WAR DEPARTMENT CHEMICAL WARFARE SERVICE, N. A, RESEARCH DIVISION, C. W. S. AMERICAN UNIVERSITY EXPERIMENT STATION WASHINGTON, D. C. September 10, 1918. MAJOR GENERAL W, L. SIBERT, DIRECTOR COLONEL G.A.BURRELL,CHIEF OP RESEARCH DIVISION MONOGRAPH NO. XV LACHRYMATORS PART II ACROLEIN. CLARENCE J. WEST FOREWORD Because of the large number of substances which have been studied for their possible use as lachrymatory, it has been thought best to issue the monograph on this subject in parts. Part I will consist of a general discussion of the question and of the methods used in their study. The following parts will discuss the various compounds in detail. These will be issued at irregular intervals as pressure of other work permits, A collective index will be issued at the end of the series. Clarence J. West. TABLE OF CONTENTS ACROLEIN Introduction 1 Preparation 2 Glycerol 2 Trimethylene glycol 25 Ethylene carbon monoxide 27 Acetylene, carbon monoxide and hy- drogen 27 Acetylene and methyl alcohol 28 Acetylene and Formaldehyde 28 Properties 30 Polymerization 30 Disacryl 30 Gum 34 Stabilization 37 Analysis • 41 Testing efficiency of canisters 46 Absorption 49 Protection 49 Physiology 50 Sensitiveness of individuals 52 Pathology 53 Treatment 54 Tactical Use 54 Shell filling 55 Bibliography 56 LACHRYMATORS PART II ACROLEIN, CHg = CHCHO INTRODUCTION British investigators report that crude acrolein (containing 25$ ether) was quite ineffective as a lachry- raator. The pure acrolein attacks the eyes and throat, but the vapors are too light. The addition of CC14 or CgH4Cl4 gave no better results. It has been suggested that the acrolein should be generated in situ in the shells. This is not feasible, because acrolein can be prepared only from glycerol, of which there is a shortage at present. The French have found that while acrolein is effective in clearing dugouts it is not as effective as stannic chloric*- which is used for the same purpose. The Offensive Warfare Committee of the Division of Offensive Chemical Research, Feb, 21, 1918, considered the work on acrolein and decided that the preparation of this compound had been so thoroughly investigated that no further 1 laboratory work on this problem was necessary. Whether this substance should be manufactured or not would seem to depend on the availability of the raw material -either glycerol or trimethylene glycol (B.M. XIII-47). I PREPARATION Owing to the pressure of other work no attempt will be made here to collect all the data on the preparation of acrolein. Only that which has been carried out by workers connected with the Research Station will be in- cluded. However, a complete bibliography ( compiled by L. H. Flett, Mass. Inst. Tech.) is given at the close of the chanter for those who desire further information. The preparation of acrolein is generally carried out by the dehydration of glycerol. The problems are the choice of a suitable catalyzer and proper temperature con- ditions . Magnesium Sulfate as Catalyzer. The moat suitable method for the preparation of acrolein seems to be that of Wohl and Mylo (Ber. 45, 2046 (1912). It consists in passing glycerol vapors through an electrically heated column packed with a catalyzing de- hydrating material, carefully maintained at a definite temperature, and removal of water and other by-products 2 Figure I. PL/iTE I DIAGRAM- of-the. APPARATUS p F\'no.\ F««<.\v«r HooJ ■S 61 ye.«T»«e F u- n T.C- TWc T m» ~ Cowpk A Copper Condenser SCALE 1 1 in, c 1 ft. f\ Retort 7&Oe.e. g, F»V it Receiver C Second *• T Cotol7*.;w£ To.be by a well controlled fractional distillation and condensatior Wohl and Mylo claim to have obtained 60$ of the theoretical yield of puie acrolein by operating the apparatus for not more than 5 hours with magnesium as the catalyst at the optimum temperature of 325-340°C. This method has teen studied by Mill liken (M-333.)* The apparatus used is shown in Figure I. "R is a copper retort of 750 cc. capacity. Into the side arm of this retort is fastened a glass tube*. This joint is best made with a glycerol-litharge cement; a cork may be used, however, and gives perfect satisfaction. Into this glass tube glycerol runs frcm a cropping funnel S of 1000 cc* capacity, which is calibrated so that the glycerol used may be read to approximately twenty-five grams. The return from the condenser A also runs into this tube. It is found best to connect the dropping funnel through an adapter and tube to make the glycerol "head" enough so that the rate varies little as the separatory funnel empties, and the adapter makes it possible to see and regulate the flow. The catalyzer (Column T) is 37 in. long and 1 5/8 in. inside diamster, with flanges at top and bottom 3/4 in. wide. The retort was made from 16 gauge copper, and the tube from 20 gauge copper. It would be a decided advantage in making up the flanges to have them of much heavier metal. The lower joint was made up with packing and six bolts. The upper joint was first made in the same way but later small steel clamps were used. For packing Johns-Mansville1s 3 "Kearsage" was used* and also graphitized packing. Water glass aids in making the joints tight. The catalyzer tube is wound first with thin asbestos paper and then with 25 feet of nichrome ribbon 0.1C0TT x .012", of resistance .42 Ohms per foot. This is held in V ' place by alundum cement, which was mixed with water and brushed on, and the whole covered with magnesia steam pipe covering. Direct current is used for heating, about 4 amperes and 47 volts giving the desired temperature. The temperature of the tube was first measured by a copoer-nIdeal" thermc couple and "potentiometer." The junstions were silver soldered. The hot junction was placed in a glass tube and this was placed in an iron tube extend- ing to near the bottom of the column. It was kept in the * hottest part of the tube near its middle which we found to be fairly uniform in temperature. Later the temperature was taken by hanging a thin thermometer in the iron tube and this gave results nearly as satisfactory. A is a copper condenser connected by copoer tubes to the catalyzer column. It is designed to return unchanged glycerol. The jacket was filled with boiling toluol which in turn was retained by a reflux condenser, R. The crude acrolein vapours are condensed in an ordinary glass condenser and run into flask B of 2000 cc. capacity. From this crude acrolein is obtained by use of a modified Hahn (Ber, 43, 422) column. It is condensed and passes into a 500 cc. side arm flask C, from which it is fractione 4 by means of a Le Bel-Henniger column. The vapors are con- densed by an iced worm in the receiver D which is also packed in ice, A Woulff bottle or a filtering bottle serve well as receivers. Any escaping gases are conducted to the hood by means of glass and rubber tubing. The catalyzer is supported in the tube by means of a disk of copper gauze of about 4 mesh which is just too large for the entrance to the retort. The retort is heated by means of a Tirrell burner with the flame about six inches high and the inner cone visible. The heating baths are filled with cottonseed oil and are heated by gas or electricity. B is kept at 120°C. C is kept 90° - 110°C* Column B is about 90°C. P is about 60°C." Results obtained by this method are given in the following tables 5 Comparison of Buns made with 99$ Glycerol and Magnesium Sulphate Catalyzer No. : of : Run : Time : in : Hrs. : Grams 99$ Glyc. : Rate •.Glycerol: : Grams : : p, Hr. : Yield %* Theory :Aver- S age :Temp. : Tube Max Temp. Tube 2 3.5 900 257 35 302 - -18.7 332° 3 10 2500 260 39 325 - -13.0 350° 4 15 3565 246 37 318 - - 5,8 331° 5 13 2850 219 35 320 - - 6.5 328° 6 20 4120 206 50 328 - - 4.9 334° B 16 2790 175 36 322 - - 6.5 335° 9** 15 2950 200 51 346 - - 5.1 355° 10 9 3500 289 40 332 - - 8.0 346° 12 14 2400 172 43 344 - - 5.0 350° 13 15 3000 200 46 335 - - 0.0 340° * These yields are uncorrected for the glycerol which dis- tilled over into flask B unchanged. This liquid was found to contain about 25$ glycerine. Correcting the above yields for this, our maximum yield is a little over 61$. gravity #9; first 910 grams. = 0.654 at 15*5°; last 450 g.f 0.857 at 18°. Catalyzer- The crystallized magnesium sulfate was de- hydrated by heating in an iron pan* A vigorous evolution of steam left a firm porous cake, which was broken up into bean-sized fragments before packing in the tube. After using, the catalyzer had a reddish brown color. The lump did not seem to disintegrate* Fracture showed that the pieces were of almost uniform color throughout except for thin incrustations of black tar in some parts of the column. The yield of acrolein decreases after the catalyst has been used for a few hours. The falling off in efficiency after 5 ot 10 hours was marked* There is no doubt, however, that a properly jacketed and carefully operated apparatus 6 may be run continuously with advantage * for many hours. Temperature.- The best yields of acrolein were obtained at temperatures between 328-346°. and with an average flow of glycerol of about 200 gm. per hour. The temperature has to be controlled accurately. Wohl and %lo state that at this temperature more than half of the glycerol is trans- formed into acrolein, while slightly higher temperatures favor the production or hydroxy- acetone, according as the primary or the secondary hydroxyl group is attacked. The hydroxyacetone at once decomposes intp acetaldehyde and formaldehyde. The rate of feeding must also be controlled, Kohler used a simpler apparatus. No yields are given. The apparatus consists of a copper tower about three feet high and four inches internal diameter, filled with lumps of anhydrous magnesium sulfate. This tower was con- nected in a vertical position to the mouth of a copper flask (about one liter capacity). The other end of the tower was connected to a glass condenser. Glycerol was bailed in the flask; the vapors passed up through the magnesium sulf- ate, where they were largely changed into acrolein. The gases issuing from the mouth of the cylinder were condensed fractionally and the final condensation fraction was purified by distillation. This product contains from 10-20$ acetal- dehyde, which cannot be conveniently removed. The acrolein obtained by this process (Mulliken) was always a yellowish product containing water and acetaldehyde. 7 It was,however, free from sulfur dioxide. Magnesium pyrophosphate as Catalyzer* The apparatus employed was that described above. The magnesium pyrophosphate was made by strongly igniting magnesium ammonium orthophosphate in a muffle furnace. The powdered pyrophosphate was moistened with about 2fo of syrupy phosphoric acid diluted with water, and the slightly damp mass compressed to cupels in a cupel mould of the type used in furnace essay work. The cupels were then broken into bean-sized pieces, and dried at 10Q°C, * The pieces were then hard stony lumps. The following table shows the results obtained from the use of magnesium pyrophosphate: Runs with MggP207 Run No. :length:Temp.: :of run:Tube : Grs.Gly- cerine used Rate Gly- cerine (g.per hr, :Yield of Acrolein :Uncor.(B.P,51-53° 18 2* 326°C 1050 420 47$ 18a 18 315 5275 292 29$ 29 2 320 1010 505 52$ 2929a 5 325 1610 302 49$ The acrolein prepared by this method was remarkable for its purity, freedom from color, and comparatively excellent keeping qualities. When slowly distilled through a Le-Bel-Henninger column nothing boiling below 51° was condensed, and acetaldehyde was apparently wholly absent* If a correction were to be made for the unchanged glyercol 8 recoverable from flask B, the yields of pure acrolein would become fully 60$ of the theoretical. One of the most striking facts connected with the use of the pyrophosphate is the high acceleration which it induces in the dehydrating reaction. The reaction if con- ducted at a temperature of 325°C. will transform about twice as much glycerol into acrolein in a given time as with magnesium sulphate, and the condensate in the recipient D remains practically colorless to the last. The most seriops drawback to the use of pyrophosphate is its short life as a catalyst. The blacknasphaltic tars forms much more rapidly than with other catalysts, and in the small apoaratus used chokes the upper layers of the column and the exit tube after a few hours. It is as tenacious as hard asphalt and difficult to remove after hardening. The product from Hun 29, which was stored in a closed closet in a bottle containing a small reduced copper gauze spiral, remained clear, colorless, and mobile two and a half months after preparation, although it had been opened and shaken many times in the interim. At the end of three months, (when last observed) the separation of a white precipitate was just beginning. Preparation of Acrolein with Bone-ash Catalysts, It was thought possible that bone-ash cupels made by- moistening the ash with enough dilute sulphuric or phos - phoric acid to give the mixture a slight acidity by 9 formation of acid calcium phosphate might give results similar to those obtained with the acidified magnesium pyrophosphate, and prove an economical substitute for this material* Bone ash without acidification is very inactive. In these experiments the cupels were mixed, compressed, and dried in the same general way as the magnesium pyro- phosphate mass. The glycerol used (a 98$ dynamite glycerin), and the apparatus, were also the same in both series of experiments. The following table shows the data secured in the bone- ash runs. In Runs #19 and #20 the cupels were moistened with phosphoric instead of sulphuric acid. The products showed less tendency to polymerize than those from experiments in which sulphuric acid was used, but an unusually heavy deposit of tenacious black tar accumulated and finally choked up the central part of the catalyzer column. The apparatus functioned most smoothly in Runs Nos. 22 and 23 in which the bone ash was moistened with 15$ of sul' - ifuric acid. The clogging and tar formation was perhaps a little more troublesome than in runs with magnesium sulphate, but the conversion was more rapid and the product was much lower in acetaldehyde. Not more than about 100 drops of distillate boiling below 51° were collected from any one of the crude products when slowly distilled through a 6-bulb LeBel-Henninger column. Increase of the proportion of sulphuric acid to 20$ increases the tar, difficulty in 10 Liun Catalyst Length of run. Temp.of Catalyst Glycer- ine used (in grs) Yield in Acrolein * Bp* Gr. of Pro- duct at 15 ♦ Boil- ing Point of Pro- duct Remarks No. in grams in°/o of theory when corrected 19 Bone ash +115to3P04 10£ hrs. 203*317° f _ 1700 333 32.5 — 0.850 51-54° Products of good stabil- ity. Much tar on ca- talyst. 20 Bone ash + 5/0H3PO4 i 380-325° 1190 225 32.5 40*5 0.847 51-54° 23 Bone ash + lSyoHgSO^ 8 hrs. 315-339° 1785 531 50.8 62.5 0.848 52-53° Products of moderate stability. Considera- ble tar on catalyst 22 Bone ash f 9i hrs. 320-330 1600 477 \ 51.0 y 60 0.846 51-53° 24 Bone ash • h20%H2S04 7jt hrs. 320-330° 1100 180 28.1 39.6 0.850 51.53° Products very insta- ble. Much tar on cata- lyst. Run 24 very irregu- lar. 25 Bone ash \ 2O/0H2BO4 2£ hrs. 320-530° 890 250 48.3 0,852 51-53° 26 Bone ash 2S04 9i hrs. 315-3250 1940 418 37.0 51.2 0.852 51-53° Products of Moderate stability. Little tar in catalyst 27 4— Bone ash -HO/0H23O4 9 hrs. 325-340° 2020 495 42.6 68.2 0.852 50-54° *Yield corrected for recoverable glycerine accumulating in flask B Results with Bone-Ash Catalysts 11 regular operation and leads to more unstable products. Thus the product of Hun 24 polymerized to a solid glassy mass (although in contact with oopper) within a month, while its duplicate No,25 gave a product that began to polymerize within a few hours, (See Page Aluminium Salta aa Catalyzer* Experiments with aluminium phosphate and aluminium sulfate as oatalyats were very unpromising (Mulliken). Engelder (B.M, IX-21, XII-32) found that aluminium oxide gave as good results as magnesium sulfate, the yields of acrolein running from 19 to 36$, AlgOg yields a smaller amount of by-product (acetaldehyde and formaldehyde) than the other catalysts and the undecomposed glycerol can be recovered• 12 Evans has made (B.M.III-20) a comparative study of various catalysts in the preparation of acrolein with the following resulti CATALYST GLYCEHGL ACHOLEIN Gm. Gm# Crude Pure KHSO4 — KgS04 500 gm,-150 220 50$ 36.5$ KHS04 K2S04 1000 gm.-250 500 73.3 36.7 KgSgO? — K2SO4 1000 gm.-300 440 73.3 52.5 MgS04 160 755 40 #0 34.2 KPO3 - HPO3 118 gm.- 80 150 45 Nag?£07 — H3P04 113 gm„-115 150 35.5 AIPO4 - 122 200 18 MggP207 - 160 100 12 Evans believes that a considerable percentage of the product obtained by means of the more familiar methods is acetaldehyde and other impurities. BY-PRODUCTS IN PREPARATION OF ACROLEIN KITH THE WOHL MYLO APPARATUS Although porous carbon may form in large quantities in the copper flask if the glycerol level be allowed to fall too low so that the metal of the side walls becomes over- heated, the loss from carbonization in most experiments was 13 very small* The loss from the formation of black asphaltic deposits on the catalyst mass was a factor of more impor- tance when acid phosphates were employed, but was not con- siderable in the experiments with the magnesium sulphate or the faintly acidified bone-ash catalysts* In experiments with magnesium sulphate, even at the lower temperatures, the loss through format.lon of free acetic aldehyde was found greater than Wohl's paper had led us to anticipate, and it is probable that the contents of Flask B may have contained appreciable quantities of condensation products of acetic and formic aldehyde. When using the acid phos - phates, on the contrary, the condensate in recipient D was free from acetaldehyde* The small condensate remaining in Flask C at the close of runs rarely exceeded 50 cc. and con- sisted of a brownish pungent oily mixture and water. It is now probable that it consisted in part of some of the compounds isolated by Prof. Moureu from the French "Papite" and described in Z-124, Small quantities of oils of similar appearance sometimes separated from the aqueous solutions in Flask B, or were obtained as small layers separating from aqueous condensates when these solutions were distilled long enough to drive over most of their water. The greater part of the glycerol not converted into acrolein by passage through the Wohl-Mylo apparatus, that could be definitely accounted for, was found in the contents of Flask B. This dark-colored solution presumably contained some of the aldehydic condensation products mentioned in the "Notes on the Mechanism of the Reaction for the Preparation of Acrolein from Glycerol" hy L* H* Flett which is given below, along with traces of the aromatic oils mentioned in the last paragraph. But it behaved on distillation almost exactly like an aqueous crude glycerol solution of the same specific gravity, as was well established by a comparison of its distillation curves, showing boiling point temperatures plotted against volumes, with similar curves obtained from distillation experiments made in the same flask in which glycerol solutions of the same gravities were distilled to dryness at the same rate. This conclusion was verified by concentrating the residues in Flask B to a gravity corresponding to an 80$ glycerol and using the concentrates for regular runs for acrolein in the Wohl-Iitylo apparatus. The acrolein yields were only a little lower than those in experiments with 80$ glycerol. The following table shows the yields of acrolein aotually obtained in certain runs, and the yields that would have been obtained if the glycerol in Flask B (calculated from specific gravity) had been all recovered as such, Obviously the composition of the solution in Flask B would not be exactly the same in different experi- ments unless the conditions of operation were identical, and any change in the design of the fractional condenser A would be likely to have an influence upon it. 15 TABLE SHOWING YIELDS OF ACROLEIN FROM CERTAIN RUNS BEFORE AND AFTER CORRECTION FOR CLYCEROL COLLECTED IN FLASK B. Run • NO. *; Amount of gly- cerol used in grams Ac colein Y ield Liquid in FJ ask Grams Actual; Yield ; in %of: theory. Yield : after • correc- ; ticn for: Flask B * ♦ Grams Speo.; Grav.: % Gly- cerol from Sp•Ur. 20 1200 225 32.5 40.5 710 1.083 33 22 1400 477 51*0 60* 745 1,066 26 23 1790 531 50.8 62,5 995 1.084 33 24 1100 180 28.1 39.6 772 1.106 42 25 890 250 48.3 553 1.091 36 26 1950 418 37.0 51.2 1220 1,110 43 27 2020 495 42.6 58.2 1200 1.120 44 29 1600 333 49.0 650 1«173 45.5 30 1600 305 32.6 1074 1.102 40 THE MECHANISM OF THE REACTION FOR THE PREPARATION OF ACROLEIN FROM GLYCEROL The following discussion was written by L. H. Plett. Acrolein. It was observed by HtLbner and Geuther (Ann. 114, 35) that, in the distillation of glycerol with potassiurr acid sulfate only water was given off at first. When this stopped an increase of temperature was necessary for the evolution of acrolein. They concluded that a glycerol potassium sulfate was formed, with an evolution of water, which broke up at a higher temperature with the evolution of acrolein. Senderens (Bull. Soc. Chem, 9, 374) gives the 16 reaction-- CHgOH K0V CH20H CHg 1 'SO 2 ' CH OH HO' CH-SO4K -I- (2H2 CH -i- 2KHS0, 1 . CH2OH HO CH-SO4K CHO ,S02 Acrolein The sulfate is then available for further action. Nef (Ann. 335, 207) after some careful research gives the reaction as follows-- ch2oh ch2oh ch2oh ch2 11 I « CHOH —» CH . —» CH H CH 1 II II CHgOH CHOH 0= C-H CHO dihydrory- hydraorylic Acrolein proplidene aldehyde Wohl and Mylo (Ber. 45, £048) suggest Hef's reaction but say "Ftir das von ihm formulierte auftreten von Methyl ene Verbindungen (mit zweiwertigen Kohenstaff) als Zwischenprodukten fehlte aber eine experimentelle Begrtln dung wahrend die Bildung von Ketonen and Aldehyden als interraedi&r enstanden Enolen durch zahlreiche Erfahrung gestiLtzt wird. ” As to the part the catalyzer plays, Wohl and Mylo say, "Ob hier ein Additionproduct der Zusammensetzung (I) 17 /Ov * 0 Mg' (HO-Mg-O-S-O v0' S glycerin (I) (II) Oder das normale basische Sal2 einer Glycerinschwefelsaure (II) vorliegt, is nicht ohne weiteres zu entscheiden. Glycerol forms two kinds of compounds with inorganic compounds, the glycerates, such as C3H3(0H)02Ca and the glycerolates, of the general formula: (Me3C3H50H3)X2 or [Me J 0H-CH2-CHS0hI3 X2) Grhn and Bockerisch (Ber. 41, 3465) and Grttn and Husraann (Ber, 43, 1291) have done some careful work on the glycerolates; they have prepared and analysed several of them. These glycerolates are crystalline compounds and intensely hydroscopic. They are very stable, as would be expected from the five-membered cyclic formation* By heat- ing, glycerol is evolved unchanged but by heating with a dehydrating agent acrolein is formed0 The addition of glycerol appears to be independent of the ability of the salt to add water, Griln and Husmann (Ber* 43, 1293) noticed that in work ing with lime acrolein was formed* They subsequently dis- tilled glycerol from calcium hydroxide and obtained only 1,7$ acrolein. This would show that there is little tendency for the formation of acrolein as a result of the 18 formation of a tri-glycerolate. Grtin and Bockerisch {Ber. 41, 3465) have prepared the compounds-- Me (OH-CH-CHgOH j3 S04)h20 Me = Co, Ni, Zn, Cu. In the normal sulfates of these metals the last water molecule is the hardest to separate so it appears in the glycerolate, Cations of the form (Me3(OH))g)** are also Relieved to form in some cases* There is no information as to the exact part of the catalyzer. If a glycerolate is formed at a high temperature there is still a question as to the formation of acrolein. Certainly a dehydrator of some kind is necessary as can be seen in the low yields given by substances which do not form hydrates. The information suggests the advisability of using a catalyzer, i.e., a glycerolate-forraing salt and a dehydrator such as a bisulfate or pyrosulfate* Almost any treatment of glycerol yields acrolein in variable quantities so the action of catalyzers yielding small results will not be considered, although some, in particular the acids, certainly must form intermediate com- pounds. Intermediate and Decomposition Products. Acetol.- Acetylcarbinal-CH3-C0-CHp,0H-(Bellstein I, 267) « Acotol is the most important decomposition product ol glycerc* outside of hydracrylic aldehyde and acrolein. It breads up ii. to formaldehyde and acetaldehyde* Nef (Ann.335,2£1) Fives the 19 reaction- CHgOH CHgOH CHgOH CHOH * COH i CO } CHHO -f CH3CHO I II i CHgOH CHg CH3 dibydrosy- acetol proplidene The above remarks about methylene dissociation should still be borne in mind a Acetol is highly undesirable* Its dissociation repre- sents glycerol losses* It condenses easily and is believed to be responsible for the tarring of the catalyzer, It boils at 1470 not undecomposed and so would certainly not pass beyond the first flask, Figure 1. It is undoubtedly all held in the first column as tar. Acetol combines at 100° with two mols. of phenyl- hydrazine to form methylglyoxal osazone. There is no reason to assert that the acetaldehyde in the product is a fair measure of the acetol formed, but this much can be said, that if it is, then the amount of acetol varies with the catalyzer* Wohl and Mylo (Ber, 45, 2046) assert that it varies with the temperature, the formation of acrolein being favored by low temperatures* Formaldehyde.- Formaldehyde comes from acetol and probably from hydracrylic aldehyde* What does not poly- merize to form formaldehyde-glycerol passes out the con- denser, carrying with it acrolein vapors. Acetaldehyde.- The source is the same as formaldehyde It forms acetaldehyde-glycerol. It is the chief impurity of the acrolein. 20 Hydracrylio aldehyde (C^HgOg). B-Hydroxyoropionic a]dohyde.- The reaction for the dissociation of hydraoryDic aldehyde is reversible. If water and acrolein are kept at 100° for some time they combine forming hydracrylic aldehyde. ch2 -f h20 I) CH I CHO Acrolein ch2oh f CH2 1 0.-C-H It was first prepared and analysed by Kef (Ann, 355, 219). It is a liquid of B.P. 90° at 18 mm* It has a sharp smell. It condenses easily and polymerises some. Its semicarbazone melts at 140° Nef (Ann. 335, 220), It reduces aramoniacal silver nitrate, but not Fehlings's solution. It is easily soluble in water but difficultly soluble in ether, A.Wohl prepared it from oxyacetal (Eer, 41, 3603) and describes it as soluble in ether and difficultly soluble in water. Hydracrylio aldehyde probably exists in the catalyzer tube and in the first and 3eoond flasks. It can be decomposed by heating it from 140° - 200° with potassium bisulfate (Nef, Ann. 335, 221), Formaldehyde-glycerine monoformal methyleneether. (Beilstein I. 468Ibchulz and Tollens (Ber, 27, 1394) obtain ed this product by heating together formaldehyde and glycerol. It has two isomeric modifications. One gives a crystalline and one an oily benzoate. 21 CHoOH HoCOH ) 2 / 20v CHO H -I- 0 HHC —, HCO or OH-CH ' C% | ( CKp ' CK20' CH2OH H2CC|/ ~ It i3 easily formed at 100°. Its toiling point is 193° (197°); the specific gravity is 1,205 at 20°. Continued heating with water doe3 not hydrolyse it* It is soluble in water and can be separated, by adding potassium carbonate, with ether. This should occur in considerable amounts in the first flask only* Glycerol formal benzoate is formed by adding sodium hydroxide and benzoyl chloride, M.P. 72°. Acetaldehyde-glycerol (CrH^qO3); Ethyl!dine; tk (3 y Tri- oxypropane; Acetoglyceral * (Bellstein 11924). - Tliis is formed by letting acetaldehyde react with glycerol. It is a liquid of boiling point 85°-87° at 18 mm, 184-188° at 760 mm, and specific gravity 10118. It is easily soluble in cold water and is decomposed by heating to 100° with ten parts of water. It does not reduce Pehllngs solution. For all that it may be easily decomposed by boiling with water, Nef found it as the chief member of this series of glycerol compounds* (Ann. 335, 214). It should be found in the first flask in appreciable quantities. 22 tf(-Acrolein-glyoerol (CgH^QO^); Ethylidinether; 4lAA Trloxypropan. CgH803 CHp=GH-(.‘E0 CHP = CH-CH °°E2 Glycerine * \ 0CH-CH20H Acrolein-glycerOl This product has not been analyzed. Nef,(Ann. 335, 216) prepared a very impure product which boiled at 102°-106° at 17 mm* The reversibility of the reaction and the re- action between acrolein and water made a pure product im- possible. The compound is hydrolysed by heating six hours with ten volumes of water at 100°. In spite of its tendency to hydrolyze, Nef (Ann. 335, 210) found appreci- able amounts of it in the decomposition of glycerol. It should be found up to the first flask only. » In Kef*s experiment the glycerol compounds were almost equal in weight to the easily volatile compounds (acetal- dehyde and acrolein) and were much greater than the acetol fraction. So they represent a considerable loss of glycerol and acrolein. A Anrole.tn-Glyoerol CfiH-ioOs; -ether (Beilsteln I. 314).- , CHp-O-CHp CHnOCHp CIJ,CO-CHPv 2 Is Is I 2 3 2vO ? s, .°-CH or CH OCH CH^CO-CKp-^ CHX 1 | | S ''O-CHp CH OCH„ 2 2 23 This is prepared by distilling glycerol with CaClg or better It is a liquid of boiling point 171°-172°; miscible in all proportions with water, specific gravity 1.145 at 0°, 1.1C at 15° and 1,097 at 18°. It combines with water at 100° to form acrolein. Nef attributes its formation to the dissociation of the chloride, giving hydrochloric acid. and $ chlor- hydrins are formed. These decompose at 440° into the isoraeres and /$ ~ dioxypropylidenes which then form acetol and hydracrylic aldehyde as glycerol does on dehydra- tion. Acrolein is then formed which unites to form acrolein-glycerol. In the presence of hydrochloric acid intramolecular addition takes place forming acrolein- glycerol . O-CHp ,0-CHo 0H2 = | —5 ch2-ch2-ch«o-ch2 o-ch-ch2-o-h 0 ch2 This may only be expected in the presence of chloride. since The presence of chlorides should be avoided/besides their formation of acrolein glycerol, they aid the formation of acetol and formaldehyde-glycerol* Crotonaldehyde.- This was found in traces by Nef (Ann. 335, 222). It comes no doubt from condensation of the acetaldehyde. Gas.- Nef (Ann. 335, 210) obtained 1.9 liters of gas in the distillation of 250 g. of glycerol; its analysis by absorption gave CO 80.3 cc., H2 12,4 cc, When acrolein 24 polymerises in ultraviolet light, it gives a gas analysing CO = 80%, Ho - ?‘%, CO2 r 5~/&, and CgHg and hydrocarbons = 10$ (Berthelot and Gadeohan (Corapt. rend. 151, 1349). Lockeman u, Liesche (J. Prakt„ Chem, 71, 493) found 02 = 1,9%, Ho «, 17.1$, CO = 63.035, CH4 s 18$, and traces of ethylene and propylene. (CHgO p* CO -J- Hg). Acrylic Acid.- Oxidation cf aorolein in the air gives acrylic acid. Propylene»- This was observed by Westphal (Ber. 18, 2931) in the use of zinc dust as a catalyzer but was not verifiedjby Nef (Ann. 335, 219) after a careful research. Phenol♦ - Phenol is found in the higher boiling residue from which it may be freed by adding alkali. It is believed to come from glycerol ether, (Linnemann and Zotta (Ann. Suppl. 8, 254) = C5H5O -j- 2 HgO. Allyl Alcohol.- This was found in the distillation from zinc dust by Westphal (Ber, 18, 2931) and in the distillation from calcium chloride by Linnemann and Zotta (Ann. Suppl. 8, 254). Acetone- Propionaldehyde.- These were found in the distillation from calcium chloride by Linnemann and Zotta (Ann. Suppl, 8, 254). TRI METHYLENE GLYCOL The preparation of acrolein from trimethylene glycol was suggested by the facts that: (1) ThOg and otho: oxides split* off HgO from ethyl alcohol yielding ethylene 25 and water: H3CCH20E —* ch2 a ch2 -i- h2o (2) Reduced metals, such as Hi and Cu will, on the other hand, split off hydrogen and give acataldehyde-- CHgCHgOH ? H2 -I- CH3CH0 (3) Certain metallic oxides exert a mixed catalytic function on alcohols, the reaction velocity of the two reactions depending on the catalyst used, ZrOg and TiOg catalyze the two reactions to about the same extent (J. Phys. Chem. 1917, 21, 676). This suggested trying to split off both water and hydrogen from trimethylene glycol. It was found that trimethylene glycol is catalytically decomposed by a mixture of and copper at a temperature of 300-400° yielding acrolein in fair quantities, consider- ing that the recovery of undecomposed glycol is good. Engelder (B.M. IX-21) believes that a 50Jo yield might be obtained. Copper alone also will decompose the glycol, giving acrolein, but AlgOg does not.MgSC4 decomposes tri- methylene glycol but no acrolein is formed. The mechanism of this reaction may be assumed to take place as follows: h2c oh • . t : . T from the bottom of which the last traces of water are also 11 r ' * i piped baok to the distilling towerras they separate. It I is now necessary to neutralize the last traces of acidic impurities in the acrolein to effectively prevent resinifi- cation. But just as acidic acrolein readily polymerises • * 11 i so does also alkaline acrolein. It is therefore necessary 29 to use a weak alkali such as sodium bicarbonate. This acrolein may be shipped in cast iron drums. PROPERTIES. A liquid with pungent smell. B.P* 52.4S. Oxidizes easily in the air to acrylic acid. It is soluble in 2-3 parts of water, HKO3 (1.2) oxidizes it to a mixture of glycollic and oxalic acids; chromic acid to COg and HCOOH ( Claus , Ann. Suppl. 2, 118). Zinc and HC1 reduces it to allyl alcohol and acropinacone (c6Hi0o2). The specific gravity is 0,84 (HgO = 1). The polymerization of acrolein is described below: POLYMERIZATION POLYMERIZATION INTO DISCARYL When acrolein is purified carefully, a perfectly clear liquid is obtained. But at the end of a variable and relatively short time, turbidity is seen to apDear, then a white precipitate* Redtenbacher (Ann. 47, 141(1883), the first author to observe it, gave to this precipitate the name of "disacryl", Gradually the precipitate becomes more abundant and more compact and encroaches upon the whole mass of liquid. The phenomenon continues in this way until the acrolein has completely disappeared. It is a curiops fact, that, although the formation of the discaryl takes place with contraction, all the glass receptacles, in which the 30 observations are made, are found to be broken sooner or later, whatever be their thickness, as if the phenomenon involved an expansion or swelling. It is observed, in addition, that the pieces of the broken vessel continue to break into still smaller fragments as long as the disacryl which adheres to them is impregnated with acrolein. This shows that the formation of disacryl develops consider- able local pressures on the walls of the vessels# Properties of Disacryl,- At the beginning of the polymer- ization this product appears in the form of light flasks and sometimes in the form of a dense, extremely fine powder, which gathers at the bottom of the vessels, and assumes, under the microscope, the appearance of a mass of small transparent spheres# Toward the end of the polymerization, it consists of light masses, having at times a fibrous texture, soft to the touch and readily pulverized between the fingers. This substance hardens in time and becomes brittle. Some- times it is obtained, also, in lustrous, hard, elastic, concretions, whose appearance recalls that of ivory; this last appearance is always observed in those parts which are in contact with the glass surface, but which possess very slight thickness. Disacryl is an inactive substance, odorless and taste- less, and appears insoluble in all neutral solvents. When first formed it is slightly soluble in concentrated nitric acid; an excess of water added to the liquid, precipitates 31 a white product which we have not yet studied. Disacryl is infusible and does not distil* Chromic acid mixture, fuming nitric acid, and the caustic alkalis attack it only very slowly. The determin- ation of disacryl presents no difficulty; it is collected on the tared filter, washed with alcohol and dried to constant weight, Cause , - The polymerization 4otd discryl seems to be due to a spontaneous modification of acrolein and not to the action of foreign substances or impurities. A sample of acrolein was purified as follows: A volume of about 6 litres was subjected to a series of three purifications in an apparatus provided with a Vigreux column, 50 cm. long. Such purification was preceded by a washing with a solution of sodium bicarbonate, followed by a drying over calcium chloride. The distillation was carried on very slowly and each time the first and last fourth was reheated; only the main portion, therefore, was subjected to further treatment. During each of these dis- tillations the thermometer did not vary by more than a half figure. Finally in this way about half a litre of very pure acrolein was obtained. Such a specimen remained clear during a period not exceeding twelve hours; then it began to become turbid in consequence of the formation of disacryl. 32 The action of the glass on acrolein tends to oppose the formation of disacryl* Acrolein kept in metallic containers changes into disacryl just as it does in the glasso The action of the air upon acrolein clearly pro- motes the formation of disacryl but air is not necessary to cause the phenomenon of polymerization into disacryl* The degree of purity possessed by acrolein is of great importance as far as it concerns the speed of poly- merization into disacryl. When acrolein is very pure, its polymerization is slow. However, it is known, that crude, that is to say, impure acrolein, such as the industry produces under the name of papite, is kept without perceptible change in darkness in metallic containers* In the course of its preparation, a mixture of water vapor is condensed, passing over at a temperature of 75°-80° instead of 52° (boiling point); a certain quantity of water is thus distilled, the excess of which is then separated by decantation, and in short, not only are the products carried by the acrolein vapor collected, but also the products carried by the water vapor. Since acrolein distilled atfter drying is unstable, it must be concluded that the stabilizing products of acrolein are capable of being carried by the water vapor but not by the dry acrolein vapor* We may add, that in order to obtain a stable product, it is not sufficient to distil the acrolein without previous drying; it is necessary to distil simultaneously a quantity of water such that the mixture 33 of the vapors passes over at a temperature of 75°-80°. It can be seen then from the practical point of view, how important it is, in order to obtain a stable acrolein, not to depart from the technical scheme of operations during the initial study of this product. The action of heat is the clearest; it accelerates the formation of disaoryl, The accelerative influence of light upon the formation of disacryl has been seen by the accumulation of the first flakes in the part of the flasks exposed to the strongest light. POL YMERIZATI ON INTO GU21. When acrolein undergoes the change which we designate under the name of "gum polymerization", no change is observed at first in the appearance of the liquid in pro- cess of transformation. This kind of change is easily revealed by measuring the density, which is found to in- crease steadily and by the determination of the properties of the residue which the product leaves behind on evapor- ation. But when the proportion of gum reached 15$, the liquid loses its fluidity and becomes more and more syrupy in the proportion as the transformation progresses* It passes through all the stages of viscosity, ending, finally in a hard, brittle, transparent mass, which keeps for a long time the characteristic odor of acrolein. The two processes of polymerization (disacryl and gum) 34 do no$ appear capable of being produced simultaneously in the same liquid. We have already observed that acrolein is transformed either into disacryl or gum, but the two transformations never seem to occur at the same time* How- ever, the change oan begin in one direction and then pro- ceed in the other. An initial polymerization into disacryl is frequently observed, soon giving place to polymerization into gum in such a way that the mass is at the same time rendered viscous by the gum and opaque by the flakes of disacryl. The substance which we designate under the name of "acrolein gum" is a product very soluble in acrolein, acetone and alcohol, slightly soluble in various organic solvents (benzene, ether), very slightly soluble in water. The gum has the property of holding acrolein strongly and consequently of diminishing the vapor pressure of that liquid. When a specimen of acrolein partially polymerized into gum is distilled, it is found that it is necessary to raise the temperature of the warming bath progressively, in proportion as the gum is concentrated; the viscous residue still retains at 100° a perceptible quantity of acrolein, and it is not eliminated entirely even when the heating takes place under a very slight vacuum. The formation of this material is, therefore, from the point of view of military use, more detrimental than that of disacryl, since the loss of activity by the initial product is., in this ease, greater than that corresponding to the weight of the polymeric 35 substanoe formed . Since the gas is not a definite product, an accurate determination cannot be obtained* Moureu considers as "gum" the residue which a changed acrolein leaves in dis- tillation when the latter is conducted with the precautions we shall indicate. The formation of the gum takes place with contraction. It is always possible by measuring the density to show very directly when the change begins and to follow accurately its whole progress. Polymerization into gum has been observed only in impure forms of acrolein. If various pure substances are added to acrolein, it is found that free bases or salts or else substances capable of becoming basic alone cause the formation of the gum. Furthermore,in cases in which polymerization into gum took place, Moureu has been able to reveal or at least to suspect the presence of basic impurities and especially of metallic salts* Acrolein can remain unchanged for a long time with a very pronounced acid reaction, but it does not iresist for more than a few minutes the action of free alkalies, even when the latter are used in proportions hardly sufficient to cause reddening of the liquid in presence of phenol- phthaleic. It is very generally to the action of metallic salts that must be attributed polymerizations into gum observed in actual practice. Anhydric iron perchloride and cupric chloride, used in the proportion of 1/500, cause the total 36 transformation in mass of acrolein in a period of time not exceeding three months. Speed of the Phenomenon.- This speed depends upon the activity and concentration of the polymerization reagents; it is possible to obtain every degree of speed by varying these two factors. At a concentration of 1$ ammonia gives a reaction of explosives rapidly. Brucine polymerizes suddenly with projections after a few minutes of contact, quinine polymerizes in.a few hours, morphine in a few weeks, pyridine in a few months. By using sufficiently weak concentrations of the most energetic reagents, their action can be regulated at will. Heat seems to promote the formation of the gum as it does that of disacryl. STABILIZATION Amyl nitrite,* Early reports by the French seemed to indicate that 6% amyl nitrite stabilized acrolein. Later observations (Z-124) showed that while it dia prevent the formation of disacryl, it seemed to promote polymerization into gum. Copper.- Kohler (B,M,I,45) reports that the stabiliza- tion of acrolein is a simple matter and consists simply in the addition of a small amount of metallic copper. This keeps acrolein in the liquid condition indefinitely whether the sample be pure or impure. Even if the polymerization 37 has started, it is immediately arrested by the addition of copper. Tubes containing acrolein with copper have been heated for a period of 24 hours at 300° without showing any change. Moureu finds that copper also prevents the formation of disacryl but it promotes the polymerization into gum. The stabilizing action of copper is due to superficially oxidized copper and not with carefully sconced copper. Perrott, Yablick and Feld (B.M, XII-II) reached a similar conclusion. From the behavior of copper it was thought that copper acted as a deoxidizer. The copper tar&ished, forming an oxide of copper, which later dis- solved (green color of the acrolein). If the cuprous salt acted as a stabilizer, it seemed simpler to add a solution of a salt. 10 drops of a saturated alcoholic solution of cuprous chloride, added to 50 cc. of acrolein, preserved it unchanged for 20 days. Perrott and Yablick (B.M, XIX-21) continued these experiments for a period of 4 months and confirmed the earlier result. 0.5 gm. cuprous chloride per liter will effectively stabilize acrolein even in the sunlight. Other experiments indicate that NagSOg , traces of SnClg (0.022 gm. per 50 cc.) or Iodine 0.4 gr . per liter are also effective in the preservation of acrolein when kept in the dark. Phosphorus and aluminium amalgam were also more or less effective. The behavior of the different agents is seen in the 38 following table: SERIES A. RESULTS: Glass stoppered bottles containing 50 e.c, acrolein kept in dim diffused light. STABILIZING ELAPSED AGENT TIME TO VISIBLE CHANGEr APPEARANCE AFTER 30 DAYS. APPEARANCE AFTER ISO DAYS. 1,Aluminum 2 days amalgam About con- verted to a white gelatinous mass, 2,Yellow phosphorus 2 days About 20$ con- verted to white solid. 3.Crystals in alcohol Clear and liquid * Sticky viscous mass. 4, 0,02 g, SnCl2 Yellow per 50 c*c. tinge in 1 day. Yellow liquid. Viscous as lubri- cating oil. 5, 0,04 g, SnClg Yellow per 50 c.c tinge in 8 hours. Somewhat viscous. Solid glue-like mass, 6, 0,05 g, SnClg Yellow per 50 c„c. tinge in 8 hours. More viscous than #5. Solid glue-like mass. 7. 0.06 g. 3nClp per 50 c.c. Consistency of lubricating oil. 8, Copper shot 21 days. Solution green and gelatinous deposit formed. Gelatinous deposii increased; some solid formed. 9, Nothing added. 1 hour. Dense white solid, 10. Cuprous chlor- ide 0,02 g, per 50 c.c* Slight amount No apparent change sediment formed when first add- ed; no visible change with time* 39 SERIES B Glass stoppered bottles containing 50 c,c< aorolein kept in direct sunlight: MATERIAL ADDED ELAPSED TIME TO VISIBLE CHANGE APPEARANCE AFTER 30 DAYS. APPEARANCE AFTER 120 DATS. Cuprous chloride No apparent change. No apparent change. Copper shot 1 day. Dense green- ish solid in 1 day. Stannous chloride 1 day. Opaque viscous mass in 3 days. Iodine 3 days. Iodine color Pleaches; vis- cous mass formed. Yellow phosphorus 1 day. Dense white solid in 3 days. An extensive study of the stabilization of acrolein has been made by Moureu (2-124), Without discussing in detail the mass of experimental data there reported, it may be said that he has found very satisfactory stabiliz- ing agents in the polyhydrin phenols: 1-1000 of pyrogallolf 1-500 of pyrocatechol or hydroquinone, 1*250 of gallin acid or 1% of tannin extract of gall nuts. From the amount of acrolein changed in one aay, he calculates the life of 40 specimens of acrolein stabilized as above to be from 14 to 27 years. The calculation, although it has no objective value, is given in order to show more strik- ingly the slowness of the change. ANALYSIS QUALITATIVE. Teague and Yoe (B.M. IX-16) have used a very dilute (0.0004N) potassium permanganate solution which has been acidulated with sulfurio acid, (1-26 gm, KM11Q4 are dissolved in 1 liter water. 10 cc* of this solution are then diluted to a liter and acidified with a little dilute sulfurio acid). The decolorization of this solu- tion indicates the "break” of the canister. Hill (B.M. III-21) found that 0.000316 gm. Klfo04 (25 cc. of a 0,01 H solution containing 2 cc, concen- trated H2SQ4 ) was decolorized in one second. Passage of laboratory air for 20 minutes caused no color ohange. It is possible to detect the odor of acrolein 2 or 3 minutes before the potassium permanganate solution is decolorized. Magenta solution (0.1 gm, per 1000 cc. water) changed in 2-3 minutes. Hill further tested a solution made by adding 0.5 gm. sodium nitroprusside, in water solution, to 15 cc. 41 piperidine and diluting to 100 ce„ Using a corftant small supply of acrolein rapor this solution underwent a color change in 50 seconds, becoming green by transmitted light, and violet red by reflected light* If diluted further, the reagent beoomes erratic and inefficient as an indicator. Tsalapotanis (Anales soo. quim. Argentina, 5, 244(1917) has used Cramer’s reagent for detecting acrolein in the following manner;l-2cc. of an aqueous solution of resorcinol and several drops of a 10$ NaOH solution are added to the liquid to be tested and the mixture is heated; after about 2 minutes a bluish green color in the case of dilute solutions and a red color in the oase of concen- trated solutions is obtained. The color disappears when acids are added and reappears after addition of alkali; the color Is quite stable in alkaline solutions. QUANTITATIVE Yoe (BeM.VII-14) devised the following method of analysis in conneotion with the testing of efficiency of canisters: A measured volume of air containing the acrolein vapor was passed through an electric furnace heated to 800° or 900°C. The carbon dioxide thus formed passed from the furnace into barium hydroxide solution contained in small glass cylinders. A large excess of barium hydroxide solu- tion was used, the excess being determined by titration 42 with oxalic acid, using phenolphthalein as an indicator. By difference, the amount of barium hydroxide that com- bined with the oarbon dioxide was obtained and from this was calculated the amount of acrolein present. From the pressure line the air was foroed first through a 40$ potassium hydroxide solution to remove the carbon dioxide of the air and then through concentrated sulphuric acid to remove any potassium hydroxide. The air was passed on through the flask and the 3/4” silica tube of the furnace, filled with small pieces of unglazed porcelain. At the other end of the furnace tube were several glass absorbent cylinders, each containing 50 cc. of 0.02 N barium hydrox- ide solution. A wet meter was connected to the last cylinder of the series, the meter being attached to the suction line. The furnace having been heated to 8000 or 9Q0°C., air was forced through the apparatus for 10 or 15 minutes or until a two-liter sample showed the absence of oarbon dioxide. The bulb containing acrolein was then broken, the pressure and suction turned on and the gaseous mixture drawn slowly through the furnace, 1/3 liter per minute, Five liter samples were taken till, by analysis, the apparatus wasjghown to be carbon dioxide free. Immediately upon starting the sampling a low Bunsen flame was placed under the beaker of water containing the flask in which the bulb of acrolein was broken. By th.V? time the apparatus was shown to be oarbon dioxide free, the 43 water in the beaker was boiling, thus insuring the complete vaporization of any acrolein which might have been in the capillary end of the bulb, the boiling point of the water being considerably above that of the acrolein. The maximum deviation from the mean value of the aorolein was 5$, This is probably no greater than one should expect when the large volumes of barium hydroxide solution and the small amounts of acrolein are considered. (0066 to ,0137 gm. of acrolein were used in the above analyses). It was found necessary to use four cylinders of baryta in series to colleot all the carbon dioxide when the latter is bubbled through at the rate of 1/3 liter per minute, Mark (B• M« IX-22) has proposed to pass the vapor of acrolein diluted with air through wash bottles containing standard potassium dichromate solution and to titrate the excess dichromate with standard ammonium ferrous sulfate (Plate II), The absorption bottles stand in a water bath at 83-87°. No. 1 contains 20 cc, N , 80 cc, HgO , and 2 cc. cone, . Bottle No, 2 contains 10 cc. KgCrg07 -{- 90 cc, HgO 2 cc. cone. HgS04 ♦ No. 3 contains 3 cc. KgCrgOy -b 27 cc, HgO -b 0,6 cc, The absorp- tion is completed in the first two bottles. The procedure, as given by Mark, follows; The wash bottles are brought to the temperature of the bath. The weighed pyknometer and contents are attached. 44 /j UbEZ) /v /Ia'/JLYS/J Of /{ C ROBB'/A/ FINCH COCK Pl./}T£ JJ .A SR'PATCP BQT TLF WATER RATH AT feS°C / HPLALL BOTTLES y» i fn f*Lf?F0RA fF D Bcl&'j 7 FLO vV MET Lf? P NO COO / PY*nOMELTEI? 2 a PINCH C OCK A C L 2. TwJ3> A steady flow of air at the desired rate is obtained and then, by opening a pinch cock on the pyknometer, a very slow stream of bubbles of air is passed through the acrolein* At the end of 30 minutes, the pyknometer is disconneoted and weighed. Air is continued to be drawn through the apparatus for five minutes more and the contents of the wash bottles are transferred to beakefs and titrated. Certain variables,such as the concentration of sulfuric acid, the temperature of the water bath, the concentration of the vapor and the rate of flow, affeet the results and are being studied. Typical results are given in the following table: Acrolein gnu Vol, of Air 1 vol,acrolein K£Crg07 cc* N. KPCrP07 cc. - 1 gm. acrolein 0.0538 70 6.97 106 0.0390 84 6.02 100 0*0360 109 4.61 104 0,0112 460 1.42 104 0,0104 515 1.31 103 0.0173 610 2.23 105 0.0130 896 1.61 101 0,0091 1187 1.14 103 Average - * - - -103.2 Calculated value for 3 atoms of oxygen to 1 molecule acrolein 102.3 45 Coith (B.M. 11-10) has applied the Higgins-Marriott method for determining carbon dioxide (J. Am. Chem. Soc. 39, 68) to the determination of acrolein. He finds that one part of acrolein can be detected in 15,000 parts of air with an error of less than 12$, TESTING THE EFFICIENCY OF CANISTERS. Teague and Yoe (B.M, IX-15; XXI-4) used the following method. Plate Hi shows the apparatus employed. A humidifier for regulating the humidity at 50$ is attached to the air 4 intake tube. The furnace is supplied with two quartz tubes (£” bore) one being £sed for entering samples and the other for effluent. These tubes are filled with pieces of unglazed porcelain. Procedure.- The tube containing the liquid is filled to the same height for each test (20 cm. has been found satisfactory). Holding the suction flow-meter at 32 L/M, the flow-meter connected with the blower is adjusted to give a concentration of 5000 p.p.m. Under present conditions about l/lO of the air is blown through the liquid. The mixture is drawn through an old canister until tests of the effluent gas show that the proper concentration is being maintained. The blower and suction are then closed and the new canister substituted. The drop in pressure across the 46 BUREAU OF MINES-W aSHINCTON D.C GaS INVESTIGATIONS ~j « C A l. E : - T drawn s- r ,4/BB4B4Tc/S 404 TSSTW5 THE EFFECTS. OF CHECKt-T gv JJi./. ~- S -7 7 CANISTERS A OAINST ACFOUN Bz-ovs/vstep 4/B 4B6l'£A TO/r PLATE HI A/ore- 4/p /A/r44£ rrt4>OiJ6£/ ry Bsgsy.4rep. 5*tvXs?t//vg rv&s KT/?pp rfAA/OMfi- r£ff ffj?P/£±/ASG 7~t/&& /cs bat.*/ Cjp/srt# C7/X/A/5 CA/4/*70£f <2 t/Ai. / T*r/ V£ 7‘£s t sc mss fi/p/v/rcs Sc/cr/o/v YV'/fsTS 54 S " A # SOP&£ a/r SucT/a/v TO NETS# E/TL i/£NTrfffL i/£/Vr StHPUVG CYL//VJ?£ZS canister at 85 L/M flow is taken on a separate apparatus at the beginning and the end of the run* The suction through the main flow-meter is now regulated to 31 L/M and that through the qualitative test at 1 L/M, making the rate of flow through the canister 32 L/M. Time is taken when-the suction and air are turned on* At intervals of about 20 minutes throughout the "run", 1-or 2-liter samples of effluent gas are withdrawn through the furnace. The samples are taken at \ liter per minute rate, measured by means of a standard wet meter, absorbed in four absorption cylinders, and analyzed according to the method herein described. The absorption cylinders each contain 50 cc. 0.03 N barium hydroxide solution. While these samples are being taken, the pressure drop through the canister is noted. At present, the qualitative test used consists of 50 cc. of very dilute (approx. 0.004 N) potassium permanganate solution which has been acidulated with sulfuric acid. When decolonization of this solution indicates the "break" in the canister, five-liter samples of effluent gas, at \ L/M rate, are withdrawn as frequently as possible until the efficiency falls below 80$. Just before each sample, several hundred c*c. of gas are drawn slowly through the furnace,(\ L/M) in order to sweep the tubes leading to the absorbent cylinders free of gas of the previous sample. The time and length of the sampling periods are noted and the elapsed times are calculated in the middle of 47 these periods. The efficiency is calculated on the basis of the nearest effluent concentration,, The observed pres- sure drops must be corrected for the drop through an empty canister. The humidity and average temperature during the test are reoorded. Moteo- The efficiency of the qualitative test has been found to be about 100$, As a oheck, however, a small bottle (which is kept in series) of the effluent gas is removed just after the decolorization of the permanganate solution, A very faint odor of gas is obtained. (If the test lasts for more than 2 hours, a fresh qualitative test solution should be inserted,) It is possible to detect the odor of acrolein 2 or 3 minutes before the potassium permanganate solution is decolorized* The odor test will be considered the 100$ efficiency point. Analysis*- The gas samples pass through the furnace which is heated tp about 900°C, The acrolein is burned and the oarbon dioxide thus formed is absorbed in 0,03 N barium hydroxide solution. The excess barium hydroxide is determined by titration with 0,03 N oxalic acid, using phenolphthalein as an indicator, A blank is used for the carbon dioxide in the air. This is subtracted from each liter, For a one-liter sample this correotion is about 1 c,c, barium hydroxide solution. The acrolein is caloulatec as parts per million at 25°C and 760 mm. pressure, 48 Calcu?ations.- In a one-liter sample, a c„e« of 0.03 N solution = 122.25 p.pam, gas at 25°C and 760 mm. pressure. 5000 p.pjm. = 11.46 mg. per liter. ABSORPTION Hill tested charcoal against 0,08$ acrolein, with a rate of flew of 540 cc, per per minute and found the following values: 34 min. 100$ efficiency 60 " 98$ " 86 " 87 " 105 n 77 u Fuller's earth gave qualitative tests from the start and after 10 minutes showed 38$ efficiency. Permutite and sawdust did not absorb at all. PROTECTION M-5 states that acrolein goes through the German mask and through the old French mask. The.French propose the following protection: Pumice incorporated with 9$ of its weight of crystallised chromic acid (not solution). Deliquesence of the acid causes im- pregnation of the pumice. Teague (B,MtIX-14) reports that the canisters tested (Bee, 10, 1917) furnished good protection against acrolein. In B.M.X-23, the U«S,Canister in use at that time is 49 reported as furnishing protection against 5000 p.p.m. for 53 minutes as compared with 28 minutes for HCN and 28 minutes for 4000 p.p.m. of chlorpicrin. Permeability tests by Perrott and Feld are given in B.M.IX-8. A Goodrich rubber sheeting MR#6 , held for 60 minutes (the same fabric held for 500 minutes against chlorpicrin.) PHYSIOLOGY Acrolein is a lachrymator and respiratory irritant; the effects on the eyes and throat occur simultaneously. In concentrations of 0.025 mg. per 1. it develops an irritating action, secretions of saliva, weeping from the eyes, nose secretions and slight narcosis. Frankel says that acrolein is highly irritating and but little antiseptic. Action on ,Man.~ One part in 200,000 acts as a laohrymator and nasal irritant, while 1 part in 100,000 is intolerable (Ph-31). The minimum effective concentration (with or without eye protection) is 1 in 100,000 (0.025 mg. per 1). The toxicity is about the same as phosgene. The gas causes nausea and heart trouble (depression and partial paralysis) and also affects the memory (M-5). Action on Mice.- Kolls, Kuhn and Todd (B.M.III-48) using a sample of acrolein boiling at 52*6° and exposing the mice for 10 minutes, found the immediate toxic concentration to be between 0.55 and 0.38 mg. per liter (that which kills 50 over 50$ of the mice exposed for 10 minutes within 48 hours) The delayed toxic point (that which kills after 48 hours and in less than 10 days) is between 0«16 and 0.17 mg. per liter. Symptoms.- In concentrations of 1 mgm. or more there was increased excitement, but lower concentrations produced depressionso All concentrations caused irritation as shown by rubbing of noses and closing of eyes. Gasping was also a constant symptom, it being only occasional with the low concentrations, but continual with the highest concentration, 4 mgm, The delayed deaths above took plaoe in from 48 to 65 hours. Concentrations of 4.1 rogms. killed all four mice within 7 min* in bottles. They showed violent excitement from outset and <£ied with convulsions* Moderate concentrations killed all within from 1-46 hours. Action on Rats.- One part in 1,000 killed in 50 minutes (Ph-3). Action on Cat.- A dose above 0.04 mg. per 1. causes such intense irritation that some days are required for recovery. With 0.2 mg. the lung irritation phenomena are not recognizable but are obviously painful and the muscles come into activity, A dose of 1.5 mg. seriously affeet3 the animal, which dies 18 hours after a 2.25 hour exposure to the gas, of pulmonary edema and bleeding of the lungs; with a dose of 1.98 mg. death sets in 2.5 hours later, (Kober and Hanson, Diseases of Occupation, p,720). 51 Action on Dogs,- Miller (BoM.X-58) reports the following symptoms: During exposure-- The dogs are very excited during the early part of the exposure, The eyes and nose are immediately irritated and the animal blinks and licks his nose the instant the gas is turned on. Lachrymation and salivation are both profuse. After a short time the animals keep their eyes tightly closed. The cornea is usually dulled. Nasal secretion is very much increased, Respiration is early affected becoming Mery slow and labored. Towards the end of the exposure the animal is usually very depressed. Subsequent symptoms,-- Within a few hours after exposure the animal develops a severe tracheal rattle, is very depressed, coughs, and has very labored respiration. The eyes aro sore and the cornea is usually dulled. With toxic concentrations death occurs in four hours to two days. With non-toxic concentrations the animal is very sick for several days and does not fully recover for a week or more* The toxic concentration may be placed at 0.35 mg. per liter. Ph-2 reports that 0*5 mg. per 1. killed in 6 hours, SENSITIVENESS OF INDIVIDUALS. Experiments to determine the smallest concentration of acrolein that can be detected by the eyes, nose, throat or lower respiratory tract (Sherwood, Snyder and Gavin, B*M* XIII-59) show that acrolein is detectable by its odor 52 at a concentration of 0.0028 mg? per liter or 1.12 pop.m. and that some individuals will detect it at even smaller concentrations, When the amount of gas is increased to 0,0077 mgc per liter (3,06 p,p*m.) irritation of the eyes and nose become prominent symptoms in the majority of tests Lachrymation does not become prominent until the concen- tration reaches 0,010 mg, per liter or 4,00 p,p,m<> PATHOLOGY Only two animals have been examined after exposure to this gas* There has been so much similarity in the lesions found, however, that a tentative description of the action of the gas will be given* The eyes are severel; affected. At the end of two or three days a purulent conjunctivitis and heavy clouding of the corneal epitheliu- have developed* As no animals surviving more than three days have been observed, the ultimate result of these lesions has not been determined. The mucosa of the mouth and nasal passages shows patches of reddening; and on the epiglottis and about the laryngeal orifice pin-head hemorrhages are found. The trachea and bronchi down to the smallest visible branches are lined by a semi-fluid purulent exudate of dirty yellowish-gray color. The enti: bronchial tree is literally dripping with pus. The suppurative bronchitis is readily differentiated from the membranous bronchitis and trachitis of mustard gas dogs. 53 The former is frankly purulent, while the latter forms a distinct membrane which can be peeled off in sheets. The suppurative condition found in acrolein dogs is associated with an extreme congestion in the tracheal and bronchial walls. The lungs are composed of normal looking ar&as of pulmonary tissue mingled with patches of atelectasis, which are deep red, firm and airless. The collapse appears to be due to the plugging of bronchi with exudate. On section these lungs show numerous droplets of pus escaping from the openings of cut bronchi and also small patches of broncho-pneumonia. Outside of these lesions of the eye and respiratory system, no noteworthy.changes have been found. TREATMENT M-25 (Chemistry of Gas Warfare) suggests the following treatment: Immediate removal from the poisonous atmosphere artificial respiration; inhalation of steam; faradic stimulation of the phrenic nerve; free blood letting; in case of obstinate spasm of the glottis tracheotomy. TACTICAL USE It has been shot in grenades and found effective. Use in shells would involve the use of a fumtgen, which would interfere with the stabilization. 54 SHELL FILLING Acrolein received in cast iron drums is a liquid under ordinary conditions and is fed without refrigeration by gravity to a chapel, in which a small feed pipe with a "constant empty space" manometer is used to fill grenades. This filling is all hand work. The filler holds the grenade in his left hand and operates the feed valves with his right# When it is full he passes it to another operator who screws the gain© into the grenade with his right hand« A third man holds the grenade in his left hand and tightens the gaine with a wrench held in his right. All these operators wear heavy rubber gloves. These grenades are quite fragile and mus$ be handled and packed with care. Here no mellinite cartridge is used in the gaine the detonating charge of mercuric fulminate being sufficient to shelter the grenade. This acrolein affects the eyes badly but its worst effect is the terribly oppressive suffocation that it produces when taken into the lungs. 55 BIBLIOGRAPHY This bibliography contains all articles dealing with acrolein that were found by Mr. L. H. Plett under that name in the more important periodicals. (Winter of 1917-1918. It contains nothing from confidential sources issued since the beginning of the war.) Pelouze "Ueber das Glycerin." Ann>19, 212 1856. Hess "Ueber einige Producte der Trockennen Distillation." Ann. 20, 9. 1836. Redtenbacher "Ueber die Zerlegungsprodukte des Glyceryloxydes durch trockene Distillation/" Ann. 47, 113. 1843. Gerhardt "Acroleine". Chimie Organique I, 778. 1853. Cahours u. Hofmann "Oxidation of Allyl Alcohol." Ann. 102, 291. 1357. Geuther u. Cartmell "Ueber das Verhalten der Aldehyde zu Sauren." Ann. 112, 1. 1859. Hubner u. Geuther Acrolein Chloride from PCle . Ann. 114, 35. 1859. L. Barth "Ueber die Einwirkung des Broms auf Glycerin." Ann. 124, 341. 1862. A, Claus "Ueber Acrolein und Acrylsaure," Ann. Suppl. 2, 117. 1862. Linnemann "Von dem Uebergang aus der Acrylreihe in die Reihe der Pettkorper und umgekehrt," Ann. 125, 310. 1863. 56 Aronstein "Ueber Einige neue Acroleins Verbin- dungen. Ann, Suppl. 3, 180. 1864, A. Claus. "Ueber Acroleinammoniak." Ann. 130, 185. 1864. Simpson Acrolein from Acetone diiodide. Z. fur Chem* 1867, 375. 1867. Schiff "Derivate des Acrolein.” Ann. Suppl. 6, 26. 1868. Wolff Acrolein from the distillation of Propylphycit (C3Hq04)‘ Ann. 150, 28, 1869. Baeyer Distillation products of acrolein ammonia. Ann. 155, 281. 1870. Taubert "Acrolein.” Jahrber. Chem. 1876,479. 1871. Claus "Ueber die Zersetzung des Acrolein- ammoniaks durch trockene Distilla- tion."Ann. 158, 222. 1871. Carstanjen Acrolein from Epichlorhydin, sodium, and Methyl Iodide. Ber. 5, 810. 1872. Linnemann u. Zotta Decomposition products of Glycerine distilled from calcium chloride. Ann* Suppl* 8, 254. 1872. Muller "Verbindungen des Acroleins mit den sauren schwefligsauren Alkalien.” Ber. 6, 1441. 1873. E. v. Meyer "Ueber eine Bildungsweise des Acroleins aus Aethylene." J. Prakt. Chem. (2) 10, 113. 1874. Linnemann "Beitrage zu ffeststellung Lagerungs- formel der allyl verbindungen und der Acrolein." J* Prakt. Chem. (2) 10, 161. 1874. Zotta "Zur Kenntniss des Glycerinflthers." Ann. 174, 87. 1875. Bruhl ' Time before Polymerization. Ber.12, 317. 1879. 57 Romburgh Derivatives of Acrolein* Bui. sec. chim* 36, 549. 1881. Gumaux et Adam "Chlorhydrate d'Acrolein.” Bui. soc. chim. 36, 23. 1881. Bartol e. Papasogli nAl«ttrcliai della Glicerine." Gaz. Chim. Ital. 13, 287. 1883. Claus "Notiz zur Darstellung von Propylene aus Glycerins." Ber. 18, 2931. 1885. Rosenthal Preparation and Oxidation of the bisulfate compound of acrolein. Ann. 233, 36. 1886. Fischer u. Tafel "Darstellung des aus Akroleinbromid." Ber. 20, 3388. 1887. Fischer u. "Verbindurgen des Phenylhydrazins Knoevenagel. mit Acrolein." Ann, 239, 196. 1887. Lederer Chlorine derivatives. J. Prakt. Chern. 42, 384. 1890 Griner "Preparation de 1'acroleine." Ann, chim, Phys■ (6) 26, 367. 1892. Beilstein Acrolein. I, 957, 1893; I, 482, 1900. V0 Meyer u. * Acrolein. I. 522. 1893. Jacobson Wohl u, Neuberg "Ueber die Darstellung des Acroleins." (Boric acid). Ber. 32, 1352. 1899, Lewin Acrolein Reactions and Tests. Ber.32, 3389. 1899. Kalle &> Co. "Verfahren Zur Haltbarmachung von Akroleinlosungen." Ger. 109, 053, 28/2/99. 1899. Kalle & Co. "Verfahren zur Darstellung eines neuen Disinfektionsmittels." K1. 30 i. Nr. 116, 974 vom 3/11/99, (21/12/1900) 1899. Kalle & Co. "Verfahren zur Darstellung von Ver- bindungen des Akroleins mit Starke, Dextrin ." Kl. 12p9 Nr* 129, 884 vom 14/4/99, (22/3/1902). 1899. 58 Kaile & Co. Same as i2p. Nr« 151, 399 vam 14/4/99, (22/5/1902). 1899. .Koch u« Fuchs "Uber den Antibaktereillen Wert des Acroleins." Cent. f. Bakter u. Parasitemk I* 26, 560. 1900. Wohl 8c. Emmerich A Chlororopionacetal from Acrolein. 'Ber. 33, 2761. 1900. Wohl. "Acrylsaure aus Glycerin." J. Prakt. Chem. (2) 61, 200. 1900. Nef Decomposition Products of Glycerine. Ann. 335, 200. 1904. Lickemann u„ "Uber die Akroleindarstellung nach liesche dem Borsaureverfahren0" J. Prakt. Chem. 71, 474 . 1905. McLeod On the Polymerss of Acrolein. Ann, Chem. J. 37, 35. 1907. Wohl Correction of Work in Ber. 32, 1352. Ber. 40r 4685 (Note) 1907." Grun and Bockerisch Glycerinates. Ber. 41, 3465. 1908. Senderens "Dehydration catalytique des composes organiques," Bui. Soc, chim. (40 3, 823. 1908. Schlotterbech Chlorine and Bromine Derivatives of Acrolein.. Ber. 42, 2559. 1909. Bergh "Uber die Darstellung des Acroleins." J. Prakt. Chem. 79, 351. 1909. Sabatier et Mailhe Change to Propionaldehyde. Ann. chim. phy* (8) 16, 72.. 1909. Hahn "Ein. neuer Fraktionieraufsatz." Ber* 43, 422. 1910. Grun u. Husmann "Glycerinate der Erdalkalien." Ber. 43, 1291. 1910. Castoro Preparation of Colloidal Solutions. Z. Chem., Ind . Koll« 6, 283. 1910. N., Iwanoff Analysis and Physiological Effects of Acrolein a Arch., f. Hyg. 73, 331.1910. Trillart Acrolein in the Incomplete Combustion of Straw. Compt. rend. 150. 340, 1910 59 Voisenet "Formation d:acroleine dans la mala- d?.e do L!amertume des vins.” Compt, rend. 150, 1614. 1910v Voisenet "La Fermentation Acrylique de la Glycerine." Compt. rend. 151, 518. 1910. Senderens "Preparation de L'Acroleine. Compt. rend, 151, 550. 1910. Mirande-Guignard "De 1'action des Vapeurs sur les plantes Vertes." Compt. rend. 151, 482. 1910. Berthelot et Action of the Utraviolet Light on Gaudechon Acrolein. Compt. rend. 151, 1351. 1910. Ostromisslensky u, "Eine neue Indigo Synthesis." Ber. Pamfelow 43, 2774. 1910. ,/ Wohl u. Maac "Zur Darstellung der Aldehyddio- cetate." Ber, 43, 3293. 1910. Wohl u. Maag "Condensation von Acrolein-mit Acetone--." Ber* 43, 3290. 1910, Auwers u. Eisenlohr "Spektrochenische Untersuchungen." J. pr. (2) 82, 114. 1910. W. Lob ' "Zur Kenntinis der Zuckerspaltung." Biochezms Z. 26>, 232. 1910, Kunz-Krause u. "Cyklogallipharsaure" C21H3603 . Manicke Arch. Pharm. 248, 405, 696. 1910, Voisenet "Sur un ferment de l'amertume des vins, agent de oeshydration de laglycerine." Compt. rend. 153, 364. 1911. Stadler "Entwicklungshemroende Wirkung." Arch. Hygiene 73, 211, 1911. Silberrad Artificial Rubber. Gummi Stg. 25, 1958-Cent, 1911, II. 1342. 1" H, Iwanoff Action of Gases on Organisms, Arch. ?. Hyg. 74, 307. 1911. Senderens Preparation de L’acroleine. Bui. soc. chim. (4) 9, 374. 1911. 60 Lespieau "Condensation of the Bromide to Malonic Acid." Ccmpt. rend. 151. 1359 o 1911. Coninck Silver Acrylate„ Bui. sci. acader. belg, 524 . 1912, Wohl u* Mylo The preparation of Acrolein. Ber. 45, 2046. 1912. Oechsner et "Sur un mod de formation de l1aero - Coninck leine," Compt. rend. 154, 1353.1910. Krestinsky et Catalytic Decomposition of allyl Nikitine alcohol. J. soc, phy. chim. R. 44, 471. 1912. Witzmann "The Preparation of Acrolein." J.Amer. Chem. Soc. 36, 1766. 1914. Voisenet "Chaleurs de formationde l?acroleine et de la metacroleine." Bui. soc. chim. (4) 17, 34. 1915. Wagner J. Russ.Chem. Soc. 16, 317. Krestownskow J. Russ. Chem. Soc. 11, 249. Lobry Rec. Trav. Chim. 4, 231, Condensation, Voisenet A New Color Reaction of Acrolein, J. de Phy. et*de Chim, (7) 3, 580. 61 CONFIDENTIAL REPOETS * B.M. 1-39 Sodium Permutit as an Absorbent* Hill* B.M, 1-46 Preparation and Stabilization of Acrolein. Kohler. B.M. 11-10 Efficiency of the Higgins-Marriott Method for Determining when Applied to the Determination of Acrolein* Coith, B.M* 11-32 Manufacture of Acrolein. McPherson. B.M. III-20 Preliminary Experiments of the Prepara- tion of Acrolein. Evans. Oct. 15, 1917. B.M. III-21 Absorption of Acrolein. Hill. 9/26/17. B.M* III-22 Permutit as an Absorbent for various Gases* Apmann and Wilson. B.M. VII- 8 Acrolein. Evans. Dec. 15, 1917, B.M. VII-14 Determination of Acrolein Vapor in Air Mixture. Yoe. Nov. 26, 1917. B.M. VII-48 Toxicity Experiments on Mice. Kolls, Kuhn and Todd. B.M. IX- 8 Permeability of Mask Fabrics against different Gases. Perrott and Feld. Dec. 3, 1917. B.M. IX-14 Report on Canister Tests made against G-4, Teague, Dec. 10, 1917. B.M# IX-15 Method for Testing the Efficiency of Canisters against G-4. Teague and Yoe. Dec. 10, 1917. B.M. IX-21 Dehydration of Glycerol. Engelder. Dec. 4, 1917. B.M. IX-22 Preliminary Report on Method of Analysis of G-4. Mark. Dec. 11, 1917. B.M. IX-36 Physiological testing of Absorbents. Crowell. B.M. X-23 Comparison of Canister Tests on different Gases. Teague. Jan,15,1918. 62 B.M* X-58 Toxicity of G-4 on Dogs, Miller, B.M. X-61 Pathological Anatomy of Poisoning by Gases, Mackenzie, B.M* XI1-32 Experiments performed in the Preparation of G~4« Engelder and Hultman. Feb.l3,191f B.M, XIII-47 Meeting of Offense Warfare Committee. Feb* 21, 1918. B.M. XIII-59 Smallest Concentration of G-4 that can be Detected by the Eyes, Nose, Throat, or Lower Respiratory Tract. Sherwood, Snyder, and Gavin. Feb. 25, 1918. B.M, XIX-21 Stabilization of Acrolein. Perrott and Yahlick* May 28, 1918. B.M. XXI-4 Method of Testing the Efficiency of Canisters against Acrolein. Teague and Yoe .• Dec* 10, 1917. M - 137 Notes on Filling Suffocating Grenades with Pappite-Aorolein. M - 25 Confidential Notes on the Chemistry of Gas Warfare* Ph- 2 Poisonous Gases and Vapors and their Physiological Effects. Physiology War Committee, Aug. 1917. Z - 124 The Stabilization of Acrolein, Moureu. Mar* 13, 1918* 63