c - c -> without discoloration. 
 
 On the other hand, a specimen dipped in acid of 
 density 1.490 ceased to be discolored by iodine after 
 a dipping of 24 hours and gave 183.7 c - c - At the 
 end of 128 hours the same specimen gave 183.8 c.c. 
 
 Nature of the reactions ; speed of reactions. The 
 durations of dipping which determine maximum nitra- 
 tion vary considerably with the degree of concentra- 
 tion of the acid. The rapid action for the density of 
 1.500 (at the most from two to three hours) becomes 
 gradually slower, and requires for the density of 1.483 
 as much as 120 hours. The corresponding nitro-prod- 
 ucts practically preserve the appearance of the original 
 cotton, but for a density of about 1.470 the action is 
 
THE NITRATION OF COTTON 85 
 
 completely modified, the cotton swells and dissolves 
 almost instantly, transforming the acid into a clear, 
 transparent collodion. If this syrupy mass is poured 
 into running water small white flocks are obtained, 
 opaque and brittle after drying, and which preserve 
 nothing of the primitive fibre of the cotton. In these 
 conditions the limit of nitration is rapidly reached. 
 Acid of density 1.469 gives after 5 minutes, 134.7 c - c - I 
 after 30 minutes, 140.5 c.c. ; after 20 hours, 139.3 c.c. 
 
 When the density of the acid falls below 1.46, solu- 
 tion does not occur; the action becomes much slower 
 and the cotton appears unattacked, but it is shown 
 upon washing that the fibres become very friable. A 
 specimen was collected in the form of a paste. At 
 the same time the yields decrease very considerably 
 below the theoretical figure, which indicates that the 
 cotton has been partially attacked. 
 
 For densities below 1.450 it is no longer possible to 
 isolate, however long the dipping, a product that is 
 not blackened by iodine. The cotton is slowly trans- 
 formed into products non-precipitable by water; e.g., 
 at the end of 15 days there is obtained by treatment 
 in water a very small residue which is intensely colored 
 black or blue by iodine, and which is of a very feeble 
 nitration. 
 
 Discontinuities in the progress of nitration. The pre- 
 ceding table shows that, under the above-mentioned 
 conditions of dipping, the degree of nitration of the 
 cotton increases more or less gradually, in accordance 
 with the concentration of the nitric acid, from 108 c.c. 
 to about 128 c.c. ; the degree of nitration then rises 
 
86 
 
 SMOKELESS POWDER 
 
 to 140 c.c., corresponding to a very small variation 
 in the strength of the acid, and it remains at this fig- 
 ure while the concentration of the acid increases to a 
 very notable extent. It again rises under similar 
 conditions to 165 c.c., then to about 180 c.c., and 
 then increases gradually to the limits of nitration of 
 gun-cotton properly so-called. 
 
 180 
 .165 
 
 140 
 130 
 
 7 
 
 1.400 
 
 1.470 
 
 1.480 
 
 1.490 
 
 The above diagram, obtained by expressing the 
 density of the acid by abscissae and the corresponding 
 percentage of nitration by ordinates, illustrates the 
 character of the progress of the reaction. 
 
 The existence of these discontinuities is of great 
 importance in relation to the establishment of the 
 chemical formulae of the nitro-derivatives of celluloses. 
 It has therefore been deemed useful to reproduce 
 
THE NITRATION OF COTTON 87 
 
 these different degrees of nitration under entirely dif- 
 erent conditions, employing sulpho-nitric mixtures. 
 
 II. SECOND METHOD OF NITRATION 
 
 We employed sulpho-nitric mixtures formed of 
 ordinary sulphuric acid (density 1.832) and ordinary 
 nitric acid (density .1.3 16). The conditions of dipping 
 were identical with those which have already been 
 described; 3 grams of wadded cotton in 250 c.c. of 
 the mixture. The temperature varied from 19 C. 
 to 21 C. 
 
 The following table presents a resume of our exper- 
 iments. The first column shows the proportion by 
 volume of nitric acid to sulphuric acid taken : 
 
 Preparation of 
 H a SO 4 by volume 
 for i vol. of HN0 3 
 A = 1.316 
 
 No. of c.c. of NO, 
 evolved from i gftn. 
 of the Nitrated 
 Product at o C. 
 and 760 m.m. 
 
 Character of Specimens 
 
 3-00 
 2.50 
 2.OO 
 1.70 
 1-50 
 
 195-9 
 190.1 
 184.6 ) 
 
 185.5^ 
 I82. 3 ) 
 
 Cotton not attacked. Soluble in 
 acetic ether and in ether-al- 
 cohol 
 
 1.40 
 1.30 
 1. 2O 
 
 164.0) 
 166.7 \ 
 166.0 ) 
 
 The cotton is very slightly at- 
 tacked (a little stringy). Sol- 
 uble in acetic ether and gen- 
 erally becomes gelatinous by 
 the action of ether-alcohol 
 
 1. 10 
 I.OO 
 
 I4I.2 ) 
 143-5 \ 
 
 Rendered gelatinous by acetic 
 ether. Is only swelled up by 
 ether-alcohol 
 
 0.95 
 
 0.90 
 
 133.3 j. 
 132.7 } 
 
 Friable products 
 
88 
 
 SMOKELESS POWDER 
 
 These results give rise to various observations. 
 
 The sharp advances in the degree of nitration indi- 
 cated in the first method are equally to be observed 
 under the second, and practically for the same con- 
 tents of nitrogen, corresponding to a yield of 130 c.c., 
 140 c.c., 165 c.c., and 180 c.c. of nitrogen dioxide, 
 approximately. 
 
 Thus the degree of nitration remains the same for 
 proportions of sulphuric acid of 2, 1.70, and 1.50; and 
 lowers abruptly by 18 c.c. for the proportion of 1.40. 
 Nitration remains stationary for the proportions of 
 1.40, 1.30, and 1. 20, and lowers again abruptly for 
 the proportions i.io and i.oo. 
 
 In order to follow this phenomenon more closely we 
 undertook dippings with proportions of acid interme- 
 diary between those for which the abrupt changes had 
 been observed. 
 
 The specimens thus obtained gave the following 
 results : 
 
 Preparation of 
 H 2 SO 4 by volume 
 for i vol. of HNO 3 . 
 A = 1.316. 
 
 No. of c.c. of NO a 
 evolved from i gm. 
 of Nitrated Product 
 at o C. and 760 
 m.m. 
 
 Character of Specimens 
 
 1-45 
 
 I7I.I 
 
 Cotton not attacked. Soluble in 
 ether-alcohol 
 
 J.I5 
 
 153-0 
 
 Cotton attacked slightly 
 
 The percentage of nitration is interpolated exactly 
 between those indicated above. These trials confirm 
 the preceding experiments, which show that the exact- 
 ness of the mixtures under the conditions of our ex- 
 
THE NITRATION OF COTTON 
 
 8 9 
 
 periments may be relied upon and that the discontin- 
 uities indicated do not arise from accidental conditions. 
 They show, moreover, that there are not, properly 
 speaking, sharp advances in nitration ; and that there 
 exists a very restricted zone of acid mixtures with 
 which one may obtain intermediate nitration. But 
 there exists, nevertheless, a discontinuity in the prog- 
 ress of the reaction ; and the following diagram, 
 obtained by expressing the percentages of sulphuric 
 acid in a mixture, as abscissae, and the yield of nitro- 
 gen in c.c., as ordinates, allow us to keep track of the 
 progress of the reaction : 
 
 180 
 
 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 
 
 It appears rational to admit that definite products 
 correspond to the periods of constant nitration, and 
 that the mixture of the two products only occurs in 
 the intervals of transition. 
 
QO SMOKELESS POWDER 
 
 The properties of the cellulose nitrates obtained by 
 the second process, in respect to solvents such as 
 acetic ether and ether-alcohol, are, for a given degree 
 of nitration, identical with those of the nitro-cellulose 
 obtained by dippings in pure nitric acid. The prop- 
 erties of cellulose nitrates appear then connected with 
 their chemical composition, and are independent of 
 the method of preparation. 
 
 We have been able to establish this fact in a more 
 rigorous manner by nitro-celluloses capable of conver- 
 sion into collodions. 
 
 The zone of collodions, according to our experi- 
 ments, is narrow. It comprises celluloses the nitrogen 
 content of which is comprised between 1 80 c.c. and 
 190 c.c., approximately. 
 
 A little above this limit however, up to 195 c.c., 
 and a little below, to 166 c.c. and often to 150 c.c., 
 celluloses occur which are transformed into jellies by 
 the action of ether-alcohol, and which filter by press- 
 ure through cloth. But it has been found that these 
 substances do not produce fluid and limpid collodions 
 useful in the arts; and on drying, collodions of low 
 nitration have been found to yield an opaque and 
 brittle film. 
 
 Now, all the nitro-celluloses capable of producing 
 collodions, obtained from widely different sources, and 
 which we have had occasion to examine, are capable, 
 in accordance with their composition, of being arranged 
 in the method that our experience has made clear to 
 us. This is shown in the following table : 
 
THE NITRATION OF COTTON 91 
 
 Collodion-cotton, Rousseau process 183.2 c.c. 
 
 Billault-Billaiidot 183.8 " 
 
 High-temperature cotton Rousseau, 1st lot 184.0 " 
 
 " " " li 2d " 189.0 " 
 Swedish gun-cotton for the manufacture of 
 
 gelatin-dynamite 1 93-O " 
 
 Swedish soluble cotton made at Vonges, 
 
 in accordance with the directions in the 
 
 aide-memoire de la marine 185.4 " 
 
 III. HIGHER LIMIT OF DEGREE OF NITRATION 
 
 The higher limit of the degree of nitration does not 
 appear in the preceding tables, which do not present 
 results relating to a density of nitric acid higher than 
 1.502. Greater densities are difficult to obtain and, 
 moreover, our first trials with an acid of 1.516 strength 
 at a temperature of 15 C. resulted in a very marked 
 attack upon the cotton. It is probable that dipping 
 in acids of so great a concentration would only give 
 good results at lower temperatures, but the upper limit 
 of nitration of gun-cotton may be obtained indirectly 
 by the use of sulpho-nitric mixtures. These mixtures, 
 which were employed for the preparation of gun-cot- 
 ton for military uses, produce products yielding from 
 208 c.c. to 212 c.c. of nitrogen dioxide per gram. 
 
 By proceeding with special care, we have been able 
 at a temperature of 11 C. to exceed these limits and 
 to obtain a product of 215 c.c. This limit remains 
 exactly the same whatever the proportions of nitric or 
 sulphuric acid employed, even when Nordhausen sul- 
 phuric acid is substituted for the monohydrate. 
 
9 2 
 
 SMOKELESS POWDER 
 
 The influence of a great excess of sulphuric acid 
 affects exclusively the rapidity of the process of nitra- 
 tion, which is thereby considerably diminished. 
 
 IV. CELLULOSE INCOMPLETELY NITRATED 
 
 The preceding results related to celluloses of maxi- 
 mum nitration, and the iodine reaction above indicated 
 permits us to regard the latter as homogeneous bodies ; 
 but it is possible, by stopping the reaction arbitrarily, 
 to obtain an idefmite number of celluloses of the 
 same degree of nitration and of variable properties 
 mixtures of different nitro-celluloses and unattacked 
 cotton. 
 
 Among these products we have made a special study 
 of those which were furnished by the sulpho-nitric 
 mixtures. On increasing the proportion of sulphuric 
 acid in the mixture we may so slow down the reaction 
 as to obtain for a given duration of dipping, say 
 15 minutes, any degree of mean nitration desired. 
 The following table shows the composition of speci- 
 mens submitted during variable times to the actions 
 of different acid mixtures : 
 
 Composition of the Mixtuies 
 
 Vol. of Nitrogen Dioxide 
 evolved per gram 
 
 HN0 8 
 
 H,S0 4 
 
 After 15 min. 
 
 After 3 hours 
 
 I volume 
 
 3 volumes 
 
 194.6 
 
 212.4 
 
 i volume 
 
 3^- volumes 
 
 182.0 
 
 212.4 
 
 I volume 
 
 4 volumes 
 
 150.8 
 
 2OO-7 
 
 i volume 
 
 5 volumes 
 
 123.4 
 
 2OO.7 
 
 
 
 126.8 
 
 2OO.7 
 
THE NITRATION OF COTTON 93 
 
 All specimens which have been exposed to only fif- 
 teen minutes of dipping are blackened more or less 
 strongly by iodine, which indicates the presence of 
 variable quantities of non-attacked cotton ; the second 
 column of the table appears to show that the excess of 
 sulphuric acid only affects the speed of the reaction, 
 and that all products tend toward the same limit of 
 nitration. 
 
 Moreover, we may verify the fact that from the very 
 beginning of the reaction the products obtained are of 
 maximum nitration, and that the results of rapid dip- 
 ping in concentrated acids are simply mixtures of non- 
 attacked cotton and products of the highest nitration. 
 
 Thus the sample which yields 126.8 c.c. treated in 
 acetic ether loses 59.44 percent, of its weight, and the 
 residue presents all the characteristics of almost pure 
 cotton ; combustion, and coloration by iodine (black 
 coloration very feeble in the presence of ferrous 
 salts and hydrochloric acid). Only 60 per cent., 
 then, of cotton is nitrated in this specimen; and if we 
 compare the total volume of nitrogen dioxide, 126.8 
 c.c., with that of I lie cotton nitrated, we obtain 
 214.7 c.c. ; which is the number that we would have 
 found if the reaction were allowed to proceed to com- 
 pletion. 
 
 This method of dipping preserves the cotton fibre 
 intact, which may be of importance in relation to the 
 manufacture of nitro hunting powders, for it produces 
 a natural and absolutely intimate mixture of gun-cot- 
 ton and ordinary cotton. Moreover, the nitrated 
 material obtained in this mixture does not differ at all 
 
94 SMOKELESS POWDER 
 
 from ordinary gun-cotton, and we may hope that it 
 possesses a stability equal to that of the latter. 
 
 V. RESUM'E AND CONCLUSIONS 
 
 Composition and formula of nitro-cellulose. The 
 first degree of complete nitration which the above-indi- 
 cated reactions allow us to determine corresponds to 
 a yield of 108 c.c. of nitrogen dioxide per gram, ap- 
 proximately. This reaction is produced with pure 
 nitric acid and three equivalents of water. It is slow, 
 and only gives a small yield, on account of the cotton 
 being attacked under the hydrating influence of the 
 acid, which is then the principal phenomenon. 
 
 As the concentration of the acids is increased, the 
 per cent, of nitration appears to advance progressively 
 to 128 c.c. ; at least this is what the nitrogen content 
 of specimens 12 and 13 of the first series (page 83) 
 would appear to show. 
 
 It would be well, however, to maintain a certain 
 reserve upon this point, because since these products 
 are insoluble in the solvents, the iodine reaction re- 
 mains as the only index of complete nitration; and it 
 is to be feared that certain traces of incompletely 
 nitrated products lower the percentage of mean nitra- 
 tion. 
 
 From 128 c.c. the percentage of nitration rises by 
 marked increases to 140 c.c., then to 165 c.c., and 
 finally to 180 c.c. Then, as the concentration of 
 acid increases, the percentage of nitration rises pro- 
 gressively, by insensible degrees, to the limit corre- 
 sponding to military gun-cotton, which yields 215 c.c. 
 
THE NITRATION OF CO 7' TON 95 
 
 In this last period, although there is no brisk change 
 in the percentage of nitration, we may note the very 
 clear transition point at 190 c.c. and 195 c.c., corre- 
 sponding to the limit of the zone of the collodions. 
 
 In order to account completely for the different 
 changes by the production of nitro-products corre- 
 sponding to definite formulae, it is necessary to quad- 
 ruple the equivalent of cellulose. The nitro-celluloses 
 which this formula leads to correspond to the theo- 
 retical yields of nitrogen dioxide per gram of material 
 indicated in the following table, and with which we 
 compare the percentages of limiting nitration ob- 
 served, corresponding to the discontinuities of which 
 we have spoken already, or else to a change in physi- 
 cal properties : 
 
 C a4 H 29 O., (;NO 3 )i 1 .. Cellulose endecanitrate. 1 Gun . cotton i 2I 4 c.c. 215 c.c 
 
 C 2 4H3 O2o(NO 2 )io - Cellulose decanitrate... I '"(203" 215 
 
 C94H 31 Oao(NO a ) 9 ... Cellulose enneanitrate. I Collodions ____ I *9<> " J 9a 
 
 C 2 4H 32 O ao (NO 2 ) 8 .Cellulose octonitrate. .. ) "(178" 182 
 
 C a4 H 3 3O2o(NO 2 ) 7 ... Cellulose heptanitrate.. \ (162" 164 
 
 V 
 
 .Cellulose hexanitrate.. V Friable cottons < 146 " 143 
 
 Ca4H 3B OaolNO a ) 6 . .Cellulose pentanitrate.. ) (128" 132 
 
 C 2 4H3 6 O 2 o(NO.,)4 . . . Cellulose tetranitrate ....................... 108 " 109 
 
 It will be seen that these formulae take suitable ac- 
 count of the production of limits of nitration and of 
 all particulars presented by the reaction. 
 
 The approximate results obtained, however, are 
 always lower than the exact volumes of nitrogen by 
 about y-J-^. The differences appear to be attributable 
 to the presence of a very small quantity of products 
 of lower nitration. 
 
 Properties of nitro-celluloses. The explosive prop- 
 erties of nitro-celluloses are in direct relation to the 
 
96 SMOKELESS POWDER 
 
 percentages of nitration. As the nitration diminishes 
 the vivacity of combustion in open air decreases, and 
 the production of carbonaceous residue becomes more 
 marked. We may thus classify at a glance nitrated 
 celluloses in one of three groups gun-cottons, collo- 
 dions, or friable cottons. The measurements of pres- 
 sures developed by these different products in a closed 
 chamber show that the force similarly diminishes with 
 the percentage of nitration. Thus a collodion-cotton 
 yielding 184 c.c. produces pressures inferior by |- to the 
 pressure furnished by Moulin-Blanc gun-cotton afford- 
 ing 2 1 1 c.c. The percentage of nitrogen constitutes 
 a true measure of the explosive qualities of a product. 
 
 Finally, we may mention that the stability of nitro- 
 celluloses decreases with the percentage of nitration., 
 with respect to reagents such as hydrochloric acid 
 and ferrous salts. For products of low nitration the 
 reaction commences when cold ; for those of mean 
 nitration a few moments' heating is required, but for 
 celluloses yielding more than 200 c.c. of nitrogen 
 dioxide per gram, the attack commences only after 
 sustained ebullition. These products appear then to 
 acquire .the maximum of stability along with the maxi- 
 mum of power. 
 
 PARIS, September, 1883. 
 
APPENDIX II 
 
 PYROCOLLODION SMOKELESS POWDER 
 
 By Professor D. MENDELEEF 
 
 (Translated from the Russian by Lieutenant John B. Bernadou 
 U. S. Navy) 
 
 THE very favorable results obtained with pyro- 
 collodion, and its adaptability to arms of all calibres, 
 depend upon its composition and properties, which, 
 for purposes of illustration, may be compared with 
 those of other materials employed as smokeless pow- 
 ders. 
 
 As to its chemical composition, pyrocollodion. may 
 be designated homogeneous,* and herein consists one of 
 its most important qualities. All previous and pres- 
 ent forms of powder did not have or do not have this 
 property to the degree here implied. From their 
 very method of preparation, black and brown powders 
 
 * Homogeneity, in its full chemical significance, is not claimed, 
 inasmuch as the composition of cellulose itself remains a matter 
 of doubt. The quality is urged from the technical standpoint, in 
 relation to the properties of other smokeless powders. It is pos- 
 sible that, a solvent may be found capable of separating pyro- 
 collodion fractionally; but pyrocollodion insoluble in ether or 
 alcohol, but soluble in a mixture of these substances, is far more 
 homogeneous than other forms of nitro-cellulose or any of the 
 nitro-glycerin powders, inasmuch as the latter are readily capable 
 of fractional subdivision. 
 
 97 
 
98 SMOKELESS POWDER 
 
 are coarse mechanical mixtures, for which any consid- 
 eration of homogeneity is out of the question. The 
 same is true for those smokeless powders containing 
 ammonium nitrate, picrates, etc.' Nitre-glycerin 
 powders may be regarded as gelatinous solutions of 
 nitro-cellulose in nitro-glycerin, which, from their 
 composition, are, chemically, non-homogeneous; more- 
 over, various solvents (alcohol, ether, acetone, etc.) 
 dissolve certain constituents out of them, leaving 
 others. 
 
 The same may be said of present-day types of nitro- 
 cellulose powders; alcohol dissolves out of them the 
 nitru celluloses of lower nitration; a mixture of ether 
 and alcohol, the collodions, leaving the excess of 
 highly nitrated cellulose undissolved. Pyrocbllo- 
 dion,* however, surrenders no part of its substance to 
 alcohol, while it is wholly soluble in a mixture of 
 ether and' alcohol. This chemical homogeneity of 
 pyrocollodion, taken in the sense in which it is stated 
 to be employed, plays an important role in its com- 
 bustion ; for there are many reasons for believing that 
 in the case of the combustion of those physically but 
 not chemically homogeneous substances, such as the 
 nitro-glycerin powders (ballistite, cordite, etc.), the 
 nitro-glycerin portion is decomposed -first, and the 
 nitro-cellulose portion burns subsequently, in a differ- 
 ent layer of the powder. f It is to be added that the 
 
 * Under the assumption that the remainder of the solvent is 
 wholly expelled from the powder. 
 
 f The experiments of Messrs. I. M. and P. M. Tcheltsov at the 
 Scientific and Technical Laboratory show that for a given density 
 
PYROCOLLODION SMOKELESS POWDER 99 
 
 homogeneity of pyrocollodion possesses a direct bear- 
 ing upon the uniformity of ballistic results developed 
 by its use. 
 
 Besides chemical homogeneity, pyrocellulose and 
 the powders prepared therefrom possess a second dis- 
 tinguishing quality, viz., that for a given weight of 
 their substance they develop a maximum volume of 
 evolved gases, the latter being measured at a given 
 temperature and pressure. This new conception in- 
 volves certain intricacies and complexities, and may be 
 discussed to some degree of fulness, in what relates to 
 nature and volume of gases evolved upon 1 the decom- 
 position of powder. 
 
 According to the law of Avogadro-Gerard, * the 
 chemical equivalents or quantities of matter expressed 
 by simultaneous chemical formulae (e.g., H a O = 18, 
 water; CO = 28, carbonic oxide; CO 2 = 44, carbonic 
 acid; N 3 = 28, nitrogen) occupy at a given temper- 
 ture and pressure a volume equal to that occupied by 
 two parts by weight of hydrogen, H 2 = 2 (its molecu- 
 lar equivalent). Consequently, if we possess the full 
 chemical equation of combustion of a substance or 
 
 of loading, the composition of the gases evolved by nitro-glycerin 
 powders varies according to the surface area of tHe grains (i.e., 
 the thickness of strips or cords), a phenomenon not to be ob- 
 served in the combustion of pyrocollodion powder. There is 
 only one explanation for this, viz., that the nitro-glycerin, which 
 possesses the higher rate of combustion (Berthelot), is decom- 
 posed sooner than the nitro-cellulose dissolved in it. This is the 
 reason why the nitro-glycerin powders destroy the inner surfaces 
 of gun-chambers with such rapidity. 
 
 * The development of this law is given in " Principles of Chem- 
 istry," by D. Mendeleef, 6th ed., 1895, chap. 7. 
 
IOO SMOKELESS POWDER 
 
 mixture of substances, of which the products are 
 gases or vapor, it is easy to calculate the volume 
 occupied by these products at a given temperature 
 and pressure. For example : the combustion of black 
 powder may be expressed typically by the equation : 
 
 2 KN0 3 + S + 3 C =K a S + 3 CO a + N 2 . 
 Mol. wt, 2 X 101 + 32 + 3 X 12 no -f- 3X44+28 = 270, 
 Volume in form of gases, 3X2 + 2 = 8. 
 
 That is, for 270 parts by weight of powder ingredients 
 8 volumes of gas* are formed, or 29.6 volumes per 
 
 * If the weights of the equivalents be expressed in grams we 
 may ascertain the volume of gas evolved in litres, when the pres- 
 sure P (in millimeters of the mercurial column) and the temper- 
 ature / (in degrees Celsius) are known. Thus as two equivalent 
 weights of hydrogen and the equivalent of each gas occupy at a 
 temperature t = o and a pressure P = 760 mm. a volume of 22^ 
 litres, then for t and P this volume becomes 
 
 22t (i + 0.00367 /) 1. 
 
 Consequently one volume, expressed in grams, occupies, approx- 
 imately, 
 
 11.1+0.4070* litreSf 
 
 / 
 where/ corresponds to the number of atmospheres, each of 760 
 
 p 
 mm. at o C. ; i.e. ,/ = - . Thus, in our example, if / 2000 C. and 
 
 p = 2500 atm., the 8 volumes of gas produced by the combustion 
 of 270 grams of powder occupy an actual volume of 
 
 8. 11-1+0.0407.2000 = a296 Htre> 
 
 2500 
 
 At o C. and a pressure of I atmosphere we attain a volume of 88.8 
 litres for 270 grams. In this manner it is easy to proceed to the 
 value ^'1000 given in the text, the actual volume of evolved gases 
 per kilogram of powder. 
 
P YROCOLLODION S>AfOJELE$$' PV 1&&E& > J ' I O I 
 
 1000 parts by weight of explosive. Brown powder 
 (cocoa) represents (in its greater progressiveness of 
 combustion and in certain other respects) a partial 
 transformation from black to smokeless powders, and 
 is characterized by the partial carbonization of its 
 charcoal (which contains much hydrogen and approx- 
 imates to a composition C B H 4 O), and by the small 
 quantity of sulphur entering into its composition. Its 
 mode of combustion may be expressed approximately 
 as follows : 
 
 6KN0 3 +2C 5 H 4 = 3 K 2 C0 3 + 7 CO + 4 H,O+ 3 N, 
 
 6X loi-f 2 X 80 = 3X138 + 7X28+4X18+3X28 = 766. 
 Volume of gases, 7X2 +4X2 +3X2 =28. 
 
 r,... = 36-5- 
 
 Or as follows : 
 
 4KN0 3 +C 5 H 4 0+S^K 2 S0 4 +K 2 C0 3 + 4 CO+2H 2 0+2N 2 , 
 Mol. wt. =516, Volume of gases = 16, 
 
 ,, F",o.. = 3 I - 
 
 It is evident, then, that the gas volume correspond- 
 ing to brown powder is nearly 34, greater than that 
 for black powder; whence originates the preference 
 generally accorded to brown powder over black. 
 
 In a similar manner we have for the complete 
 (typical) combustion of nitro-glycerin, 
 
 4 C 3 H 5 N 3 9 = i2C0 2 +ioH 2 0+6N, + O 2 . 
 Mol. wt., 4 X 227 = 12 X 44 + 10 X 18 + 6X28 + 32 = 908 
 Volume of gases = 12X2 +10X2 +6X2 +2 =58. 
 
 r,... = 6 3-9- 
 
 If we present in the same manner the decomposi- 
 tion of a type of nitro-cellulose of high nitration, such 
 
IO2 S-MGiK&L&SS POWDER 
 
 as Abel's trinitro-cellulose, C 6 H 7 (NO 2 ) 3 O 6 , we obtain 
 the following: 
 
 2C.H 1 N 1 O n - 3 C0 2 + 9 CO + 7 H 2 + 3 N 2 . 
 Mol. wt., 2 X 297 = 3 X 44+ 9 X 28+7X18+3X28 = 594. 
 Volume of gases = 3X2 +9X2 +7 X 2 +3 X 2 =44. 
 
 ^,000 = 74- 1. 
 
 If the typical combustion of this nitro-cellulose be 
 presented by the equation 
 
 2 C 8 H 7 N 3 O 1J = ioCO 3 + 2CO + 7H 2 + 3 N 2 , 
 
 that is, if we assume that water is not formed, and 
 that the oxygen combines wholly with the carbon, 
 then the F" 1000 remains unchanged, as the volume of 
 gas formed (22 X 2 = 44) remains as before. Therefore 
 we need not stop to consider how the oxygen is dis- 
 tributed between the carbon and hydrogen in the 
 products of combustion, as the value of FJ 000 does not 
 vary.* 
 
 If, for nitro-cellulose of high nitration we substitute 
 Eder's pentanitro-cellulose, C 12 H 16 (NO 2 ) B O 10 , which 
 
 * It is another matter if a portion of the oxygen continues in 
 combination with the nitrogen, or if the oxygen proves insuffi- 
 cient to convert all the carbon and hydrogen into gases; that is, 
 if hydrocarbons are formed; but this becomes a case of incom- 
 plete combustion. Such conditions have a certain bearing upon 
 the combustion of smokeless powders, especially when the sol- 
 vent is not completely expelled; but, on the one hand, the quan- 
 tity of hydrocarbons formed is relatively small, and on the other, 
 they are formed (as also compounds of carbon and nitrogen, as 
 cyanogen), in relatively small quantities, for all powders, even 
 when the latter contain an excess of oxygen. In considering 
 type forms of combustion there is no need of investigating second- 
 ary conditions of this class, especially as by so doing we are 
 diverted from the direct study of the general problem. 
 
PYROCOLLODION SMOKELESS POWDER 1 03 
 
 corresponds to a content of 12.75 P er cent, nitrogen, 
 and to the composition of. ordinary nitro-cellulose 
 employed for smokeless powders, we obtain a greater 
 evolution of gas, for 
 
 2C 12 H lb N 6 20 - CO, + 2 3 CO + i5H 2 + 5 N 2 . 
 Mol. wt., 2X549 44 -h 23 X 28 + 15 X 18+ 5X28=1098. 
 Volume of gases =2 -)- 23.2 -f- 15.2 +5-2 = 88. 
 
 . y tm = so.i. ' 
 
 The increase in volume of gas hereby realized is due to 
 the fact that the quantity of carbonic acid evolved is 
 diminished, while that of carbonic oxide is increased, 
 which causes an increase of total gas volume, since, 
 for the equation 
 
 and for the equation 
 
 C+O =CO, F,... = 7i.4. 
 
 If we descend to a lower nitration and consider 
 Eder's tetranitro-cellulose, C 12 H 16 (NO 2 ) 4 O, , we have 
 no longer the case of complete combustion; for 20 
 equivalents of oxygen are required to convert 12 of 
 carbon and 16 of hydrogen into gaseous products and 
 vapors, while there are but 18 of oxygen available, un- 
 less we assume the products of combustion as CO, 
 H 2 O, and H 2 .* It is known, however, that in the case 
 
 *In the latter case (without formation of carbon) the decompo- 
 sition would be: 
 
 C, a HiflN 4 O, B = I2CO + 6H a O + 2H 2 -f 2N 2 . 
 Mol. wt., 504 = 12 X 28 + 6 X 18 + 2x2 + 2X28 = 504. 
 Volume of gases = 12X2 +6X2 +2X2 + 2X2 =44. 
 
 F.ooo = 87.5. 
 
 But typical combustion, according to such an equation, is prac- 
 tically impossible; carbon and hydrocarbons are formed, and the 
 
IO4 SMOKELESS POWDER 
 
 of combustion of carbohydrates low in oxygen, the 
 latter combines with the hydrogen, 'from its greater 
 affinity for that substance, leaving a part of the carbon 
 deposited in an uncombined state. Consequently, such 
 conditions do not correspond to complete combustion. 
 The formation of CO 2 shows that there is a certain 
 excess of oxygen in pentanitro-cellulose ; whereas 
 typical combustion, corresponding to maximum gas 
 volume, requires all carbon to be converted into car- 
 bonic oxide. Such typical combustion is afforded by 
 pyrocollodion, the composition of which corresponds 
 to the formula C 30 H 38 N 12 O 49 , as is shown by the follow- 
 ing equation : 
 
 5 C.H 10 0.+ I2HNO,= Q.HJNCg.A, + I2H,O. 
 
 Cellulose Nitric Acid Pyrocellulose Water 
 
 combustion that actually does occur is intermediate in character 
 to that expressed by the above and by the two following equa- 
 tions : 
 
 2C 13 H 16 N4O ]8 = C 4 + 2oCO -f i6H 2 O + 4N 2 . 
 2C ia H 18 N 4 18 = 22CO + MH a O -f- C 3 H 4 + 4 N 2 . 
 For the first, FJOOO = 79'4; for the second, 81.4; the mean of 
 the two latter is 80.4; of all three, 82.4. This quantity is close to 
 that afforded by pentanitro-cellulose and pyrocollodion. In this 
 manner may be explained the phenomenon that upon the combus- 
 tion of nitro-cellulose containing a little less than 12.5 per cent, 
 nitrogen, or of pyrocollodion containing a certain quantity of un- 
 evaporated solvent (which is equivalent to a lowering of nitra- 
 tion), results are obtained that approximate to those produced by 
 pyrocollodion powder, although velocities and pressures are 
 somewhat lowered. The existence of this phenomenon depends 
 upon the homogeneity of the powder; whence it follows that it is 
 better to have a content of a little less than 12.5 per cent, nitrogen 
 with a homogeneous powder than a content of above 12.5 per cent, 
 with the powder non-homogeneous, and that the best results are 
 developed by homogeneous pyrocollodion of nitration N = 12.4 
 per cent. 
 
PYROCOLLODION SMOKELESS POWDER 10$ 
 
 In typical combustion it corresponds to the follow- 
 ing equation : 
 
 C JO H 38 N 12 4 , = 3 oCO + i 9 H,0+6N,. 
 
 Mol. wt., 1350 = 30X28 + 19 Xi8 + 6 X 28 = 1350. 
 
 Volume of gases =30X2+19X2+6X2 = no. 
 
 ^ooo = 81.5- 
 
 Before proceeding further, we desire to call atten- 
 tion to the fact that, whereas for brown powder we 
 realize a volume of 34 approximately, we have here a 
 volume of 81.5, whence, judging from volumes of 
 evolved gases, pyrocollodion should prove 2\ times 
 more powerful than brown powder. Actual experi- 
 ments show that the powders stand in about this rela- 
 tion to one another. In units of energy per unit for 
 weight of explosive we have 
 
 47 m.m. R. F. about 
 
 9-in. gun, 
 
 12-in. gun, 
 
 From this it is evident that our computed value of 
 
 F 1000 is in complete accordance with actual experimental 
 
 results. 
 
 In this manner we may consider it as proven that, 
 for a given temperature and pressure, pyrocollodion 
 develops a greater volume of gases (and vapor) than is 
 developed by black or brown powder (for which 
 F" 1000 = 30), and even greater than is afforded by pow- 
 ders prepared from the more highly nitrated forms of 
 nitro-cellulose, and by the nitr o-glycer in powders * 
 
 * The rapid, simple and novel method of comparing the force of 
 explosives herein employed was first suggested and used by me 
 
 Pyrocollodion 
 powder 
 
 Brown 
 powder 
 
 Ratio 
 
 22O 
 
 8l. 5 
 
 2.7: I 
 
 22 3 
 
 90.7 
 
 2.5 :i 
 
 2IO 
 
 93 
 
 2.3:1 
 
106 SMOKELESS POWDER 
 
 In order to establish the full significance of the 
 above deduction, it remains to show (i) that from the 
 standpoint of practical applicability we can foresee no 
 other material capable of developing as great a value 
 for F" 1000 as pyrocollodion ; (2) that the physico-chem- 
 ical and ballistic qualities necessary in a smokeless 
 powder are developed by pyrocollodion, not in a less, 
 but in an equal or greater degree than by other mate- 
 rials employed as smokeless powders ; (3) that the 
 estimate of ballistic efficiency of a smokeless explosive, 
 through consideration of its volume of evolved gas, 
 without regard to conditions of temperature, leads us 
 
 in 1892. It was developed through comparison of the composition 
 of smokeless powders and of their products of combustion and of 
 the results of experimental firings made at the laboratory. Al- 
 though I see clearly that not only the volumes of products of 
 combustion, but also their temperatures, must be taken into 
 account in a complete analysis of phenomena attending the 
 decomposition of smokeless powders, nevertheless I purposely 
 give preference in these investigations to phenomena relating to 
 volumes of gases evolved, not only on account of the simplicity 
 of the latter and their direct accordance with ballistic results, but 
 for the reason that with present methods for estimating tempera- 
 tures developed by explosives (and these methods are unreliable) 
 it becomes necessary to make numerous arbitrary assumptions 
 (especially in relation to specific heats of gases and vapors at high 
 temperatures); while for volume calculations we have definiteness 
 of composition as a starting point; and if there be any assump- 
 tion to be made, it relates to the distribution of the oxygen be- 
 tween the carbon end hydrogen, which, from the chemical stand- 
 point, is not so material and so far as it relates to volume it 
 is of little significance. In all cases, however, I present but the 
 elementary comparisons of performances of powders, as the fuller 
 treatment of the subject does not constitute the object of my 
 investigation. 
 
PYROCOLLODION SMOKELESS POWDER IO/ 
 
 into no error, although it would at first appear that 
 temperature would have a direct effect upon the prac- 
 tical qualities of a powder. 
 
 Since smokeless powder was discovered, so many 
 schemes were set afloat for meeting general demands, 
 that up to the present time there remain "as open 
 questions which form of smokeless powder is the 
 superior, and whether new and still more efficient forms 
 may not be looked for in the future. In order to re- 
 ply to these inquiries, it will -be necessary first to glance 
 over the compositions of materials capable of conver- 
 sion into smokeless powders under the following as- 
 sumptions: (i) That they leave no solid residue after 
 combustion, and that their gases exercise no injurious 
 effect upon the metal of guns; (2) that they undergo 
 no change upon keeping for long periods of time, and 
 contain 'no volatile ingredients; and (3) that they may 
 be readily prepared in quantities sufficiently abundant 
 for practical needs. 
 
 There are but few elements capable of producing 
 gases that do not act upon metals, and, generally 
 speaking, it is useless to try to find others besides 
 hydrogen and nitrogen, and their compounds with 
 oxygen and carbon, that do not act upon gun-cham- 
 bers at the temperatures of combustion of powder. 
 Therefore, in general terms, the composition of those 
 mixtures or compounds suitable for conversion into 
 powder may be expressed as 
 
 C,H M .N,0,. 
 
 The energy imparted to the projectile is derived from 
 
108 SMOKELESS POWDER 
 
 the conversion^of the mass of the powder into gases, * 
 the transformation being accompanied with the pro- 
 duction of great heat. These fundamental conditions 
 serve to limit the number of materials that are capable 
 of conversion into smokeless powder, the limitations 
 arising not only ffom the above-named practical re- 
 quirements, but also from the chemical impossibility of 
 existence of many bodies which, if obtainable, would 
 decompose in the manner requisite for efficient ballis- 
 tic action. Thus, e.g., there does not exist, nor can 
 we look forward to the existence of, such a polymer 
 (in the solid or liquid form) of hydrogen as H , which 
 would decompose into hydrogen, H a , with a corre- 
 sponding production of heat.f 
 
 If we may not look for explosives among the sim- 
 plest chemical combinations of the elements, we may 
 perhaps find them among those compounds of nitro- 
 gen and hydrogen which stand in the same relation to 
 ammonia, NH 3 , as the hydrocarbons do to methane, 
 
 * And is not derived from an external source of energy, such as 
 the tension of a spring or the physical compression of gases, as 
 in the Giffard gun. 
 
 f If such a substance existed, then its decomposition, according 
 to the equation H a = wH a , would afford a weight in for a volume 
 2; that is, KJOOO would be equal to 1000, the greatest possible 
 value. For nitrogen under similar conditions we would have 
 Fiooo = 71.4, or less than for pyrocollodion. But the existence 
 of such a polymer is highly improbable. If argon (vid. Mende- 
 leef, " Principles of Chemistry," 6th ed., 1895, p. 749) were the poly- 
 mer of nitrogen, Na, its conversion into nitrogen could only be 
 accomplished through the absorption of heat; i.e., it would find no 
 place in the category of " explosive " bodies (to which ozone pos- 
 sesses a relation). 
 
PYROCOLLODION SMOKELESS POWDER 
 
 CH,. We should thus have, corresponding to ammo- 
 nia, the series N M H M + 2 (e.g., diamide N 2 H 4 ; triamide 
 N 3 H & ; etc.) and the series N W H M , N M H W _ 2 , etc. As 
 the representative of the latter we have, for n = 3, the 
 nitro-hydric acid of Curtius, N 3 H, which actually is a 
 very explosive body, and which forms salts, e.g., with 
 ammonium N 3 (NH 4 ) = N 4 H 4 , which is also explosive, 
 decomposing into the gases nitrogen and hydrogen 
 with the evolution of heat, although ammonia itself is 
 not susceptible of explosive decomposition, but ab- 
 sorbs heat in the reaction. If such compounds could 
 be easily prepared, and if they possessed the qualities 
 necessary to an efficient smokeless powder, such as 
 non-volatility, good keeping quality, progressiveness 
 in combustion, etc., they would prove especially suit- 
 able for conversion into smokeless powders, as the 
 corresponding values of F 1000 would be greater than 
 for other powders. Thus we should have for nitro- 
 hydric acid, 
 
 2N 3 H = H,+ 3 N 2 . 
 Mol, wt. 2X43= 2+3x28=86. 
 Volume of gases, 2 -j- 3 X 2 = 8. 
 ^1000 = 93-0- 
 
 For N M H M the volume would be still greater, as 
 F ]000 133.3. But even if such products could be con- 
 veniently prepared from readily procurable materials it 
 would be useless to consider them as available for 
 conversion into smokeless powders, for the reason 
 that they do not decompose through gradual or 
 progressive combustion, which is indispensable in a 
 
HO SMOKELESS POWDER 
 
 smokeless powder, but detonate or decompose with 
 extreme suddenness; hence, while they might prove 
 suitable for filling mines or shells, they are unadaptcd 
 for use in cannon. This property of progressive com- 
 bustion or decomposition in successive strata is pos- 
 sessed only by those substances containing both com- 
 bustible ingredients, and ingredients capable of effect- 
 ing progressive combustion, such as carbon and 
 hydrogen, which are consumed by the oxygen held in 
 close proximity to them but which is not in direct 
 combination with them. 
 
 The " combustion " of a powder is the union of 
 the carbon and hydrogen of the mixture or com- 
 pound with the oxygen that it contains, and with 
 which it is in association, but not in direct combina- 
 tion. From what has been said already, it is evident 
 that if the powder is to be smokeless and produce the 
 maximum volume of gas, FI OOO , it must evolve no 
 other gases than carbonic oxide, CO, water vapor, 
 H a O, and nitrogen, N a . If hydrogen be evolved, with- 
 out the formation of the corresponding quantity 
 (equal volume) of carboni,c acid, free carbon may re- 
 sult, i.e., the powder will not be wholly smokeless 'on 
 account of insufficiency of oxygen. If the combus- 
 tion, as indicated by the equation, reveals carbonic 
 acid or free oxygen (without the corresponding vol- 
 ume of hydrogen), an excess of oxygen is evident, 
 and F^oo,, will not possess its maximum value. 
 
 We have, therefore, in the case of a composition or 
 mixture CH 2w N ? O r , the maximum volume F" 1000 for 
 typical smokeless combustion, corresponding to two 
 
PYROCOLLOD1ON SMOKELESS POWDER HI 
 
 conditions: (i) When the content of oxygen, r, is 
 just sufficient to convert the carbon into CO, and 
 the hydrogen into H 2 O, i.e., when r n + ;// ; (2) 
 when the content of hydrogen is relatively great, as 
 F 1000 for H 2 O equals HI.I, i.e., more than for nitro- 
 gen and for carbonic oxide, for which F 1000 f{p 
 = 71.4. Moreover, all substances of the composition 
 QH 2w N^O M+wt will develop volumes between 71.4 and 
 HI.I, if the decomposition products be CO, N 2 and 
 H a O alone, as is required for rendering F 1000 a maxi- 
 mum. Our problem becomes, therefore, the com- 
 parative examination of those bodies rich in hydrogen, 
 for which F looo may be greater than for pyrocellulose 
 (81.5). We must ask: Are there not known sub- 
 stances, or mixtures 'of substances, rich in hydrogen 
 suitable for smokeless powder? To answer this query, 
 let us examine various definite compounds and mix- 
 tures. 
 
 Among the carbon compounds a large content of 
 hydrogen is characteristic of methane (marsh gas), 
 CH 4 ; also among the nitrogen compounds, or the 
 ammonium derivatives. 
 
 Hydrocarbons of the limiting (saturated) series 
 C w H 2M + 2 * form nitre-compounds, and may, therefore 1 , 
 produce explosives. To methane itself, CH 4 , there 
 correspond mononitro-methane, CH 8 (NO 2 ); dinitro- 
 methane, CH 2 (NO 2 ) 2 ; trinitro-methane or nitroform, 
 CH(NO a ),, and tetranitro-methane C(NO 2 ) 4 . These 
 
 *See "The Principles of Chemistry," by D. Mendeleef, 1891, 
 Longmans, Green & Co., London, Vol. I, p. 344. J. B. B. 
 
112 . SMOKELESS POWDER 
 
 substances are volatile as well as explosive, but all 
 represent a deficiency or an excess of oxygen. As 
 shown by V. Meyer and Professor Zalinski, the ex- 
 plosive properties of mononitro-methane are espe- 
 cially great when it is combined with potassium or 
 sodium to form the metallic salts, CH 2 KNO 2 and 
 CH 2 NaNO 2 , which represent, so to speak, first homo- 
 logues of the salts of nitric acid, since CH 2 NaNO 2 
 NaNO 2 equals the homologous difference CH 2 . 
 Experiment shows that this substance belongs to 
 the category of detonating explosives, and is, there- 
 fore, unsuitable for use in guns (but suitable for 
 mines). 
 
 If the decomposition proceeds without formation of 
 free carbon (although there be but little oxygen), it 
 should be as follows: 
 
 2CH 3 NO 2 = 2CO + 2H 2 O + H 3 +N a . 
 
 If it be thus, then F 1000 = 98.3, which is very great. 
 But, as has been said, the substance is unfit for use in 
 guns on account of its tendency to detonate. Be- 
 sides, like other nitro-methanes, it is volatile, and for 
 this reason is further unadapted. 
 
 The little known, but doubtlessly explosive, dinitro- 
 methane contains an evident excess of oxygen, devel- 
 oping on combustion, CO, + H U O +"iC) 2 +N 2 , which 
 corresponds to the relatively small volume F 1000 = 66. 
 It is evident that the excess of nitrogen and of oxygen 
 combined with it in the NO 2 , according to the known 
 principle, does not increase but rather diminishes 
 F 1000 . The same is true for nitroform, CH(NOJ 3 = 
 
PYROCOLLODION SMOKELESS POWDER 113 
 
 CHN 3 O 6 ,* and for tetranitro-methane, the discovery 
 of which is due to the skill of our eminent savant, 
 L. N. Shishkov. Both of these substances contain too 
 much oxygen to develop maximum gas volumes. A 
 large value of fj 000 would be characteristic of mixtures 
 of products of nitration and of hydration (substituting 
 the water radicle for hydrogen, H + OH) derived 
 from methane, as : 
 
 4 CH,NO, + CH,N,0 4 - 
 
 Mol. wt., 4x61 + io6=5X28+;x 18+3x28 = 350. 
 
 Volume of gases, 5X2+7X2+3X2 =30. 
 
 ^000 = 85. 
 Or: 
 
 Nitro-methane Methyl alcohol 
 
 4CH,NO S +CH 4 = 5CO +8H,O+2N,. 
 Mol. wt.,4 X 77+32 = 5x28+8x18+2X28 = 340. 
 Volume of gases =5x2+8X2+2x2 =30. 
 F; ooo = 88.2, etc. 
 
 But such mixtures, although possible from the chem- 
 ical standpoint, are unsuitable for use as powder, be- 
 cause their constituents are in part volatile ; and this, 
 apart from the consideration that liquid explosives are 
 prone to detonation, which is more to be dreaded than 
 formation of smoke, as detonation destroys the guns. 
 
 * As the typical decomposition of nitroform, we have 
 
 4 CHN 3 8 =4C0 2 + 70 2 + 2H 2 O + 6N a . 
 Mol. wt., 4 X 151 =4X44 + 7X32 + 2X18+6X28 = 604. 
 
 Volume of gases 4 X 2 +7X2 +2X2 +6X2 =38. 
 f The mixture 7CH 4 -+- 3C(NO 2 ) 4 presents such a composition, 
 etc., but such mixtures are all as practically unsuitable for pow- 
 der as mixtures of mono- and dinitro-methane. 
 
H4 SMOKELESS POWDER 
 
 , Among the closely allied derivatives of methane as 
 a hydrocarbon rich in hydrogen, the development of 
 a large gas volume may be looked for from substances 
 presenting the composition CN 2 H 6 O 4 . Such a com- 
 position is possessed, for example, by the mixture 
 of a molecule of formic aldehyde, CH 2 O (or of one 
 of its numerous polymers), with ammonium nitrate, 
 NH 4 NO S , or the hydroxyl derivative of methylamine 
 (i.e., CH 3 NH 2 in the form CH 2 [OH]NH 2 ) in com- 
 bination with nitric acid, HNO 3 . The typical de- 
 composition of such a compound, if realized, would be 
 expressed by the equation : 
 
 CN 2 H 6 4 = CO+ 3 H 2 +N 2 . 
 Mol. wt., 1 10 = 28 + 3 X 18 + 28 = 1 10. 
 
 Volume of gases = 2 +3X2 + 2 10. 
 ^1000 = 9-9- 
 
 But such a compound either cannot be produced, or 
 else is obtained only with great difficulty ; or, as a mix- 
 ture of ammonium nitrate with the polymers of formic 
 aldehyde (e.g., glucose, C 6 H 12 O 6 = 6CH 2 O), it develops 
 undesirable qualities, such as hygroscopicity, a charac- 
 teristic of all mixtures containing ammonium nitrate, 
 and is therefore unsuited for use as smokeless powder. 
 
 Hence, after searching through all the possible com- 
 binations of the simplest derivatives of methane, we 
 are unable to find among them (as also among sub- 
 stances containing no carbon) any suitable for employ- 
 ment in practice as smokeless powder, although we 
 find compounds developing larger volumes of gas than 
 pyrocollodion, and which may prove suitable for use 
 in mines. 
 
PYROCOLLODION SMOKELESS POWDER 11$ 
 
 If we turn from substances containing one atom of 
 carbon to those with two, three, etc., atoms to the 
 molecule, we shall find, other conditions being the 
 same, smaller values of Fi 000 , the volume decreasing 
 the farther the limit is departed from, as is illustrated 
 in the following table of possible, little volatile, com- 
 pound ethers of nitric acid* and their hypothetical 
 nitro-compounds, corresponding to the series of alco- 
 hols, C.HJNO,),. 
 
 3 C 2 H 4 (N0 3 ), + 2C.H.O, F 1000 = Tftfr 1000-86.2 
 
 C 3 H 6 (N0 3 ) 3 " - Mi 1000-84.3 
 
 QH B (N0 3 ) 2 -f 4 C 4 H,(N0 2 )(N0 3 ) 2 " - T ^T 1000-83.7 
 C 6 H 10 (NO,) 2 + 4 C 6 H 8 (NO a ) 3 (NO s ) a ' - T Wo 1000-82.7 
 
 etc., etc. 
 
 The possible, yet up to the present, hypothetical, 
 nitric ethers of nitro-glucol, although capable of devel- 
 oping large values of F 1000 , and adaptable on account 
 of their non-volatility, possess no advantages over 
 derivatives of the higher alcohols, such as glycerin 
 and mannite, materials that are readily obtainable 
 as they are widely disseminated throughout nature. 
 We shall therefore fix our attention upon the latter, 
 first, as they present in their analogues substances ex- 
 tremely rich in hydrogen, capable of producing large 
 values of F 1000 ; and second, because they are easily re- 
 
 * Considered by themselves these ethers of diatomic limiting 
 (saturated) alcohols C,zH 2w (NO 3 )2 consume into CO and H 2 O only 
 forw = 3. For greater values of there is a deficiency in oxy- 
 gen; for n = 2, an excess. We have chosen them as an example 
 on account of their slight volatility, and because they approximate 
 nitro-glycerin and nitro-mannite in composition. 
 
Il6 SMOKELESS POWDER 
 
 acted upon by nitric acid, forming the explosive com- 
 pound ethers, nitro-glycerin, C fl H & (NO 3 ) 3 = C e H 6 N s O 8 , 
 and nitro-mannite, C 6 H 8 (NO 2 ) 6 O fl = C e H 8 N 6 O 18 . Both of 
 these nitro-derivatives are easily prepared. The for- 
 mer was first employed as an explosive by the re- 
 nowned Russian chemist N. N. Zinin, at the time of 
 the Crimean war, and subsequently by V. F. Petru- 
 shevski, in the sixties, before the discovery and very 
 general employment of Nobel's dynamite and other 
 nitro-glycerin preparations; the cause of their gen- 
 eral use being the ease with which the base material 
 glycerin was obtainable in nature, while the reac- 
 tion with nitric acid (admixed with sulphuric) was 
 easily effected, i.e., the manufacturing process was a 
 simple one. 
 
 Nitro-mannite, isolated and investigated by N. N. 
 Sokolov, professor at the Medico-Chirurgical Academy, 
 is also easily prepared, but not in its lower degrees of 
 nitration. This circumstance is important, for the 
 reason that the readily manufactured nitro-glycerin 
 and nitro-mannite are not themselves available for use 
 in guns, although very well adapted for detonating 
 effects. They correspond, moreover, to relatively 
 small values of F 1000 , as they contain an excess of 
 oxygen : 
 
 C.H.(NO,). affords F 10M = 63.9. 
 C.H.(NO S ). " F mo = 6i. 9 . 
 
 But as these substances contain an excess of oxygen, 
 they may be admixed with others containing a defi- 
 ciency thereof, and which they consume, evolving car- 
 
PYROCOLLODION SMOKELESS POWDER I I/ 
 
 bonic oxide and developing relatively large volumes, 
 F" 1000 ; while by admixture with such substances, low in 
 oxygen, or not containing it, their detonating qualities 
 may be caused to diminish, or made to vanish, as in 
 dynamite, by combination with an inert base (tripoli, 
 magnesia, etc.), whereby the tendency of nitro-glyc- 
 erin to detonation through shock is .diminished. In 
 this manner, by admixture with a combustible sub- 
 stance, nitro-glycerin powders are formed. If we take 
 cordite as an example, we find that on account of 
 its excess of oxygen it produces a relatively small gas 
 volume ; we may, therefore, select a mixture of nitro- 
 glycerin and collodion (assumed as C 6 H 8 [NO 2 ],O 6 ) 
 such as ballistite and determine for it F 1000 on the as- 
 sumption that it shall develop only CO, H 2 O and N,. 
 
 2C 8 H 6 N 3 9 + 7 C 6 H 8 N a 9 = 4 8CO -j- 33 H,O+ ioN,. 
 Mol.wt.2X227-f 7 X 252 =48X28-1-33X18-}- 10X28 2218. 
 Volume of gases =48 X 2 +33X2 +10X2 =182. 
 
 Therefore, if nitro-glycerin powders contain only 
 the quantity of nitro-glycerin necessary to produce 
 H 2 O and CO, then the volume of gases evolved by 
 them is almost the same as that developed by pyro- 
 collodion. It is evident, then, that neither nitro- 
 glycerin nor its mixtures, when employed as smoke- 
 less powder, evolve volumes of gases greater than pyro- 
 collodion, and that admixture with other substances, 
 of whatever kind they may be, although homogeneous 
 from the mechanical standpoint, are still far less homo- 
 geneous than any single substance, and that it is use- 
 
118 SMOKELESS POWDER 
 
 less to seek for nitro-glycerin powders capable of ex- 
 ceeding pyrocollodion powders in point of magnitude 
 of F 1000 , apart from other considerations. This ap- 
 plies also to nitro-mannite, the source of preparation 
 of which is far less common than glycerin, and to 
 many of the hydrocarbons analogous thereto, as glu- 
 cose, starch, cellulose, etc. If all of the six atoms of 
 carbon in mannite are in the same combination as in 
 the limiting (saturated) alcohols: 
 
 C,H,A= 
 
 CH 3 (OH)CH(OH)CH(OH)CH(OH)CH(OH)CH 2 (OH), 
 
 then in glucose, C 6 H 1U O 6 , one atom of carbon should 
 be combined as in aldehyde, 
 
 C 6 H 1Q 6 = 
 
 COH CH(OH)CH(OH)CH(OH)CH(OH)CH a (OH), 
 
 and, therefore, if mannite represents a hexanitrated 
 product (compound ether as derived from an alco- 
 hol), glucose represents only a pentanitrate. Mate- 
 rials such as cellulose, starch, and the like, of compo- 
 sition C e H 10 O 6 , may be regarded as the preceding alco- 
 hols anhydrous, arranged as follows with reference 
 to the di-alcoholic groups: 
 
 CHCH, 
 
 C 6 H 10 O 8 = COH CH(OH) CH(OH) CH(OH) ^~ , 
 
 O 
 
 whereby it appears that there are only three com- 
 plete alcohol groups remaining out of the six in man- 
 nite. Therefore, "in the latter case, we should expect 
 to find only trinitrated products, which is actually 
 what occurs. If such a scheme throws light on the 
 
PYROCOLLODION SMOKELESS POWDER I 19 
 
 matter from one standpoint (as relates to the number 
 of hydroxyl radicles giving rise to nitric ethers), it illu- 
 minates it obliquely from another, which is of consid- 
 erable importance to us. In all aldehydes, beginning 
 with the formic and acetic, a tendency to polymeriza- 
 tion is to be noted, due, doubtless, to the property 
 of aldehydes of entering into various combinations 
 (with H 3 , O, NaHSO 3 , etc.); hence the composi- 
 tion C 6 H lft O 6 , containing an aldehyde grouping, 
 should also possess this property, so far as relates 
 thereto. We may, therefore, safely assume that the 
 molecular composition of cellulose, judging from its 
 properties, is polymerized, i.e., it is of the form 
 C 6W H 10W O 6M , where n is probably very great. If we as- 
 sume n 5, the cellulose becomes of composition 
 C 30 H 60 O 2B , and, for the highest degree of nitration, 
 C SO H 3B (NO 2 ) 16 O 2 ,. But pyrocellulose has a composi- 
 tion C, H 88 (NO 9 ) 19 O aB ; therefore, the number of inde- 
 pendent nitro-celluloses (nitric ethers) may be very 
 large. 
 
 This is very important in the conception that the 
 nitration of cellulose may be carried up to any desired 
 degree, and for known concentrations a mixture of 
 nitric and sulphuric acids neither dissolves nor reacts 
 upon a product of nitration. Again, cellulose is the 
 most widely disseminated in nature of all the hydro- 
 carbons possessing the composition C,H 10 O 5 , consti- 
 tuting the tissues of all plants and prepared from time 
 immemorial in great masses as cotton, flax, paper, 
 etc., while its products of nitration present an unal- 
 terable material suitable for conversion into smokeless 
 
120 SMOKELESS POWDER 
 
 powder. This side of the matter needs no further 
 elucidation, but it must be remembered that before 
 the development of pyrocollodion, it was stated that 
 the higher the nitration of the cellulose the higher the 
 explosive produced, and that in manufacturing powder 
 from highly explosive nitro-cellulose (of composition 
 about C 30 H 36 [NO 2 ] 14 O 25 ), collodion (of composition about 
 C 30 H 40 [NO a ], O a6 ) was added, for the reason that higher 
 nitro-celluloses in the form of filaments or dust were 
 easily detonated (hence their employment for mines), 
 while the latter property was reduced or caused to dis- 
 appear after gelatinization, of which collodion was 
 easily susceptible and for which purpose it was added. 
 The introduction of pyrocollodion changed existing 
 views upon the subject, showing that maximum force 
 for nitro-cellulose was not to be sought from the high- 
 est degrees of nitration (i.e., for maximum content of 
 nitrogen and oxygen), but that it obtained for that 
 mean degree of nitration present in pyrocollodion. 
 For the latter material F" ]000 = 81.5 ; while for nitro- 
 cellulose of maximum nitration, C 6 N 7 (NO 3 ) 3 O 5 , F 1000 = 
 74. i . The above represents only one side of the theo- 
 retical investigation of materials suitable for smoke- 
 less powder; but other considerations are also in ac- 
 cord, as will be shown later; and we have, therefore, 
 gone considerably into detail, the more urgently since 
 it has been necessary to struggle with prejudice, 
 harmful to success in such a new field as that of 
 smokeless powders. 
 
 Among the possible materials proposed, apart from 
 mixtures of such different bodies as ammonium nitrate 
 
PYROCOLLODION SMOKELESS POWDER 121 
 
 and various organic substances (such mixtures were re- 
 jected in practice), must be considered the nitro-com- 
 pounds corresponding to benzol and its derivatives, 
 naphtalin, etc., as coal-tar constitutes an abundant 
 source for their production in large quantities, and 
 they are easily nitrated. From the class of the so- 
 called "aromatic compounds" derived from benzol, 
 C 6 H 6 , it is useless to expect smokeless explosives de- 
 veloping large volumes of F" 1000 , although many are 
 high explosives, beginning with Melinite or picric 
 acid, C 6 H 3 (NO,),OH, which constitutes a powerful 
 material, although far from the best, for torpedoes 
 and explosive shells; and since some of the first 
 smokeless powders were mixtures containing picric 
 acid. The cause of the small gas volumes F 1000 devel- 
 oped by the aromatic compounds is due to their com- 
 position, as they are all low in hydrogen. This view 
 may be illustrated by a few examples. 
 
 To benzol there correspond bodies the general com- 
 position of which may be expressed by the formula 
 C 6 H,_ a (NO a ) a . If a equals 1,2 or 3 (these substances 
 are known and easily obtained), the oxygen content is 
 insufficient to consume the carbon into CO and the 
 hydrogen into H 2 O, although explosion occurs with 
 the formation of carbon (smoke, soot) and of hydro- 
 carbons. 
 
 Total smokelessness could be realized from mixtures 
 of highly nitrated products, as : 
 
 4 C 9 H a (N0 2 ) 4 + C 6 H 4 (NO,) a = 3 oCO +6H a O+ 9 N a . 
 Mol. \vt., 4X258+ 168 = 30X28 + 6X18+9X28=1200. 
 Volume of gases 30 X 2 +6X2 +9X2 =90. 
 V 71 
 
 ' 1000 / D* 
 
122 SMOKELESS POWDER 
 
 If, instead of such non-existing highly nitrated ben- 
 zols, pyroxylin be employed (as in the American 
 smokeless powders of Munroe and other inventors), a 
 gas volume approaching that developed by pyro- 
 collodion is realized : 
 
 6C,H 6 NO a -f 26CeH7N3Ou =19200 + io6H 2 O 
 Mol. wt., 6 X 123 + 26X297 = 192 X 28+106 X 18+42X28=8460. 
 Volume of gases = 193 X 2 + 106 X 2 +42X2 =680. 
 Tiooo = 80.4. 
 
 The matter assumes a different aspect if ammonium 
 nitrate, NH 4 NO 3 , be employed for converting the car- 
 bon and hydrogen of nitro-benzol into CO and H a O. 
 This substance contains a large quantity of hydrogen for 
 a'relatively small content of oxygen (but for this re- 
 sult a large quantity of NH 4 NO 3 must enter into the 
 mixture), and greatly increases the volume of gas 
 developed, as is evident from the following: 
 
 2C 9 H.NO a + I3NH 4 NO 3 = isCO + 3iH 2 O + I4N 8 . 
 Mol. wt., 2 X 123 + 13 X 80 = 12 X 28 + 31 X 18 + 14X28 = 1286. 
 Volume of gases = 12 X 2 + 31 X 2 +14X2 =114. 
 
 F"iooo 88.7- 
 
 But mixtures of this salt must always be avoided if 
 a satisfactory smokeless powder is to be produced, as 
 it is soluble in water, as well as hygroscopic, arid pro- 
 duces with viscous, oily materials only coarse mechan- 
 ical mixtures. 
 
 Similar results are to be obtained from other aro- 
 matic substances, and we may refer by way of ex- 
 ample to mixtures of pyroxylin with picric acid, 
 C 6 H,(NO 2 ) 3 OH, and ' nitro-naphtalin, C 10 H,(NOJ, 
 such compounds having been recently experimented 
 
PYROCOLLODION SMOKELESS POWDER 123 
 
 with (but abandoned in practice) as smokeless pow- 
 ders. Among other disadvantages, they develop 
 smaller gas volumes, F ]000 , than pyrocollodion, on 
 account of their relatively small content of hydrogen. 
 
 After an examination in the above manner of the 
 composition and properties of all possible materials 
 capable of employment as smokeless powders, we 
 arrive at the following deductions in relation to the 
 volume of gases (measured at given temperature and 
 pressure), F 1000 , developed by their combustion: 
 
 i. Only substances containing nitrogen, carbon, 
 hydrogen and oxygen are capable of entire conversion, 
 as required for smokeless powders, into gases that do 
 not react upon gun-metals. Hence all other explo- 
 sive substances (e.g., fulminate of mercury, chloride of 
 nitrogen, etc.) containing haloids, metals, phosphorus, 
 etc., are unsuitable for use in gunpowders. 
 
 2. When the combustion of carbon results in the 
 formation of carbonic acid gas, CO 2 , a less volume of 
 gas is formed than when carbonic oxide, CO, is the 
 resultant product ; and as the former requires more 
 oxygen than the latter, the increase of oxygen or 
 nitrogen (if, as is usually the case, the oxygen enters 
 into combination with the aid of the elements of nitric 
 acid) is injurious, instead of useful, although there 
 exists full conversion into gases as is required for 
 smokeless powder. 
 
 3. The greater the quantity of hydrogen in the 
 powder, other conditions being equal, the greater the 
 gas volume, F 1000 , corresponding to the combustion of 
 the powder; and, therefore, substances derived from 
 
124 SMOKELESS POWDER 
 
 the limiting (saturated) series of hydrocarbons are 
 more suitable than bodies of the "aromatic" series 
 for smokeless powders. 
 
 4. Not any of these explosive materials not contain- 
 ing carbon (as N,H, NH 4 NO a ), that evolve large vol- 
 umes of gas F 1000 and decompose upon ignition, are 
 such as will not detonate, i.e., evolve their gases so 
 rapidly that they crush the walls of guns ; whence it is 
 useless to consider them as materials adaptable for 
 conversion into smokeless powders. 
 
 5. Some of the materials containing but little car- 
 bon and much hydrogen may prove suitable for use as 
 powders or powder mixtures, evolving large volumes 
 of gas upon combustion ; but, so far as known, they 
 are either volatile, or liable to decompose spontaneously 
 and detonate, or else they are prepared with difficulty 
 from mixtures not widely disseminated, so that at 
 present it is useless to look for materials for smokeless 
 powder from among them. 
 
 6. Nitro-glycerin itself develops but a small gas 
 volume (Fi 000 = 63.9), as it contains an excess of oxy- 
 gen. It may be employed in mixtures to form smoke- 
 less powder, and its mixtures with nitro-cellulose, such 
 as Cordite and Ballistite, which are practically homo- 
 geneous from a physical standpoint, develop gas 
 volumes, F,, 00 , a little less than that evolved by pyro- 
 collodion (although such mixtures erode guns, as 
 already stated). 
 
 7. Cellulose, C 6W H 10W O BM , is a substance widely dis- 
 seminated in nature and of general industrial employ- 
 ment; by its non-volatility, insolubility, durability, 
 
PYROCOLLODION SMOKELESS POWDER I2J 
 
 etc., and by the readiness with which it is nitrated (as 
 it contains much hydrogen), it constitutes a superior 
 base for smokeless powders. 
 
 8. Among all the forms of nitro-cellulose capable of 
 smokeless combustion, the maximum gas volume, F, 000 , 
 corresponds to C 30 H 88 N la O 49 (= 12.44 P er cent, nitro- 
 gen), which is pyrocollodiori, for which F" 1000 = 81.5, 
 and which is capable of complete gelatinization in a 
 mixture of ether and alcohol, in which form it is com- 
 pletely free from any tendency to detonate. In the 
 first place, it is the most suitable of all the nitro- 
 celluloses; in the second, it is the most rational and 
 readily obtainable form of smokeless powder, destined 
 to supplant not only other smokeless powders, but also 
 to replace, by reason of its greater homogeneity- and 
 its combination of qualities, other pyroxylin powders. 
 
 Pyroxylin powder is a mixture of nitro-celluloses, 
 of higher nitration, such as C 6 H 7 (NO 2 ) 3 O 6 , and of lower, 
 as C 6 H 8 (NO 2 ) a O 5 ; pyrocollodion is a definite homo- 
 geneous single form of nitrocellulose. By changing 
 the proportional relation of contents of the high, or 
 insoluble, and low, or soluble, nitro-cellulose, it is 
 evidently possible to make the pyroxylin approach 
 the pyrocollodion powders ; but (as has been shown 
 in recent years, especially in cannon-powders) the limit 
 of improvement of these forms always falls short of 
 pyrocollodion. The latter is homogeneous and un- 
 changeable, while pyroxylin powders vary according 
 to their composition. However, from its origin it is 
 in no wise different from the perfected powder of 
 Vieille (although considerably different from the 
 
126 SMOKELESS POWDER 
 
 original form thereof), presenting instead of a mix- 
 ture, from the chemical and mechanical standpoint, 
 a homogeneous limiting mass of the composition 
 C, H 38 N J3 O 49 , which is required in order that the pow- 
 der may create upon combustion the maximum volume 
 of vapor and gases. It is certain that henceforth 
 pyroxylin powder will continue to approximate to 
 the pyrocollodion until the two become identical. In 
 brief, pyrocollodion represents the Russian limit of 
 modification and improvement of the French pyroxy- 
 lin powders, the development of which marked an 
 epoch in ordnance progress, but which has not hitherto 
 presented an invariable and constant relation of the 
 elements entering into its composition. In this light 
 pyrocollodion powder may well be styled Franco- 
 Russian. Begun in France, it has been completed in 
 Russia. 
 
APPENDIX III 
 
 THE NITRATION OF COTTON 
 By M. BRULEY 
 
 WITHIN recent years different varieties of nitro- 
 cellulose which hitherto have possessed only very 
 restricted uses have had their fields of application 
 broadened, especially in their relation to the manu- 
 facture of explosives. 
 
 Photographic collodions, celluloid, artificial silk, 
 etc., have nitro-cellulose for their bases. On the 
 other hand, besides gun-cotton, employed for military 
 purposes, as for the charging of torpedoes and other 
 similar appliances, nitro-cellulose enters into the com- 
 position of various kinds of smokeless powders, gum- 
 dynamites, cartridges for use in presence of fire-damp, 
 etc. 
 
 Each of these applications demands, so to speak, a 
 special variety of nitro-cellulose, and the government 
 explosive-factories entrusted with their preparation, 
 so far as relates to their use as explosives, have been 
 gradually forced to meet requirements differing widely 
 as to character. 
 
 For a long time the Moulin Blanc factory, which 
 was established for the production of naval gun- 
 
 127 
 
128 SMOKELESS POWDER 
 
 cotton, had scarcely more to do than to produce gun- 
 cotton of maximum nitration for military uses. For 
 this the well-known mixture of three parts sulphuric 
 acid by weight of 65.5 Baume" and one part by weight 
 of nitric acid of 48 Baume, has always been employed. 
 
 Since then the Moulin Blanc works and those at 
 Angouleme, where the manufacture of gun-cotton was 
 established in 1887, have had to produce other kinds 
 of nitro-cellulose, which it became necessary to man- 
 ufacture practically in a regular manner. 
 
 The theoretical researches of M. Vieille, described 
 in his note of September 13, 1883, inserted in Vol. II 
 of the 4< Memorial des Poudres et Salpetres " (p. 212 
 et seg.), classified different varieties of nitro-cellulose 
 in the manner following, in accordance with the 
 formulae for their chemical composition : 
 
 Vol. of Nitrogen Dioxide 
 Disengaged per gram 
 
 Cellulose endecanitrate. . . 214) 
 
 I gun-cottons. 
 " decanitrate 203 ) 
 
 enneanitrate 190 
 
 octonitrate 178 V collodions. 
 
 " heptanitrate 162 ) 
 
 " hexanitrate 146 \ 
 
 " pentanitrate 128 V friable cottons. 
 
 " tetranitrate 108 ) 
 
 Gun-cotton, as well as friable cottons, are insoluble 
 in a mixture of alcohol and sulphuric ether, while 
 collodions, on the contrary, are soluble in such a 
 mixture. 
 
 Indeed, it is hardly possible to isolate each one of 
 
THE NITRATION OF COTTON I2Q 
 
 these varieties, and one always finds mixtures of two 
 or more neighboring varieties. 
 
 Thus in practice, the products obtained are classed, 
 not into cellulose endecanitrate, cellulose decanitrate, 
 etc., but into gun-cotton properly so-called, for which 
 the content of nitrogen is between 210 and 200 c.c., 
 into superior and inferior collodions, with nitrogen 
 content ranging between 190 and 180 c.c. for the 
 former and from iSoto 170 for the latter; and friable 
 cottons. 
 
 Gun-cotton, properly so-called, is theoretically in- 
 soluble in ether-alcohol. In reality it always contains 
 small quantities of less highly nitrated soluble products. 
 Collodions deposit in this solvent an insoluble residue, 
 due to the presence of non-attacked cotton, or friable 
 cottons, or even more highly nitrated cottons. 
 
 It is for this reason that solubility in ether-alcohol 
 is determined the same time that nitration is estimated. 
 
 The different soluble collodions are distinguishable 
 from one another, moreover, by the greater or less 
 viscosity of the solution obtained. As this property 
 is one that may possess practical importance in the 
 employment of collodions, it appears useful to study 
 it simultaneously with the solubility and the content 
 of nitrogen. 
 
 A series of researches were undertaken at the An- 
 gouleme powder-works for the purpose of determining 
 the practical method of obtaining a chosen product 
 from the series of those just enumerated. These 
 researches do not claim to possess the scientific value 
 of those made in 1883 by M. Vieille upon the nitra- 
 
130 SMOKELESS POWDER 
 
 tion of cotton. The resources at the powder-works 
 laboratory, especially in relation to personnel, are too 
 limited to justify the hope of rigorous precision for 
 results obtained. But apart from their utility as a 
 guide in the manufacture of a desired product, they 
 seem to possess a certain interest in relation to the 
 conditions controlling the nitration of cotton. 
 
 I. METHOD OF NITRATION AND MODE OF REPRE- 
 SENTING RESULTS 
 
 M. Vieille in his researches in 1883 had employed 
 two methods of nitration. In one he used mixtures 
 containing only nitric acid and water. In the other 
 he added to these bodies variable proportions of sul- 
 phuric acid. 
 
 The latter method has been 'alone employed in 
 practice; it is the one exclusively used in the trials 
 that constitute the object of the present work. 
 
 Apart ffom nitrous vapors, which are always present 
 in appreciable but very feeble quantities, and foreign 
 matters, which are never found in the acids employed 
 except in negligable quantities, every mixture em- 
 ployed for dipping cotton for the production of any 
 one of the various nitro-celluloses contains the three 
 following elements: monohydrated sulphuric acid, 
 H 2 SO 4 ; monohydrated nitric acid, HNO 3 ; and water, 
 H 3 0. 
 
 If we desire to classify the infinite number of mix- 
 tures that may be obtained from all possible combin- 
 ations of the three elements, in groups giving rise 
 
THE NITRATION OF COTTON !$! 
 
 finally to products of the same kind, it is necessary to 
 have recourse to a graphic method of representation of 
 these mixtures. 
 
 In expressing the relations of the proportions of 
 the two elements to the third, each mixture is de- 
 fined by only two numbers. It is easy, then, by em- 
 ploying one of them as an abscissa and the other as 
 an ordinate, to express by the ordinary system of 
 rectangular coordinates any mixture. determined by a 
 single point. 
 
 The different mixtures giving rise to the same 
 product will thus be grouped by zones, and a knowl- 
 edge of these zones^ill facilitate the practical obtain- 
 ing of a desired nitro-cellulose, whether new acids are 
 employed to produce it or whether spent acids from a 
 previous manufacture are employed for the purpose. 
 
 II. RESULTS OBTAINED BY M. VIEILLE 
 
 We shall reconsider, in the order of ideas just in- 
 dicated, the results obtained by M. Vieille in 1883, 
 and mentioned in the preceding note. 
 
 In all experiments under the second method, sul- 
 phuric acid of density 1.832 was employed, corre- 
 sponding, according to the special tables, to a degree 
 by areometer of 65.5 Baume and to a percentage of 
 water of about 8 ; and, upon the other hand, to nitric 
 acid of density 1.316 which, if it be supposed free from 
 all traces of nitrous vapors, should mark 34.6 by the 
 Baume areometer and contain 50$ of water. 
 
 Applying these figures to the volumetric propor- 
 
132 
 
 SMOKELESS POWDER 
 
 tions of each mixture given in the table on page 87. 
 the following new table is formed : 
 
 No. of Ex- 
 
 periment 
 
 Proportions of Components 
 
 Per cent, of Ni- 
 trogen Dioxidein 
 Product ob- 
 tained 
 
 Observations 
 
 H 2 S0 4 
 
 HNOg 
 
 H a O 
 
 I 
 
 100 
 
 13.1 
 
 21,7 
 
 1959 
 
 
 2 
 
 IOO 
 
 15.5 
 
 24-3 
 
 igo.I 
 
 
 3 
 
 IOO 
 
 19-5 
 
 28.2 
 
 184.6 
 
 
 4 
 
 IOO 
 
 22.9 
 
 3L7 
 
 185.5 
 
 
 5 
 
 IOO 
 
 26.0 
 
 34-8 
 
 182.3 
 
 Solubility 
 
 6 
 
 IOO 
 
 27.8 
 
 36.5 
 
 164.0 
 
 not deter- 
 
 7 
 
 IOO 
 
 30.0 
 
 38.6 
 
 166.7 
 
 mined 
 
 8 
 
 IOO 
 
 32.4 
 
 41.1 
 
 166.0 
 
 
 9 
 
 IOO 
 
 35-4 
 
 44.1 
 
 141.2 
 
 
 10 
 
 IOO 
 
 39-o 
 
 47.6 
 
 143-5 
 
 
 ii 
 
 IOO 
 
 41.0 
 
 49.8 
 
 133.2 
 
 
 12 
 
 IOO 
 
 43-3 
 
 61.9 
 
 132.7 
 
 
 Under the system of representation adopted, by 
 which the proportions of the two other elements are 
 expressed as ratios to 100 parts by weight of H 2 SO 4 , 
 the figures of the preceding table may be graphically 
 expressed as in Fig. I, page 133. 
 
 All the points representing the mixtures employed 
 lie upon the same straight line ab, since they are 
 formed with the same nitric acid marking 34.6 by the 
 Baum areometer. This line is inclined at 45, since 
 the acid chosen contains as much water as mono- 
 hydrated acid; it starts from the point a correspond-, 
 ing to eight per cent., there being exactly that much 
 water in one hundred parts of the sulphuric acid to 
 which the one hundred parts of the other components 
 are expressed as ratios. 
 
 Outside of this line there exist an infinite number of 
 other points corresponding to acid mixtures the action 
 
THE NITRATION OF COTTON 
 
 133 
 
 of which upon cotton was not studied by M. Vieille. 
 By employing nitric acid of increasing strengths succes- 
 sively, along with sulphuric acid of 65.5 Baum6, mix- 
 tures will be obtained represented by points upon 
 the lines ac, ad, etc., which start from the common 
 point a and are more or less inclined according to the 
 
 Q 
 
 b 
 < 
 
 o 45- 
 
 so- 
 
 9 
 
 020 
 
 o 
 cc 
 t!5- 
 
 5 O- l'o 15 20 25 30 35 40 45 50 
 
 PROPORTION OF WATER TO 100 PARTS OF SULPHURJC ACID. 
 
 FIG. i 
 
 quantity of water in the acid employed. The practical 
 limit is the line af corresponding to 48 Baume", the 
 strongest obtainable in current manufacture, and which 
 contains about 10 per cent, water. 
 
 These are the regions left aside in previous studies, 
 the exploration of which is now undertaken from the 
 
134 SMOA'ELESS POWDER 
 
 standpoint of practical results to be obtained industri- 
 ally from the corresponding acid mixtures. 
 
 III. EXPERIMENTS IN NITRATION. (FIRST SERIES) 
 
 In this first series of experiments the base material 
 was the chemically pure absorbent cotton employed in 
 pharmacy. The acid mixtures were prepared direct 
 from the three principal elements, sulphuric acid, nitric 
 acid and water. For this purpose a carboy of sul- 
 phuric acid of 65.7 Baum was drawn from current 
 supplies. On the other hand a certain quantity 
 of nitric acid of the highest possible strength had 
 been collected at the nitric-acid factory, and from this 
 nitrous vapors were completely expelled by passing 
 carbonic-acid gas through it for a number of hours at 
 a temperature of about 70 C. After expulsion of 
 fumes this acid marked 47.7 Baume. 
 
 The selected sulphuric and nitric acids were kept in 
 carboys carefully sealed and provided with a com- 
 pressed-air emptying system, the air used being care- 
 fully freed from all traces of moisture. As all the ex- 
 periments could not be carried on at the same time, it 
 was necessary to provide that the elements entering 
 into them should be always identical. 
 
 The water present in the acids was that shown by 
 the special table corresponding to their areometric de- 
 gree, this quantity being checked by direct analysis, 
 and was 6.5 per cent, for the sulphuric and 10.5 per 
 cent, for the nitric acid. 
 
 The mixtures employed for the experiments were 
 determined by adding to the quantities of the two 
 
THE NITRATION OF COTTON 135 
 
 acids taken the quantity of water necessary to bring 
 the total weight up to 1.2 kilos. The percentage 
 compositions were deduced from these weights and 
 that of the water in the acid components. 
 
 Each mixture was divided into three lots of 400 
 grams each, of which two were employed for dupli- 
 cate experiments and one was held in reserve. 
 
 The 400 grams of mixed acid were introduced into a 
 large-mouthed flask provided with an air-tight stopper, 
 and an equilibrium of temperature established; then 
 the four grams of absorbent cotton were added and 
 the flask violently shaken. Under these conditions 
 nitration took place almost instantaneously, and it 
 may be said that each fibre of cotton was instantly 
 surrounded with IOO times its weight of liquid acid. 
 
 The flask was then corked and placed under a 
 stream of water maintained at a constant temperature 
 of from 18 to 19 C., which was that of dipping, the 
 cotton remaining the whole time in the acid bath. 
 
 Two dippings were made for each mixture studied, 
 one of them corresponding to reaction double in length 
 of time to that of the other. The admitted durations 
 of reaction were six, twelve, and twenty-four hours, 
 according to the kind of nitro-cellulose that was sup- 
 posed should be obtained. 
 
 It was not considered necessary to study reactions 
 of longer duration, difficult to realize in practice. In 
 certain cases it was deemed sufficient to control results 
 by reconducting the experiment in a bath of the same 
 composition, and allowing the reaction to continue 
 for forty-eight and even up to sixty hours. 
 
136 SMOKELESS POWDER 
 
 Twenty-five mixtures, corresponding pretty nearly 
 to the whole region embraced by those obtainable in 
 practice, were thus experimented with ; that is to say, 
 by maintaining as proportions of nitric acid and water 
 expressed in ratios to one hundred parts of sulphuric 
 acid by weight, from 10 to 60 per cent, of the former 
 and from 10 to 45 per cent, of the latter. 
 
 In fact, for nitric acid below a certain limit, reac- 
 tions became too slow. On the other hand, from a 
 standpoint of economy, its proportion cannot be too 
 much increased on account of its relatively high price. 
 As to water, its lower limit is fixed by the strength 
 of the strongest acids obtainable in current manufac- 
 ture. 
 
 The twe'nty-five mixtures studied represent suffi- 
 ciently well, then, the region that may be explored in 
 practice ; they are spaced apart regularly, so that by 
 the examination of the products obtained, proper ac- 
 count may be taken of the phenomenon of nitration 
 throughout the region explored. 
 
 The percentage of nitrogen in each of the products 
 thus obtained was determined by Schloessing's method, 
 described at the close of M. Vieille's paper above re- 
 ferred to. Solubility in ether-alcohol and viscosity in 
 that medium were determined under the methods em- 
 ployed at the Angouleme powder-works and described 
 at the end of the present paper. 
 
 The following table presents a resume of all results 
 obtained from the analysis of specimens produced in 
 this first series of experiments : 
 
THE NITRATION OF COTTON 
 
 137 
 
 
 
 >. 
 
 o 
 .2 
 
 K" 
 
 a 
 
 n 
 
 
 </> 
 
 c 
 
 3 ^VS. 
 
 MIOOMOO vni^-vn 
 
 
 
 
 9 
 
 -f 
 
 "0= a 
 c/5;3'~ 
 
 N M T^-O ^ M N ir> 
 
 M O 
 
 "o 
 
 
 c 
 
 hi- G 
 
 OONCOW 000000 
 
 .eaction 
 
 
 SSo-2 
 g * 
 
 M >-i ^ C^ ^ O O^ f^ 
 
 MMQoor^ MOO 
 
 C4C4MHIM C) M C4 
 
 ation of B 
 
 
 >, 
 
 I- 
 
 > 
 
 oo xr>CT>O'-ir->. CJNO co O to 
 
 M TfvO 00T)n OC^vO 00 IH rt 
 
 M N e^ M 
 
 3 
 
 Q 
 
 '/) 
 
 3 **". 
 
 in M ^}-CQ M <O tr>o O O C^ O^O oo M vr>O \O tr> O ^ w O 
 
 rt 
 
 <M 
 
 
 
 33 
 
 <S 
 
 "5= a 
 <3"" 
 
 ONMcnt-avo or^invo-i-omoo-j-voc<-)i-<M^fTi-r^o 
 
 Tj- MOOO O^OOC* Ooo O O> O O O O^ 
 
 O 
 W 
 C 
 
 
 c 
 
 co O O M -too <N co *t <N O N ^O aoocovo O OOoooooo 
 
 a! 
 2 
 
 
 g"<"io 
 
 rt Oo 
 
 COMMCTCOIHCT> vOc r >Ol^*1-Ol'^u->M'*l-t~.'3-O"">OOr^ 
 OMMOr^r^c> OO^t^>cot~>.r^inc>or^Oooi-iO>OooO 
 
 O 
 tfl 
 
 
 g ^ 
 
 
 "3 
 D 
 tf 
 
 
 >. 
 
 Sen" 
 > 
 
 oo r^ M co M "^-o O vo co m 
 MOoo^l-O winoo M N m 
 
 M M W M M 
 
 
 
 
 
 co t^ OinOQ^OMCOinrtMCOrf M co 
 
 
 3 
 
 C 
 
 S 
 o 
 
 'o'-^ c 
 OT2- 
 
 xn co McovO-OOcocor^ r^O coco co m 
 
 in f**. w O^ ^ ON O^ ON r^ QN ON ON ON ON O^ 
 
 
 
 ?*0 
 
 O W vO Tj- Tj-CO * -^-O O Tj- ^ O 'd- O 'I- 
 
 
 
 o v^, <j 
 
 ~ BiM G 
 
 ^t ON cowooOOONi^MOcoOco m in 
 O ON OONr^cor^OmONOr^-Oco ON ! 
 
 
 
 ^w rt '" 
 
 
 
 
 
 v. 
 
 O coco d co N ^ N ONOO miHOT}-Mrfr^i-iinMOCNi-iOr^ 
 
 ixture 
 
 
 rt 
 
 ON -^-00 ^ Tf Tj- CO TJ- CO ONOO T rf CO CO Tt ON ON TT t ONCO ON ONCO 
 NNMMt-i(NvNCOa<MCOCOCOCO'^-CONCOC<COl-iMMC<M 
 
 X 
 
 
 Is 
 
 M MO rj- 1^ T^ M vr> co O ONM coONr^comex Ooo como ^ON 
 
 | 
 
 
 JJ o 
 
 g^ 
 
 ONO>ONONONONOONON ONCO ON O> ON ONCO CO CO ONCO ON ON O> ON Tf 
 
 in'Tj-cOM MMMco>nTrjNco--tin'^-'^-cocoe^WMi-it-( 
 
 o. 
 
 s 
 
 
 Jl'C'O 
 
 O. * w .., g- *'.. 
 
 
 
 
 V5^<J 
 
 M 
 
 
 ! 
 
 1 
 ? 
 
 s 8 
 
 ^^-3 
 
 
 
 S 
 
 ^^^tr;!=!^XX^^S^^^XX^SS 
 
 - >> r H aSRHttggR : ?HB|gn 
 
138 SMOKELESS POWDER 
 
 Inspection of this table shows that mixtures 'in 
 which the percentage of nitric acid is low, and no- 
 tably mixtures V and VI, produce after twelve hours 
 reaction an incompletely nitrated product. Even 
 after twenty-four hours the nitration is not com- 
 plete; a fact verified by certain complementary ex- 
 periments for these latter. They will be discarded, 
 then, as out of the category of practical mixtures. 
 
 The graphic representation of the other mixtures, 
 by the method above indicated, is reproduced below 
 (Fig. 2). By the side of each point representing one of 
 them, the corresponding number is indicated in Ro- 
 man numerals; and the principal qualities of the prod- 
 uct taken from the preceding table are expressed in 
 ordinary figures, the figures chosen for each reaction 
 being those corresponding to the longest duration of 
 reaction. The first number is the percentage of ni- 
 trogen; the second, the solubility; the third, when 
 given, the viscosity. 
 
 If, as indicated at the beginning of this article, the 
 products be divided into gun-cottons, collodions and 
 friable cottons, overlooking certain small quantities 
 of accompanying products that may be mingled with 
 them, it will be seen that mixtures giving rise to 
 products of the same kind may be grouped in quite 
 distinct zones. The lines which limit these zones are 
 nearly parallel to one another. They depart a little 
 from the straight line and are slightly inclined to the 
 co-ordinate axes; percentages of nitrogen and solu- 
 bilities alone have served for tracing them. Vis- 
 cosities, which refer to collodions only, are pre- 
 
THE NITRATION OF COTTON 
 
 139 
 
 sented only as indicators, and are far from being the 
 same for all products in a given zone. 
 
 GO- 
 
 56- 
 
 50- 
 
 45 
 
 O lU 
 
 cc. I 
 035- 
 I O 
 
 5 10 15 20 25 30 35 40 45 
 
 PROPORTION OF WATER TO 100 PARTS OF SULPHURIC ACID (MONOHYDRATE) 
 FIG. 2 
 
 Without being able to discover, in results obtained 
 for these viscosities, the expression of any well-de- 
 fined law, it may be remarked, however, that they are 
 
140 SMOKELESS POWDER 
 
 in general, greater, according as the corresponding 
 percentages of nitrogen themselves are greater. Thus 
 it is to be remarked that the mean viscosity for the 
 zone of lower collodions is 65 s , while for the higher 
 collodions the mean attains 
 
 IV. EXPERIMENTS IN NITRATION. (SECOND SERIES) 
 
 The preceding results were obtained, as indicated, 
 with pure wadded cotton. In general this is not the 
 base material employed, but spinning waste, bleached 
 and freed from grease, which costs less. The fibres 
 of the cotton are more or less twisted and entangled, 
 and, notwithstanding the effects of carding, there 
 always remain at the time of dipping small agglom- 
 erations, into which the acid penetrates with diffi- 
 culty. In order to estimate the influence of the base 
 material employed upon results obtained, a second 
 series of trials were undertaken under conditions 
 similar to those of the first series, except that the 
 refuse cotton from current manufacture was substi- 
 tuted for the absorbent wadding. During this series 
 of experiments the temperature was maintained dur- 
 ing the whole time of the dipping at about from 7 to 
 8 C. The acid mixtures were prepared as indicated 
 above. The sulphuric acid employed marked 65.8 
 Baume, and contained only 5 per cent, of water. 
 The nitric acid, freed from nitrous vapors, marked 
 47. i Baume and contained 1 5 per cent, of water. The 
 per cent, composition of each of the twenty-five mix- 
 
THE NITRATION OF COTTON I4 l 
 
 tures employed was calculated on the basis of these 
 figures, and verified for a certain number of cases by 
 direct analysis. As above stated, 4 grams of cotton 
 were immersed in 400 grams of acid. The absorp- 
 tion of the cotton by the liquid proceeded as rapidly 
 as with the wadding. 
 
 The table on page 142 presents the results ob- 
 tained in the second series of experiments. 
 
 More clearly than in the preceding series, it is to 
 be remarked that for a certain number of mixtures, 
 especially for those in which the relative proportion 
 of nitric acid is low, prolongation of the reaction in- 
 creases the percentage of nitration. We shall dis- 
 card mixtures V and VI as heretofore, since the re- 
 action, to be complete, would have to be prolonged 
 too far. 
 
 Figure 3 (page 143) represents graphically the 
 results of the second series of experiments. 
 
 As in the preceding series, different mixtures, pro- 
 ducing products of the same kind are grouped in 
 zones, and these zones are sensibly of the same form 
 as those of Fig. 2; still again, but this time perhaps 
 a little less clearly, however, the viscosity of the col- 
 lodions appears to increase with the percentage of 
 nitration. 
 
 V. EXPERIMENTS IN NITRATION. (THIRD SERIES) 
 
 A third series of dippings was finally undertaken, 
 the conditions of which were practically identical 
 with those obtaining in practice. Bleached English 
 
142 
 
 SMOKELESS POWDER 
 
 
 
 >> 
 
 V 
 
 - 7 -', 
 
 
 't 
 
 3 
 
 
 
 SB 
 
 a** 
 
 c c 
 WJ5- 
 
 N r^ m r^ O O 
 
 M HI CO ^ t^ M 
 
 M in M 
 
 'o 
 
 _0 
 
 
 a 
 
 aj C 
 
 S d 
 2 & 
 
 O 00 O O> 00 -O 
 
 M co M o* r^ oo 
 
 M TT N O ON 
 
 W N M M W M 
 
 V 
 
 rt 
 
 "o 
 
 
 X 
 
 1" 
 
 0> aw O oo M 
 O r^ r^co o co 
 
 M 
 
 VH 
 
 O) 
 
 P >^^P 
 
 Wt^OWcomO ^incnMN-^-OMTj-MONONininin 
 
 Q 
 ed 
 u 
 
 
 
 E 
 
 "o c 
 
 t5g- 
 
 r^r^<NWOON. OvCM'i-MinNr^corfMcneovOoovd'^- 
 O WinC^ ooOOCT'^W C^ONCOO O ON 
 
 T3 
 
 
 I -2 
 
 P<OOOOT}- 1 ^-P< rf ^-OO NO N NO>O NvOOO O ^ ^ O O 
 
 1 
 
 rt 
 
 
 o 2 -o 
 
 rt O<j 
 
 Tt-ino^i-M con O t~.oco TJ-OO MOO o mc>oo O r^^-Qs^o 
 OO O O N N O Ooooi^m^f'l-r^O^inor'.iHoo OO O 
 
 8 
 
 
 2 Z 
 
 
 Results 
 
 
 >, 
 
 [- 
 
 > 
 
 m M o oo 1-1 r^ o 
 
 O t>- ON CO T^- 
 
 
 B 
 
 a ^v. 
 
 coO ^t mMinMMMinor^CT'O^coooci N 
 
 
 6 Hou 
 
 2|j 
 
 t~O O COO O O O O N O l^-O N O ^ m c 
 OIH O^ coOOOfOt-i O^ON Oco O O 
 
 
 
 C "S CT 
 
 &SOH 
 OOJfcU 
 
 OON CO OOONNTf'*OOO^'OOOO 1 ^- O 
 
 ^N co r->ONr^ocoMcoOMoooc7>Tt- i^ 
 
 
 
 2^: 
 
 
 * 
 
 
 u 
 
 t^t^coo Ooo mN moo N -^-r^N too N w t^TtmrtNooo 
 
 ixture 
 
 
 (4 
 
 gg 
 
 O Tfco CON N N cotoO^a^^ coco Tf m o O <* -^-co oo oo r^ r^ 
 
 CONMMMNNCONNCOCOCOCOTtfOCOCONCOiH(NMMl-i 
 
 f<5 
 
 
 .g-g 
 
 NONMNNOOOCOOOO^O-^-N *^-O CO O>O N tO twO 
 
 B 
 
 
 
 inmr^co CT^cococoooo rj-o r^r^o tmtoo r^r^oooo co 
 
 
 
 i^< 
 
 
 O 
 
 a 
 8 
 
 
 u 
 
 J.'u T3 
 
 3 a'r; 
 
 8 
 
 O 
 
 
 ffl 
 
 
 
 Number 
 
 <U V 
 
 s*.i 
 
 ^, O -> 
 S 
 
 -H HM * **-^, ^ * 4 1 H H-( *s> K> HH t-H HH ^ ^ HH -H HH k>< K> t * HH 4 K. K^ 
 
 HH^^^^^I^^^k^|^^H>l>K.^)^^^^^_ ( HH^ > ^ > 
 
 -" >>-- **^*>>*x^* 
 ^x x xx 
 
THE NITRATION OF COTTON 
 
 H3 
 
 spun-cotton waste from the powder-factory supplies, 
 was employed as the base material. This was dipped 
 in different acid mixtures formed, as indicated above, 
 
 XI 
 
 o 
 
 5 10 15 20 25 30 35 40 45 
 
 PROPORTION OF WATER TO 100 PARTS OF SULPHURIC ACID (MONOHYDRATE) 
 FIG. 3 
 
 from distilled water and the sulphuric and nitric acids 
 of which part had been employed in the preceding 
 experiments. The proportions for dippings were al- 
 
144 SMOKELESS POWDER 
 
 ways the same; 4 grams of cotton to 400 grams of 
 mixed acids; but the duration of immersion of the 
 cotton in the acid bath was reduced to eight minutes. 
 After such immersion at a constant temperature (12 
 to 13 C. approx.) the cotton was removed, drained, 
 lightly pressed, and placed in an earthenware crock 
 in which the reaction completed itself, the pot being 
 placed in a stream of water at the current tempera- 
 ture. 
 
 The results from this third series of experiments 
 are given in the table on page 145. 
 
 These results may be expressed graphically in the 
 manner shown in Fig. 4, page 146, neglecting mix- 
 tures V, VI and XXV, which only produced partially 
 nitrated products. 
 
 Still again, the products obtained allow the mix- 
 tures to be grouped into zones analogous to those 
 resulting from the two other series of experiments. 
 In each of these zones viscosities are variable and ap- 
 pear not to follow a well-defined law,, although high 
 viscosities accompany the highest percentages of 
 nitrogen. 
 
 VI. VARIOUS EXPERIMENTS 
 
 The experiments of which a resume has just been 
 given exhibit the influence of acid mixtures upon the 
 products obtained. Here, indeed, lies the principal 
 element entering into the whole manufacture of nitro- 
 cellulose. But other, secondary causes may equally 
 influence final results, and it has appeared useful to 
 
THE NITRATION OF COTTON 
 
 '45 
 
 
 
 r. 
 
 i 
 
 \r> <*t M M a 
 Tfr t> u-> t^ r^ 
 
 
 e 
 
 a 
 
 
 
 EC 
 
 i>>* 
 
 "o'- 2 c 
 w2- 
 
 u-> N O en rj- O 00 OO CO 00 t^. rf W 
 
 w -^- M u-> o o d M M M o m r^ o 
 M N O OO M O N CXD O t^. 
 
 c 
 o 
 
 
 W 
 
 U O 
 
 2 * 
 
 NQOO vOOCOCO<N 00 O M OO 
 
 ir>*rr<. Nt^-i^vOMco M ^ d wr>. 
 Oto-^ r^r^.vr>'^-^i-r>. to \o oo O 00 
 
 C4MM MMMMMM M M M MM 
 
 Pi 
 
 c 
 c 
 
 
 >, 
 
 1 - 
 > 
 
 vo 100001-10 u->>-ic^o^r^ 
 
 O ^OO^ 1 ^-'^- -rtirM>r)O 
 
 N M MM 
 
 rt 
 u 
 
 OB 
 
 A >>>.. 
 
 OO^Tt-MMTj-o^ MoOTfoocoOmxnoooocoootoNO^r^O 
 
 rt 
 
 g 
 
 i O 
 
 E 
 
 "o- c 
 w3- 
 
 r^NN>r)OC>O r^vn-i-vOMTj-Mor-^^MMcnOMinoQ 
 O^iO WMO O^O^C^ Q^O M O^ O^ n O^OO C7^ N vO M 
 
 2 
 
 
 c 
 
 j> s 
 
 NOOONNOOO NNvOONNOOMOOO TOO O O T O 
 
 in 
 
 
 N .J 
 B "o o 
 
 r^ \r> o^ cooo N M - >O O ~t W TO c")co o m O^ c^ico N M N tr> 
 O\ O O O w TOO O^co <o r^- m T N r^co xn o^O O oo o O r^ 
 
 U 
 
 .V 
 
 
 2 * 
 
 
 General 
 
 
 I- 
 
 > 
 
 to N r^o co co O oo oo r-N 
 O uiMcovn coTr^*ncv 
 
 
 
 
 
 O OO N oo M r-H T to vnoo O O TOO O^O N TT*r>O M w 
 
 
 
 
 33 
 
 .'o c 
 
 BS* 
 
 OMC^iriOO^C^ iDvOr^Nr^vOMTOvOWvOmTOtoo 
 & \r> M M oo O^ OOO O> T M O^ O> N O>oo O^ W O W 
 
 
 
 11 JJrT 
 
 C4OCOCOOOT NvOTvOTTOMOMOTOOToOTO 
 
 
 
 tuot/iC 1 o 
 ouZd 
 
 Otootocjcncn <OTT'- | c r 5TMTiovor^.cor^Of r >Ooo 
 
 O^ O O O 01 Too Oco O i^ in T N t^oo 10 OO O oo OO O 
 
 
 
 x rt 
 ^Id 
 
 
 
 
 w 
 
 oo to co M o coco M O cooo o coo TmcoTr*-mTmM OO 
 
 ixture 
 
 
 1 
 
 oo TOO co'O 1-1 O cow or^Tcoco IOTCOCO cococococooo r^ 
 
 NC'IMMMNNCOaNCOCOCOCOTCOacOCJCOMWMMM 
 
 5 
 
 
 o 
 
 O OW COM TOOoomOOmcoO M M TmTt^oo cow O 
 
 ition o 
 
 
 11 
 
 vomr^oo oooo r^co r^Tt^r^i^r>Tniooo r^r^ccco co 
 \nTcOM MMCMcoinTC4cOTinTTcococc>JMMM 
 
 p 
 
 
 -iia 
 
 . , 
 
 
 
 U 
 
 
 ^t< 
 
 M 
 
 
 Number 
 
 ot Order 
 of 
 Mixtures 
 
 HH ^^^^^XX^^^^^^XX^^S 
 
 B- >> r XZK$>>X%XZX% 
 
146 
 
 SMOKELESS POWDER 
 
 study them also. Among these causes, duration of 
 reaction has been briefly touched upon already, as 
 well as the nature of the base material. To these must 
 
 xvi XI 
 
 /172.8 
 95.3 
 45 
 
 XV 
 
 5 10 15 20 25 30 35 40 45 
 
 PROPORTION OF WATER TO 100 PARTS OF SULPHURIC ACID. 
 FlG. 4 
 
 be added temperature during dipping and reaction, as 
 well as the later effect of the manipulations that ni- 
 trated cotton may undergo after the operation of 
 
THE NITRATION OF COTTON 147 
 
 dipping. These different effects will be examined 
 successively. 
 
 Base materials. Cotton wadding nitrates more 
 readily than spun-cotton waste, which nitrates with 
 the greater difficulty the more it is tangled, the 
 coarser its threads, and the more knots it con- 
 tains. On the other hand, it equally appears that the 
 previous preparation of the waste destined for dip- 
 ping exercises a very sensible influence upon the vis- 
 cosity of the collodions obtained. Thus it is that two 
 cottons of absolutely different origin, submitted un- 
 der identical conditions to .the action of a common 
 acid mixture, produce collodions possessing the same 
 mean percentage of nitrogen, but with viscosities 
 varying from what they ordinarily are to double this. 
 
 Duration of reaction. The appended table, in 
 which two or three durations of reaction correspond 
 to each mixture, shows that for a same final product, 
 the reaction should be the more prolonged the lower 
 the proportion of nitric acid in the mixture. 
 
 Some experiments were made to follow more 
 closely the progress of the reaction; a resume of them 
 is presented in the following table. These experi- 
 ments refer to three acid mixtures from the first 
 series, the numbers of which are recalled in reference 
 to results obtained; the conditions of dipping and re- 
 action are those of this series; the temperature of 
 the reaction was about 20 C. approximately: 
 
148 
 
 SMOKELESS POWDER 
 
 
 Mixture XIII 
 
 Mixture XVI 
 
 Mixture XXI 
 
 of 
 Reaction 
 
 Nitrogen 
 c.c. MO 2 
 
 Solubility 
 in* 
 
 Nitrogen 
 c.c. NO a 
 
 Solubility 
 
 \n% 
 
 Nitrogen 
 c.c. N0 3 
 
 Solubility 
 in 
 
 I hour 
 
 165 8 
 
 91.7 
 
 186.8 
 
 94-9 
 
 206.4 
 
 10.9 
 
 2 hours 
 
 1 66 8 
 
 95-5 
 
 189 
 
 95-0 
 
 209.4 
 
 8-3 
 
 4 
 
 167.8 
 
 93-0 
 
 191.8 
 
 96.2 
 
 209.2 
 
 6.8 
 
 6 " 
 
 167.8 
 
 94.8 
 
 198 
 
 94-1 
 
 2IO.2 
 
 6.7 
 
 8 " 
 
 166.8 
 
 95-4 
 
 lgl.8 
 
 96.7 
 
 210.2 
 
 5-6 
 
 12 " 
 
 
 
 
 
 210.8 
 
 7-4 
 
 24 
 
 166.8 
 
 98.1 
 
 194 
 
 96.6 
 
 2IO.6 
 
 10.6 
 
 These experiments comprise the three principal 
 types of nitro-cellulose studied; superior and inferior 
 collodions and gun-cotton. They show that for the 
 two former a maximum of nitration is hardly ob- 
 tained before the end of two hours of reaction; for 
 gun-cotton, on the other hand, from eight to ten 
 hours are required. If the reaction be prolonged 
 beyond these limits, the solubility has a tendency to 
 increase. In these experiments viscosity could not 
 be measured; but the preceding tables seem to show 
 that no relation exists between it and duration of re- 
 action. 
 
 Temperatures of dipping and reaction. A certain 
 number of dippings were made under conditions iden- 
 tical with those of the second series of experiments, 
 at the three very different temperatures of i, 12 
 and 25 C. These temperatures were maintained 
 throughout the whole duration of the reaction. 
 
 Results obtained are grouped in the following 
 table: 
 
THE NITRATION OF COTTON 
 
 I 49 
 
 
 
 Temperature during Dipping and Reaction 
 
 \l 
 
 "a 
 
 ft IT) 
 
 
 iC. 
 
 12 C. 
 
 25 c. 
 
 .H 
 -}** 
 
 1* 
 hrs. 
 
 Nitro- 
 
 Tor 
 
 in c.c. 
 
 Solu- 
 bility 
 in* 
 
 Vis- 
 cos- 
 ity 
 s. 
 
 Nitro- 
 gen as 
 NO, 
 in c.c. 
 
 Solu- 
 bility. 
 in* 
 
 Vis- 
 cos- 
 ity 
 s. 
 
 Nitro- 
 gen as 
 NO, 
 in c.c. 
 
 Solu- 
 bility 
 in* 
 
 Vis- 
 cos- 
 ity 
 s. 
 
 
 6 
 
 190 o 
 
 95-9 
 
 180 
 
 191.8 
 
 97.8 
 
 104 
 
 193.2 
 
 97-7 
 
 56 
 
 in 
 
 12 
 
 2O8.2 
 
 1-5 
 
 
 207.4 
 
 1-7 
 
 
 207.8 
 
 I.O 
 
 
 VI 
 
 12 
 
 2OI.6 
 
 4-4 
 
 
 205 4 
 
 4.4 
 
 
 2O7-6 
 
 3.6 
 
 
 /n 
 
 6 
 
 177.6 
 
 84.9 
 
 22 
 
 189.0 
 
 92.9 
 
 60 
 
 193.2 
 
 97-7 
 
 36 
 
 X 
 
 6 
 
 178 8 
 
 95-1 
 
 145 
 
 185.4 
 
 96.8 
 
 72 
 
 184.4 
 
 96.2 
 
 6-1 
 
 til 
 
 6 
 
 167.2 
 
 78.9 
 
 38 
 
 172.0 
 
 92-9 
 
 36 
 
 172.0 
 
 96.1 
 
 27 
 
 Other experiments were made to the same end 
 with wadded cotton under the conditions of the first 
 series. The results were as follows: 
 
 
 g 
 
 Temperature during Dipping and Reaction 
 
 ."S 
 
 II 
 
 2C. 
 
 15 C. 
 
 26 C. 
 
 ^5 u 
 
 rt u 
 
 
 
 
 
 *a 
 
 hrs. 
 
 Nitro- 
 
 N0 a 
 in c.c. 
 
 Solu- 
 bility 
 in* 
 
 Vis- 
 cos- 
 ity 
 s. 
 
 Nitro- 
 gen 
 NO, 
 in c.c. 
 
 Solu- 
 bility 
 in* 
 
 Vis- 
 cos- 
 ity 
 s. 
 
 Nitro- 
 gen. 
 NO, 
 in c.c. 
 
 Solu- 
 bility 
 in* 
 
 Vis- 
 cos- 
 ity 
 s. 
 
 XII 
 
 6 
 
 183.2 
 
 96.5 
 
 
 186.8 
 
 96.7 
 
 
 189.2 
 
 97-3 
 
 
 
 
 12 
 
 183.2 
 
 97-4 
 
 
 186.8 
 
 95-1 
 
 
 188.2 
 
 94-4 
 
 
 XIII 
 
 6 
 
 167.4 
 
 93.8 
 
 
 169.0 
 
 93.2 
 
 
 173-4 
 
 93-0 
 
 
 1 1 
 
 12 
 
 165.6 
 
 90.7 
 
 
 166.0 
 
 94-5 
 
 
 168.6 
 
 94.1 
 
 
 XV 
 
 6 
 
 188.8 
 
 96.5 
 
 43 
 
 192.8 
 
 98.1 
 
 48 
 
 194.0 
 
 Q3.3 
 
 22 
 
 
 
 12 
 
 188.8 
 
 97-4 
 
 37 
 
 191.8 
 
 97-0 
 
 26 
 
 193.6 
 
 98.1 
 
 20 
 
 XXI 
 
 12 
 
 208.8 
 
 8.9 
 
 
 209.2 
 
 9.2 
 
 
 208.6 
 
 7.1 
 
 
 1 
 
 2 4 
 
 209.8 
 
 6-7 
 
 
 209.2 
 
 7.0 
 
 
 208.4 
 
 8.4 
 
 
 These experiments seemed to show that, so far as 
 collodions were concerned, increase of temperature 
 during dipping and reaction increases the percentage 
 of nitrogen and the solubility, but sensibly diminishes 
 
SMOKELESS POWDER 
 
 the viscosity. In other' words, a low temperature re- 
 tards the reaction. In what relates to gun-cottons, 
 the influence manifests itself less distinctly; it is ad- 
 mitted, however, that at high temperatures the solu- 
 bility has a tendency to increase. 
 
 Subsequent manipulations. The different cellu- 
 loses are submitted before use to a number of 
 manipulations; first to washings, which are necessary 
 to remove the last traces of acidity, and for certain 
 ones to pulping, which serves to facilitate subsequent 
 operations. 
 
 Washing is effected by a more or less prolonged 
 boiling, either in pure water or in an alkaline solu- 
 tion; the pulping by beating-engines similar to the ap- 
 paratus employed in the manufacture of paper. 
 
 These operations, which are absolutely mechanical, 
 have no influence upon the chemical composition of 
 the final product, and therefore upon the percentage 
 of nitrogen; but it is different for solubility and vis- 
 cosity, which are purely physical properties. 
 
 Experiments were made with the view of ascertain- 
 ing how various kinds of nitro-celluloses acted during 
 these operations. Results are grouped in the tables 
 on page 151. 
 
 These various experiments were made with quite 
 large quantities of gun-cotton, consequently more or 
 less homogeneous, so that taking of samples should 
 suffice to explain certain anomalies to be noticed in 
 the progress of phenomena observed. From the 
 figures in the preceding tables it may be concluded, 
 on the one hand, that the two operations of washing 
 
7W.fi: NITRATION OF COTTON 
 
 
 
 
 
 fe 
 
 M 
 
 
 
 
 
 
 
 
 
 
 
 
 33 
 
 I 
 
 
 
 
 
 vg 
 
 S> CTU " 
 
 ? 
 
 
 
 
 S 
 
 oO u 
 b^ c 
 
 1 
 
 
 
 
 
 
 
 
 
 
 
 
 h 
 
 
 
 
 
 
 S-. 
 
 C9 
 
 
 
 
 
 x> 
 
 
 
 
 
 a 
 
 3 c 
 
 (N 
 
 
 
 
 3 
 
 
 S"" 
 
 
 
 
 
 33 
 
 C 
 
 
 
 
 
 S 
 
 J J 
 
 . 
 
 
 
 
 N 
 
 0" 
 
 i 
 
 
 S5 
 
 *o 
 
 
 .ti^.S 
 
 s 
 
 ^ 
 
 C 
 
 be 
 
 
 * 
 
 
 2 
 
 COTT 
 
 1 
 
 
 S>R 
 
 CO 
 
 o 
 u 
 
 2; 
 
 !> 
 
 
 
 "o 
 
 a 
 o 
 
 12 
 
 3 
 
 ffi 
 
 "o 
 
 C/) 
 
 f 
 
 M 
 
 
 
 > 
 O 
 
 c 1 
 
 ^ 
 O 
 
 O 
 2 
 
 
 
 H 
 
 0_ 
 
 JssJ 
 
 o 
 
 1 
 
 fc 
 
 o 
 
 (3 
 
 z 
 
 Cu 
 
 S 
 
 rt 
 
 
 
 
 _1 
 
 c 
 
 *- 
 
 
 4_i 
 
 
 & 
 
 < 
 
 3 
 
 c/) 
 
 2^ 
 
 
 
 OH 
 b. 
 
 h 
 
 O 
 
 h 
 
 U 
 M 
 fc 
 h 
 
 3 
 1 
 
 0- 
 o 
 
 100 Houi 
 
 S 
 
 
 EFFECT o: 
 
 c 
 
 
 
 1 
 
 W 
 
 H 
 
 
 * 
 
 
 
 
 3 
 
 
 
 
 
 j-j 
 
 ^^ 
 
 a 
 
 
 5 c 
 
 M 
 
 
 
 
 52 
 
 3-- 
 
 M 
 
 
 
 
 3 
 O 
 
 I 
 
 
 
 
 
 X 
 
 G 
 
 ^ 
 
 
 
 
 
 
 o'd'j 
 
 
 
 
 
 
 
 .^Z c 
 
 
 
 
 
 
 
 S-w. 
 
 N 
 
 
 
 
 M 
 
 .s 
 
 >/> 
 
 
 
 
 3 
 O 
 
 1 
 
 
 
 
 
 33 
 
 G 
 
 
 
 
 
 8 
 
 &c,U- 
 
 o O o 
 fiS5a 
 
 8 
 
 
 
 
 
 55 
 
 
 
 
 
 1.2 ' 
 
 J 
 
 
 g 
 
 
 
 
 
 
 1 
 
 vO 
 en 
 
 
 
 pi 
 
 O 
 
 
 
 fe 
 
 r^ 
 
 i_t 
 
 "o 
 
 g 
 
 
 
 
 bfl 
 
 c 
 
 '5. 
 
 EC 
 
 co 
 
 I." ' 
 
 
 
 I 
 
 
 oO o 
 S5 a 
 
 
 
 o 
 
 
 55 "" 
 
 
 c 
 o 
 
 
 ^ 
 
 
 '5 
 
 
 ^w. 
 
 
 
 3 
 
 Q 
 es 
 
 If 
 3 
 
 1" 
 
 8 
 
 u 
 
 EC 
 
 c 
 
 CD 
 
 
 
 Si <> 
 
 
 a 
 
 
 |tfu 
 
 3*" 
 
 o 
 
 3 
 "O 
 
 
 ^ 
 
 <N 
 
 PH 
 
 
 fc, 
 
 
 "o 
 
 
 S>. 
 
 ^ 
 
 
 
 V) 
 
 is 
 
 co 
 
 "tfl 
 
 g 
 
 c8 
 
 
 S 
 
 33 
 
 M 
 
 c 
 
 O J 
 
 CO 
 
 
 
 oOo 
 
 g 
 
 
 
 55 
 
 N 
 
 
 
 
 
 vO 
 
 
 
 
 * 
 
 
 
 J3.S 
 
 M 
 
 
 g 
 
 
 
 
 
 EC 
 
 S o 
 
 00 
 
 
 
 fgs 
 
 s 
 
 U 
 
 I 4 
 
 CO 
 
 a 
 
 !S c 
 
 CM 
 
 'jx 
 
 H" 3 
 
 
 "3 
 
 On 
 
 c/5 
 
 
 
 G 
 
 CO 
 
 i 
 
 P| 
 
 i" 
 
 
 u bx 
 
 N r^ 
 
 O en 
 
 
 "a. ' 
 
 CO Ht 
 
 \r> \rt' 
 
 
 ^3 
 
 M 
 
 
 
 a. 
 
 
 
 
 Sc* 
 2 aw- 
 
 OOO 
 N 
 
 OOO 
 
 
 2| 
 
 
 
 "o 
 
 tx 
 
 VI 
 
 3 
 . 
 
 0_ 
 
 
 c 
 
 
 
 
 
 c 
 .2 
 
 
 
 3 
 O . 
 
 33 w 
 
 00 
 
 en i^ 
 
 
 3 
 
 Q 
 rt 
 
 C 
 
 3 
 . 
 
 ^t O 
 loco 
 
 en M 
 
 CO M 
 
 1 
 
 J2 <" 
 
 M en 
 
 
 "rt 
 
 O 
 
 
 
 V 
 
 c 
 
 5 
 
 JD 
 
 !2 
 
 3 
 . 
 ES:? 
 
 O 
 
 r^ XO 
 
 CM O 
 
 O>co 
 
 O 
 
 
 
 
 3 
 
 a 
 o 
 
 3 
 
 O 
 
 t en 
 
 u 
 
 O . 
 
 co co 
 
 
 
 
 33" 
 
 en t^* 
 
 tH M 
 
 o 
 
 Jf 
 
 
 
 g 
 
 (A 
 
 
 
 .1 
 
 3 
 
 W O 
 rtco 
 
 M M 
 
 M 
 
 > 
 
 00 
 
 
 
 
 3 
 
 ^^ 
 
 en oo 
 
 
 !* 
 
 35 
 
 en i- 
 
 
 52 
 33 w 
 
 en vn 
 
 r^co 
 
 rrln 
 
 M M 
 
 
 o 
 
 
 
 < 
 
 ' 
 
 imens 
 
 
 
 ! 
 
 3 O 
 "1 *^ 
 
 
 
 
 
 M ci 
 
 en rt 
 
152 SMOKELESS POWDER 
 
 in warm water and pulping have the effect of increas- 
 ing the solubility of gun-cottons; on the other hand, 
 that these two operations diminish to a marked de- 
 gree the viscosity of the collodions. 
 
 VII. RESUME AND CONCLUSIONS 
 
 As already stated, the results of the experiments 
 just described complete in a certain sense those ob- 
 tained by M. Vieille in 1883, since they relate to a 
 series of mixtures that had not been studied up to 
 this time. 
 
 Apart from the theoretical interest they may pos- 
 sess, they also possess a practical utility. 
 
 When it is desired to produce a certain definite 
 nitro-cellulose, the first thing to determine upon is 
 the composition of the dipping mixture to be em- 
 ployed. Heretofore, the proportions of the different 
 elements, whether new or spent acids, were calcu- 
 lated somewhat arbitrarily. The experiments, of 
 which a resume has just been given, permit a more 
 methodical procedure in such a case. 
 
 These are only laboratory experiments, it is true; 
 and in practice a thousand causes, more or less well 
 understood, arise to influence final results. Some 
 supplementary experiments have been made to as- 
 certain how these known causes tend; but it would 
 be rash to assert that it would be possible, on the 
 strength of the data afforded by these experiments, 
 to obtain with certainty a product of which all the 
 qualities were predetermined. 
 
THE NITRATION OF COT 7 ON I 53 
 
 However, the great similarity of the results ob- 
 tained in the three series of experiments, each of 
 which approaches more nearly than the preceding 
 to conditions obtaining in practice, justifies the be- 
 lief that, although no one of them may enable us to 
 calculate the proportions of a mixture capable of pro- 
 ducing a certain product, nevertheless, the general 
 trend of the phenomenon of nitration is that indi- 
 cated by the position and relative importance of the 
 different zones above referred to. 
 
 The knowledge of this progress in the phenomenon 
 of nitration permits us, then, after a preliminary trial, 
 which, besides, is not undertaken by chance, to 
 modify, if necessary, the proportions of the mixture 
 first taken, in such a way as to arrive at a desired end. 
 
 Different considerations serve as guides in the 
 preparation of the preliminary mixture. While still 
 keeping within the zone corresponding to the prod- 
 uct desired, the relative proportions of nitric acid 
 and water- may be varied between quite wide limits. 
 It is desirable, from the standpoint of economy, to 
 diminish the quantity of the most expensive element; 
 that is to say, the nitric acid; in certain cases, how- 
 ever, to develop certain qualities in the final product, 
 we may be led to increase it. Finally, if spent acids 
 of known composition are to be employed, one mix- 
 ture may be found more advantageous than another, 
 according to the relative quantities of acids to be 
 taken. 
 
 The acids used, nitric acid or spent acids, a'ways 
 contain nitrous vapors to a small extent. These 
 
154 SMOKELESS POWDER 
 
 nitrous vapors may, if present in sensible quantity, 
 falsify conditions upon which it was thought reliance 
 could be placed, and this fact must be borne in mind 
 in calculating the trial mixture. But in practice the 
 proportion does not exceed from i to 2 per cent., and 
 within these limits it may be admitted that the pres- 
 ence of nitrous vapors does not change results sen- 
 sibly. 
 
 It is necessary, besides, to bear in mind that on ac- 
 count of the reaction that takes place in the dipping- 
 vats, and which produces water, the proportions of 
 sulphuric acid, nitric acid and water in the vats are 
 no longer the same as those of the original mixture. 
 These alone are the conditions which affect the de- 
 gree of nitration and are determined beforehand ac- 
 cording to the considerations just indicated. The 
 others, which must also be known, since they serve 
 in preparing the mixtures properly so-called, are 
 easily deduced by means of corrections, which de- 
 pend upon the size of the vats, the relative weight of 
 cotton employed each time, etc., etc.; and they are 
 determined in practice, by analysis, before and after 
 dipping, of a certain number of mixtures most com- 
 monly employed. 
 
 When the formula for a mixture permitting de- 
 sired results to be obtained is thus found, the choice 
 of raw material to be dipped is also to be thought of, 
 as well as temperature of dipping, duration and tem- 
 perature of the reaction, etc., etc. 
 
 Here also, the experiments recalled above may 
 serve as a guide to results to be obtained. 
 
THE NITRATION OF COTTON 155 
 
 There always exist sensible differences between the 
 industrial manufacture of a product and the labora- 
 tory process which serves as a basis for it. A thou- 
 sand causes, varying from day to day, arise to modify 
 the qualities of the final product. At first one pro-' 
 ceeds only by guess-work; it is only in the end, as 
 the result of the processes followed with method and 
 perseverance, that one becomes able to define with 
 exactness the effect produced by each one of a num- 
 ber of stated modifications. 
 
 With this line of thought a number of experiments 
 were made at the Angouleme laboratory upon the 
 production of nitre-celluloses. Much still remains 
 to be done, but the results obtained already permit a 
 more methodical procedure than in the past. 
 
APPENDIX IV 
 
 THE DEVELOPMENT OF SMOKELESS POWDER* 
 By Lieutenant JOHN B. BERNADOU, U. S. Navy 
 
 THE systematic development of improved ballistic 
 properties from progressive explosives constitutes 
 one of the most important ordnance problems of the 
 present day. The idea is beginning to gain ground 
 among us that hereafter we must look to the powder, 
 as well as to the gun, in our efforts to increase the 
 rapidity of flight and the penetrative power of pro- 
 jectiles; that we must consider the source of energy 
 at our disposal conjointly with the apparatus whose 
 function it is to convert that energy into useful work. 
 
 So long as the art of powder-making remained at 
 a standstill as it practically did for several centuries 
 while the practices of alchemy, rather than the prin- 
 ciples of chemistry, may be said to have controlled 
 the manufacture of all explosives, the best that could 
 be done was to follow the progress of mechanics in 
 efforts to effect ordnance improvement; guns were 
 built and powders were found to fire from them. To- 
 
 * Abstract of lecture delivered before the U. S. Naval War 
 College, July 20, 1897. 
 
 156 
 
THE DEVELOPMENT OF SMOKELESS POWDER I 57 
 
 day, however, not only is the composition of powder 
 undergoing modification, but new explosive com- 
 pounds, the development of which is based upon 
 chemical discovery, are coming into general use; the 
 results of investigations into the chemical and physical 
 properties of explosives, systematized and coordi- 
 nated by the methods of mathematical analysis, have 
 so increased our knowledge of ballistics, that de- 
 signers of ordnance are forced to accept new condi- 
 tions as factors of prime importance in the attain- 
 ment of ballistic effect. 
 
 For purposes of comparison, the old forms of pow- 
 der, such as black gunpowder, may be regarded as 
 imperfect mechanical mixtures of particles of the ma- 
 terials of which the powder is composed; the new ex- 
 plosives, as very intimate mixtures of the atoms of 
 those elements from the union of which into molecules 
 the substance of the explosive is formed. When the 
 old powder is employed as a fine dust, it burns with 
 great speed and violence; when agglomerated into 
 grains* it burns in a slow, progressive manner. If the 
 grains of the old powders become disintegrated before 
 they are completely consumed, through effects of heat 
 and gas pressure developed in the bore of the gun, the 
 grains crumble away; pressures become violent and 
 regular progressive combustion ceases to obtain. 
 Similar, but not identical, conditions exist for the 
 new explosives. In the form of dust they burn with 
 
 * ^y " grain " is to be understood any regular form, flat strips, 
 rods, cubes, etc. 
 
158 SMOKELESS POWDER 
 
 exceeding rapidity and great violence; when all the 
 particles are decomposed simultaneously, they deto- 
 nate; by building them up into dense grains they may 
 be made, under favorable conditions, to burn pro- 
 gressively. 
 
 It was obvious from the start that many advan- 
 tages were to be obtained by the substitution of the 
 new explosives for the old as progressive powders. 
 The former burned up completely, leaving no resi- 
 due. Many of them made no smoke. Other condi- 
 tions being equal, these two qualities alone would 
 have been sufficient to justify their general adoption. 
 But other conditions were not equal. Up to a few 
 years ago the fact remained that no positive, certain 
 means of making nitro-explosives burn progressively 
 had been found. All known precautions could be ob- 
 served in the preparation; they could be built up into 
 dense grains with the greatest possible care; yet, 
 every now and then a charge of the powder would 
 detonate; that is, instead of burning progressively, in 
 accordance with the finished form of its grains, it 
 would burn as the dust from which the grains were 
 built up. A gun would be shattered, perhaps a life 
 or two lost, and then all confidence in the new ma- 
 terial would disappear, the chosen line of develop- 
 ment would be abandoned; no fundamental facts 
 would be left unchallenged to anchor new hopes 
 upon. 
 
 Until some way could be found, then, of firing a 
 nitro-compound from a gun with positive assurance 
 that detonation would not occur, there could be no 
 
THE DEVELOPMENT OF SMOKELESS POWDER 1 59 
 
 change from the old powders to the new. This as- 
 surance was, however, obtained by the discovery that 
 nitro-cellulose, colloided and formed into grains of 
 regular size, would in all cases, if ignited in a closed 
 space, burn away in a progressive manner, at. a rate 
 proportional to the form and dimensions of the grains 
 and to the conditions of their confinement. Two 
 proofs made the fact certain that colloids would not 
 detonate; first, that the grains of colloid powders 
 which were shot out of the gun without being com- 
 pletely consumed, preserved the original shapes, in re- 
 duced dimensions, of the grains of which the powder 
 charge was primarily composed; second, that not tens, 
 nor hundreds, but thousands of rounds of colloid pow- 
 ders, fired in guns or exploded in closed vessels, de- 
 veloped in every case pressures that could be shown 
 to correspond rationally, in accordance with the 
 theory of progressive combustion, to size and form 
 of grains and to dimensions of gun-chamber or ex- 
 plosion-bomb. To make these facts certain,, pres- 
 sures were carried up beyond the 33,ooo-lb. limit al- 
 lowed for cannon; and in explosion-bombs to well 
 beyond 100,000 Ibs. per square inch. 
 
 Now, as the only certain means yet found of avoid- 
 ing detonation and of assuring progressive combus- 
 tion is through the colloiding of nitro-cellulose, and 
 as nitro-cellulose alone can be colloided, it follows 
 that we are definitely limited in our choice of ma- 
 terial for progressive powders to a certain prepara- 
 tion of nitro-cellulose. It is true that a number of 
 substances, such as nitro-glycerin and nitrates of 
 
l6o SMOKELESS POWDER 
 
 metallic bases, may be distributed in minute particles 
 throughout the body of the colloid, and that im- 
 munity from simultaneous detonation may be se- 
 cured for the particles so distributed; but they all re- 
 main uncombined in the colloid the nitrates, as the 
 sand or minute shells in the body of a sponge; the 
 nitro-glycerin, as the water in the pores of the 
 sponge. It is desired to emphasize by this compari- 
 son the fact that all the new powders, without excep- 
 tion, must be built up from some form of colloid 
 nitro-cellulose, whether they contain other ingredi- 
 ents or not. Thus, in the case of those powders con- 
 taining nitro-glycerin we may reduce the percentage 
 of nitro-glycerin to zero; that is, we may eliminate 
 it. If we were to remove the colloid nitro-cellulose 
 from such a powder we would have remaining nitro- 
 glycerin, which would detonate in the gun upon the 
 attempt to fire. 
 
 Nitro-cellulose, which is usually prepared by dip- 
 ping cotton into nitric acid, possesses a property 
 which the cotton, before dipping into acid, does not, 
 i.e., of dissolving in a number of substances. One 
 of these is acetone, a volatile fluid, with a character- 
 istic pungent and aromatic odor, somewhat suggest- 
 ing common alcohol in appearance and properties. 
 Another solvent for a different kind of nitro-cellulose 
 is a mixture of ether and alcohol. If the clear liquids 
 which constitute the solutions in these substances be 
 evaporated, there will be obtained, not the nitrated 
 cotton, in its original fibrous form, but first, a 
 syrupy liquid, then a jelly, and finally, as dryness is 
 
THE DEVELOPMENT OF SMOKELESS POWDER l6l 
 
 approached, a solid translucent mass, varying in color 
 according to the variety of nitro-cellulose from which 
 it is prepared, from a straw-yellow to a chocolate- 
 brown, and generally suggesting, in its various forms, 
 tortoise-shell. To such a substance the appella- 
 tion colloid, from its glue-like consistency, has been 
 applied. 
 
 The general name for the material produced by 
 steeping cellulose into nitric acid is nitro-cellulose. 
 One of its common forms is gun-cotton. Chemists 
 are well aware that there are many different kinds of 
 nitro-cellulose, but just how many there are no one 
 has as yet even been able to predict, the exact com- 
 position of cellulose and of its nitro-derivatives re- 
 maining among the yet unsolved mysteries of nature. 
 Three forms were originally assigned to it, the mono-, 
 di- and tri-, just as there were the three forms of 
 mono-, di- and trinitro-glycerin. A later investi- 
 gator (Eder) succeeded in proving the existence of 
 six. The authority of to-day upon the subject, whose 
 views are now generally accepted (Vieille), has form- 
 ulated eight. Now, just as there are many varieties 
 of nitro-cellulose, so there are many varieties of col- 
 loids. The nitre-celluloses themselves all look alike; 
 in their common pulped form they suggest fine white 
 flour. They can be distinguished from one another 
 with ease by the readiness with which some of them 
 go into solution in certain solvents, while others re- 
 main undissolved in these solvents like so much sand. 
 The fact that they possess such different properties 
 is accounted for in practice by proven differences in 
 
1 62 SMOKELESS POWDER 
 
 chemical composition. It suffices here to state that 
 there are a number of different varieties of the sub- 
 stance which form a number of different colloids. 
 The question of composition will be considered later 
 in relation to the gases resulting from the combus- 
 tion of nitro-compounds. 
 
 We are familiar with the kind of nitre-cellulose 
 used for detonating purposes gun-cotton. We are 
 also familiar with a form of colloid in common use to- 
 day as a material. I refer here to celluloid, now very 
 generally employed for the manufacture of a great 
 number of useful articles. The nitro-cellulose from 
 which celluloid is prepared may be made by steep- 
 ing cotton in weak acids, and is rather a combustible 
 than an explosive; it is a very different substance 
 from the high explosive, gun-cotton, which is pre- 
 pared from cotton by the use of strong acids. One sol- 
 vent used to make celluloid is a mixture of ether and 
 alcohol; the same solvent has no effect upon gun-cot- 
 ton, to dissolve which acetone must be used. We 
 have, then, two different types of colloid to start 
 with celluloid (formed from weakly nitrated cellu- 
 lose by the use of ether-alcohol), and the acetone col- 
 loid of gun-cotton. It may be stated here that these 
 two types of colloids represent all that is important 
 in relation to colloid material for the manufacture of 
 smokeless powder, as the matter has been understood 
 up to a yery recent date. The various colloids of the 
 eight varieties of gun-cotton above referred to range 
 themselves under one or the other of these two types. 
 We thus have two classes of colloid to experiment 
 
THE DEVELOPMENT OF SMOKELESS POWDER 163 
 
 with, as gunpowder, and all the information we pos- 
 sess in relation to them is the fact that, however they 
 may burn in the gun, yet they will not detonate. 
 
 Suppose that a number of rounds of powder are 
 prepared from the two colloids, how will they act 
 when fired from a gun of a given calibre? Let us as- 
 sume that we have at our disposal the instruments 
 commonly employed for the measurement of muzzle 
 velocities and of bore-pressures, the chronograph and 
 pressure gauges; we will then have, as a basis of com- 
 parison, first, the ratios V/P of velocities to pres- 
 sures*; second, our personal observations of other 
 phenomena attending explosion. 
 
 Actual practice shows that the best results ob- 
 tained for the two powders in a given gun, by vary- 
 ing weights of charge and dimensions of grain, would 
 be about as follows: 
 
 Gun-cotton-acetone colloid V/P = 2100/15-19; 
 
 Celluloid nitro-cotton f colloided with ) r r/r> i c. 
 
 ether-alcohol \ V/P = 2100/16. 
 
 Inspection, of the results shows the existence of a 
 pressure-range of from fifteen to nineteen tons for 
 the acetone powder; this means that while no detona- 
 tion would occur, yet that pressures would jump be- 
 tween certain limits. Such a phenomenon is often 
 observed in the tests of brown powders for heavy 
 
 * A convenient expression for comparing, in a given gun, the 
 ballistic properties of different progressive powders, ^repre- 
 senting velocities in foot-seconds, and P, pressures in tons per 
 square inch. 
 
 f Commonly cal'ed " soluble nitro-cellulose." 
 
164 SMOKELESS POWDER 
 
 guns. A powder of this character would be unsuitable 
 for general f use by reason of pressure irregularity. 
 
 Upon firing the celluloid powder another and per- 
 haps worse inconvenience would be met with. Con- 
 siderable smoke would be developed and the interior 
 of the bore would be found lined, after each round, 
 with a heavy coating of soot, which, after one or two 
 shots had been fired and the gun had become heated, 
 would ignite after each succeeding round upon in- 
 gress of fresh air on opening of the breech, thereby 
 producing flaming at breech and muzzle. 
 
 Obviously, as gunpowder, neither the one nor the 
 other form of colloid is suitable. If they are to be 
 employed they must be improved. Irregularity in 
 pressures from the gun-cotton-acetone colloid is due 
 to brittleness; if we are to use this material we must 
 devise some means of toughening it. The celluloid 
 does not contain enough oxygen to consume its sub- 
 stance into gases; to use the latter we must put more 
 oxygen into it. 
 
 The original phases of the colloid powder question 
 thus present themselves. Neither the one nor the 
 other form of colloid proving suitable for direct 
 manufacture into powder, people began to try to im- 
 prove them by combining them, or by adding foreign 
 substances to them. It was well understood that no 
 form of gun-cotton, however highly nitrated, con- 
 tained enough oxygen to effect its own complete 
 combustion the conversion of its carbon into the 
 higher oxide of carbon, CO 2 . The first idea of the 
 experimenter was to add enough oxygen to the nitro- 
 
THE DEVELOPMENT OF SMOKELESS POWDER 165 
 
 cellulose to thus complete its combustion, and this 
 unfortunate attempt led to many years' delay in the 
 development of smokeless powder. It started in- 
 vestigators off upon a wrong track; complete com- 
 bustion was one thing, and the work necessary to de- 
 velop highest velocity with lowest bore-pressure, an- 
 other. 
 
 With the purpose, then, of improving ballistic 
 qualities of powders by causing them to consume 
 completely and to develop regular pressures, experi- 
 menters in different countries began to try the effect 
 of introducing into the colloid various foreign sub- 
 stances, generally, oxidizing agents, such as ni- 
 trates of metallic bases; sometimes, when the mix- 
 tures became too violent in their action, a substance 
 rich in carbon, called a deterrent, was added. Such 
 work was a good deal like groping in the dark. 
 There was no method in it. But there was one great 
 incentive to keeping it up, viz., the fact thereby es- 
 tablished that the addition of these nitrates to the 
 colloids actually increased the velocity developed for 
 a given bore-pressure, whatever the inconveniences 
 attendant upon the employment of these mixtures as 
 powders may have been.* 
 
 To establish a comparison between the ballistic 
 efficiencies of the two types of pure colloids above 
 
 * The increase in muzzle velocities for a given bore-pressure to 
 be attained by incorporating certain quantities of metallic ni- 
 trates, nitro-glycerin, etc., into the body of the colloid, con- 
 stitutes a special phase of development of progressive powders, 
 which will be discussed in a subsequent paper. 
 
1 66 SMOKELESS POWDER 
 
 cited, and of certain compound colloid powders into 
 the substance of which metallic nitrates or other oxy- 
 gen-carriers are incorporated, the tabular record of 
 performances of powders of these classes is submitted 
 on page 167. Note is made therein of objectionable 
 features developed for each of the explosives named. 
 
 Referring to the table we find powders A and B 
 with properties as already described. The K, BN, 
 and cordite all make good ballistic showings, giving 
 velocities greater by about 300 ft. sec. for a de- 
 veloped pressure than the former. 
 
 The last line of the table shows that each of the 
 powders possesses certain unfavorable qualities which 
 militate against its adoption for service use. The 
 gun-cotton acetone colloid develops irregular pres- 
 sures; the ether-alcohol colloid of soluble nitro- 
 cellulose deposits soot; the K and BN produce some 
 smoke and bore-deposit. Cordite contains a volatile 
 liquid, nitre-glycerin, which develops great heat 
 upon combustion. Now the development of highest 
 velocity at lowest pressure is most important, even 
 if obtained at the expense of the production of cer- 
 tain partially unfavorable conditions; but there was 
 a further incentive to progress at this stage of de- 
 velopment the fact there had been found a form of 
 pure colloid, unadulterated by admixture with other 
 substances, which, while developing high velocities 
 at moderate pressures, possessed the full round of 
 good qualities necessary in a service powder. 
 
 The end sought for in the development of gun- 
 powder is the attainment of the capability of deliver- 
 
THE DEVELOPMENT OF SMOKELESS POWDER l6/ 
 
 * o rt 2 
 
 ~ ~ "^ G ~ 
 T3 CQ 3 "- '- 
 I-" S C 
 O *j u T3 
 
 Jj 'v V 2 G 
 
 p " !2 "e 
 
 I ' 6 '_c J3 .H ^ 
 
 O 'r* i il -4-J ^ 
 
 , v w *o 
 
 O.^ W ^ > rt 
 
 >.So o ^S 
 
 . ^ c 
 
 U, H-> JJ 
 
 fc, S 
 
 ^.S E 
 
 W 
 
 3 
 .'S 
 
 C u 
 
 O v- 
 
 O 
 
 " bfi 8 
 
 'a, ^ 
 ' a; E 
 
 O 4J o 
 
 w^ w 
 
 p 
 
 C/) 
 
 o 
 d. 
 1) 
 
 Q 
 
1 68 
 
 SMOKELESS POWDER 
 
 ing most accurately in a given interval of time the 
 greatest number of most powerful blows; this result 
 to be effected with minimum risk to gunners and with 
 least possible injury to gun. Hereby is implied the 
 fulfilment of a number of important independent 
 conditions, no one of which may be overlooked in the 
 effort to successfully accomplish the object sought. 
 These conditions correspond to qualities possessed 
 by a powder, and may be given, with their opposites, 
 a general grouping under headings as follows: 
 
 TABLE II 
 
 QUALITIES OF POWDERS 
 
 
 i 
 
 2 
 
 3 
 
 4 
 
 5 
 
 Positive < 
 
 Non-liability 
 to 
 detonation 
 
 Devel'pment 
 of minimum 
 heat 
 
 Formation of 
 minimum 
 residue 
 
 Good keep- 
 ing quali- 
 ties 
 
 Maximum 
 propulsive 
 effect 
 
 Negative -j 
 
 Detonation 
 
 Erosion 
 
 Smoke and 
 bore-deposit 
 
 Decomposi- 
 tion 
 
 Low value 
 of V/ P 
 
 The first condition is of paramount importance 
 and limits us to the employment of a colloid material. 
 The second represents the fact that the greater the 
 heat developed the greater the wearing away of the 
 inner surface of the bore. The third requirement 
 means obviation of bore-deposit, which operates to 
 reduce rapidity of fire by necessitating more or less 
 frequent sponging, and to diminish accuracy of prac- 
 tice through the formation of smoke. The fourth re- 
 quirement, stability, is all-important, when we re- 
 member that the ship must carry safely her store of 
 powder for at least a cruise, and for the reason that 
 
THE DEVELOPMENT OF SMOKELESS POWDER 169 
 
 when powders begin to decompose they lose their 
 homogeneity and crumble which means an end to 
 regular-pressure development. 
 
 Let us now take up requirement fifth the attain- 
 ment of maximum propulsive effect and see what it 
 leads to. What causes the expulsion of the projec- 
 tile from the gun? The expansion of powder-gases. 
 What limits our employment of the expansive force 
 of these gases? The attainment of the limiting bore- 
 pressure, this limit being commonly set at fifteen tons 
 per square inch. What represents the greatest 
 amount of work in the form of velocity? The great- 
 est amount of gas-expansion in the gun behind the 
 projectile. 
 
 We desire, then, the greatest amount of gas ex- 
 pansion, but we are limited as to the rate of this ex- 
 pansion; that is, we must not let it develop in the 
 'bore a pressure greater than fifteen tons in the gun- 
 chamber or upon the interior walls of the piece. This 
 is tantamount to saying that to produce maximum 
 velocity we require the evolution, at a suitable lim- 
 ited rate of expansion, of the greatest possible 
 volume of gas, the expansion of the gas constituting 
 under these conditions the propelling impulse. 
 
 It may be stated here that we possess, within limits, 
 the power of controlling the rate of combustion of 
 colloid powders by varying certain conditions re- 
 lating to combustion, the principal of which are (i) 
 the size of the grains of which the charge is com- 
 posed, (2) the volume of the powder-chamber, (3) 
 the length of the bore of the gun, (4) the weight of 
 
SMOKELESS POWDER 
 
 the projectile. Granting, then, that we possess 
 within limits the capability of controlling the rate of 
 evolution of the powder-gases, we are led to the fol- 
 lowing conclusion: that the best smokeless powder 
 is that stable colloid which, for a given weight of its 
 substance, evolves in the bore of the gun at the most 
 suitable rate of evolution and expansion, the greatest 
 volume of gas, the said evolution being accompanied 
 with the development of the least heat. 
 
 We are led by this deduction to regard the action 
 of the powder from a new standpoint. Besides con- 
 sidering what a powder is composed of, we must now 
 consider what gases it is converted into upon decom- 
 position, and what volumes of these gases it gener- 
 ates. Most important of all, it leads us to conduct 
 experiments for the purpose of ascertaining what 
 colloid will, for a given weight of its substance, liber- 
 ate the greatest volume of gases. 
 
 Here was one starting point for a series of experi- 
 mental researches culminating in the discovery of an 
 efficient smokeless powder. Another line of experi- 
 mental approaches to the same end connected the 
 efforts to toughen the substance of colloid films con- 
 taining sufficient oxygen to effect their own com- 
 plete combustion, with a view to the attainment of 
 increased regularity in developed pressures; a third 
 related to fhe determination of the causes of certain 
 ballistic, phenomena hitherto unexplained, e.g., that 
 acetone colloids developed, in some cases, greatly 
 improved ballistic qualities, after the lapses of periods 
 of from six to twelve months from time of manufac- 
 
THE DEVELOPMENT OF SMOKELESS POWDER IJI 
 
 ture. These several lines of experimental investiga- 
 tion proved in the end to converge towards the at- 
 tainment of a common result the development of a 
 special form of nitro-colloid that possessed the 
 toughness and therefore the regularity of burning of 
 celluloid, and that contained enough oxygen to con- 
 vert its substance, upon ignition, into wholly gaseous 
 products which the celluloid did not and that 
 liberated not only the greatest volume of powder- 
 gases at the most suitable rate, but the greatest 
 volume of gases that can be evolved by any colloid 
 at any rate of combustion, whether these colloids 
 contained or did not contain nitro-glycerin, metallic 
 nitrates or other substances. 
 
 The questions now present themselves: in what 
 ways do we possess the control of kind and amount 
 of gases evolved upon the combustion of nitro- 
 cellulose and of its colloids; how does the constitu- 
 tion of these gases vary; what are they? 
 
 It has been already stated that the exact chemical 
 formulae for cellulose and its nitre-derivatives have 
 not yet been written. That for cellulose may be ap- 
 proximately expressed as C 6w H IOM O s , where n is an 
 undetermined or indeterminate numerical quantity. 
 When cellulose is steeped in nitric acid or in a mix- 
 ture of nitric and sulphuric acids, it is converted into 
 the substance the composition of which may be ex- 
 pressed as a.sC 6n H. ton . an O 5n (NO 2 ) an . These two expres- 
 sions, C 6 H Io;i O ew and C 6w H IOW _ flW O 5 (NO 2 ) tfw , may be 
 considered in comparison with one another. Bnt 
 contain the same quantity of carbon; the q-.?.::t:i- 
 
172 SMOKELESS POWDER 
 
 ties present of the other elements are changed by 
 nitration; an atoms of hydrogen are displaced by 
 an equivalents of a combination of nitrogen and 
 oxygen (NO 2 ). This additional oxygen from the 
 NO 2 acts to supply the energy that converts the 
 cellulose into an explosive, and enters into its sub- 
 stance in combination with a certain quantity of nitro- 
 gen. Why it carries the nitrogen with it we do not 
 know, but it is a chemical fact that it does so the 
 fact upon which the designation of the new ex- 
 plosives as nitro-derivatives or nitro-explosives is 
 based. 
 
 Upon the ignition of the nitro-cellulose or its 
 colloid the nitrogen is set free; the hydrogen com- 
 bines with its equivalent of oxygen and appears in 
 the air as steam; the remaining oxygen unites with 
 the carbon to form gaseous oxides of carbon. If 
 there be not enough oxygen to consume all the car- 
 bon into gas, part of the latter is deposited as soot; 
 this result was obtained in the attempt to employ 
 celluloid as gunpowder. If there is enough oxygen 
 to consume all the carbon into gases, we have, as 
 products of combustion, a mixture of the gaseous 
 oxides of carbon, of which there are two, CO 2 and 
 CO. 
 
 The property possessed by carbon of combining 
 at high temperatures with intense energy with 
 oxygen, to form gaseous oxides, is the fact upon 
 which the practical development of explosives de- 
 pends; there are many explosives that contain neither 
 
THE DEVELOPMENT OF SMOKELESS POWDER 
 
 carbon nor oxygen, but with these we have as yet no 
 practical relations in ordnance matters. 
 
 It has been stated already that the attempt to ob- 
 tain what is called complete combustion for nitro- 
 cellulose colloids by incorporating oxidizing agents 
 into them had misled investigators, who confounded 
 the attainment of complete combustion with the de- 
 velopment of maximum velocity at lowest pressure. 
 " Complete combustion " means the conversion of all 
 the carbon into the higher oxide of carbon car- 
 bonic acid, CO 2 a dense gas about 1.9 times as 
 heavy as the air, the formation of which is accom- 
 panied with the development of a high degree of heat. 
 The complete combustion of carbon into carbonic 
 acid gas with the corresponding evolution of a great 
 amount of heat is the characteristic of the combus- 
 tion of nitro-glycerin. The lower oxide of carbon, 
 CO, is a gas much less dense than carbonic acid, pos- 
 sessing a density of about 1.4 times that of air. Sup- 
 pose that we have a given weight of a compound of 
 carbon and oxygen, with the elements taken in such 
 proportions as to produce, on ignition, complete 
 combustion into carbonic acid gas, CO 2 , with the ac- 
 companying evolution of a large amount of heat; 
 let us also suppose that we have an equal weight of 
 compound of the same elements in such proportions 
 as to develop, on ignition, the lower oxide of carbon, 
 carbonic oxide, CO; then, the latter compound, de- 
 veloping the lower oxide, would liberate a volume of 
 gas nearly 1.9/1.4=1.36 times greater than the 
 former. The greater heat produced by the forma- 
 
174 SMOKELESS POWDER 
 
 tion of the carbonic acid gas would cause the volume 
 of that gas evolved to be the more expanded, but, 
 at the same time, and this is a fact of crucial impor- 
 tance in the present work, the greater heat would 
 cause the gas to be generated at a more rapid rate, 
 in fact, at an extremely rapid rate; and what we re- 
 quire is a low, .regular rate of gas evolution to pre- 
 vent our exceeding at any time the set limit of per- 
 missible bore-pressure. 
 
 In our effort to generate from colloid nitrocellu- 
 lose the greatest volume of gas at the most gradual 
 rate of expansion, we must seek, then, (i) to avoid 
 the formation of CO 2 when CO may be formed in 
 lieu thereof; (2) to avoid the formation of free car- 
 bon; and (3) to generate the maximum volume of the 
 lower carbon oxide. If we give due consideration 
 to the amounts of water and of nitrogen formed 
 simultaneously with the oxides of carbon, we shall 
 find that the form of nitro-cellulose developing the 
 greatest volume of gas at the most suitable rate, cor- 
 responds to the formula C3 H3 8 (NO 2 ) 12 O 25 , which 
 breaks up on decomposition into 30 CO + 19 
 H 2 O + 12 N. 
 
 This material is a new type of nitro-cellulose,* de- 
 veloped by experiment to meet ballistic require- 
 
 * This special form of nitro-cellulose, which corresponds to a 
 content of nitrogen of 12.44 P er cent, and which was first devel- 
 oped in Russia by the eminent chemist, Professor D. Mendeleef, 
 has been independently developed at the Torpedo Station, 
 through the study of effects of variation of times of immersion, 
 temperatures of nitration and of washing, and strength of acids 
 employed in the nitration of cellulose. 
 
THE DEVELOPMENT OF SMOKELESS POWDER 1/5 
 
 ments, which contains just enough oxygen to convert 
 its substance into a gaseous ' body. Its formation 
 from cellulose depends upon strengths of nitric and 
 sulphuric acid, temperatures of reaction and time of 
 immersion of the cellulose in the acids from which the 
 material is prepared. With ether-alcohol it forms a 
 colloid that possesses, on the one hand, the tough- 
 ness, and therefore the capability of development of 
 regular pressures, of celluloid; and on the other the 
 capability of consuming into wholly gaseous prod- 
 ucts that characterizes the gun-cotton acetone 
 colloid; while as a powder it develops, with present 
 types of guns, excellent values of V/P of about 
 2400/16. Briefly, it may be described as a celluloid 
 containing enough oxygen to convert its substance 
 (when it is consumed out of contact with the atmos- 
 phere) wholly into gaseous products. 
 
 It was stated in a preceding paragraph that the bal- 
 listic effect produced by a progressive explosive de- 
 pended directly upon the volume of gas it evolved 
 upon combustion, but was not directly dependent 
 upon the attainment of complete combustion. As- 
 suming total conversion from solids into gases, and 
 non-liability to detonation, pyrocellulose was shown 
 to be the form of nitro-cellulose best adapted for con- 
 version into smokeless powder. As this material con- 
 tains only enough oxygen to convert its carbon into 
 carbonic oxide, CO less than gun-cotton, which 
 converts its carbon partly into carbonic acid gas, 
 CO 2 , and partly into carbonic oxide, CO the at- 
 tainment of maximum efficiency from nitro-cellulose 
 
SMOKELESS POWDER 
 
 was thus shown to be accomplished through a reduc- 
 tion from a maximum to a mean in the quantity of 
 oxygen capable of being incorporated into nitro- 
 cellulose. 
 
 On the other hand, it was stated that the incor- 
 poration of certain quantities of oxygen-carriers 
 (nitro-substitution compounds other than nitro-cel- 
 luloses, such as nitro-glycerin and nitrates of metallic 
 bases) into colloid nitro-cellulose, led to the attain- 
 ment of an increase in initial velocity of projectile for 
 a given developed bore-pressure. As nitro-glycerin 
 furnishes a surplus of free oxygen to aid in complet- 
 ing the combustion of the gases from the nitro-cellu- 
 lose, while the nitrates surrender oxygen on applica- 
 tion of heat, it would appear in this case that the at- 
 tainment of a more complete combustion led to im- 
 provement in ballistic effect. 
 
 We are thus brought face to face with a seeming 
 contradiction how, on the one hand, we must re- 
 move oxygen; how, on the other, we must add oxy- 
 gen to a progressive explosive, in order to obtain 
 maximum ballistic effect therefrom. In order to 
 reconcile these apparently contradictory statements 
 we must consider the manner of decomposition of the 
 explosive in both cases. 
 
 One chief characteristic of pyrocellulose is its 
 homogeneity. It represents no mixture of explosives 
 and combustibles, such as are presented by other 
 forms of powders, and it is converted directly by com- 
 bustion into a set of gaseous decomposition products 
 that may not be varied in amount and kind. Under 
 
THE DEVELOPMENT OF SMOKELESS POWDER 1/7 
 
 these conditions the ballistic effect of the expanding 
 gases from pyrocellulose may be referred to quantity 
 of charge, area of ignition-surface, weight of projec- 
 tile, calibre of gun, and volume of powder-chamber. 
 Other conditions affecting developed pressure and 
 velocity are bore-friction and resistance of the pro- 
 jectile to rotation through inertia. 
 
 The gun may be regarded as a gas-engine in which 
 the walls of the chamber and bore form the cylinder; 
 the projectile, the piston. The expanding powder- 
 gases perform work by imparting velocity to the pro- 
 jectile, the inertia of which they overcome just as gas 
 by its expansion in the cylinder overcomes the inertia 
 of the piston and the parts linked thereto. In the en- 
 gine the gas is admitted alternately, first at one end of 
 the cylinder and then at the other; in the gun it is ad- 
 mitted in rear of the projectile but once, so that the 
 gun is an engine of a single stroke. In the engine 
 the steam is admitted into the cylinder through a 
 valve, and, after the lapse of a period of time less than 
 that required for a full stroke, admission is cut off 
 and work for the rest of the stroke is performed ex- 
 pansively; in the gun the charge of powder consti- 
 tutes both the gas itself and the valve that admits the 
 gas for each grain of powder may be considered as 
 a notch of opening of a valve; the more grains there 
 are the greater the ignition surface, the greater the 
 rate of emission of gas, or the greater the number of 
 notches the valve is open. 
 
 The action of nitro-cellulose powder-gases in im- 
 parting motion to the projectile is that of the gas in 
 
1^8 SMOKELESS POWDER 
 
 the engine cylinder. The decomposition products 
 are evolved at a high pressure, and act to propel the 
 projectile, just as the gas or vapor drives the piston 
 in an engine. Thus far the two cases are in parallel; 
 they differ in that the space occupied by the gas in 
 the gun is constantly increasing, both through the 
 effect of the motion of the projectile along the bore 
 and from the increase of chamber space due to the 
 melting away of the powder charge, while space in 
 the engine cylinder is increased through the motion 
 of the piston and through connection with the valve 
 before cut-off. As shown in Table I, the ballistic 
 value of gun-cotton colloided in acetone (for a given 
 
 2 100 
 gun) was ; that of colloided soluble nitro-cellu- 
 
 lose, containing not enough oxygen to convert its 
 carbon wholly into carbonic oxide, CO, was j- . 
 
 Under similar conditions of firing, pyrocellulose de- 
 
 2400 
 veloped a value of V/P = ^-. It may be urged 
 
 that the ballistic superiority of the latter colloid as 
 compared with that of the two former is not wholly 
 attributable to character and volume of evolved gases, 
 as the acetone colloid is brittle, and that prepared 
 from soluble nitro-cellulose is somewhat brittle, and 
 deficient in oxygen, while pyrocelluloid is of a tough, 
 leathery consistency, capable of withstanding high 
 pressures without premature disintegration. Never- 
 theless, as these colloids prove inferior to the pyro- 
 colloids for lower pressures of about 10 tons per 
 
THE DEVELOPMENT OF SMOKELESS POWDER 
 
 square inch, at which the effects of brittleness are not 
 perceptible, and for which they all -afford pressures 
 regularly proportional to develop velocities, it re- 
 mains that magnitude of volume of evolved gases is 
 a factor of prime importance in the attainment of bal- 
 listic efficiency. 
 
 COMPOSITE POWDERS 
 
 The results of incorporating an oxidizing agent or 
 oxygen-carrier into colloids merit special study. Sup- 
 pose that a given nitro-cellulose be colloided and 
 formed into strips of a number of definite thicknesses. 
 If these strips be collected separately and dried, we 
 may prepare from them series of rounds, each series 
 composed of different weights of strips of some one 
 thickness. If the length and breadth of the strips be 
 great in relation to their thickness, we need consider 
 only the latter element of dimension in relation to 
 their mode of combustion.* 
 
 * We have (Glennon, Interior Ballistics, chap. VI, pp. 59, 60) 
 
 where y is the fractional part of the least dimension of the grain 
 burned up to any moment; (>(y) the fractional part of the whole 
 grain burned up to the same moment; and a, y, and ju, constants 
 depending upon the form of the grain. 
 
 If the grain be a rectangular parallelepiped with a square base, 
 and the altitude as the least dimension, we have 
 
 2X + X* ~ 
 
 a = i -f 2x, A. = - , ju = 
 
 i -f zx' 
 
 where x is the ratio of the altitude to the'side of the base." 
 
 Applying the above to the present case we find that if the alti- 
 
180 SMOKELESS POWDER 
 
 Upon firing series of rounds of the several powders 
 from a given gun we obtain the following results as 
 to their manner of explosive action: 
 
 I. Strips of over a certain mean thickness will be 
 only partly consumed in the bore; the unconsumed 
 remnants will be projected burning from the gun, to 
 be quenched in the cool outer air, where they fall un- 
 consumed to the ground and may be picked up at 
 various distances from the piece in front >f the muz- 
 zle, possessing the original form (in reduced dimen- 
 sions) of the grains of which the charge was origin- 
 ally composed. Such powders develop low bore pres- 
 sures and afford low muzzle velocities. In point of 
 work performed they are equivalent to smaller 
 charges of quicker powders. It may be remarked 
 that no work is done in raising the temperature of 
 the unconsumed portions of the grains, for if the tem- 
 perature of the latter be raised but a few degrees, the 
 ignition point of the explosive is reached and its sub- 
 stance would wholly disappear. 
 
 2. Strips of under a certain mean thickness are 
 totally consumed in the gun. They develop high 
 pressures for low velocities. The thinner the strips 
 the less the weight of charge required to develop the 
 
 tude be considerably diminished (jc approaches zero) we have the 
 case of the thin plate; and that the constants approach the values 
 
 a= i, A = o, JLI = o, 
 
 or <p(y) = y. 
 
 But y depends alone on the thickness of the plate, therefore 
 the speed of combustion of a plate is a linear function of its least 
 dimension. 
 
THE DEVELOPMENT OF SMOKELESS POWDER l8l 
 
 limiting permissible pressure, on account of the 
 greater initial surface presented by the thinner strips, 
 which occasions a high initial gas development. 
 
 3. A certain mean thickness of strip will be found, 
 for which, at a set limit of pressure, a minimum 
 weight of powder will develop the greatest velocity 
 that can be developed at that pressure. If strips of 
 other thicknesses develop practically identical veloci- 
 ties and pressures for the same pressure limits, it will 
 be by burning greater weights of powder. Such a 
 powder may be designated a maximum powder, for 
 the material from which it is prepared and for the gun 
 from which it is fired. 
 
 Suppose, then, that colloided gun-cotton of nitra- 
 tion N= 13.3 develops in a given gun a maximum 
 
 value VIP = , what will be the effect of incor- 
 porating into such powder a certain quantity of nitro- 
 glycerin, or of metallic nitrates such as barium and 
 potassium nitrates? Assume that during the process 
 of colloiding the requisite quantity of nitrates be uni- 
 formly incorporated throughout the substance of the 
 pasty mass, which is subsequently formed into strips, 
 as before. For this material we shall find that the 
 
 maximum powder develops a value of V/P- f 
 
 2100 
 as against for the pure colloid, a gain in velocity 
 
 of 300 ft. sec. for a given pressure; in energy, ( J, 
 
 of about 30 per cent. 
 
1 82 SMOKELESS POWDER 
 
 If, in lieu of nitrates, we incorporate nitro-glycerin 
 into the colloid, we will obtain a pasty mass that can 
 be worked conveniently into the form of rods or 
 cords, whence the name " cordite," applied to one of 
 its best known types. Cordite, as used in England, 
 consists of 
 
 Nitro-glycerin 58 parts 
 
 Gun-cotton 37 parts 
 
 Vaseline 5 parts 
 
 Such a powder, fired under the above conditions, 
 develops a value of V IP = - approx. 
 
 There is one characteristic of powders, such as the 
 K and the French BN, containing nitrates, to which 
 attention is to be directed. The nitrates contained in 
 these powders exist in them in a state of suspension; 
 in an undissolved state. For the BN the microscope 
 reveals minute crystalline particles uniformly dissem- 
 inated throughout its mass; the barium nitrate em- 
 ployed in the K powder is insoluble in the colloiding 
 agent, acetone, and is also insoluble in the colloid, in 
 which it is held in a state of suspension and of uni- 
 form distribution. 
 
 In the case of the nitro-glycerin powders it is 
 known that the nitro-cellulose is not in true solution 
 in the nitro-glycerin. In this connection the follow- 
 ing quotation from an authority upon nitro-glycerin 
 powders, Mr. Hudson Maxim, may be cited: 
 
 " In the very early smokeless powders, especially 
 those made of compounds of soluble pyroxylin (gun- 
 
THE DEVELOPMENT OF SMOKELESS POWDER 183 
 
 cotton) and nitre-glycerin, it was supposed that the 
 nitro-glycerin actually held and retained the pyroxy- 
 lin in solution, but it has since been learned that the 
 nitro-glycerin is held by smokeless powders, whether 
 made from high- or from low-grade gun-cottons, in 
 much the same manner as water is held by a sponge; 
 in fact, the pyroxylin exists in smokeless powders in 
 the shape of a very minute spongy substance, and the 
 nitro-glycerin is held in a free state within the pores 
 cf this sponge." 
 
 " It is possible even with powders containing as 
 little as 25 per cent, of nitro-glycerin, to squeeze out 
 the nitro-glycerin in a pure state by subjecting a piece 
 of this powder to great pressure between smooth 
 steel plates." 
 
 The quantity of nitro-carrier (nitrate or nitro-sub- 
 stitution compound other than nitro-cellulose) consid- 
 ered necessary to the production of good ballistic re- 
 sults, as exemplified in certain known powders, may 
 be tabulated as follows: 
 
 TABLE III 
 
 Variety of .,._ , Per cent, of Nitro-carrier 
 
 Powder in Given Wt. of Powder 
 
 Cordite Nitro-glycerin 58 
 
 Maxim Nitro-glycerin 10 to 25 
 
 BN Barium- and Potassium nitrates 21 to 25 
 
 K Barium nitrate 14-25 
 
 t 
 
 The composition and ballistic properties of the 
 three classes of explosives pure colloids, colloids 
 containing metallic nitrates, and colloids containing 
 nitro-glycerin may be compared as follows: 
 
1 84 
 
 SMOKELESS POWDER 
 TABLE IV 
 
 Pure Colloid 
 
 K 
 
 BN 
 
 Cordite 
 
 Gun-cot- 
 ton, 85.00 
 Soluble ni- 
 tro-cellu- 
 lose, 10.00 
 Sod.carb, i.oo 
 
 Solvent, res- 
 ins, etc., 4.00 
 
 Gun-cot- 
 ton and 
 soluble 
 nitro-cel- 
 lulose, 
 balanced 
 
 Barium 
 nitrate, 
 
 Calc. carb 
 
 -84,25 
 
 14.25 
 , 1-50 
 
 Insol. nitro- 
 cellulose, 38.67 
 
 Soluble nitro- 
 cellulose, 33.23 
 
 Barium 
 nitrate, 18.74 
 
 Potassium 
 nitrate, 4.54 
 
 Calc. carb., 3.65 
 Volatile, 1.29 
 
 Nitro-gly- 
 cerin, 58 
 
 Gun-cot- 
 ton, 37 
 
 Vaseline, 5 
 
 100 
 
 100.00 
 
 100.00 
 
 IOO.I2 
 
 TABLE V 
 
 Type 
 
 Pure 
 Colloid 
 
 Metallic Nitrate 
 
 Metallic Nitrate 
 
 Nitro-glycerin 
 
 Manner 
 
 
 Solid undis- 
 
 Solid undis- 
 
 Undissolved 
 
 of incor- 
 poration 
 of oxy- 
 gen-car- 
 rier 
 
 
 solved parti- 
 cles, uniform- 
 ly distributed 
 throughout 
 colloid ma- 
 trix 
 
 solved parti- 
 cles, uniform- 
 ly distributed 
 throughout 
 colloid ma- 
 trix 
 
 particles 
 held in sus- 
 pension like 
 water in 
 sponge 
 
 V 
 
 2IOO 
 
 2400 
 
 2400 
 
 2400 
 *5 
 
 P 
 
 15-19 
 
 15 
 
 15 
 
 Remembering what has been said in relation to the 
 ballistic performance of the varieties of powders cited, 
 we are led to the following conclusion: 
 
 That minute particles of an oxygen-carrier uniformly 
 incorporated into a nitro-colloid and Iwld in suspension 
 
THE DEVELOPMENT OF SMOKELESS POWDER 1 8$ 
 
 in an undissolved state throughout tJte body of the latter, 
 render more progressive tJw combustion of the nitro- 
 colloid into which they are incorporated. 
 
 For convenience of reference I shall refer here- 
 after to the oxygen-carrier held in suspension in the 
 colloid as the accelerator. Viewed in the light of the 
 principle here enunciated, the several powders we 
 have been considering are all similar variants of the 
 pure colloid. The remark of the compounder, " that 
 a little nitro-glycerin certainly does help the powder 
 along," is now the more readily comprehensible. 
 
 The methods commonly employed for co-ordina- 
 ting natural species may be applied, by way of illus- 
 tration, to the classification of the various types of 
 progressive explosives, to establish their relations to 
 one another, and to indicate the lines alohg which 
 advances have been effected. 
 
 We shall next consider how the accelerator acts to 
 develop increased velocity without developing in- 
 creased pressure. 
 
 i. It has already been shown how it is possible 
 with powders containing as little as 25 per cent, of 
 nitro-glycerin to squeeze out the nitro-glycerin in a 
 pure state by subjecting the powder to great pressure 
 between smooth steel plates. 
 
 If it be possible to extract nitro-glycerin by appli- 
 cation of pressure from powder in which it is incor- 
 porated, then there will be a tendency to flow in the 
 direction of least pressure from the instant of ignition 
 of a charge to that of its complete combustion. This 
 would mean, first, flow from within outwards in the 
 
} 
 
 
 
 
 
 1 86 
 
 
 
 SMOKELESS POWDER 
 
 
 
 
 
 go A-a , 3 o ' 
 
 
 
 
 
 l-o'E-f -f .-S^i rt'|J o ^."Sl-? 
 
 
 
 
 
 A.S-c 2 ^^ 'vO 012 *)'?.- 1 ^ V'5^3 o *" 
 
 
 
 
 
 3-gU pj -^-tig^gC-SS s'o-r.ti'o 
 
 
 
 
 
 
 
 
 
 "X3 
 
 ^'5 C> c3f8o S i-^Tj n 
 
 
 
 
 J3 
 
 rt Srt'S^ 2 ao"S " 
 
 
 
 
 ^ 
 
 w -^ *"* S M c 8 .2n*c2o f o*^ 
 
 
 
 
 
 U 
 o 
 
 ^^haco g |a-B'c!2rta 
 
 i 
 
 
 
 w 
 
 ^i ,'Cg , j. 
 
 
 
 
 2 
 
 
 ^ 
 
 
 
 15 
 
 o ^ <Q *K **^ ^ w goVS" ju 
 
 ^ 
 
 
 
 
 
 c 
 
 
 
 u 
 
 fZn^Q^gOv* rtuO* -t ^C 4J 
 
 rt 
 
 
 
 ^ 
 
 <j85i^3rtS > a-s 'c^rt'S 
 
 i 
 
 
 
 
 "*8 -S .i'sto'g. 
 
 
 
 
 
 
 ||ic2 ||| IS-gftlS 
 
 s 
 
 
 00 
 
 W 
 
 
 w^.Sl'S o^^ J3S.o.^'| 
 <j o c o o (jH^i J> KG C o O4i 
 
 
 V 
 
 -5 
 
 
 00 
 
 
 f 6 o w 
 
 
 
 s 
 
 
 ^rS8 g-^^g 
 
 8 
 
 
 PH 
 
 
 QJ^-^C fy C >C U > 
 
 G 
 
 
 
 
 O fe tp V* t_ 
 
 u 
 
 > 
 
 H 
 
 3 
 
 pa 
 
 ESSIVE I 
 
 
 111! II, III! ' " 
 
 yS.UM (,00^5 
 
 loped in F 
 
 < 
 H 
 
 PROGR 
 
 Colloids 
 
 3 1 i|*= 
 Ss -S ^ ^|Scl 
 
 > 
 ii 
 
 t) 
 
 4J 
 
 
 
 
 **^bi 3^i_,DjH*d3 
 
 a 
 
 
 
 I 
 
 <W^^ O-OOJ30 
 
 fe 
 
 
 
 CM 
 
 g f S.sf8 S 8 
 
 p-ssig l..g Q 
 
 ^ 
 
 
 
 
 (X-^^^^ p- mfe 
 
 
 
 
 
 
 i ' "^ 
 
 c/: 
 
 
 
 
 c-a'n'rt c'o 5-^ 
 
 "rt 
 
 
 
 
 o - Q QJ *- ^ ^ *5cS"^O!d efl r) 
 
 C 
 
 
 
 
 Ilibli s^liil 
 
 O 
 u 
 
 
 
 a 
 
 fe 
 
 ^^ 2* J 'o- C^ U 'e tu[i < l,^-'o. ^ 
 
 H 
 
 
 
 S-o 
 
 --( ^ 
 
 * 
 
 
 
 n 
 
 w f Jf !. ij 
 
 
 
 
 **3 
 
 OQtX^ ^.Soo;na.rt 
 
 
 
 > 
 
 
 
 S , S c u 
 
 
 
 'a 
 
 C 
 
 <u 
 O 
 
 'o -2 'o " 'S S 
 
 oj 3 41 rt ti /u 
 
 a c/) a n S 
 
 
THE DEVELOPMENT OF SMOKELESS POWDER l8/ 
 
 gun-chamber, where a relatively large proportion of 
 the nitre-glycerin would be consumed; second, flow 
 in the direction of the windage, where the quantity of 
 nitre-glycerin consumed would also be relatively 
 great. Such action accounts for the rapid erosion of 
 the surfaces of gun-chamber and rifled bore when 
 powders containing nitro-glycerin are employed. 
 
 2. The eminent Russian chemist, Professor D. 
 Mendeleef, developer of smokeless powder in Russia, 
 in a paper upon pyrocellulose powder, says: 
 
 " The chemical homogeneity of pyrocollodion 
 plays an important part in its combustion, for there 
 are many reasons for believing that upon the com- 
 bustion of those physically but not chemically 
 homogeneous materials, such as nitro-glycerin pow- 
 der (ballistite, cordite, etc.), the nitro-glycerin is de- 
 composed first, and the nitro-cellulose subsequently 
 in a different layer of the powder. The experiments 
 of Messrs. T. M. and P. M. Tcheltsov at the Scientific 
 and Technical Laboratory show that for a given 
 density of loading the composition of the gases 
 evolved by nitro-glycerin powders varies according 
 to the surface area of the grains (thickness of strip), 
 a phenomenon not to be observed in the combustion 
 of the pyrocellulose powders. There is only one ex- 
 planation for this, viz., that the nitro-glycerin, which 
 possesses the higher rate of combustion (Berthelot), 
 is decomposed sooner than the nitro-cellulose dis- 
 solved in it. This is the reason why nitro-glycerin 
 powders destroy the inner surfaces of gun-chambers 
 with such rapidity." 
 
1 88 SMOKELESS POWDER 
 
 We conclude from the above that the nitro-glycerin 
 incorporated into a colloid burns more rapidly than 
 the nitro-cellulose forming the colloid. More nitro- 
 glycerin is consumed with one part of the charge than 
 with another. During the first period the products 
 of combustion evolved in chamber and bore are 
 largely those of nitro-glycerin; during the second, 
 those of nitro-cellulose. 
 
 Moreover, as both materials exist in an uncom- 
 bined state, although in one of intimate admixture; 
 as both decompose wholly into gases, while each con- 
 tains sufficient energy to continue its own decompo- 
 sition, once that decomposition is begun, there is no 
 reason why the rates of the two decompositions 
 should be equal; it would rather appear that each 
 substance should decompose at the rate peculiar to 
 itself, so far as it was able, under existing conditions 
 of heat and pressure, to effect a separation of its sub- 
 stance from the mixed mass of the powder. 
 
 Conditions point, therefore, to there being two in- 
 tervals in the decomposition of the charge, during one 
 of which a maximum quantity of nitro-glycerin, and, 
 during another, a maximum quantity of nitro-cellu- 
 lose is burning. 
 
 In what follows it is not intended to attempt more 
 than to indicate the mode of progressive combustion 
 as implying the superimposition of maxima and 
 minima of effort. This may be represented graphic- 
 ally in the present case as follows: 
 
THE DEVELOPMENT OF SMOKELESS POWDER 189 
 
 NITRO-GLYCERIN NITROCELLULOSE 
 
 The result of the combination of the conditions 
 here indicated would be the imparting of a double im- 
 pulse to the projectiles due to the successive occur- 
 rence of two maxima of acceleration. Considered as 
 to their limit of possible range, the successive im- 
 pulses may occur incrementally, so that the accelera- 
 tor may be expressed in the form 
 
 where p' represents the pressure due at any instant to 
 the combustion of the nitro-glycerin; p" , that due to 
 the nitro-cellulose. 
 
 The projectile may be regarded as receiving a third 
 impulse, resulting from the chemical combination of 
 the gases evolved by the nitro-glycerin and the nitro- 
 cellulose. According to the researches of Messrs. 
 Macnab and Ristori (Proc. Royal Soc., vol. LVI, 
 p. 8), the decomposition products of nitro-glycerin 
 are 
 
 CO, 
 
 CO CH 
 
 57.6 - - 
 
 O H 
 
 2.7. 
 
 N 
 18.8 
 
 H 2 0. 
 
 20.7 
 
190 SMOKELESS POWDER 
 
 And from the same source we obtain the decomposi- 
 tion products of nitro-cellulose (N = 13.3) as 
 
 CO 2 CO CA 4 O H N H 2 O 
 
 29.27 38.52 0.24 0.86 13.6 16.3 
 
 What may be called the third impulse would repre- 
 sent the combination at a high temperature of mul- 
 tiples of decomposition products developed in the 
 ratios 
 
 CO 2 CO CH 4 O H N H 2 O 
 
 '.6 2.7 18.8 20.8 
 
 .#[29.27 38.52 0.24 0.86 13.6 16.3] 
 
 These phases may be indicated graphically as fol- 
 lows: 
 
 COMBINATION OF PRODUCTS Of 
 COMBUSTION OF NITRO-GLYCERIN 
 NITRO-GLYCERIN AND NITRO-CELLULOSE AND NITRO-CELLULOSE 
 
 Accelerated colloids of K and BN types containing 
 metallic nitrates are next to be considered. We may 
 assume that the nitro-colloid into which minute par- 
 ticles of a nitro-carrier of this type are cemented itself 
 burns in approximation to the law of decomposition 
 of the colloid. This state of affairs is similar to, 
 
THE DEVELOPMENT OF SMOKELESS POWDER IQI 
 
 though not identical with, the preceding; in the 
 former, both nitre-glycerin and nitro-cellulose are 
 able to effect their own decomposition, evolving 
 gases that recombine; in the latter, the nitro-cellulose 
 alone possesses this property, the metallic nitrates 
 surrendering their oxygen through the effect of heat 
 developed during decomposition of the colloid. The 
 successive reactions may be represented as follows: 
 
 UNION OF GASEOUS DECOMPOSITION 
 NITRO-CELLULOSE PRODUCTS 
 
 DECOMPOSITION OF BARIUM 
 NITRATE 
 
 Instead of three maxima of effort there are two 
 maxima and one minimum, the maxima representing 
 the combustion of the nitro-cellulose and the subse- 
 quent combination of the gases therefrom with the 
 oxygen of the barium nitrate; the minimum, the ab- 
 sorption of heat expended in decomposition of the 
 barium nitrate. 
 
 A comparison of the diagrams shows that the proc- 
 esses of combustion in the case of colloids containing 
 nitro-glycerin and of those containing metallic ni- 
 trates are similar. Both represent aggregates of 
 work resulting from successions of independent de- 
 compositions. For such powders an element of time 
 enters into our conception of chemical action; what 
 the ultimate products of combustion are depends 
 upon the order of occurrence of successive evolutions 
 
192 SMOKELESS POWDER 
 
 of various volumes of different gases at high tem- 
 peratures.* 
 
 The base of the projectile is subjected to a series of 
 impulses due to the development of successive waves 
 of pressure; the result is an increased initial velocity 
 for a given developed pressure, the acceleration be- 
 ing sustained throughout a comparatively longer 
 period of time. 
 
 Those familiar with experimental development of 
 ordnance during recent years remember a type of 
 multi-charge gun the construction of which seemed 
 based upon a favorable combination of correct prin- 
 ciples, but which was rejected on trial, as its practical 
 disadvantages were found to outweigh by far its ad- 
 vantages. I refer to the Lyman-Haskell multi-charge 
 gun, a weapon supplied with a number of pockets dis- 
 tributed along the axis of the bore. In each pocket 
 a charge of powder was placed; it was supposed that 
 the projectile, by uncovering successive pockets in its 
 flight, would cause their contents to ignite and thus 
 furnish successive accelerating impulses to ihcrease 
 its velocity. 
 
 From what has been already said in relation to the 
 principle of successive combustion, it will be seen 
 that the employment of charges of accelerated pow- 
 der, like those above described, in a gun of present- 
 day type, represents the limiting extension of the 
 multi-charge principle. In relation to their succes- 
 sive combustions, the nitro-glycerin and the nitro- 
 
 *See extracts from paper by Prof. Mendeleef, p. 33. 
 
THE DEVELOPMENT OF SMOKELESS POWDER 1 93 
 
 cellulose may be considered as sub-charges, con- 
 tained in independent chambers or pockets, or dis- 
 tributed throughout a very large number of small 
 pockets. 
 
 The principle already stated, as established by the 
 study of the ballistic action of accelerated or com- 
 posite powders, may be now amplified as follows: 
 
 Minute particles of an oxygen-carrier, uniformly in- 
 corporated into a nitro-colloid and held in suspension 
 throughout the mass of the colloid in an undissolved 
 state, act through their independent combustion in such 
 a manner as to render more progressive the combustion 
 of the colloid into which they are incorporated. 
 
INDEX 
 
 PAGE 
 
 Accelerator 185 
 
 Acid, nitric, table showing action and results of mixtures of . . . 83 
 
 Nitrohydric 109 
 
 Sulpho-nitric, table showing action and results of mix- 
 tures of 87 
 
 Acid bath, table showing composition of, in relation to yield of 
 
 nitrogen dioxide 132 
 
 Alcohol, absolute, as a solvent 52 
 
 Action of, in facilitating solution 43 
 
 Alkalies, action of, on cellulose 46 
 
 Ammonium compounds as materials for explosives 108 
 
 Ammonium nitrate in smokeless powders 122 
 
 Attainment of maximum propulsive effect 169 
 
 Ballistic efficiencies of pure colloids and compound colloids 
 
 containing accelerators, table showing 167 
 
 Ballistic properties and composition of explosives, table show- 
 ing 184, 186 
 
 Ballistite 117 
 
 Benzol derivatives as materials for explosives 121 
 
 Cellulose 5 
 
 Action of alkaline hydrates on 46 
 
 As a base for smokeless powders 125 
 
 Composition of 57 
 
 Conditions governing formulation of 66 
 
 Constitution of 21 
 
 Dinitrate 14 
 
 Electrolytic strain in 50 
 
 Hexanitrate 12 
 
 195 
 
196 INDEX 
 
 PAGE 
 
 Cellulose hydrate, composition of 57 
 
 Hydration of 45 
 
 Incompletely nitrated 92 
 
 Molecule, theory of 37 
 
 Nitrates 74 
 
 Nitrates, composition of 11 
 
 Nitrates, volume of nitrogen dioxide disengaged per gram 
 
 by 128 
 
 Pentanitrate 13 
 
 Structure of 21, 59, 60, 61 
 
 Structure of polymeric forms of 63, 64 
 
 Tetranitrate 13 
 
 Thiocarbonate 48 
 
 Trinitrate 13 
 
 Type 69, 70 
 
 Cocoa powder, gases evolved by 101 
 
 Cold, action of, in accelerating solution 51 
 
 Solubility of nitro-celluloses in the 43 
 
 Solubility of nitro-hydrocellulose in the 41, 42 
 
 Solubility of pyrocellulose in the 43 
 
 Collodion cotton ? 
 
 Collodion-pyroxylin 7, 1 3 
 
 Collodions 16 
 
 Table showing effect of washing and pulping on viscosity 
 
 of 151 
 
 Colloid 161 
 
 Colloidization 71 
 
 Colloids, accelerated 190 
 
 Colloids, compound, containing accelerators, table showing 
 
 ballistic efficiencies of 167 
 
 Pure, table showing ballistic efficiencies of 167 
 
 Combustion, complete 173 
 
 Of grains 179 
 
 Of powder 110 
 
 Complete combustion. . : 173 
 
 Composite powders 179 
 
 Cordite 166 
 
 Cotton, friable 7 
 
 Nitration of 127 
 
 Researches upon nitration of 81 
 
 Definitions 5 
 
INDEX 197 
 
 PAQE 
 
 Deterrents 165 
 
 Dinitro-methane, explosive properties of 112 
 
 Ether as a solvent 55 
 
 Ether-alcohol as a solvent 51, 55 
 
 Experiments, various 144 
 
 Explosives, ammonium compounds as materials for 108 
 
 Benzol derivatives as materials for 121 
 
 Hydrocarbons as materials for 109 
 
 Tables showing composition and ballistic properties of. 184, 186 
 
 Friable cottons 7, 16 
 
 Gases, calculating volume evolved 99 
 
 Evolved by ballistite 117 
 
 Evolved by black powder 100 
 
 Evolved by brown powder 101 
 
 Evolved by nitro-cellulose 102 
 
 Evolved by nitroform 113 
 
 Evolved by nitro-glycerin 101, 124 
 
 Evolved by nitrohydric acid 109 
 
 Evolved by pentanitro-cellulose 102 
 
 Evolved by pyrocollodion 105 
 
 Evolved by pyroxylin 122 
 
 Evolved by tetranitro-cellulose 103 
 
 Evolved by trinitro-cellulose 102 
 
 Gun-cotton 7, 12, 162 
 
 Discovery of 1 
 
 Table showing effect of pulping on 151 
 
 Table showing effect of washing on 151 
 
 Gun-cottons 16 
 
 Hydration 75 
 
 Hydrocarbons, as materials for explosives 109 
 
 Hydrocellulose 6 
 
 Amorphous 50 
 
 Mercerization ; 47, 71 
 
 Mononitro-methane, explosive properties of 112 
 
 Multi-charge principle 192 
 
 Nitration 5, 71 
 
 Experiments in 134, 140, 141 
 
 Higher limit of degree of 91 
 
 Of cotton 127 
 
 Of cotton, researches upon 81 
 
198 INDEX 
 
 PAGE 
 
 Nitration, table showing results of, for different periods of 
 
 time 137, 142, 145 
 
 Temperature in 35, 36 
 
 Nitro-cellulose 5, 161 
 
 Composition of 94 
 
 Decomposition-products of . s 190 
 
 Formula of 94 
 
 Gases evolved by 102 
 
 Insoluble 5 
 
 Of high nitration 5 
 
 Of low nitration 5 
 
 Of mean nitration 5 
 
 Soluble 6 
 
 Nitro-celluloses, properties of 95 
 
 Solubility of, in the cold 43 
 
 Nitroform, explosive properties of 112 
 
 Gases evolved by 113 
 
 Nitrogen dioxide, table showing yield of, in relation to com- 
 position of acid bath 132 
 
 Volume disengaged per gram of cellulose nitrates 128 
 
 Nitro-glycerin 116 
 
 Decomposition-products of 189 
 
 Gases evolved by 101 
 
 Gas volume developed by % 124 
 
 In smokeless powders 182 
 
 Nitro-hydrocellulose 6 
 
 Insoluble 7 
 
 Of high nitration 6 
 
 Of mean nitration 6 
 
 Of low nitration 6 
 
 Soluble 7 
 
 Solubility at freezing temperature 41, 42 
 
 Solutions 38 
 
 Nitro-hydrocelluloses, constitution of 34 
 
 Nitro-mannite 116, 118 
 
 Nomenclature 4 
 
 Oxygen-carriers 32 
 
 Pentanitro-cellulose, gases evolved by 102 
 
 Polymerization 71 
 
 Powder, black, gases evolved by 100 
 
 Cocoa (brown) , gases evolved by 101 
 
INDEX 199 
 
 PAGE 
 
 Powder, pyroxylin 125 
 
 Smokeless, development of 156 
 
 Smokeless, origin of 1 
 
 Smokeless, ammonium nitrate in 122 
 
 Smokeless, cellulose as a base for 125 
 
 Smokeless, nitre-glycerin in 182 
 
 Smokeless, pyrocellulose as a base for 175 
 
 Smokeless, pyrocollodion as a base for 97, 125 
 
 Powders, colloid 28 
 
 Composite 179 
 
 Table showing qualities of . 168 
 
 Table showing quantity of nitro-carriers in certain 183 
 
 Propulsive effect, attainment of maximum 169 
 
 Pyrocellulose 7 
 
 Adaptation for smokeless powders 175 
 
 Solubility in the cold '. 43 
 
 Pyrocollodion 97 
 
 As a, base for smokeless powders 125 
 
 Gases evolved by 105 * 
 
 Smokeless powders from 97 
 
 Pyroxylin 7 
 
 Gas volume developed by 122 
 
 Powder 125 
 
 Pyroxylin and nitro-naphthalin 122 
 
 Pyroxylin and picric acid 122 
 
 Smokeless powder, development of 156 
 
 Origin of 1 
 
 Smokeless powders, ammonium nitrate in 122 
 
 Cellulose as a base for 125 
 
 Nitro-glycerin in 182 - 
 
 Pyrocellulose as a base for 175 
 
 Pyrocollodion as a base for 97, 125 
 
 Solubility, temperature in relation to 41, 44 
 
 Solution, action of cold in accelerating 51 
 
 Table showing action of nitric acid mixtures, and results 83 
 
 Showing action of sulpho-nitric acid and mixtures, and 
 
 results 87 
 
 Showing ballistic efficiencies of pure colloids and com- 
 pound colloids containing accelerators 167 
 
 Showing composition and ballistic properties of explo- 
 sives 184, 186 
 
200 INDEX 
 
 PAGE 
 
 Table showing composition of acid bath in relation to yield 
 
 of nitrogen dioxide 132 
 
 Showing effect of temperature during dipping and reac- 
 tion 140 
 
 Showing effect of pulping on gun-cotton 151 
 
 Showing effect of washing on gun-cotton 151 
 
 Showing effect of washing and pulping on the viscosity of 
 
 collodions .. 151 
 
 Showing qualities of powders 168 
 
 Showing quantity of nitro-carrier in certain powders 183 
 
 Showing results of nitration for different periods of time. 
 
 137, 142, 145 
 
 Temperature, in nitration. . . ... 35, 36 
 
 In relation to solubility 41, 44 
 
 Table showing effect of, during dipping and reaction 149 
 
 Tetranitro-cellulose, gases evolved by 103 
 
 Tetranitro-methane, explosive properties of 113 
 
 Theory of cellulose molecule . 37 
 
 Trinitro-methane, explosive properties of 112 
 
 Viscosity 27 
 
 Of collodions, table showing effect of washing and pulping 
 on . . 151 
 
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 SO CENTS ON THE FOURTH 
 DAY AND TO $t.OO ON THE SEVENTH DAY 
 OVERDUE. 
 
 JUN 28. 
 
 CO 
 
 IAN -6 1959 
 
 LD 21-100m-7 ! '39(402s) 
 
YB 53795 
 
 3G1224 
 
 UNIVERSITY OF CALIFORNIA LIBRARY