Document 65KDExyvOGNjGv1QxKbEeKoyR

RIPCRT WO. 2755 FINAL UFOtT CM KCL SCATIHOaS FCR JOB NO* 171-360 mfcHQia. R. J. Good March 13, 1952 RESEARCH DEPARTMENT Phosphite DlvWoa MUNS Q6b494 Report No. 2755 FINAL REPORT ON HCL SCAVENGERS FCR TRANSFORMER HRAMOLS Job No. 171-360 File No. 141-32 Work Started: November 30, 1948 Prepared By: R. j. Oood Work Completed: November 13, 1951 Report Submitted: March 13, 1952 Chemlata: a. m. Eiianburg S. J. Oood H. B. Richards, Jr. Heotor Flores R. L. Crenshaw R. R. Knight 20 copies were msde of this report and distributed as follows: No. 1 No. 2 No. 3 No. 4 No. 5 No. 6 No. 7 No. 8 No. 9 No. 10 No. 11 * No. 12 VNo. 13 "Nu. If No. 15 No. 16 No. 17 No. 18 No. 19 No. 20 Research File R. L. Jenkins - St. Louis H. 7. Moss J. L. Christian - T. R. Smith - St. Louis I. T. Durrett - F. P. LsBelle H. F. Weaver - St. Louis C. A. Hochealt - St. Louis N. N. T. Samaras - Dayton J. R. Wilson - St. Louis B.-Q. Johnson - St. Louis I. P. Rucker E. E. Hardy H. 0. Tittel II. Or'IiUei------------------ A. M. KLlenburg R. J. Oood Paul Logus - St. Louis T. H. Wheelock - St. Louis Extra ' Extra This la Copy This report and the information contained herein is the property of i MONSANTO CHEMICAL COMPANY MONS 088495 TABLE OF CONTENTS I. INTRODUCTION Page 1 II. SUMMARY.................................................................................................................... 1 III. REFEREES............................................................................................................. 2 IV. EXPERIMENTAL WORK............................................................................................... 3 A. APPARATUS USED, AND METHODS................................................................. 3 (1) Tensile Strength Screening Test............................................... 3 () Electrical Equipment....................................................................... 6 B. RESULTS.................................................................... 6 (|) Tensile Strength Screening Tests.............................................. 6 (a) Estimate of Accuracy and Significance of Tests.... 9 (1) Electrical Tests.................................................................................. 11 1. (a) Significance, Accuracy and Interrelation of 1 . Electrical Tests.......................................................................11 (50 Hydrolysis Tests of Organometallic Compounds...................... 13 (1} Westlnghouse Cu-Corrosion Test.................................................... 13 (f) Compatibility of Scavengers...........................................................14 (0) Odor of Phenoxypropene Oxide in Pyranol.............................. 14 V. DISCUSSION............................................................ 14 a. Tentative criteria for a "good scavenger"..................................... 14 B. DISCUSSION OF COMPOUNDS TESTED............................................................. 15 (1) Active Scavengers of Known Composition................................... 15 (2) Active Scavengers of Unknown Composition..................... 16 (}) Proof of Non-validity of the Broad Claims of the G. E. Tetraphenyl Tin Patent................................................................ 16 C4) Verification and Extension of Previous Results..................17 (|) Results Reported by Westlnghouse................................................ 18 VI. CONCLUSIONS............................................................................................................... .... " VII. RECOKljiENDATIONS...................................*............................................................. 19 VIII. DESCRIPTION OF RECOMMENDED PROCESS............................................................19 IX. PATEN^ STATUS......................................................................................................... 19 x. cost Estimates....................................................................................................... 22 XI. MATERIALS SPECIFICATIONS, ANALYTICAL PROCEDURES................................22 XII. TOXICITY AND HAZARDS.............................................. XIII. ACKNCi|lEDGEMENT................................................................................ 22 22 MOMS 068496 TABLE Of CONTENTS (CON'T) XIV. APPENDIX...................................................................................................................23 A. SUBSTANCES FOUND TO BE LESS ACTIVE THAN TETRAFHENYL TIN, BY TENSILE STRENGTH SCREENING........... ............................... .. 23 B. COMPOUNDS COVERED BY THE BROAD CLAIMS OF THE G. E. TIN TEIRAPHENYL PATENT,FOUND TO BE ACTIVE SCAVENGERS..................24 C. (fODE OF MANUFACTIHERS........................................................................... D. PYRANOL SPECIFICATIONS......................................................................... E. lIlANUFACTUREHS1 DATA ON PHKNOXYFROPENE OXIDE............................ F. MANUFACTURERS' DATA ONDIBUTYLDIPHENYL TIN................................ 32 G. NOTEDOOK PACES........................... .... .......................... ............................... 33 H. HOT SCAVENGER TEST METHOD (AS WRITTEN FCR TRANSMISSION TO OTHER UBCMATORIKS.)......................................................... 34 MONS 038497 - 1- X. Introduction Pyranol 1488, which is a mixture of trichlorobenzene (40J?) and Aroclor 1260 (60), ia a very stable, non-flammable dielectric cooling and insulating medium for transformer* However, in the course of operation of a trans former, a power surge or prolonged overload may cause an arc to occur through the cooling liquid, and the transformer may fail. An electric arc, by reason of its high temperature, causes Pyranol I486 to break down into carbon and HC1. The HC1 attacks the paper insulation of the windings, weakening it or causing it|to crumble, and also corrodes all metal parts of the transformer. Scavengers,ere added to the transformer liquid, to react with the HC1 as soon as it is evolved and thus prevent or localize damage to the transformer. Tetraphenyl tin is the compound used for this purpose at present} it is covered byOeneral Electric patents. i The object* of this project was to find HC1 scavengers that function at least as well asi and preferably better than, tetraphenyl tin. Patents in this field willjgreatly improve Monsanto's position if and when we start selling dielectric^ directly to the transformer industry without subjection to 0. E. patents. I II. Summary ( A total ofi73 new suggested compounds have been screened during the phase of the study Reported herein. Of these, 16 were found to be as good as or better thaA tin tetraphenyl according to our tests. For the tabulated re sults, see | Table 1 and Appendix A, The protection from attack by HC1 afforded by the proposed scavenger to paper of the type used as transformer insulation was taken as a criterion of effectiveness as an HC1 scavenger. The tensile strength of test papers was measured, after a standard exposure to HC1 dissolved in Pyranol con taining the proposed scavenger This was compared with the tensile strength of similarly treated papers protected by tin tetraphenyl* The most satisfactory of the new scavengers were: poly (alkyl-alkoxy) tin compounds pf the type of Advance Solvents Stabilizers Nos. 3 and 52, and Paraplex G+-60. Samples of the two Advance Solvents stabilisers have been fca 'arded 4o the Meetinghouse Electric Co. for their evaluation. The PatentjDepartment has expressed the opinion, {based on our observation that tetralauryl tin la completely inoperative as a scavenger under the test condilions) that the broad claim of the G. E. patent on tin tetraphenyl is invalid'and will not dominate our patent on dibutyldiphenyl tin. Meetinghouse has expressed considerable interest in dibutyldiphenyl tin and phenoxy propene oxide. (These were found to be effective scavengers In previous w6rk in this laboratory, reported in Interim Report No. 2353.) ' A drum of pyranol I486 with 0.1255C dibutyldiphenyl tin scavenger was sold to Mestinghouse in 1950. A tank car of Pyranol I486 with 0.205JC phenoxypropene oxide scavenger (lot E-l) was shipped to them from the Mm. G. Kruminrich plant on Jan. 4# 1952. HUNS 088496 ~2- III. Referehces (A) ftreviolus Monsanto Technical Reports (1) Short Form Report No. 2052, Anniston, "Stabilisation of Pyranol", B. F. Jackson, June 5, 1946 (2) Central Research Report No. 374, "Interim Report on Preparation of Orjgsnometallic Compounds to Test as Stabilizers for Polyvinyl Chloride", E. W. Gluesenkamp, April 16, 1945 (3) Cejntral Research Report No. 503, "Fundamental Study of Stabilization of Polyvinyl Chloride", Miss Hogg,. April 9, 1948 (4) SlJort Form Report No. 2209, Anniston, "HC1 Scavengers for Pyranol, Testing Procedure", T. H. Cleveland, April 20, 1948 (5) Central Research Report No. 525, "Preparation of Organic and Metallic Organic Compounds Primarily for Test as Polyvinyl Chloride Stabilizers", E.fw. Gluesenkamp, Oct* 1948 (6) Interim Report No. 2353, Anniston, "HC1 Scavengers for Transformer Pyjranols", R. R. Knight, May 2, 1949 (7) Snort Form Report No. 2483, Anniston, "Purification of Organometallic Compounds for Pyranol Scavenger", H. Flores, May 5, 1950 (B) Literature (8) Clark, "Nonflammable Dielectric Organic Compounds", I. and E. Chem., 29, 698-702 (1937) (9) Oilman and fowne, JAC5 61, 739-43 (1939) (10) Oilman, Ed. in Chief, "Organic Chemistry, an Advanced Treatise", Vcjl. I, pp. 1071-2, Wiley and Sons, Inc., New fork, 1943* (11) Clark, "Electrical Insulation, a Field for Chemical Exploitation", Ctyem. Eng. News, 2, 2976-8 (1948) .. (12) "Stabilizers" (brochure), Advance Solvents and Chemical Corp., New York. 1950 (13) "Ejpichlorohydrin", Shell Chemical Corp. Technical Bulletin SC: 49 35 (This gives Shell's analytical method for determining epoxides.) (C) Patents (14) U. S. Pat. 1,931,373 and 1,931,455, "Dielectric Material for Electrical Devices", F. M. Clark, assignor to G. E., Oct. 17, 1933* These two patents cover mixtures of chlorinated biphenyl and chlorinated benzene a^ dielectric and coolant in transformers and capacitors. (15) U.j S. Pat. 2,468, *>44* "Stabilized Halogenated Compositions and Electrical Devices", F. U. Clark, assignor to G. E., April 26, 1949- T^is patent covers tin tetraphenyl as an HC1 scavenger for Pyranol add its broad claim, if valid, would dominate our patent on dibutyldiphenyl tin. (16) British Patent 601,359, "Improvement in and Relating to Liquid Hydrocarbon Compositions" -- to British Thompson, Huston Co., May 4, 1^48 (17) Brjitish Pat. 418,230, "New or Improved Compositions Comprising Chlorinated Organic Substance", to I. 0. Farlen, Oct. 22, 1934* ' British patent 601,359 covers phenoxypropene oxide as an HC1 scaven ger for Pyranol. British patent 418,230 is believed to have been cited against the corresponding application by G. E. in this country. (18) uj S. Pat. 1,235,339, "Uninflammable and Electrically Insulating Liquid". Georges Lepine, July 31, 1917 MQNS 088499 -3- (19) U.1 S. Pat. 1,696,641. "Stabilised Chlorinated Rubber", Carlton Elite, . assignor to Chadeloid Chemical Co., Dec. 18, 1928 . (20) d/s. Pat. 1,944,274, "Process Of Treating Chemical Compounds", P. L. Saizberg, assignor to du Pont, Jan. 23, 1934 - (21) U.j S. Pat. 2,036,274, "Insulating liquid", H. D. Holler, assignor to Weptinghouse, April 7, 1936 ' (22) U.1 S. Pat. 2,005,840, "Insulating Material", R. Engelhardt, assignor to' I. G. Farben, June 25, 1935 (23) U.1 S. Pat. 2,073,009, "Fireproof Material", R. Engelhardt, assignor to' 1. G. Farben, Mar. 9, 1937 (24) uJ S. Pat. 2,105,406, "liquid Insulating Composition," F. M. Clark assignor to G. E., Jan 11, 1938 (25) U. S. Pat. 2,105,407, "liquid insulating Composition", F. M. Clark, assignor to 0.' E., Jan. 11, 1938 - (26) U. 3. Pat. 2,143,685, "Electric Device and Dielectric Materials Therefor", F. M. Clark, assignor to 0. E., Jan. 10, 1939 (27) U.J S. Pat. 2,116,604, "Stable Composition Comprising Chlorinated Substances", 0. Meyer, assignor to I. G. Farben, July 18, 1939 (28) U. 3. Pat. 2,169,872, "Liquid Halogenated Compositions", F. M. Clark et'.al., assignors to 0. E., Aug. 15, 1939 ' (29) U.[S. Pat. 2,217,173, "Stabilized High Film Strength Lubricating Oil, B. H. Lincoln, assignor to Continental Oil Co., Oct. 8, 1940 (30) OJS. Pat. 2,377,630, "Stabilized Dieleotric Composition", J. L. Hyke, assignor to Sprague Specialties Co., June 5, 1945 (31) U. S. Pat. 2,388,529, "Apparatus for Absorbing Decomposition Products", F.1 M. Clark, assignor to 0. S., Hov. 6, 1945 (32) U.'3. Pat. 2,453,493, "Halogenated Hydrocarbon Composition", F. M. Cj et.al., assignor to Q. E., Hov. 9, 1948 (33) Bt .ah Patent Application No. 12,969, "New Organo-Tin Compounds ar teslns Stabilized Therewith", G. P- Mack, et.al; application, Me .6, 1947 IV. Experts A. Ampj ' al Morki :bxu.usa Uoeseeod, anda Mmeetchnowdsr sile strength screening test. See Figure 1, sketch of apparatus treatment of papers for tensile strength test. The gas burst manifold on the left was the means of admitting reproducible ntltles of dry HC1 to the system. The solution of the scavenger ar test, in Pyranol 1488, was contained in the three-necked flask, thsre treated with HC1. After a fixed reaction time, the Pyranol blown over into the. test tube containing the papers. The papers e in the form of strips 1" X 7", with the ends punched,, and they e hung on the rack in the test tube, separated by glass spacers. fhs series of operations was as follows) HONS 088500 TOSPPO SNOW -5- l (l) Dfy HC1 (from a HgSO* drying train) was swept through the 273 ml bjilb, with the exit gas being passed into the beaker of water, until the lack of bubbles in the beaker indicated that all air had been sirept out* At the same time, the space above the Fyranol was evacuated several times and nitrogen admitted* Also a slow sweep ojr nitrogen was passed through the test tube containing the papers, a\, a reproducible rate and for a reproducible length of time. (2) The HC1 was diverted from the sampling bulb, (at stopcock 1) and tpen turned off. The flask was evacuated, and then with stopcocks 4[, 9 and 11 closed the line was opened from the bulb to the flask. Stopcock 6 was opened, and the mercury was used as a piston to displace the HC1 from the bulb. (The mercury was allowed to rise oply to stopcock 3, which was then closed, as was stopcock 7.) (3) Tho high-speed agitator motor was started. The Hoesch-type impeller gfcve very rapid absorption of the gaseous HC1. Nitrogen was ad- ' mitted via stopcock 7 to keep the internal pressure up to atmospheric pressure. If the nitrogen was admitted intermittently, and the seal on the agitator was tight, it was possible to follow quali tatively the rate of absorption of the HC1 by the drop in pressure indicated by the manometer. This gave a qualitative indication of tpe rate of reaction of the scavenger with HC1. , (4) After 20 minutes agitation, the motor was turned off and enough j Pyranol was blown over into the test tube to cover the paper strips completely. The papers were allowed to soak in the Pyranol for 1 1/2 h^urs. They were then removed from the Pyranol, soaked 15 minutes in benzene followed by 15 minutes in methanol, and placed on a paper towel to dry. Tfhen dry, l/2 inch was cut from each end, leaving a strip 1M x 6M. Tthe tensile strength of the strips of paper were determined for us aft the Southern Research Institute, Birmingham, Ala. A Scott 1P2 Serigraph was employed, having a jaw separation of 3 inches and a rate of travel of 34*5 seconds for a load of 40 lb. Those papers t;hat did not break under 40 lb were cut to a narrower width and a suitable correction factor applied. Before testing, the papers wjere humidified at 65/6 R.H. and 70#F for a minimum of 48 hours. Since papers were treated in sets of six, the tensile strengths reported are the average of six breaks. In the earlier work of R. R. Knight, the gas buret and manifold on tne left of figure 1 were omitted, and only the nitrogen inlet, the vacuum line and manometer were employed. In the schedule of operations, after the flask was evacuated nitrogen was admitted to ar predetermined pressure (30 cm Hg below atmospheric pressure). Dry HC1 was then admitted until the pressure in the flask was the same as atmospheric. The valves were closed and the agitator started. There was one other difference in the procedure, in that tjhe nitrogen sweep of the test tube containing the papers was ( Omitted. MONO, 086502 -6- ( The feas buret and manifold were added to the Bystem because it was feltjthat with the method formerly employed, the variable rate of absorption of HC1 into the Pyranol was causing different total quantities of HC1 to be drawn into the system in different runs. In particular, scavengers which effected the surface tension would affect the rate of absorption* Also, scavengers that were particular ly r^pid in their action would cause more rapid absorption, and would thusjbo exposed to a greater quantity of HC1; they would consequently be penalized in comparison to slower-acting scavengers* The nitrogen sweep of the test tube was added because it was deemed illogical to take pains to remove oxygen from the space above the Fyranol in the flask, and do nothing about the oxygen in the test tube. The nitrogen sweep was found to have a very marked effect, (higher tensil strengths were obtained) but it was later found that an oxygen sweep had the same effect* It was eventually decided that the variation in tensile strength with gas sweep of the test tube was probably due to tne drying of the papers by the stream of gas. Water adsorbed on cellulose fibers might be expected to catalyze the degradation re action, both chemically and by promoting adsorption of HC1 molecules on the cellulose* No specific experiments were preformed to test this hypothesis, however* ` (2) Elec rical equipment ' (a) eneral Radio Capacitance Bridge type 716-C (b) eneral Radio Megohm Bridge type 544**BS4 (c) eneral Electric resistivity test cell for liquids (d) alsbaugh dielectric test cells 1 (e) iachar Isotemp oven (f) ielectric Strength test equipment, consisting of the following.' ,l) X-ray transformer, 8 KVA, 140 KV secondary voltage, 220 V. primary 2) Sensitive circuit breaker 3j General Radio Variac, type 50B 4) General Electric Liquid Testing Receptacle, Catalogue No* 2248-09* Electrodes 1" diameter, 0*100" gap* [5) Oeneral Electric A. C. voltmeter, type AR-2y53* Electrical tists were made on materials that gave favorable results in the tensile strength screening* , B. Results r; (l| Tensile Strength Screening Tests Table 1 lists compounds found to be as good as or better than tetraphenyl tin, or for which quantitative proof was desired that the compound was Inoperative. Compounds that were inferior to tetraphenyl tin are listed in Appendix A. The sources of the materials are given in Appendix C* Tensil strength data are reported as per cent loss in tensile strength. MOMS 086503 -7- TSgm gg<M-.rB Mnc ttst Table !+ 6 X 10 34 31 150 151 115 133 vn 104 122 74 ru 75 124 81 106 105 ' 108 53229 * . * e 69636 67635 . 67635 65004 65003 65604 67635 65004 67635 HUNS Gb65G* Table I-c Setraerr Soarnn Cone .1 tetralawyl tin r (Also 11) Diphenyl tin dichlorlde *. * . no an Ubut7l tin dlchlcrile a a ' '' ' NX 0.0$ 0.5 0.5 1.0 1.0 1.5 oT&5 0.50 1.0 1.5 1.0 i.d 1.0 1.0 i loss tana .sir. 105 90 UO 96 100 74 ltt 100 Prearart drop Lndieatad1 121 icm reactiai wltb 109 the HC1 0 IU 104 0 Vast No. KE s. . i r. ' 129 133 109 1U "ijr 134 111 135 152 i2 138 67535 * ` 671>J6 67635 67636 67636 67636 67636 Ditmtjl diphenyl tin Ttraphony1 tin FhaacxvTTOoena oxide Aluainua isopropyl*t Ifflril n * * * Table I-d XT NT S Tic ft (Also Si) * 0.125 0.125 0.10 0.1 c.oj 4.1 0.1 0.075 0.05 0.05 77 73 51 34 Vary rapid action 77 77 Kapid actioo 42 103 125 113 85 66 66 1 i. 60 | k iP . 54 52 49 51 6*004 65003 a . " mqns oaasus i I -9(i Lot a tensile strength of papers after exposure to Pyranol containing no scavenger, but which had been treated with HC1. Let b tensile strength of papers after exposure to pure Pyranol which had not been treats^ with HC1. Let x tensile strength of papers which had been exposed to Pyranol containing the stated concentration of the scavenger under test, and which had been treated with HC1. I | $ loss in tensile strength 100 This method j of treating the data applies equally well to the data obtained by Xnlght (with admission of HC1 by manometer pressure instead of by gas buret)* The only difference 1b that the correct value for the constant a was not so certain as jthen the gas buret was used. Also, the statistical fluctuations in a, and the systematic trends over a period of months, appear to have been more pronounced with the earlier method. The tensile[strength tests may be divided into 4 groups of experiments t (1) To establish the method, Table 1(a) (2) To discover new scavenger, (a) the earlier work of Knight; (b) the tests of the author; both in Table 1 (b) (3) To prove the broad claims of the 0. E. patent on tin tetraphenyl to be invalid, Table 1 (c). (Negative results in this phase of the study are listed in Appendix B.) (4) To verify and extend previous results (e.g. on dibutyl ( diphenyl tiA). Groups 2, 3 and 4 will be discussed under V, "Discussion*. (a) Estimate accuracy and significance of the method. It"is estimated that the precision of the tensile strength results in the middle range, as expressed in the table, is about $ to 15$. (In the earlier work, the precision may have been about 20 to 40$.) Near 0 and 100$ on the scale, the precision is better, probably about 3 to 5$. The termhaccuracy" has little meaning in these tests except as to comparative values. In the comparison of results on different materials, particularly when the runs were near together in time, the accuracy is probably about 5$* t MUMS 088506 - 10 - At 0.5% tin betraphenyl, the actual tensile strengths were very nearly the same in tests with and without the Na sweep. The percent loss was much lower when the Na kweep was omitted, because the value of a (the no-scavenger tensile strength) was very much lower, 3*5 lb, vs. 25.5 lb when the Na sweep was em ployed. (The1 value of b ranged from 51 to 53 lb.) It is not worth while to attempt at present to explain the difference between the curves in figure 2, since a variety of unknown effects are certainly involved -- such as the rate reaction of (scavenger with HC1, and the mechanism of attack of HC1 on the cellulose. | A 30 minute sweep of Na, with back pressure 2 cm, was adopted as the standard. A shorter sweep gave intermediate results. From figure 2 it may be seen that results depended strongly on Na sweep; it is thought that small fluctuations in the flow, whether or not they showed up in the back pressure, were probably re sponsible fok most of the statistical fluctuations in the tensile strength results. ! The two points in figure 2 indicated by circles, (tests 64 and 86) were ob tained with Ithe Na sweep. This indicates the range of uncertainty, referred to above, in the comparisons of tests run at relatively far apart times. We may take it (that a scavenger that gave, at 0.1$ concentration, a tensile strength losjs less than r?5%t was better than tetraphenyl tin. If the loss was above 85*, the scavenger was poorer. A number of different three-necked flasks were used for the 1-1 flask con taining the Pyranol. In the earlier work of R. R. Knight, no effort was made to use flasks of matched volumes. In the first work of the author, the same flasks were used, but in all the runs listed in Table 1 (except those of Knight), matched flasks were used, having the same volume within about 5 cc. The diversity or volumes of the flasks led to a much greater per cent variations in the volumk of the gas space above the fyranol, and hence larger variations in the partial pressure of HC1 from the solution. This contributed some uncertain ty in the constant a, and decreases the quantitative value of the early ex periments . i The quantity! of Pyranol employed was in most cases 1400 g. This quantity was chosen as standard because it gave the minimum gas space above the liquid, and hence caused the largest fraction of the HC1 to dissolve in the Pyrenpl. 1000 g. was used in the first experiments (and a few of the later tests when the quantity of scavenger was very limited) and several tests were run with 1300 g. Qualitative comparisons of loss in tensile strength are valid as between runs using different quantities of pyranol; but quantitative comparisons bf active scavengers should be made only with reservations. This is chielfly because of the question of molar excess: at a given scavenger concentration (say 0*l), the HC1 might be in a molar excess when 1000 g. of Pyranol, containing 1.0 g. of scavenger, was used; while the scavenger would be in excess when 1400 g. (i.e. 1.4 g. of scavenger) was used. This would not necessarily be automatically taken into account when the per cent tensile strength loss was calculated. (The difference between 1300 and 1400 g. of Pyranol is probably not serious.) ' HUNb 088507 -n - l; Tests 45 find 46 in Table 1, show the effect of the tine of agitation of the Pyranol solution after admission of the HC1. Evidently the reaction of tetraphenyl tin la very far from instantaneous. Consequently, for slow-acting scavengers, the'tlme of agitation plays a part in determining the precise value of the tensile strength result. Ibis factor was probably one of the lesser sources of error. (2) Electrical tests The results pf the various electrical tests that were made on these materials are shown in Table 2. (a) Thi significance of the tests per se is as follows) Dielectric strength is the measure of the ability of the fluid to prevent an arc between (charged or current carrying components of the transformer. The precision of this test is relatively poor, and variations of 5 to 10 Vsv on a single sample are not uncommon. The mechanism of dielectric breakdown of llquldo Is very poorly understood St present. It 1b known that n large number of factors are Involved, euch as the rate of voltage rise, and the (pacing, configuration, surface preparation, etc., of the electrodea. The effects of ionic contaminants, surface active agents, etc., are no doubt great, but there is little real knowledge. In the tests'listed in Table 2, only ethyl silicate can be regarded as at all ( elgnificantljf different from the other four materials; and it is not signi ficantly different from the specification value, 35 kv. (See Appendix D) Eoelstlvitv is the measure of the specific DC resistance of the fluid. It it in roughly an inverse relation to power factor (PF), which measures the AC conductivity* The Inverse relationship actually holds only when the conduction is due to ionic contaminants, and when the ionic materials are of the same kind In the different samples compared. When the ions are different (and of course each different additive (rings its own type of contaminant), all one can say is that an increase in PF generally accompanies a decrease In resistivity. Dow resistivity (or high PF) is harmful In a transformer because it leads to excessive heating of the fluid. In addition to specific electrode action, thermal degradation of the liquid and solid insulating material occurs, leading to yet lower resistivity and eventual arcing and failure. The accuracy of PF and resistivity measurements is probably better than 5% The measurements after 96 hr. at 100* (see Appendix D) are quite significant. Pyranol Itself always increases in resistivity (possibly because of water being driven off). A lowering of the resistivity shows that the additive will be degraded by heat. This would lead to loss of the protecting power of the scavenger and more important, the further loss of resistivity in eervlce, mentioned above. i ( '! ' HONS 08850a Tehle 2 Electrical Tesla Scnouer XDC* 1 Strength, KT 3-sletivity, can cs *10--9 After 96 ten. at. ioo* 60 Cycles lesist- * IvitT FF cne Tetra- H -thienyl tin Bl-d -thlerrrl tin Stabiliser 6-le 8UJ7I Silicate Fbenoij-prrjptne oxide Kane (Lot T-2) Dlbutyldiphenyl tin, greet. 0.1 0.1 UT 0.1 0.1 TZ-- 0.2 0.4 0.12$ --i--------- ----------a---------- a-----------1 "735 1st eit turn dibutyldlpbenyl .125 tin distillation Icrldlne " Paxeplex G-60 46 # n 34 oooe Stabilizer Ho. 21 a I6B ;t . 35a Aniline 0.1 0.1 0.1 0.1 0.1 1900 1950 0.65 95D 3.4 1353 1400 530 6300 90 0.65 n UOOO 0,50 3400 "ffrn-- 12 ------ 555---------- ~T5--------- 2.1 2250 1970 1.75 3.3 --T5---------- 566 22 53 3 12.7 210 ?J.0 1480 < 2.0 7.7 554 / 5.1 f 45 . - enarfca IB p. 504^7 54510 fyranol lot T-2 used f realstirtty testa 65044 56538 . 50407 50420.3 61580 59834 This aaterial aas shipped to Veatiagboua 61560 59838 Solution developed a slight yellow color on hastily. Booted only 72 txa. Scaranger treated with sttspulyoa earth baiore^ 61572 64375 65005 6&166225 69677 None Advance Sclvsits Stabiliser! 3 3. solnticn earth-treated 3 3, solnticn earth-treated ' 52 >2, solution earth-treated 52 52, solution airib-treated 0.1 0.1 1.0 1.0 0.1 0.1 1.0 1.0 4970 20.7 1760 1.6 K>35 21 z33 - 1.9 8.3 1.7 4.4 3.8 2.8 59 5.3 1 'I 74822-6 HONS 0dd509 - 13 - While the resistivities listed in Table 2 are generally far above 0. E. specifications (of Appendix D)> it should be noted that the observed re sistivities bf Pyranol 1467 have ranged from barely-meeting-specification (100 x 10* o^im-cm.) to above 10,000 x 10^ ohm-cra. It is clear that scaven ger? that arfe more harmful to the resistivity will have less of a chance of finding rcceptance. (3) Hydrolysis tbsts of Organometalllc Compounds Very rough testB were made of the susceptibility to hydrolysis, on several scavengers* Information was desired on this point because transformer fluids are ordinarily exposed to some moisture in handling; it would be very un desirable tojhave to insist on special precautions against moisture. Solutions of.about 0.1 to 1$ concentration of scavenger in Pyranol were ex posed, in open test tubes, to air of about 45 and 60 RH. This was quite a mild test, sb in addition, in separate tests a drop of water was added to about 5 or 10 cc of solution and the tube shaken* The tubes were examined for signs of a wskimH jin the surface, or a precipitate. Scavengers fpund highly susceptible to hydrolysis; Dibutyl tin diethylate Diphenyl ^in dibutylate Aluminum butoxide Tetrabutyl titanate Tributyl borate ' Scavengers obly slightly to very slightly susceptible to hydrolysisi Tetraethyl silicate Advance Solvents Stabiliser 3, 21 and|52 Diphenyl tin diethylate Aluminum isopropylate It is thought that the lower susceptibility of diphenyl tin diethylate and aluminum iso^ropoxide may be related to their low solubility. The homologues in the first]list were procured in hopes that they would bo as rapid scaven gers, and soluble as well. Rapid they were, but also hydrolyzable. The Advance Solvents stabilizers 3 and 52 are polymers formed by splitting out alcohol from|compounds of the type dibutyl tin diethylate. By virtue of their polymeric nature, they were, apparently, able to stay almost entirely in solution irrespective of the attack of water. Ethyl .silicate showed only a very faint hAse on the third day after a drop of water had been added. (4) lestinghouse Copper-Corrosion Test Westlnghouse in 1949 turned down three very active conveners which we sug gested to th^m (Monsanto products known as Methasan, Sopanox and SA326) on the basis of[the following test; A burnished strip of copper is heated at 100* in a solution of the proposed scavenger in Pyranol, for four days. It was required jthat there should be no darkening of the copper. . muais oaasio -U - Compounds thjat were satisfactory under this test were : Dibutyl diphenyl tin Phenoxypropene oxide Aluminum lisopropylate Diphenyl tin diethylate Paraplex G-60 Di-i -thienyl tin Ethyl silicajte end also ecridine did not tarnish the strip; but with ethyl silicate a haze developed in the solution, and with acridine the solution developed a yellowish color. With a solution of dimethyl glyoxime, the submerged pdrtion of the strip remained bright, but the portion above the liquid was badly tarnished. Tetra-l -thienyl tin produced a net/ effect: a thin layer cjf a bright yellow or brassy color appeared on the copper. It was as if sdme tin had plated out. The strip was left in contact with the solution fojJ about 18 months at 25*; when examined again, the brassy color had been replaced by a brown tarnish. (The strip in contact with di thienyl tin [remained bright.) Tri-{-thienyl stibine produced a dark film on the copper. J (5) Comnetibilitjy of Scavengers: Phenoxypropene oxide was **ound to be compatible with tetraphenyl tin in solution in fyranol. (Information on this point was requested b^ Meetinghouse.) As would be expected, no precipitation from solution wad observed, nor was there interference in HC1 scavenging action. (6) Odor of Pheroxypropene Oxide in Pyranol: Solutions of 0.1$ and 0.3$ Phenoxy propene oxice in Pyranol 1488 were tested by a sniffing panel of four chemists It was observed that the odor at 0.3$ is slight but perceptible when compared with straight Pyranol. No distinguishable difference was noted at 0.1$. On heating to !0*C, the trichlorobenzene odor practically completely mashed the phenoxypropene oxide odor even at 0.3$. It was concluded that it would be very difficult to detect the odor of phenoxy propone oxide unless a comparative sample of straight Pyranol was also sniffed V. Discussion A. Tentative c:i.*iteria for a "good scavenger" * (l) A good scav^tinger must reac-t. rapidly with HC1 in the Pyranol medium to give an innocuous pr<oduct. T. L. Dakin of Meetinghouse has expressed the opinion (see trip report no. 722) that tetraphenyl tin is not a sufficiently rapid-acting scavenger Some time ago they arced several transformers in order to obtain an on-the-,Jfb test of tetraphenyl tin. Dakin said the results were quite equivocal to whether tetraphenyl tin .gave any appreciable protection, Hence they ere really the most interested in scavengers that gave more and quicker protection. Unfortunately, this positive criterion, that the scavenger be active and rapid, tendi to be incompatible (in terms of chemical structure) with most of the other (Negative) criteria listed below. For example, the amines react very rapidly with HC1, and can be had with very low equivalent weight. But they are generally harmful to the electrical properties of fyranol, and are far from bej.ng sufficiently stable. 0dd5li - 15 (2) Thu electrical properties of Fyranol (resistivity, power factor and dielectric strength) cannot be seriously affected. This eliminates materials that ionise strongly or that cannot be freed from ionising contaminants. (3) The scavenger must be stable indefinitely at elevated temperatures. No definite temperature^limits can be set, but we can guess that temperatures above 130* are quite unlikely. The temperature 100* for 96 hours i3 specified by 0. E. for an accelerated^ test. (See Appendix D) (4) The scavenger must be unreactive towards all materials normally found in trans formers and' in the equipment used in handling Pyranol. This is the reason for the copper-corrosion test (see p. 13). Water is included under this heading, because it is ordinarily difficult to exclude water totally from the systems for handling Pyranol, and it is absorbed from the atmosphere when Fyranol is exposed to moist air. (5) The scavenger must not have a high vapor pressure from solution in Pyranol at sievet&c temperatures. As a rough guess, we may exclude any substance with a normal boiling point below 100*; and boiling points above 150* should be satisfactory. (6) The scavenger must be soluble in Pyranol. The solubility of tetraphenyl tin, about 0.5% |at 25*,is considered to be too small, because the scavenger tenda to como out- of solution at low temperatures and because it prevents the use of higher concentrations for extra protection. B. Discussion of Compounds Tssted Active Scavengers of known composition Tetra-| -thienyl tin and dl-( -thienyl tin. These compounds were found to be considerably more active than tetraphenyl tin. In the only electrical test made, they had no harmful effect on the dielectric strength. Tetra-(-thienyl tin was unsatisfactory according to the cupper-corrosion test. (b) Ethyl orthosilicate. This was a very rapid-acting scavenger. It failed to meat the thermal stability and hydrolysis teste, though it did not fall down extremely badly. <o> Tetra-2-furyl tin was a very active scavenger. There was not sufficient sample left after tho tenelle strength screening test (and testing at St. LouIb as a polyvinyl chloride stabilizer) to run any further test. The yield in the preparation had been very low. (a) Dibutyl tin diethylate. Very rapid action as a scavenger. Very susceptible to hydrolysis. <) V-butyl glypydyl thioether. The presence of sulfur in the compound is highly undesirable, as leading to unstablllty and reactivity toward copper. No furthor [teats were mode on this material. (f) Advance Solvents Co. poly alkyl alkoxy tin stabilizers, Nos. P-1276, 3 snd 52. Very active and rapid. Only very slightly susceptible to hydrolysis. They did poorly on the thermal stability test, and were harmful to the electrical [properties. Treatment with attapulgus earth produced a satis factory resistivity in the case of No. 3, but did much lees good for No. 52. muns aaasu - 16 - No. P-1276 vjas a special sample prepared by advance Solvents Co., to specifi cations t approximately the tetramer of ,'fiuC[SnBu^O| nBu. It was not as soluble as Ifos. 3 and 52. Since Nos. 3 and 52 were the only ones commercially available at present, the further tests beyond the tensile strength screening were made oii them only. The composition of Nos. 3 and 52 is not known beyond general formula, R0[5nRa2l ftR. (g) Meta-tolueni-bis-ethylene urea and p~aminoa2obenze are of molecular types known to be jrelatively unstable and harmful to the electrical properties of Pyranol. No further tests were made on them. (h) Carbide and Carbon G-18 (glycidyl oleate) and A-5 ("glycydyl ether of di phenyl ol propane")* .. (CjHe) (jQ-O-CHjCHGH,), Although the stated composition of A-5 sounds like the mono-ether, that sub stance would probably have too high an equivalent weight to show much scaven ger activity. Hence we write the di-ether. Ho further tests were made on these, beyond the dielectric strength of 0-18, which was satisfactory. Since A-5 was active, it is anomalous that Epon 1164-1 (see Appendix A) was found inactive. (It gave a 9 55* loss in tensile strength in a test by the early method.) Possibly this was because of the lower accuracy of the earlier teste. (2) Active Scavengers of Unknown Composition Two of the three materials listed here are believed to be wholly organic polymers. dha manufacturers have not disclosed their composition, to our knowledge at this writing. Since the Patent Department advises that we will have difficulties obtaining legal protection when we cannot specify the com positions, these materials were not pursued as thoroughly as others. (a) Paraplex 0-60. This appears to be very rapid in its action. It does not affect the electrical properties seriously, and stands the thermal stability test satisfactorily. (b) Advance Solvents E6B. A little more harmful to the resistivity then was Paraplex G-60. Stands thermal stability test well. E6B is believed to be an epoxy compound. (c) Advance Solvents No. 21. This is described as the cadmium salt of No. E6B. It caused a serious drop in the resistivity and did poorly on the thermal stability test. (3) Proof of Nod-validity of the Broad Claim of the 0. E. Tetraphenyl tin Patent The broad claims of ,U. S. Patent 2,468,544 which might.be considered to dominate our patent on dibutyldlphenyl tin are as follows( *2. Anormally liquid dielectric and insulating composition consisting essentially of halogenated aryl hydrocarbon and containing in solution by weight about one-fourth to one per cent of organo-metallic oompound having the formula in which U is a metal chosen from the group con sisting of tin, lead, and mercury; R is an organic, radical chosen from the group consisting of aliphatic and aromatic radicals; combinations of such radicals ant' radical containing substituents; and x Is an integer in thq range of 2 to 4* MGNS 0*8513 - 17 3. A formally liquid dielectric and Insulating composition consisting essentially of halogenated aryl hydrocarbon and containing in solution by weight about one-fourth to one per cent of organo-metalllc compound having the formula MRXY in which U is a metal chosen from the group consisting of tin, lead, and mercury; R is a radical chosen from the group consisting of aliphatic and aromatic radicals, combinations of such ridicals, and such radical containing substituents, x is an integer in the range of 2 to 4, and I is an inorganic substituent chosen from the grcu: consisting of halogen and hydroxyl attached directly to the metal. 10. A liquid dielectric and insulating composition including chlorinated diphenyl as an ingredient and containing in solution by weight about 0.0$ to 1 par cent of an organo-oetallic compound of a metal selected from the group consisting of tin, lead, and mercury, the metal being attached directly to a carbon atom of an aromatic radical.* Claims 2 and 3 clearly cover tetralauryl tin. The results of the tensile strength screening tests, table 1-c, show that in the range up to 1.0}, tetra lauryl tin ijs Ineffective as a scavenger. At 1.5}, it shows some activity, though this Iresult has not been confirmed. Kharasch's scries (see literature reference 10) predicts that ease of cleavage by |HC1 decreases with increasing alkyl chain length in alkyl tin compounds, we can be confident that a yet higher homologue, say tetraoctadecyl tin, would be even less active, or would show activity only at concentrations considerably higher than 1.5}. The increase in equivalent weight ggf g would certainly rejduce the scavenger activity, and the Kharasch effect makes an additional contribution in the same direction. Hence we can conclude that claims 2 and 3 of the 0. E. patent are not valid. ' The data on Biphenyl tin dichloride and dibutyl tin dichloride prove the nonvalididity of other of the broad claims, including Ho. 10 in particular. It may be seen from Table 1-c that their action is, if anything, distinctly harmful. Ip the absence of HC1, no harm was done. Since the pressure drop indicated by1 the manometer was greater than that when no scavenger was used, the two substances must have reacted with the HC1. A possible mechanism is the following reaction' R.SnClgl* 2HC1 -------------- > H.(R.SnCl4) -------------- 2RH + SnCl4 | rapid slow The chloroalkyl stannic acid postulated should be strong enough to attack the paper. If a substance such as this is an intermediate in the reaction of tetraphenyl tin with HC1, it must decompose rapidly. In the present ease, it must persist for at least two hours, all the while increasing the effective concentration of HC1 in the solution. (4) Verification and Extension of Previous Results Table 1-d shoes, first, that dibutyldiphenyl tin is very slightly less effective as| a scavenger than is tetraphenyl tin. The difference, is to small, however, to be considered significant. MUNS 088514 - 18 - Phenoxypropeis oxide is definitely more active than tetraphenyl tin. The difference ii certainly meaningful in thia oaee. , Aluminum isobropylate is considerably more active than tin tetraphenyl. It was also a notably rapid scavenger. Westinghouse raised the objection that the final reaction product of this compound with excess HC1 would be AlCla> and AlCla Is'known to catalyse the decomposition of Pyranol with the evolution of yet more HC1. The single experiment at 0.05# was undertaken in order to teat this objection. The result was negative, i.e., aluminum isopropylate was still beneficial at the lower concentration, although KC1 was evidently present in excess. These experiments were not pursued, because of the other drawbacks to aluminum isopropylate> its low solubility and appreciable susceptibility to hydrolysis. |Besides, it was felt that a very extensive series of experi ments would be necessary to convince Westinghouse, and the prospects of ob taining good|enough proof were by no means certain. Acridine was!found to be a rapid-acting scavenger{ but, eurprielngly, at low concentrations it was found to be harmful. There is no obvious mechanism for this effect. In addition, acridine gave poor results in the thermal stability teat by producing a decrease in the resistivity, and the solution became colored. (5) Results Reported by Westinghouse . Westinghouse has reported that, as scavengers, dibutyldlphenyl tin and phenoxy- propene oxide are satisfactory. Negotiations are being carried on at present through the ^ales department. Westinghouse is carrying out a study of the kinetics of the reaction of these and presumably other compounds with HC1. This is partly with the object of verifying our report to them that tetra- lauryl tin is a non-scavenger, and that the broad claim of the 0. E. tetra phenyl tin patent is invalid. They have stated that they are still Interested In new rapid scavengers. We have sent them samples of Advance Solvents Stabilisers Nos. 3 and 52, as well as tetralauryl tin. Aa mentioned |on p. 3, Westinghouse purchased 55 gallons of Pyranol with di butyldlphenyl tin scavenger in January, 1950, and a tank car of pyranol with phenoxypropsne oxide scavenger in January, 1952. VI. Conclusions | Phenoxypropsne oxide and dibutyldlphenyl tin are the materials that have shown up best, acceding to the criteria listed under 2-A. As might be expected, we must compromise and sacrifice extremely rapid scavenging action in order to satisfy the other criteria. Particularly in view of Westlnghouse's acceptance, we may conclude that these compounds constitute a satisfactory end for our search. i MGNS uaobi5 - 19 - Next best, after the above two scavengers, rank the organic polymers (see VB(2) ) Paraplex G-60 was the most extensively tested, and would very possibly be satisfactory to Westinghouse Advance Solvents &6B and Carbide and Carbon 0-18 and A-5; might very possible be found satisfactory on more extensive tests. These have not been pursued because the Patent Department advises us that we will probably be unable to obtain legal control of the materials in the absence of knowledge, of their structures* i. Thera are various drawbacks to the other materials discussed above -- In acme cases absolutely prohibitive, In others only possibly* or *probably* fatal. It is not prjofi table at preeent to attempt to rank these materials in order of merit. VII. Rscoamendatlbns In view of the favorable reports from Weatlnghouse on phenoxypropene oxide end dibutyldiphenyl tin, it is recommended first that we devote no more time to the search for nlew scavengers. Second, if fpr any reason it should bs decided tc. reopen the project, it is recommended {that we concentrate first on the polymeric type scavengers. lither we should offer said scavengers to Westinghouse after a minimum of further testa of electrical properties, etc.,* or else we should attempt to duplicate these materials ourselves. The latter suggestion is made because it ia be lieved that Paraplex 0-60 is not specifically protected by patents. Eoha and Haas (its manufacturer) are reported to have declined to reveal its composition, and to hava intimated that the proprietary formula and *know how* are their only protection. (The Paraplex resins are believed to be of the polyester type.) It is reportjed that some epoxidlzed polyester resins ere being prepared it Dayton. These should| be investigated if the scavenger Job is reopened. . VIII. Description of Recommended Process Does not app, 4IX. Patent Stati The Patent Department has expressed the opinion that, in view of the inactivity of tetralauryl tin and diphenyl tin dichloride as scavengers, the broad claims Nos. 2, 3 and 10 of the 0. E. tetraphenyl tin patent are Invalid. Hence we are free to sellj Pyranol protected by dibutyldiphenyl tin. . They have also expressed the opinion that the 0. S. application of Q. E., corresponding to British Patent 601,359, covering phenoxypropene oxide as an HC1 scavenger, will never issue. Bence we are free to eell Pyranol protected by phenoxypropene oxide. The following patent applications have been filed, and those that have been issued at the date of this writing ere noted. MONS 088516 - 20 - 1. Case A-36 Russell L.. Jenkins - Ser. No. 691,652 - Filed Aug. 19, 1946 STABILISATION OF HALOGENATED CROANIC COMPOUNDS WITH DIBUTYL DIPHENYL TIN This application covers the use of dibutyl diphenyl tin as a scavenger for Pyranol. Patent No. 2.578,359. Issued Dec. 11, 1951 2. Case A-J54 Edgar E. Hardy - Ser. No. 38733 - Filed July 14, 1948 COMPOSITIONS OF MATTER COMHUSINO HALOGENATED ORGANIC COMPOUNDS. Tills application covers the use of blguanide and its substituted de rivatives as scavengers for Pyranol. 3. Case A-fe6 Elwood F. Jackson - Ser. No. 49986 - Filed Sept. 18, 1948 DIELECTRIC WITH N,N'-l-3 DIMCRPHOLINO ISOFROFANOL AS SCAVENGER I This application covers the use of morpholine and its substituted de rivatives as scavengers for Pyranol. Patent po. 2.572,808. Issued Oct. 23, 1951 4. Case A-J58 Edgar E. Hardy - Ser. No. 53695 - Filed Oct. 9, 1948 COMPOSITIONS OF MATTER COMIRISING HALOGENATED ORGANIC COMPOUNDS. Patent No. 2,532,616. Issued Dec. 5, 1950 This application covers the use of ainc dithlocarbamete and derivatives thereof as scavengers for Pyranol. Case A j59 Edgar E. Hardy and Robert J. Slocombe - Ser. No. 53696 - Filed Oct. 9,1 1948 DIELECTRIC COMPOSITION OF HALOGENATED AROMATIC COMPOUND AND ALUMINUM ISOPROP'PLATE AS A CORROSION INHIBIT. Patent B.566.195. Issued Aug. 28, 1951. This application covera the use of aluminum lsopropylste as a scavenger for Pyrjanol. 6. Case A-t>0 Russell L. Jenkins - Ser. No. 54258 - Filed Oct. 13, 1948 DIELECTRIC COMPOSITION OF HALOGENATED AROMATIC HYDROCARBON AND GANIC ANTIUONN'Y COMPOUND AS A CORROSION INHIBIT. Patent B,566,208. Issued Aug. 28, 1951. This application covers the use of stibines ss scavengers for Pyranol. Case A-61 Edgar E. Hardy - Ser. No. 57968 - Filed November 2, 1948 COMPOSITIONS OF MATTER COMPRISING HALOGENATED GANIC COMPOUNDS. This application covers phenoxy propene oxide, butadiene monoxide and styrene oxide as scavengers for Pyranol. MONE 088517 - 21 - l 8. Case A-62 Edgar E. Hardy - Ser. No. 67662 - Filed Dec. 28, 1948 CHLORINATED AROMATIC DIELECTRIC WITH SCAVENGER Patent tjo. 2,566,196. Issued Aug. 28, 1951 This application covers the use of aminated-N-phosphoryl-o-amino biphenyl as a scavenger for Pyranol. I 9. Case A-63 Edgar E. Hardy - Ser. No. 67663 - Filed Dec. 28, 1948 COMPOSITIONS OF MATTER COMPRISING HALOGENATED ORGANIC COMPOUNDS. This application covers the use of diphenyl guanidine as a scavenger for Pyranol.| 10. Case A-64 Edgar E. Hardy, Elwood F. Jackson and Robert J. Slocombe Ser. No.1 67664 - Filed Dec. 28, 1948 COMPOSITIONS OF MATTER COMPRISING HALOGENAIED ORGANIC COMPOUNDS This application covers the use of pyridine and its substituted derivatives as scavengers for Pyranol. i 11. Case A-65 Edgar E. Hardy and Elwood F. Jackson - Ser. No. 67665 Filed Dec. 28,' 1946 COMPOSITIONS OF MATTER COMPRISING HALOGENATED CROANIC COMPOUNDS. ( This application covers the use of dimethyl glyoxime as a scavenger for Pyranol.| 12. Case A-66 Edgar E. Hardy and Elwood F. Jackson - Ser. No. 67666 Filed Dec. 28,' 1948 COMPOSITIONS OF MATTER COMPRISING HALOGENATED ORGANIC COMPOUNDS This application covers the use of o-anisidine, p-anisidine, p-nitro aniline and 2,4-jdiamino diphenylamine as scavengers for Pyranol. 13. Case A-67 Russell L. Jenkins - Ser. No. 67388 - Filed Dec. 27, 1948 COMPOSITIONS OF MATTER COMPRISING HALOGENATED ORGANIC COMPOUNDS. This application covers the use of diphenyl tin diethylate as a scavenger for Fyrdnol. 14* Case A-82 Buford H. Hartaog - Ser. No. 110,518 - Filed Aug 16, 1949 COMPOSITIONS OF MATTER COMPRISING HALOGENATED ORGANIC COMPOUNDS Patent l|o. 2.582,200. leeued Jen. 8, 1952 Thla application covers the use of ethyl silicate as a ecavenger for Pyranol. 15. Caaa A-83 Arthur M. Ellenburg - Ser. No. 109,576 - Filed Aug. 10, 1949 COMPOSITIONS OF MATTER COMFRI3INO HALOGENATED ORGANIC COMPOUNDS. ( Patent l|o. 2,573.894. Issued Nor. 11, 1951 This application oovera the use of thienyl tin as a Beaver-*!- for Pyranol. HUNS 088618 -22 - 16* A-106 irthur M. E^lenburg and Robert J. Good - Ser. No. 260,1$6 - Filed Dec. 6,|l951 COMPOSITIONS OF MATTER COMfillSING HALOGENA TED AROMATIC COMPOUNDS. This spill cation covers the use of a polyalkyl alkoxy tin compound as a scavenger for Pyranol. An application is in preparation on tetra-2-furyl tin as a "new composition of matter", having utility as an HC1 scavenger. . X. Cost Estimates Not applicable XI. Material Specifications, Analytical Procedures Sob Appendix D for specifications for various pyranols. See Apendices E and F for the manufactures' specifications on dibutyldlphenyl tin and phenoxypropene oxide. See literature reference 12 for method of analysis for epoxides. The electridal tests mere made according to the following methods (see the Di electrics file)< (a) "Procedure for Testing Dielectric Media for Dielectric Constant and Power Factor", 9-1)4-50, by R. J. Good. (b) "Procedure for Dielectric Strength Testing", 10-4-48, by R. R. Knight. (This has since been revised, mainly by the addition of more detail in the in structions, py R. J. Good, 12-27-50-) XII. Toxicity and Hazards Does not apply XIII. Acknowledgement The author wishes to express his thanks to those who carried out the earlier phases of this study (see references 1 to 7)J to Dr. R. L. Jenkins and Dr. B. E. Hardy For their active interest and encouragement throughout the project) and to Centrjal Research at Dayton for their helpful suggestions and for the synthesis of compounds that would otherwise have been unavailable. MGN5 088519 I - 23 APPENDIX A Compounds found to be less active scavengers than tetraphenyl tin, by tensile strength screening test* I lead monometHyl triisobutenyl succinate Carbezole Anethole laoeafrol lugenol Santover 0 Hydroquinone Benzil Stabilizer JC X XI 12 * ill 50B n XI 14 Plastoflex VS Dioctyl dloctoxy ellane Tetrabutyl titanate Tribntyl bori te Tetre-2-ethylhsxyl silicate Dimethyldiethoxy ellane liethyl tiietnoxy ellane Amyl * Phenyl * * Tetrabutyl el llcate p-oonylphenoxypropene oxide Tri-p-ehlorophenyl etibine Methyl-9,10-epoxyetearate Dimethylphenylglycldate Slntol T (condensation product of toluene and eulfur by Priedel-Crafts reaction) Epon 1164-1 ( x,x'-bis-glycydyl biphenyl) 2-chlorothioj hene . 2,5 dlchloroihlophene 2,3,4,5, tatrachlorothlophene 4,4' diphenyl methane-bis-ethyleneurea Melamine Triphenyl phosphine 2,7 dlhydroxj naphthalene Stannic octyl phosphate Oleic acid Dlbenzyl Tin olsate Camphene j Antimony sail of t-dodecyl mercaptan Antimony Trl< xlde Source HD I r EE IK 11 IK IK A A A A A HD AL AX, HD AL AL AL AL AL HD HD HD HD HD S sir sir sv he H or IK EK IK HA T EX T SK H H and also B MQNS 088520 APPENDIX B Compounds covared by the broad claims of the Q. E. tetraphenyl tin patent, found to be]active scavengers. Diphenyl mercury tlbutyl * Tetrabutyl tl; Trlphenyl tin chloride Tetra - (j tflpt;enyl) tin Tetraniethyl tin Tetrabenayl tin ! B B B S IS iS S Source APPENDIX C Code of manufacturers- A AL B CC EK F H It MA 11D I UK VC H RH S sv TEC Advance Solvents Co. Anderson Laboratories, Adrian, Michigan J. T. Baker Carbide and Carbon Eastman Kodak Fischer Scientific Co. Harehsw Monsanto Monsanto, Anniston Monsanto, Dayton Monsanto, Everett Metal and Thermit Co. Heilly Tar and Chemical Co. Rohm and Haas Shell Chemical Co. Socany-Vacuum Tennessee Eastman Corp. moms 0085^1 - 25 APPENDIX D No. 1488 Pyranol Specifications Composition i Viscosity |at 37.8*C. - 60* No. 1482 Fyranol (Aroclor 1260) by weight . - 40* No. 1478 Pyranol (trichlorobeneene) by weight - 54" 2" Saybolt Universal Ornvity atj 15.5/15/5*C. Color (APlj) I Acidity | -1.560 - 1.568 - 150 maximum - Less than 0.01 mg NaOH/gram Free chlorides (H-6845480) - *10 ppm max. Fixed chlorine I Burn Pointj - 59.1/G minimum - None up to the boiling point Pour Pointj - Lower than -32*C. Distilling Range 1st Drop Below 1270 *C. 9056 point - 200*C. minimum - 1+0% maximum - 385 - 400*C. Corrosion (H-3954201) - After heating with aluminum for six hours at 200-220*C., the aluminum must not be corrodei either on visual or weight Inspection and the Pyrsnol should meet the following apeciflcationst- Color.................... 200 APH Acidity................. 01 Free Chlorides. .10 ppm max. Condition...... Clear Chemical Stability - (l) No generation of free chlorides when refluxed with distilled water in presence of air for 6 hours. (2) No free chloride generation when heated with distilled water in sealed glass bomb at 150*C. for 6 hours. Arc-formed Gases - (l) Less than 1.0 total combustible ( (oxygen4f*ee liquid at 25*C.) gases including carbon monoxide, hydro- gen and volatile hydrocarbons. MUAIS 088522 - 26 - Dielectric Constant at 1000 cycles at 100*C. . , at 25*0. 1.5 - 3.8 .0 - 4.3 Analysis of Products of Fractional Distillation When fractionally distilled, the distillate shall be separated into 10-cc boiling fractions. Ho frac tion on analysis shall shoe lass than one chlorine atom for each hydrogen present In the mixture. This shall be determined by chemical analysis. Ho fraction of the distillate so col lected shall give more than 1.5$ total combustible geses on arcing (Including carbon monoxide, hydrogen and volatile hydrocarbons). Refractive!Index at 25*C. 1.6137 - 1.6147 Ho. 1467 Fyranol Specifications Composition Tin Tetraphenyl content Viscosity it 37.8*C Sp. Cr. at 15.5/15.5*0 Color APHA Acidity, mgm. NaOH/gm Free chlorides Fixed chlorine Condition ! Burn point I Pour Point I Distilling Range 1st dtp 40$ 1 90$ , Resistivity at 100*0 Dial. Constant et 100C cycles 100*0 : 25*C Refractive |lndex at 25*0 later ppm I Corrosion (H-3954201) 59.5* 0- S. 1482 (1260) 40.5* 0. B. 1478 (TCB) plus 0.125$ Tin tetraphsnyl by wt. 0.125$ t 0.01$ 52-56 Saybolt Universal Seconds I.56O-I.568 150 max. 0.01 max. 0.1 ppm max. 59.1$ min. Clear Hone to boiling pt. Loser than - 32*0 200*0 min. 270 or higher 385-400*0 100 x 10 ohra om minimum 3.5-3.8 4.W.J ""S ""5 1.6137-1.6147 <30 - after heating with aluminum foil for 6 hours at 200-220*0, the aluminum must not be corroded either on visual or weight inspection and the Pyranol should meet the following epeciflcatione. - 27 ( Arc-formecj gases Color Acidity Free chloride Condition 200 A.P.H. .01 Uax. 5 ppm Max Clear (oxygen>iree liquid at 25*C) Lees than 1.0$ total combustible gases including carton monoxide, hydrogen and volatile hydrocarbons. Dielectrid Strength at 25*C I at 100*C Greater than 35 KV. * 35 KV. ' Electric^ Stability after heating 1467 Pyranol for 96 hours at 100*C in a closed container, the resistivity shall not be less than 25 x lQ^ ohm ems at 100*C or have a decrease in resistivity over the original greater than 10$. The container shall contain 85$ of its volume as 1467 Pyranol and 15$ of its volume aa air. Analysis of Products of Fractionni Distillation ( When fractionally distilled, the distillate shall be separated into 10 cc. boiling fractions. No fraction on analysis shall show less than one chlorine atom for each hydrogen present in the mixture. This shall be determined by chemical analysis. No fraction of the distillate so collected shall give more 1.5$ total combustible gases on arcing (including carbon monoxide, hydrogen and volatile hydrocarbons). }Jo. 1478 Pyranol Specifications (Trichlorobenzene) Composition a technical mixture of TCB Viscosity at 37.8#C. 28 to 32 Sabolt Universal Seconds Gravity at 15.5/15.5*0 1.460-1.477 Refractive Index at 25*C 1.5680-1.5715 Color | 15 APH Max. Free Acidity less than 0.01 mg. NaOH/gm. Free Chloride 0.10 ppm max. Fixed Chlorine 58.5$ minimum Burn point None to boiling point pour Poirot (Freezing Point) Lower than 10 *C Condition! . Clear Distilling Range (Corrected) 1st drop 205#C. min. 5$ point ( 50$ jpoint 210#C. min. 213#C. min. ' 90$ point 215*C. min. Water j 75 ppm Dielectric Strength Greater than 30 KV HQNS 088524 Corrosion Arc-Formed Gases Chemical Stability - 28 - after heating with aluminum foil for 6 hours at 210*C 4 10#C, the aluminum must not be corroded either on visual or weight inspection, and the Pyranol should meet the following specifications! Color 200 APH max. Acidity (mg NaOH/gm) 0.01 max. Free Chloride (ppm). 0.10 max. Condition . Clear Total combustible gases (Including carbon monoxide, hydrogen and voltatile hydro carbons) --less than 1.5*. (1) No change In color or acidity, no formation of free chlorides when refluxed with water In the presence of air for 6 hours. (2) No free chloride generation when heated 6 hours In a glass sealed bomb with distilled water at 150*C. (3) No free chloride generation on exposure to direct sunlight for a minimum of 30 days at normal tempera ture. No color or acidity increase from original value. No. 1482 Fyranol Specifications (Aroclor 1260) Burn Point (ASTM D-92) Viscosity at 98.7'C. (ASTM D-88) Specific Gravity at 90*C/15.5*C- (asTM P-287) Pour Poln (ASTM D-97 Color Acidity NaOH/g Chlorides I Sulfates (ASTM D-117) Svaporaticn at lOO'C.for 6 hrs. Distillation Range (ASTM D-20) Refractive Index, 25*C. Stability Fixed Chlorine Content (Carlus) Resistivity 100*C Greater than 350*C. 73-78 S.U.S. 1.550-1.560 30*C 4*C. 150 aFh max. less than .01 0.10 ppm max. None less than 0.2* 10* Point 370-377*C. 50* Point 377-385*0. 90* Point 385-400*C. 1.6460 + .0005 There sKall be no liberation of chlorine or chlorides when the material is heated to 100*C. in a glass vessel in contact with air for periods of at least one month. 60* 4 1/2* 500 x 10' ohm cm. min. MONS 0885^5 Corrosion1 j ! , Dielectrii Strength 50*C. 100*C. Water conient ppm. - 29 after heating with aluminum for six hours at 210*C 10*C., the aluminum must not be corroded either on visual or weight inspection and the Pyranol should meet the following specifications Color APH 150 max Pres chlorides 0.1 ppm Acidity mg NaOH/gm .01 max Condition Clear 30 K.V. min. 30 X.V. min. 33 max. MGNS 088526 - 30 - APPENDIX E OLYCIDYL PHENYL ETHER (Phenoxypropene Oxide) __oH HCHC Molecular Formula *1----------- Dolllntr Pojnt Moltinp Point Flash Point (Tan Open Cup) Specific Oravity (vac.) 245*0 at 760 mm. 3.5*0 175 *F = 79*C dO z dio 4 d22 4 Molecular Weight 150.17 1.1273 1.1183 1.1092 1 Refractive |lndex d22 1.1001 4 a2 1.0913 4 Coefficient of Expansion (22.5*0' (27.5*C; n20 1.5314 ' D ! 'f 1.5269 1.5428 1 Refractive Dlperaion Viecoeitv n2 1.5290 (iy-nc) x 104 - 159.2 t*C Centlpolee 20 7.05 30 4.88 MONS 08852 7 - 31 - l I. Surface Tension t*C Dynes/Cm. Solubii ity at 20*C,fo 20 41.7 0.24 compound in water 0.4 water In compound 12.9 compound in octane 10.0 octane in compound Miscible with acetone and toluene Vapor Pressure t*C P mm. of Hg. 120 11.41 130 18.25 040 28.30 150 42.67 160 62.69 170 89.99 180 126.47 190 174.3 200 235.9 ( 245 760 log p(mm.) 6.97812 - 1657.89 tm`rsCTBB Toxicity ] j | 1 , 1 : I j Teste on animals show glycidyl phenyl ether to te a toxic material and to produce phenol type burns after exposure to the skin. The vapor hazard at normal temperatures is not serious due to its low vapor pressure. Adequate precautions should be observed in handling this material to prevent spillage. If accidental spillage does occur, the affected area should immediately be washed with water and then alcohol. Shell Development Company Emcryville, California Bulletin ]J>S-4719 moms 088526 - 32 - APPENDIX r Dlbutyl Diphenyl Tin Formula (CHjCHjCHjClijJjSn (C6H5)2 Formula Weight -- 387 Contains - 30!% Sn A clear slightly greenish liquid Insoluble in water. Soluble in most organic solvents. Boiling Point - 1750 at 2 mm. Refractive Index - 1.563 at 17.50 C. Density 1.19 Metal and Thermit Corp. Now York, N. Y. Bulletin, 1949 MQNS 088529 47,957 47,977, 9 47,990 50,402, 7 50,418, 9 50,420, 1,2 ,3,7 50,432, 7,8 50,440-50 53,201-7,9 53,210, 8 53,226, 8,9 53,230, 5 53,242 54,501, 2,3,7 54,510, 1,2 56,588 1 57,389 57,729 1 57,749 59,801, 3,4,5,7,9 59,010-3,7,19 59,820, 8,9 59,831, 3,4,7,0 - 33 - APPENDIX G Notebook pages. 59,840, 1, 8 61,053' 61,084, 7, 9 61,090 61,556, 8 61,560, 1,4,5,7,9 61,570, 1,2 61,500, 8 61,600 61,658, 9 61,660, 2-4 61,686 62,755, 8 62,766, 7 62,785 64,357-9 64,360-2,4,8,9 64,370-5,7-9 64,381-3,7,9 64,727 65,003-5,8 65,010 65,024 65,037 65,044 66,225 66,705 66,711 67,608, 9 67,615, 9 67,620, 3,4,8,9 67,635-7 69,635-7 70,833 72,111 74,882-6 I MUNS 088530 - 34 APPENDIX H Monsanto Chemical Company Research Department - Phosphate Division Anniston, Alabama HCL Scavenger Test Method Scope? This is| the method used here for screening suggested HC1 scavengers for transformer Pyranols. It is intended to give a semi-quantitative comparison of scavenger activity, using tetraphenyl tin (an accepted scavenger) as a reference compound* Equipmentt (See sketch, p. 4 of main report) - ^ Heating bath with thermostat attachments 1 1-1. 3-necked flask with 35/25 spherical Joints 1 Flat |endad teat tubes, 2 l/2 i.d. x 10" 1 No. 13 rubber atopper 1 High-Upeed atirrar motor 1 Stirrer bearing with lubricated seal to Impeller shaft 1 Hoesah-type Impeller , 1 Adaptjer for Introducing gas 1 Adapter for bottom take-off of liquid from flask. 1 Olasa rod support (see sketch) 1 Glass weight (see sketch) 10 Olasa ring spacers 1 Oas-aampling bulb, approximately 275 ml capacity 1 Separatory funnel, to be used as mercury reservoir and levelling bulb 1 Hater aspirator 1 U-tube mercury manometer 1 Beaker, 600 ml or larger B 2-way stopcocks, 2 m bora 2 3-way stopcocks, 2 mm bore (Note11 Stopcock lubricant should not ba a silicone. "Celvacene Medium* Is satisfactory.) .. Glass tjubing, B mm o.d., for T's, connections, etc., (see sketoh) Tygon tubing to connect glass tubing .. 1 Sulfuric add drying train for HC1 6 Strips of 3 mil manlla cable wrapping paper (manufactured by John A. Manning), cut In strips 7" x 1*, with hole punched 1/2" from each and. (Tensile strength about 50 lb per inch under the tests described below.) Chemicals - MONS 088531 [ - ?* - ( Plrocsdure s (1) A solution of the scavenger under test, in Fyranol 1466,is prepared* 1400 g of I this solution are placed in the 1-1 flask, and allowed to come to the temperature of the bath. The papers are suspended on the arm of glass rod' support, separated by the spacer rings, and threaded on the glass weight (with spacers). The mercury is adjusted so that only the bulb and tubing down to stopcock 6 contain mercury. The bulb is supported on a ringstand at kich height that when the gas-sampling bulb is full of mercury, up to stopcock 3, the level in the bulb ie above stopcock 3. The beaker is filled with water. (2) HC1 par the drying train is swept through the gas-sampling bulb, with the exit gas being passed into the beaker of water, until the lack of bubbles indicates| that all air has been swept out. At the same time, the space above the Pyranol in the 1-1 flask is evacuated several times and dry nitrogen la admitted after each evacuation. A slow sweep of dry nitrogen is passed through the test tube containing the papers. The rate and duration of this sweep must be reproduced as closely as possible. In the tests described in Report/2755, the time was 30 minutes and the rate wae suoh as to give a back-pressure of 2*0 cm on the manometer. (3) The HC1 is diverted (at stopcock 1) from the sampling bulb, stopcock No. 5 is opened and No. 4 closed, and then the HC1 may be turned off* The flask is evacuated, stopcock No. 9 is closed and No* 7 is turned, opening the bulb to the flask. Stopcock No. 6 is opened, and the mercury is allowed to displace the HC1 from the bulb up to stopcock No* 3. (4) Stopcock 7 is next closed, and the high-speed agitator started* The Hossch-typs impeller gives vexy rapid absorption of the HC1 into the fyranol, by projecting streams of liquid through the gas space* Nitrogen is admitted from timej to time, to keep the pressure inside the flask up to atmospheric pressure* (5) Aft. 20 minutes agitation, th. motor la turned off, aid enough Pyranol i>.blown over (with nitrogen) Into th. test tub. to cover the paper atrip. completely The paper, ar. allowed to .oak In the Pyranol tar 1.5 hour., Tiny ar* [then removed from the Pyranol, aoaked for 15 minutes in baniene followed py 15 minute, in methanol, and placed on a paper towel to dry. Whan dry. 1/2 inch la cut from aach end, leaving atrip. I" x 6". (6) The tensile strength* cfthe .trips of paper are determined on a Scott IP2 Serigraph, having a Jaw separation of 3 Inches and a rate of travel of 34.5 seecjnd. for a load of 40 lb. Those papers that do not break unds- a 40 lb. load are cut to a narrower width, and a suitable correction factor applied. Before testing, the papers are humidified at 65H EH and 70*F for a minimum of 48 hours. The tensile strength is reported as the average of alx breaks. I MUMS 08053.; - 36 - Reporting Re; With every!series of scavengers tested, there should be run (l) at least two sets oir papers treated with Pyranol containing neither HC1 nor scavenger (let b average tensile strenght from these tests). (2) at least three sets of papers treated with Pyranol containing no scavenger, but which had been treated with the HC1. (Let a average tensile strength ft*om these tests.) ! Thun if x f average tensile strength of papers that had been exposed to Pyranol containing the scavenger under test, and which had been treated with HC1. ! y % 1^)83 in tensile strength * Alternatively, the values of a, b and x may be reported. Standardization) The accompanying figure shoes a series of tests using the method described above) with varied tetraphenyl tin concen tration. The three circles were obtained iA a closely-spaced series of tests, and the two squares injanother closely- spaced series. The point re presented by the triangle', at 12 tetraphanyl tin, was ob tained a considerable time earlier then the rest of the points, and perhaps it should not be taken too seriously. (When the qitrogen sweep was omitted, tie dashed curve was -A-- 1.0 obtained. The experimental points for this curve are not shown. The differences between the two curves was due to different values of tha quantity a: about 25 lb. when the nitrogen sweep was imployed, and less than 5 lb. when"it was omitted. At 0.52 tetra- phenyl tin thc tensile strengths (x's) were very nearly the same when the nitrogen svjeep was employed and when it was omitted.) 088533 - 37 - I Notes and Dlacuaalon This Is not intended as a "final" form of the scavenger screening test. It was actually a compromise between the method as It existed when the author started on the project, and a form toward which it was evolving. While It was held constant during the last two major series of tests, several ex periments Mere made (e.g. the variation of the nitrogen sweep) which pointed to acme further changes. Certain changes would, no doubt, have im proved the quantitative nature of the method. While only one each of the 3-necked flask, test tube, glass rod support, etc., are specified, it is found convenient to have at least two and preferably three, in lorder to carry on aami-simultaneous experiments. The 3-nscked flasks should be of matched volumes, to within 10 ce or closer If possible. adl 3/27/52 R. J. Good II ( MGNS 088534