Citation
The nature and reactions of halogen addends of the silver fluorocarbon carboxylates ..

Material Information

Title:
The nature and reactions of halogen addends of the silver fluorocarbon carboxylates ..
Creator:
Crawford, George Homer, 1928-
Publication Date:
Language:
English
Physical Description:
78 leaves : ; 28 cm.

Subjects

Subjects / Keywords:
Bromine ( jstor )
Carboxylic acids ( jstor )
Copyrights ( jstor )
Fluorocarbons ( jstor )
Halogens ( jstor )
Iodides ( jstor )
Iodine ( jstor )
Pyridines ( jstor )
Silver ( jstor )
Solvents ( jstor )
Chemistry thesis Ph. D
Dissertations, Academic -- Chemistry -- UF
Fluorocarbons ( lcsh )
Genre:
bibliography ( marcgt )
non-fiction ( marcgt )

Notes

Thesis:
Thesis (Ph.D.)--University of Florida, 1954.
Bibliography:
Bibliography: leaves 75-77.
General Note:
Manuscript copy.
General Note:
Vita.

Record Information

Source Institution:
University of Florida
Holding Location:
University of Florida
Rights Management:
This item is presumed in the public domain according to the terms of the Retrospective Dissertation Scanning (RDS) policy, which may be viewed at http://ufdc.ufl.edu/AA00007596/00001. The University of Florida George A. Smathers Libraries respect the intellectual property rights of others and do not claim any copyright interest in this item. Users of this work have responsibility for determining copyright status prior to reusing, publishing or reproducing this item for purposes other than what is allowed by fair use or other copyright exemptions. Any reuse of this item in excess of fair use or other copyright exemptions requires permission of the copyright holder. The Smathers Libraries would like to learn more about this item and invite individuals or organizations to contact the RDS coordinator (ufdissertations@uflib.ufl.edu) with any additional information they can provide.
Resource Identifier:
021879268 ( ALEPH )
13378977 ( OCLC )

Downloads

This item has the following downloads:


Full Text












THE NATURE AND REACTIONS OF

HALOGEN ADDENDS OF THE

SILVER FLUOROCARBON CARBOXYLATES












By
GEORGE H. CRAWFORD


A DISSERTATION PRESENTED TO THE GRADUATE COUNCIL OF
THE UNIVERSITY OF FLORIDA
IN PARTIAL FULFILMENT OF THE REQUIREMENTS FOR THE
DEGREE OF DOCTOR OF PHILOSOPHY









UNIVERSITY OF FLORIDA
June, 1954

























ACDIWLI


The author wishes to express his appreciation to

Dr. J. H. Simons for the encouragement and advice provided during this investiptin. The author also wishes to acknowledge the sponsorship of Minnesota Mining and bniufacturing Company, without which this work would not have been possible.


















TA3= OF CONiTMTS


Phew

IZBT 07l 'L31 . . . . . . . ... . . . . . . . . . . . i
LTST OF rUB TOW . .. . ... . . . * *l

I . 11'JODUONO. . . . . . .. .. 0 *

Ii. lPIARATION, ISOLATIONp AND 33XNM CATION
or n OUM S * * *. . � .. .. . . . *. 13

Detrmination of Xxtent of Reaction
AnalysJi of Reaction Mixtures
Analysis of the Caplexes in Solution
Isolation of the Solid Coplazoo

in. usE r mu COmI3IEB IN TE Su ESIS0

General Discussion
Substitution of Pyridine for X in
the Complex-Foraing Reaction
An Improved Method for the Synthesis
of Fluorocarbon Mco-Zodides
Reactions of (S) and the Complex with Phosphorus

rV. RIAT SE W=TIC C0DU S . . . . . � � � � . � � 63
Preparation of Fluorocarbon Nitriles
from (8) and Cyanogen
Synthesis of Fluorooarbon Mono-Ioides
from Fluorooarbon Carboxylic Acid Anhydrides

V. DTJO0USION (W UL.. ..... . . . � . . . � . � 67
Vi. NXOGR . .. .. . .. .. .. . 75

v i i . V I T A . . .. . . . , . . . . . . . 7 8


iii


















LaST OF TABLES


Table Page

1. Source and Purity of Materials Used in the Investiptlt s Herein Reported....� �� �� ......� �� �. 23

2. Preliminary Experiments Determining the Stolehicaetry of the Reaction of Silver Aoetforate with Iodine or
Bromine in Various Solvents.. ............... .... .... 24

3. The Stolobicmetry of the Reaotimu of Silver Salts of Fluorocarbcn Carbwylio Acids with Iodine or
Brmine in Fluorooarbn Solvents................... ... 26

4. Measurement of Excess Iodine Present in Solutions of Reaction Mixtures by the Color Comparison Method.... 27

5. Preparation and Analysis of teaetim Mixtures of Silver Salts of Fluorooarbon Carboxylic Aoids and
Iodine or Bromine
A, Preparation ........ ..... ..... ......... ............ 35
B. Analysis .... ........... .. .... ...... ...... ......... 36

6. Analysis of the Complexes in Solution in Fluorooarban Solvents
A. Relationship Between Acid Number, Oxidizing
Power, and Recoverable Silver... ........... 40
B. Comparison of the Stabilities of the
Complexes in Solution......................... 41

7. Preparation, Isolation, and Analysis of the Solid complexes ......... ........ ........... 46

8. Preparation and Analysis of the Pyridine Addends of the Silver Fluorooarbon Carb5ylate................52

9. Improved Synthesis for Fluoroearbon Iodides

B, Procedure 2. ...... .......... . ................ 59









V

LIST O TABIX--CContinued


Table Page

10. Proparstioa of FluorocarbmC Nitriles by the Dooarbwlatiom of Silver Salts of Fluorooarbon
Carbezylio Aoi&s in the Presenoe of Cyanogen
A. 2xerimuntal Dat................................. 66
33. Analytioal Daa.......................��� 66

11. Preparatimn of Fluorooarbcm Io&ide from Fluorooarbon
Carboxyllo Acid h des............................. 70


















L3ST (W ILLUSTRATIONS


Figure


1.

2.


Page 30 6o



















INfRODUCTIcOI


At the current stage in the development of the chemistry of the fluorocarbons and their derivatives, it is highly desirable to enlarge and expand the knowledge of the fluorocarbon carboxylic acids and their related compounds. Toward this end, the specific problem posed to the author and dealt with in the work herein reported is as follows: isolate, establish formula, and characterize as to physical and chemical properties the intermediates in the reactions between the silver salts of the fluorocarbon carboxylic acids and the halogens or halogen-like substances.

The fluorocarbon carbaxylic acids have already proven themselves to be valuable starting materials in the synthesis of fluorocarbon derivatives. Their acid chlorides and anhydrides have also been adapted to a number of synthetic methods through modifications of procedures familiar to the organic chemist. In addition, several processes have been developed involving decarboxylation of the metallic salts of these acids, either with or without other materials being present. A knowledge of the nature of the halogen addends of the silver fluorocarbon carboxylates is essential to the explanation of the mechanism of certain reactions of this type and is contributory to the understanding of the general reactions of the fluorocarbon

1












oarboxylic acids and their derivatives. More important, however, is the value of this knowledge in the prediction of entirely new routes to fluorooarbon derivatives hitherto unavailable through the synthetic approaches used in organic chemistry.

The literature contains a number of papers relating to the copounds produced through the interaction of the metallic salts of the organic carboxyllo acids with the halogens. There is considerable disagreement as to the nature of the intermediate substances, and many of the conclusions drawn regarding reaction mechanisms appear conjectural. Some reactions of mixtures of the salts of the fluorocarbon carboxylic acids with the halogens have also been investigted. In the development of synthetic procedures involving the use of these mixtures, somae investigators have suggested mechanisms apparently based upon some of the earlier conclusions regarding the silver salts of the orgnic oarboxylic acids. The suggested mechanisms involved intermediates which were never isolated or identified and whose existwas purely speculative. In addition, extrapolation from orgnic to fluorocarbon chemistry has been demonstrated to be, at best, highly hazardous, even when based upon completely substantiated argnic data. It was evident that a detailed quantitative investigation should be undertaken in order to eliminate the confusion existing in this area.

The intermediates in the reaction between the silver salts of the fluorocarbon carboxylic acids and the halogens were found, by methods described in the experimental section of this dissertation, to be addendum compounds of the composition (eCOO)2AgI where 0 is a










3

fluorocarbon radical and X is a halogen. The properties found for these substances serve to demonstrate again both the similarities and the fundamental differences between the conventional organic compounds and the fluorocarbons and their derivatives. Information is provided which is useful in the chemistry of both silver and the halogens. Halogen-like properties of other materials substituted for X in this compound are clarified. The value of these addendum compounds in the synthesis of the derivatives of the fluorocarbons is also demonstrated, in part, in the experimental section of this dissertation.

A number of approaches to the problem of characterizing these intermediates might have been suggested.

The problem could be attacked through purely theoretical methods based on previously-determined thermodyramic data. This approach was discarded since the data necessary to such an investigation were not available.

The molecular structure and configuration of the intermediates might be determined through the use of X-ray diffraction, infrared spectra, and other physical means. This method was also discarded due to the fact that preliminary investigations had indicated that the substances were not adapted to such an approach.

A third possibility was the synthetic approach, involving

preparation and isolation of the intermediates under rigorously controlled conditions, followed by their characterization through chemical reaction. This method was adopted since both the information in the literature and that gained through preliminary investigation indicated











that it should be most productive of conclusive results.

Historical background. --With the original preparation and investigation of trifluoroaoetic acid by Swarts,3 it became evident that here was a new series of acids vhose characteristics showed every promise of proving to be markedly different from those of their organic analogs. However, only since the developaent of the Slmons eleotrochemioal proeoess7 have the fluorocarbon oarboylic acids become readily available for research purposes.

Before considering the reactions of the silver salts of these acids with the halogens, the earlier work involving the silver salts of the orgsani oarboxylic aoids should be cited.

The first reference in the literature is to the reaction

between silver benzoate and bromine. PJligot,27 in 1836, combined these materials in benzene solution and obtained meta-bromo benzols acid as a product.

SBidni35,36 conduoted the first extensive investigtions into these reaotions, using the silver salts in combination with iodine or bromine, and arrived at oonclusions which, as far as they vent, have found acoeptanee among most of the subsequent investigtors. Simcnini reported that the silver salt of a hydrocarbon oarboxylio acid reacted vith halogen in a 1:1 equivalent ratio, giving a complex of the formula (RCOO)2AgI and liberating a mole of silver iodide. This complex, he continued, vould deoarboxylate upon heating, giving esters of the formula RCOOR and liberating carbon dioxide and silver iodide.

Kleinberg,2 in reviewing the work prior to 1947, ccmoluded











5
that the products of reactions involving these starting materials depend upon five factors, namely: (1) the nature of the silver salt,

(2) the temperature of reaction, (3) the presence or absence of other reactive materials, (4) the nature of the solvent, and (5) the equivalent ratio of silver salt to halogen. Kleinberg states that the reaction appears to fit a definite pattern only when the equivalent ratios of silver salt to halogen are 1:1 or 1:2. According to Oldham and Ubbelohde,25 however, a third category exists, involving an equivalent ratio of 3:14.

The above-cited work of Simonini falls into the 1:1 group, as does that of Prevost29"34 who, in a series of papers, elaborated upon the earlier work of Simonini. He described the addition of the complex intermediates to olefinie cumpounds to give the halogen-containing esters followed by di-esters which could then be bydrolized to give glycols. He contended that the complex first added across the double bond, giving the halogen-containing ester and liberating RCOOAg which then underwent an exchange reaction with the halogen in the ester to produce the di-ester, with the liberation of silver halide. He found that the proportion of di-ester formed was directly dependent upon the solubility of the silver salt in the solvent employed. Prevost also investigted the reaction between the complex and butadiene, and coneluded that the addition was 1, 2 rather than the 1, 4 exhibited by halogens. He described the iodinating action of the complex formed with silver benzoate and iodine and refers to the "metallic" character of halogens in such complexes. Prevost confines himself almost










6

entirely to the use of silver benzoate and iodine as starting materials. Hersohberg19 and Wieland and Fisoher4 also investigted these reactions and arrived at similar onclusions.

The reactions of the silver salts of the dicarboxylic acids with the halogens have been investipted..j'6'v0' 2'43' Both 1:1 and 1:2 silver salt to halogen equivalent ratios have been employed.

InvestSgstione involving a 1:1 ratio were first carried out by Birnbaum and Gaier, who reacted the salts of malnio, malic, suecinie, tartaric, famric, and maleic acide with iodine and obtained the corresponding acids or anbydridees, carbon dioide, and silver ioide. Wieland and Fischer4 obtained laotones along with the acid or anhydridce. Similar results were obtained by other investiators.

When equimolar quantities of halogen are heated with the silver salts of the dicarboylio aids carbon dioxide, silver halide, an the corresponding dl-hallde are obtained. For example, Boohemuller and Hoffman6 obtained 1,,-dibrrnobutane from silver adipate and bromine.

This secnd general classifioation, that involving an equivalent ratio of 1:2, is represented in the work of Birokenback, Goubeau, and Berninger;3 Booheauller and Hoffmn;6 and others23,26 According to these Investiators, an intermediate substance of the formula RCOOX apparently is formed by the following reaction:

RCOOAg + X2 - ICOOX + AgX

Upon heating, this material will deoarboxylate according to the following equatien

RCOOX heat, PI + 002










7

The work of Bochemuller and Hoffman can be cited as being

typical of the investigtions involving the 1:2 equivalent ratio. In the process of chlorinating allyl chloride in acetic acid solution, these investigtors obtained a dichloro ester by a process for which they postulated the following mechanism:

(1) C3CoaH + C1 -- CH3C Cl + HC1

(2) CE3CO2Cl + cH2=(m2C1 C93CO2CCHClC2Cl

The intermediate acetyl hypochlorite was never isolated, nor were the corresponding hypobromites or hypolodites in subsequent investigtions by these and other workers. They found that equilibrium (1) could be shifted to the right by the use of a silver salt in preference to the free acid. They found that the reaction mixture would take up another mole of silver salt, with the formation of complexes of the formula (RCOO)2AgX. These are, of course, the previously described Simonini ccomplexes.

Bochemuller and Hoffman6 describe the thermal decomposition of the acyl hypoiodites as occurring by two paths, one giving the alkyl iodide as above and the second giving the ester by the following reaction:

2 RCOOX -heatRC0CR CO2 + X2

Upon reacting bromine with silver benzoate in boiling carbon tetrachloride, these investigtors obtained bromobenzene in an 80 percent yield. Luttringhaus and Schaede23 disagreed with these findings, maintaining that the degradation was unsatisfactory.

Bohemuller and Hoffman;6 Birokenback, Goubeau, and Berninger;3










8

and Usobakov and Tohistcw4l investioted the reactions of the acyl hypohalownites with olefinio substances.

Dirokanbaok and coqorlmrs' reacted the filtrate of the reaction between equizolar quantities of silver acetate and Iodine at
-800 in other with oyohexene. He obtained the aetate of 2-iodol-oyelchezanol. Kleinberg22 states that this is undisputed evidence of the presenoe of CM3COOI in solution. These results would seem to indicate that the ester-producing reaction is due to the presencoe of RO001 rather than the Sixonini ooplex as was concluded by Prevost.

Carlsobn9 was able to isolate a substance of the foul RIOO(Py)I fren the reaction of pyridine with solutions resulting from the reactions of silver salts with iodine.

O1an and Ubbelohde25 characterized a third clas of ocpounds, the Iodine triaoyls, which were obtained when an equivalent ratio of 3 silver salt to 4 iodine was employed. They postulated the following equation for their produotion:

3 RCOOA + 2 12)- 3 A81 + 1(ocoM)3

These compounds contain tervalent iodineo, as opposed to the univalent positive iodine said to be present in both the Simonini complex and the acyl hypoblogenite. They were able to isolate the iodine triaoyls by chilling filtered solutions of the triaoyl in carbon tetrachlorid, or petroleum ether and collecting the material which crystallied on a filter septum. They postulated a number of free-radical mechanisms for the thermal decompositions of these substances and present oonsiderable analytical data in support of their findings. It










9
appears that the reaction products of the iodine triaoyls are not substantially different from those of the Simonini. omplexes or aoyl hypobaloganites.

Olam26 was not able to obtain the corresponding bromine triaoyls. He investigated the reactions of bromine with monobasic aliphatic aoids, monobasio arcmatic acids, monbasio aliphatio monoketo aoids, dibasio aliphatic acids, and ditlbasio aromatic acids. His results were similar to those previously obtained by other investigators.

The first Investigation into the reactions of the silver salts of the fluorocarbon oarboxylio acids was undetaken by SwartL,40 who reacted silver trifluoroacetate with iodine in benzene solution at 600. He gives the following equation, which is representative of the overall reaction:

2 CFC0r2Ag + 2 12 + 2 C6N6- (F3CC )O + E20 + 2 Ag! + 2 06151 Thus, the reaction mixtures of the silver salts of the fluorocarbon oarboxylio acids with iodine were demonstrated to possess iodinating power similar to that previously observed in the organic carboxyllo acids.

Simons and Drioe38 developed a process for the production of fluorocarbon iodides through the Aeoarboxylation of the silver salt of a fluorocarbon oarboxylio acid in the presence of exoess iodine in a higi-bolling fluorocarbon solvent, and were wanted a patent. No attempt was made to postulate a mechanism for this reaction.

Henne and Zimmer18 investigated the action of mixtures of










i0

silver trIfluoroacetate .and bromine or iodine on toluene. They found that ring halogenaticn occurred below WOO�, and advanced the following equatios for the reaotion:

Mr CO 2A9 + k - Agi + ar5C02X

7o2X + 13C6H5 - - CCo323 + C63x

In this mechanism, the lodiating entity is the CF3COOX. They state that the positive halogen "hangs" from a polarizable atom attached to one or two stroagly-active methforyl groups.

Haszeldine and oo-4rorkers,l'14 in a series of papers, have reported Investieptions of the reactions of a number of the salts of the fluorine -ocntaining monocarbomylio acids with the halogens.

In the first of this series, Haszeldinell describes the deoarbexylation of silver trifluroacetate in the presence of iodine to give trifluoracethyl iodide. He states that the intermediate in the reaction is 303COOI. This substance, he cotewes, is formed when iodine is taken up by silver trifluoroacetate in a 2:1 equivalent ratio.

In the second of this series, Easzeldine and Sharpe,1 after obtaining some results somewhat inconsistent with the above postulate, suggest that an equilibrium exists between the hypohalogenite and a Simacini-type oaeVlex in which the iodine and the iodine trifluoroacetate are ecupeting for the silver salt.

(730"2A + MJ30OOI - (C73C02)2ASI
If a readily iodinated substance, such as phenol, is added, the iodine triflurosoetate is removed and the proper amount of silver iodide is











11

precipitated in accordance with the equation cited under the work of Renne and Zimmer.18

Haszeldine12 was unable to isolate either iodine trifluoroacetate or the Simonini complex. He used nitrobenzene as a solvent in the above experiments.

Haszeldine12 brominated and iodinated a number of aromatic compounds by this method and devoted considerable space to the discussion of ionic and free-radical mechanisms for this and related reactions.

In the third paper in this series, Haszeldine and. Sharps13

describe the decomposition of the silver salts of the fluorohalogenoacetic acids in the presence of the halogens to give the fluorohalogeno alkyl halide in the same manner as with the fluorocarbon oarboxylic acid salts. They also describe the preparation of higher fluorocarbon mono-iodides from the higher acids. They discuss the preparation of mthforyl halides by the same reaction, using the potassium or sodium salt as an alternative to the silver salt. A number of miscellaneous reactions and attempted reactions are described.

In the fourth paper in this series, Haszeldine and JanderIl

report the synthesis of fluorocarbon nitro and nitroso compounds by the deoarboxylation of the silver salt of acetforic acid in the presence of nitrosyl chloride. They give the following generalized equation for the reaction:










12

cr3.(ar2),co2Ag Nl Age + cT3.(c',2)nCooNO

(3 . (72)MO h Cr.(C,2)n o + ar3.1(o2)nNo2 + C2, etc.

This Involves an intermediate which is analogous to the triflucrcoetyl bypohalogenite previously postulated. This intermediate, like the others, was never isolated or characterized. Dmnual had previously synthesized fluorooarbom nitro and nitroso eompomnd. by another method.

Cady and Kelloeg,7 in an interesting paper, describe the isolation of an unstable substance having the formula C73C0CV from the reaction of aaetforic acid with fluorine in a olosed glass system. Menefree and Cad?4 extended this work to the preparation of C0 and C FTWO. The hypofluorites were deteoted by explosions which could be Initiated in the reactor. Acetforyl bypofluorite and propicmforyl hypofluorite were identified through the determination of molecular weights and the analysis of &eomposition produots.

Thus, the existence of substances of the formula OCOC has

been established. However, the existence of the analogous hypohalogenites of the general formula C00X in the reaction mixtures of silver salt with halogen remains to be proved.


















PUiPARATION, ISOLATI( , AND IDENTIFICATION OF HE COMPLEXiS Determination of Extent of Reaction Discussion.--The stoichiometry of the reaction between the

silver salt of a fluorocarbon carboxylic acid (hereafter designated S) and a halogen (X) is of paramount importance, far it provides an essential clue as to the nature of the product formed at temperatures below that of decarboxylation and in the absence of other reactive materials. Thus., if the reaction leads to the formation of a substance of the forwla 000X (where 0 is a fluorocarbon radical) by a path of the type

(S) + X2 - O COOX + AgX

one mole of (8) should consume two equivalents of X. If the reaction leads to the formation of the fluorocarbon analog of a Simonini complex by a path of the type

2 (8) + X2- (QCOO)2Ag + AgX

one mole of (S) should take up one equivalent of X. If, as Haszeldine subsequently contended, an equilibrium exists of the type

(S) + COOI -- --"(GOO)2AgI

free iodine should be present during the entire process, its color remaining relatively constant or becoming progressively more intense,

depending upon the value of Keq. (No values for Keq were provided.) 13










14

In order to obtain accurate and reproducible results, a solvent must be employed which can in no way react chemically with the materials under investigation. The solvent should be clear, non viscous, reasonably low-boiling, and capable of dissolving the active intermediate under consideration.

In the preliminary investigations, a number of different c on organic solvents were tried and subsequently discarded as failing to meet one or more of the above specifications. Erratic, non-reproducible results could usually be attributed to the occurrence of secondary reactions involving the solvent. Among these are oxidative attack upon the solvent by the intermediate, halogenation of the solvent by the intermediate, hydrolysis of the intermediate by the solvent itself or by traces of water in solvents having a high affinity for water, and secondary ccmplexing between the solvent and

(s) or the intermediate.

Ether, chloroform, carbon tetrachloride, and pentane were tested and discarded. Of these, pentane gve the most consistent results, checking closely with those obtained using a solvent consisting of fluorocarbons or their inert derivatives. The results of these preliminary investigtions appear in Table 2.

Fluorocarbons, fluorocarbon nitriles, and fluorocarbon cyclic oxides were used as solvents. These materials are ideally suited to this investigation because of their inertness to chemical attack, complete lack of color, lo viscosity, and the fact that a particular











15

solvent could be chosen having a boiling point suited to the experiment being performed. The low solubilities of halogens and (S) in these media do not constitute a serious drawback. The rate of reaction of

(S) with X is decreased and becomes partially dependent upon the stirring rate. However, since the investigation is not a study of rate phenomena, this is not a deterrent factor. On the contrary, it is desirable to have the reaction proceed at a reasonable and orderly rate, as contrasted to the violent manner in which it occurs in solvents such as ether, in which both (S) and X are highly soluble.

The solvents consisting of fluorocarbons or their derivatives were subjected to rigorous purification and deactivation before being used in experiments. The procedure for this is described in the experimental part of this section.

The source and purity of the fluorocarbon carbozylic acids and other materials employed in the following investigtions are described in Table 1.

Experimental.--Determination of the stoichiometry of the reaction between (S) and X was carried out using the apparatus shown in Figure 1. The silver salts of the fluorocarbon carbaxylic acids were prepared by refluxing the acid with excess AgO in aqueous solution, followed by filtration and removal of water under vacumm. The salts were purified by repeated recrystallization from anhydrous ether. Final ether removal and drying was effected by warming in VaOUO.

Since the solubility of X in the fluorocarbon solvent is low,











16

it is impractical to make up the complexes by addition of a standard solution of X to a slurry of (S). The following techniques were adopted.

Iodine (analytical reagent grade) was resublimed through P205 and. collected on a water-cooled ooldfinger in the sublimation flask. The apparatus was transferred to the drybax and the iodine removed from the coldfinger and stored in an appropriate container. Samples of the iodine were transferred in the drybox to weighed 3" or 4" test tubes. The tubes were then quickly sealed, using a gas oxygen torch. The tube and contents plus the scrap were weighed on the analytical balance and the net amount of iodine per sample determined.

Analytical reagent grade bromine was dried by distillation from P205, and weighed samples were made up in the same manner as with iodine.

The following fluorocarbons and fluorocarbon derivatives (hereafter desigaated 0F) were used as solvents:

#1. N and iso-pentforane mixture (C5F12) -- This material

was washed with anhydrous ether for removal of fluorocarbon hydrides, refluxed with EM , refluxed with OF3C0OAg + I, treated with sodium thiosulfate, dried over W12, and fractionated. The fraction, boiling 29-310, was taken. The fluorocarbon solvent, thus freed of any impurities which might react with the intermediate, was stored over P205 in sealed glass tubes.

J2. Nonforane mixture (C9i20) -- This material was treated as above, and distilled. A rather wide fraction, boiling 118-130�,











17
was obtained. This was stored over P205.

J3. Butf fl nitride (01 27N) -- This material was treated as above and used for a solvent in the fluorocarbon mono-iodide synthesis of Simons and Brice .38 It was distilled; and the fraction, boiling 165-1800, consisting of various isomers was obtained. The material was stored over P205.

J4. 'Fluoroehemical 0.75" (a cyclic oxide of the formula C87160) -- This ma aerial was treated in the same manner as the butforyl nitride, and fractionated. The fraction, boiling 100-101.50, was stored over P205.

The procedure used for carrying out controlled reactions of

(S) with X under absolutely anhydrous conditions and for determining the stoichiometry of that reaction was as follows: The reactor (Figure 1) was thoroughly dried by alternate evacuation and admission of dry air. A sample of (S), weighed to be somewhat (usually in the vicinity of .1 gram) les than that required for a 1:1 equivalent ratio with one cf the previously weighed X samples, was carefully poured into the graduated reactor vessel D. This was accomplished by lowering D a few inches, while maintaining a stream of dry air through the system from L and paring (S) from the weighing bottle directly into D. Great care was taken to avoid getting any (S) on the inside of the joint. D was replaced and the drying process repeated to remove any water that might have entered the system. The weighing bottle was then reweighed and the net amount of (8) determined.











18

The vial of I was dropped in through the top while the air stream from the system was mintained. Breakage of the vial could be assured by shooting it in with a sling-shot arrangement attached to the rim of the tube. The solvent was then admitted from B, removing any (S) that might have clung to the sides of D. Magnetic stirring was started, and a reasonable time was allowed for the reaction to go to completion.

If all color was removed, more halogen could be adled. If, on the other hand, the characteristic red-brown or red-purple color due to free bromine or iodine remained, enough (S) was added to bring the equivalent ratio up to, or perhaps very slightly in excess of, 1:1. Completion of reaction could then be noted by the disappearance of all color in the solution when this point was reached.

In every case free halogen remained until additional (S) was added,., after which a few minutes' stirring removed all trace of color. The numerical data pertaining to these experiments appear in Table 3.

Thus it is demonstrated that bromine and iodine are coeumed

quantitatively to the 1:1 equivalence point, and that immediately after passing that point the color of free halogen is evident. Next it is desirable to determine if the amount of free halogen rem ins relatively constant as X is added, due to the establishment of an equilibrium of the type cited under the work of Haszeldine, or whether all the halogen added past the end point is present in the free state. Only me unassailable method for such an experiment appears to exist. That is to bring a mixture of (8) and X2 to the 1:1 equivalence point and then to










19

add known quantities of standard X solution past that point. A sample of the same solvent not containing any reactant materials (hereafter referred to as the blank) is then titrated with the same X solution to the point at which its color matches that of the above solution containing the reactants. The ratios of added solution to original solvent in the reaction mixture are compared to the corresponding ratios in the blank at the point at which the colors are observed to match. Thus, if all halogen added to the reaction mixture in excess of the 1:1 stoichiometric point is present in the free state, the values for these ratios should coincide at each point of identical color intensity. The tabulated values for the above ratios appear in Table 4. The degree to which they coincide as successive amounts of iodine-containing solvent are addod is evident.

A saturated solution of 12 in "Fluoroohemical 0.75" (solvent 14) was prepared and standardized against thiosulfate. The violet solution had a normality of only .00134.

AgT was prepared from A IO3 and KI and dried by warming in vacuo in the absence of light. This was to be added to the blank in order to provide identical visual conditions. However, it was discovered that AgI itself is capable of removing smll quantities of iodine from solution. The extent of this phenomenon was measured by titrating 2.5074 g. AgI in 10 oc. of #4 solvent with the standard iodine solution. At 65 ac. a noticeable pink color appeared. Thus, the ASI decolorizes 12 in a 121 to 1 equivalent ratio of AgI to 12 in the W solvent.











20

The experiment was repeated using reagent grade nitrobenzene. A .026 N solution of iodine in this solvent was prepared and standardizod against thiosulfate. A suspension of 4.3059 g. AgI in 10 cc. pure nitrobenzene decolorized 14.90 cc. of 12 solution. Detection of color change is difficult in a colored solvent such as nitrobenzene; hee the 12 solution was added until the color was obvious. The nitrobenzene solution was then filtered and back-titrated w1th .5 cc. of .1000 N thiosulfate to a starch end-point. The AgT was found to decolorize 12 in a 54 to 1 equivalent ratio of AgI to 12 in nitrobenzene.

This phenomenon, though small, is just another of the complicating factors which can lead to erroneous results, if not properly recognized.

The experiments were conducted using an amount of AgI in the blank proportional to that which would be produced by the complexforming reaction.

The color-change characteristics of free iodine in W solvents, plus the extreme dilution of the solution used, enabled this investigator to make measurements of high precision with a minimum of human error. The results obtained were in remarkably close agreement with theory.

Results of the above measurements.--The object of these measurements has been to establish the stoichiometry of the reaction of (8) with 1. This was accomplished through the addition of X to (S) in a OF solvent and observation of the point in the addition at which free halogen appeared and the manner in which the color intensity increased











21

past this point. Similar results were obtained for (S) from silver trifluoroscetate through silver valerforate and for bromine and iodine. It was demonstrated that after the 1:1 equivalence point is reached in the addition of I to (S), free halogen appears in what bad been, up to this point, a colorless solution. Furthermore, the dLeepening of the color as 12 was added past this point progressed linearly with the addition of 12, paralleling the color intensity deepening of a simultaneously-run blank.

It was therefore demonstrated that when the reaction is carried out at temperatures below that at which decarboxylation can occur, in a completely inert solvent, and under strictly anhydrous conditions, the reaction occurs in a 1:1 equivalent ratio and that any halogen added in excess of this is present only in the free state.

The latter measurement was conducted using iodine only. Similar color deepening past the end point was observed with bromine, but the color characteristics of bromine in the solvent used and the high

volatility of bromine (resulting in its escaping from solution into the air space above) prevented precision observation of color change.

Some of the investigators have arrived at conclusions as to

the extent of the reaction by weighing as AgX the precipitate obtained at various stages during the halogen addition. This involves the erroneous assumption that all the material precipitating during the reaction is silver halide. It will be shown that the precipitated material is, in reality, a mixture of AgX and (QCOOAgX. The proportion of the latter in the precipitate is inversely proportional










22

to its solubility in the solvent employed. Then, when the mixed precipitate is removed from the reaction mixture by filtration, and prosumably washed and dried prior to weighing, the (QCO0)2ASX is converted to AgI, A9103, and free acid by a reaction of the type

(9000)2A6% + E20-*2 OCOOOE + AgOX followed by

3AOX - 2 AgX + P-403

as will be shown in the following section.

The danger of confusion is further increased by the great similarity in appearance between AgI and the mixture AgX + (QCO0)2Ag. In the preliminary experiments (see Table 2) and in subsequent investigtions, (S) was stirred with 12 under anhydrous conditions and in the absence of a solvent. A yellow, amorphous powder resulted, having the overall composition QCOOAgI. This material cannot be visually distinguished from AgI. The addition of iodine in excess of the 1:1 equivalent ratio produces the red-purple color of the free halogen.

The phenomenon of iodine absorption by silver iodide, though small, is still a complicating factor. It was demonstrated to be over twice as large in nitrobenzene as in "Fluoroohemioal 0.75."













TABLE 1

SOURCE AND PURITY CF HATERIAIS USED IN THE
INVESTIGATIONS HEREIN REPORTED


Compound Source


Aoetforic acid Propicnforic acid Butyrforic acid Valerforic acid Caproforic acid Bromine

Iodine

Pyridine


Pyridine Phosphorus (red) M-di isopropyl benzene 2-butyl benzene Silver oride


M. M. & M. Co. M. M. & M. Co. M. M. & M. Co. M. M. & M. Co. M. M. & M. Co.

Wallinckrodt Mallinakrodt Mathieson

Eastman Kodak

(DP)

Mallinckrodt

Dow

Koppers Fischer


Purity 95-98%

90-95% 75-85% 95-98%

90-95%

Anal. R. Anal. R. U.S.P.

Spectral Grade Tech. Tech. Tech. C.P.


Further Purification fraotionation fractionation fractionation fractionaticn fractionation described on p. 16 described on p. 16

fractionation and
drying

drying


dessication fractionation

fractionation







TAB NE 2
PLIMIABY EXIRINTS MD M ING WE 0TOICXRTI0 CF RAION CF SILVE AfrfCRAE WI IODINE O M0PMM IN VARIOUS SOLVENTS


(S) g0 bt.3q. I g0 Mjq. Solvent Reaotion Temperature Extent of Time & Conditions Remotion 4w Mixt. 250 - 1000 CFCOOAg 55 250 12 63 500 BP-170 7 days tube-shaking inomplete 250 cc.

OF3COOAg 60 270 added to above reaction mixture 12 hours 250 - 350 0 omplete tube-sakiiiiiiplet


10 days 15 min. 12 hours instantaneous reacted as titrated


250 - i000
tube-shaking

260 - 270
tube-shaking

0
21� - 26�
stirring

2.0 - 380 stirring

260 - 270
stirring


inoomplete


complete i oc plete complete

end-point vague


C730OOAg 2.3360 10.57 12 2.5071 19.82 ether reacted an 250 - 270 93% of 2:1
CFCOO ~ ~ ~ ~ ~ ~ 4 00.010571 .01 ..2~ titrated, stirring


cr3COOAg








CF3COOAg
ar3C00A9


22.1 6.8


11.5764


13.8905 2.6689


100 30.7 52.38


62.48 12.07


12 12 12 33r2 12


25.4 3.89
(approw) 6.14799 9.3254 1.9151


200 30.7
(approx) 51.02


11.65


15.08


W mixt. BP-17o 100 co.

(HC13 32 cc.


47 oo.

ether 120 co.

ether 57 co.


I










(S)




037COOA9
OJ',COOAg C3F7COOAg C3F7COOAg


7.6 7o .668
3 .2031





10.3668


1 .201.6


3.6



1I4.0'4 43.8 32.29 3.75


I

r2 B'2







added


go

4.6136 1.8619 5.5533


4.5029 to above


TANI 2--Cntinume


M-4. Solvent Refation Tim.

57.67 other reacted as 89 o0. ttrated. 11.65 other 15 Min.
53 08.

43.9 pene 30 mi.
100 Go.

35-.45 penten 5 hours
100 c.

reaction mixture 2 min.


Temperature & conditions 260 - 27o
stirring

24� - 250
stirring

250 -260
stirring
0 0i
25 -27o
stirring

250 - 260
stirring


Extent of Reaction

83% of 2:-.1 inomplte





inoomplete oomplete


li II





TABL 3
TH ST0ICBI(KMY OF TBE PMATON OF SII1W SAlM OF FIMOCJBON CAIWOIMC ACIIB WITH iQIJIN a~ NmCL IN PIMOCA1RBOU SOLVM71

(S) g. M.Ej. X g. M.Eq. Solvot Tamp. t). Extent Of TiM "C Ratio
CF3COQAg 28.0 126.7 12 16.0 126.00 #3 250 cc 3 bira 25-27 .993 complete
--- 12 .20 1.57 same 2 hre 25-26 1.007 Incmplete .C37000-A 5.0772 15.81 12 2.3148 18.23 #3 30 cc 2 hre 25-26 1.153 Incomplete
C0317C00NA .8081 2.52 -- same -- sam. 10 min 25-26 .995 complete


c3. Co0oo +.6053 20.85 Br. 1.8668 2.6 #1 25 cc 6 tLr 24-26 1.121 in mplete
CF3COOAC .5675 2.57 -- same --- fl 25 cc 18 hre 24-26 .998 cmplete

. C9FTCOAg 2.5) 79.5 12 10.858 85.6 #3 150 cc 15 hr. 25-25 1.07 incomplete C37C00Ag 2.00 6.2 -- same --- same 15 rin 23-25 .999 complete
C5FPuCOOAg 42.o 99.8 32 17.00 15.9 #2 220 cc 24 b a 24-26 1.55 incCuplete,
7.
C5 1OQA9 15.80 37.5 - sam --- Sam 3 bra 214-26 .969 complete

6. CFOOA9 8.61.18 39.10 Br2 .795 6.67 #5 150 so 6 bra 24-26 .9 inomplete
CI'COOA8 1.935 8.74 - s am --- sam 6 hro 24.e6 .975 complete


In the second part of each of the six eperme ah indicated reactant was added to the reaction mixture,


mi. an additional amount of the


NOTE:





TABi 4


MV F E]ZISS ICVINE PRESENT IN SOLDTIONB OF REACTION )I1JRES BY TIE COLOR COMPARISON MHMOD


REACTION MIXTURE BLANK
strrng 0 Ratio (R) 00. Ratio (B8) B Contents Tm Soln. titer Contents Soln. titer Added i=iv. Added sov. CF3COOAg, 1.103 g. (4.99 M.Eq.) 20 min-. 25 .500 AIy, .34 9. 14.00 .700 1.22
20 miin. 25 1.00 9.60 1.18 1.18 12, .6323 g. (4.99 M.Eq.) 20 min. 25 1.50 (f', 20.00 co. 9.25 1.64 1.09
20 min. 25 2.00 10.05 2.14 1.07 G7, 50.00 cc. 20 min. 25 2.50 ...... 8.64 2.58 1.03
cJ' COO, .5437 9. (2.46 M.q.) 20 min. 25 .500 AgI, .28 g. 12.05 .602 1.20
20 min. 25 1.00 9.00 1.05 1.05 12 .3124 g. (2.46 M.q.) 20 min. 25 1.50 a7, 20.00 cc. 9.6o 1.53 1.02
20 min. 25 2.00 10.10 2.04 1.02 or, 50.00 cc. 20 mi.. 25 2.50 9.28 2.49 1.00 c275C00, 1.4171 g.(5.23 M.Eq.) 30 min. 25 .500 AgI, .614 g. 9.90 .495 .9p
20 min. 25 1.00 lO.45 1.o 1.01 I2 .6641 g.(5.23 M.Eq.) 20 min. 25 1.50 W, 20.00 cc. 10.65 1.55 1.03
20 min. 25 2.00 9.60 2.03 1.01 G7, 50.00 cc. 20 rain. 25 2.50 9,25 2.49 1.00 3?7COOAg", . , 0 1 .(3.3 M.FQ.) 30 Qrn. 25 .500 f .36" . 9,50 .q,4-- .95
20 min. 25 1.00 10.25 ,985 .98 3 .3986 g.(3.14 M.1q.) 20 min. 25 1.50 97, 20.00 cc. 10.55 1.51 1.01
20 min. 25 2.00 10.00 2.02 1.01 G7, 50.00 0o . 20 min. 25 2.50 9.85 2.51 1.00






TABLE 4--C ntimed


C. Ratio (R) c. Ratio (B) B Contents Stirring Soln. titer Contents Soln. titer
Added______ Addled so7 R C4F9COOAg, 2.2482 g. (6.06 M.Eq.) 45 mn. 25 .500 Ag, .50 g. 10.0 .500 1.00
20 rin. 25 1.00 10.50 1.05 1.05 12, .7694 g. (6.06 M.Eq.) 20 min. 25 1.50 Q!, 20.00 cc. 10.10 1.53 1.02
20 rin. 25 2.00 9.40 2.00 1.00 50.00 cc. 20 min. 25 2.50 10.33 2.52 1.01
C5F11cooAg, 2.5829 8.(6.13 M.Eq.) 45 min. 25 .500 AgI, .50 g. 10.00 .500 1.00
20 min. 25 1.00 9.01 .951 .95 , .7791 g.(6.13 M.Eq.) 20 min. 25 1.50 W , 20.00 co. i0.41 1.47 .98
20 min. 25 2.00 10.05 1.97 .99 GF 50.00 cc. 20 min. 25 2.50 9.63 2.46 .99














KEY TO FIa= 1


A.

B.

C.

Wt. D.




G.


H.
I.

I'�

J.

K.

L.

N.

MI.

N.


Drying tube Burette Manometer #1 Manometer #2 Graduated reactor vessel Magnetic stirrer Stopcock Filter-weighing bottle Trap

Saran flex Joint #1 Saran flex Joint #2 Trap

Stopcock Dry-air supply Vacuum Vacuum 5-liter sas bulb







Fig. i


OD-


E -


G


L M


H


C' M'

I I


N -,,,,-,-,1











31

Analysis of Reaction Mixtures

Discussion.--Further information as to the nature of the halogen addends of the silver fluorocarbon oarboxylates can be obtained through the analysis of the reaction mixtures far total oxidizing power and recoverable acid.

A comparison is made between the total available oxidizing power for the couplexes formed with silver acetforate through silver caproforate and bromine or iodine. From these data important ocnolusions as to the formulae and properties of the complexes may be drawn.

Recoverable acid is important from the standpoint of determining whether or not all the (S) is consumed in the reaction forming the readily hydrolizable complex. Also, it serves to determine the extent of side reactions such as deearboxylition, which could eliminate acid from the system. The extent of the occurrence of the deearboxylation reaction, if any, was also checmd by the observation of any pressure changes occurring in the system during the initial reaction and by measuring the weight increase of an ascarite tube attached to the system, as shown at M in Figure 1. No appreciable pressure rise or weight change could be detected for iodine, although the bromine reaction did give some evidence of gs evolution.

In a few cases, recoverable silver was determined as a check on the analytical method to be used in subsequent determinati ne of the silver content of the purified solutions and isolated complexes, in which this factor is highly critical.













Riperimental.--The reactions were conducted as before, using the apparatus in Figure 1 and a Q solvent.

Total oxidizing power was determined by adding XI to the

colorless reaction mixture, shaking until the reaction appeared complete, adding water, and immediately titrating the liberated" iodine with standard thiosulfate.

Liberated acid was determined on separate reaction mixtures by treatment with water, followed by titration with standard base to a phenolpthalein end point. In other cases, a considerable excess of

(8) was employed. The acidic solution resulting upon the addition of water was distilled into a closed receiver. More water was added and the process repeated until all the free acid was obtained. This was then titrated to a phenolpthalein end point with standard base.

Recovered silver was determined by the Denigses cyanide method.

Experimental data and analytical results appear in Table 5.

Results of the above measurements.--The results obtained lead to the following conclusions:

(1) In the case of both bromine and. iodine, one equivalent of oxidizing power is available for each equivalent of halogen consumed.

(2) The total oxidizing power is somewhat low for the reaction mixture of silver acetforate and X, but closely approaches the theoretical value as the molecular weight increases.

(3) The reaction time for (S) with X appears to increase with molecular weight of (8) and in going from Mr2 to 12.

(4) When the total acid, both liberated and present as the













silver salt, was titrated in the reaction, mixture, an equivalent of acid was found for every equivalent of (S) originally present. In cases in which the original excess of (8) was large, AgCE was preoipitated due to the reaction of the excess Ag ion in solution with the titrating base. However, when the liberated acid was distilled from the mixture in order to separate it from the unreacted (S), one equivalent of acid was found in the distillate for each equivalent of X originally present in the reaction mixture. Thus, the acid liberated from the reaction mixture upon hydrolysis with water is independent of the amount of (S) present in excess of the 1:1 equivalent ratio.

(5) In several instances the precipitate obtained upon treatment of the reaction mixtures with water was separated, washed, and vacuum dried. This precipitate was observed to possess oxidizing power when treated with I solution. The precipitate was presumed to contain AgIO3.

(6) In every case the precipitate obtained did not seem to be as subject to decomposition by light as is AgX obtained by more conventional methods. Samples of the precipitate obtained after treatment of the reaction mixture were washed and vacuum dried. These were then analyzed according to the Deniges cyanide method and were found in every case to be almost pure AgI.

From the above findings, the equation for the reaction of the complex with KI may be written

(e0o)2ASI + 2 KI- . AgX + 2 GC00K + 12










34

while the reaotim with water may be written as before, i. e.,
(ecoo)2Agt + 20 -a--AgOX + 2 GCOOH

3 AgOX v. 2 AgI + AgXO3






TANI 5
P ARATION AND ANALYSIS OF REACTION )MMUS OF SILVNI SAIM[ OF
FPUMORARBON CA1MCBY1C ACEII An) IOMIPI (M ] OMINE

A. Prepiratio


Mean Rw.atioa
Taimp. Time
(s) go X.Eq. x 9. M.Bq. Solvent cc, 00 (Min.)

CF3COOAg 1.098 4.97 12 .6305 4.96 F# 2 11 25 90 CF3COOAg .R. , 6.69 12 .828 6.49 ) # 2 12 26 90 C2F5COQAg 2..558 9.44 12 . 6 8.73 QF3 20 N 3I00 ...C cAg 2.7140 8.45 12 1.59 8.06 BF 3 15 26 480 C1479cooAg 4.8305 13.0 12 1.5945 12.55 QF3 30 25 600 C5FlCOOALI 2.3930 5.66 '-2 . 6908 5.439 3 3 30 27 840 C5F9lCOOAg 5.2500 7.71 12 .9105 7.17 OF 5 37 26 360 CycoaAg 1.6-452 7.44 Br2 .5916 7 4 r .. 2 140 25.... CFCOOAg 10.5751 47.8 Br2 3.7293 46.67 0; #3 1,50 26 480 c9C1 O0Ag 1.9688 6.195 T-2 .7343 5.782 B)' 4 75 25 300 C397COOA9 1.8218 5.675 12 .6975 5.1+92 or ~ 4 75 26 300








B. Analie


Total Od4izmn
Power Recovered Acid Recovered Ag
f4% Of Theory Based on % of Theory Based on M.Bq. Theory M.. Orig. X Orig. Ag 14.B. Orig. I OrIg. Ag
4.1168 83.00 4.92 98.9 98.8
5.3487 82.41
7.9459 91.02
7.66 9.5 .. 8.36 103 98.9 12.0355 9.5.1 . H,,
4.. .9798 91.5

7.68 107 99.6 7.45 1o 96.6 76 77.13 96.4a 95.7 54(t ) .... 97.3 95.0

96.0 89.6 6.18 107 99.6
__ _ 5|, .....52 100 97.3 _5.61 102 98.9













Analysis of the Complexes in Solution

Discussion.--In the following series of experiments, clear solutions of the complex in the GF solvent are analyzed for the same quantities determined for the entire reaction mixtures in the previous section. The complicating factors due to precipitated AgI, unreacted

(S), and undissolved complex are eliminated. The acid number, total oxidizing power, and recoverable silver are obtained as functions of the amount of complex present in solution.

Since the complexes are only partially soluble in the 9F solvent, the greatest significance will be in the comparative values of the three measured quantities. After the samples of solution are removed., the residual solution and precipitated mixture can be titrated and the suiation of all the quantities compared to the theoretical totals.

If the formula for the complex is (C00)2AgX, and if its reactions with KI and 120 are represented by the previously-cited equations, then the total oxidizing power in equivalents should equal the acid number, while the recoverable silver should be one half that value.

Also included in these experiments are two designed to determine the relative stability of solutions of the complexes containing bromine as compared to those containing iodine.

Observations of the behavior of the complexes of bromine have indicated that they are less soluble in the solvents employed, and that the solutions are less stable, losing their measurable activities











38

if allowed to stand for any period of time.

Total oxidizing power was chosen as the parameter for the measurement of the stability of (QCOO)2Ag r in solution, since this quality has been demonstrated to be reasonably dependable and independent of excess (8).

Experimental.--The complexes were prepared as before, using the apparatus in Figure 1 and $everal of the different fluorocarbon solvents. As soon as the reaction was complete, the precipitate was allowed to settle until no particles remained suspended in the clear supernatant liquid.

A special pipette was constructed with a stem sufficiently long to reach into the apparatus from the top and remove 10 ca.

samples. A tiny chamber containing calcium chloride was sealed into the stem to eliminate the possibility of contamination of the solution by the operator. The lower end was further constricted to prevent loss of the dense, non-viscous solution during transfer.

The pipette was dried by passing a stream of dry air through it. The top was removed from the reactor, while maintaining a stream of dry air from L. The pipette was partiaily inserted and allowed to remain suspended in the air stream for a few minutes to remove water from its outside surface. It was then lowered into the supernatant liquid and the sample obtained. The pipette and sample were then removed, the sample rapidly transferred to an erlenmeyer flask and titrated immediately. This process was repeated until the solution level was too low to permit removal of samples without the danger of











39
their being contaminated with precipitated material from the bottom of the reactor. In the last two experiments a measured time was allowed between removal of samples for titration. The reactor vessel D was then removed and the remaining material titrated directly in the vessel. Experimental and analytical data appear in Table 6. In Part A of this table, values for oxidizing power, liberated acid, and recovered silver for samples removed from reaction mixtures as doscribed above are tabulated. In Part B, values of total oxidizing power for samples removed from reaction mixtures at measured time intervals are tabulated. The time variation of this value for samples of solutions containing (05OO)r and (%7O0)2Agr may be compared.


Results of the above measurements.--The tabulated experimental data and results clearly indicate that the theory cited on page 33 is accurate and that the complexes in solutions separated from precipitated material react with water and KI in accordance with the suggested equations.

The complexes formed with iodine are evidently quite stable in solution, since no marked decrease was noted in the measured quantities during the period of measurement.

On the other hand, the solutions of complexes of bromine

rapidly lose their oxidizing power on standing. Furthermore, their initial solubilities in the solvent are somewhat lower than those of the iodine complexes.

For these and other reasons, iodine was used more frequently than bromine in the foregoing and following measurements.





TABIR 6


ANAUIYBS 0F THE C(OWL IN SOLlfION IN FIMOPCABDON SOEn'IM3

A. Relatio ship Between Acid N=br, Oxidizing Power and Rveable Silver

Cotet of th Oxidizing idM No. I Iecoverable
Reaction. Mitrssample f Power (X4Dia.) IAg. (M.Ea.)()
.e. .........(cc.) . ) .)(oA)I Nt5. o (17


1. CF 3COOAg,

12.0
OF #2,


2. C3F 7CooAg

12,

OF#3,


.8517 g. (3-.76 M.Eq.) 6 59. (3.c. M.A.) 65 cc.


1.600 s. (3.30 M.1k.) .43 g. (3.23 M.Eq.) 65 cc.


10 20 10
residue
total


4 4i


10
10 10 10
10 residue total


.00993
.0196 .00990
3.17.
3.2111


.2976 .2857 .2877

1.2598 2.2726


3-. C59jPO0A49 1.71 9. (4.06 M.Eq.) 10 .0516 10 .050
12 .740 g. (3-.732 .Eq.) 10 10
QF #3, 65 c. rsdu 3.3067 total 3.5.60


Insufiient material in selution for


3.729 3.729+


1.17
1.16+


.11477 . 91 .11.5 ..89
.11.98 0521. .1187 .522 .1436 .5o 2.6869 2.13
3.522. 1. 6
.o226 .438 .0240 .475


3.7296 1.128 3.9374 1.109


NOTE: In obtaining totals for oxidizing power and recoverable silver, appropriate values are
substituted for those not determined.





TABLE 6--Continued
B. Ccmusrieen of Stabilities of Complexes in Solution

Oxidizing Aige of Contents of the Sample Power Sample
Reaotion Mixtures (cc.) (M. j.) (min.) . c2F5coAg, .986 g. (3.611 M.]E.) 5 .3050 15 5 .2910 633 12, .71 9. (3.520 .zf.) 5 .2950 1008 residue 2.6170 oP #4, 65 cc. total 3.58

5. c2P5co0A, 1.4121 g. (5.211 M.Dq.) 10 .4010 10 10 .2650 130 Br, .3987 9. (4.987 M.Eq.) 10 .1830 250 10 .11Wo 370 OF #4, 120 cc. 10 .1010 490 10 .0050 1330 residue 3.5350 total .63oo













Isolation of the Solid Complexes

Discussion.--The problem of isolating the solid complexes from their solutions was not easily solved. It was necessary to obtain solutions absolutely free of precipitated AgI and (QCOO)2AgI, and to devise a method for separating the complex from the solvent without decomposing it in the process. The entire operation would, of course, have to be carried out under strictly anhydrous conditions.

A number of attempts to remove the solvent from the complex by distillation were mde. This invariably resulted in the decomposition of the complex, even when the distillations were carried out under vacuum.

It was discovered that when clear solutions of the complex in a low-boiling QF solvent were chilled to -780, white crystals precipitated. These were presumed to be the complex, and the apparatus (Figure 1) was adapted to effect their separation by this method. The complexes, thus removed by crystallization and filtration, were weighed and analyzed for total oxidizing power, acid equivalent, and recoverable silver.

Experimental.--The entire apparatus was very carefully dried by alternate evacuation and admission of dry air. The complexes were prepared by the previously-described method, employing as a solvent the low-boiling fluorocarbon mixture (#il). During the preparation of the complex, carried out in the flask desi&iated D in Figure 1, the remainder of the system was subjected to more rigorous vacuum drying, with special attention being paid to the filter-weighing bottle G.











4~3

The latter was specially constructed, using a "fine" grade

sintered-glass disc as the filter. It was designed to be sufficiently light for use as a weighing bottle (weight 55.4701 g.) and sufficiently compact to fit inside a 1-liter dewar; yet it had a capacity of 42 cc., which enabled it to hold enoug solution to assure that a measurable amount of complex could be obtained. It was connected to the system by means of 10/30 standard taper joints, as shown in the diagram. Teflon plugs were provided for insertion in the inside of the male joints on the bottle during the weighing process.

After the reaction had reached completion, the precipitate was allowed to settle. G was chilled in dry ice-acetone. Stopcock F was then opened and the solution allowed to pass into G. Trapped air in G was vented by opening stopcock K. Any suspended particles were removed from the solution by a glass-wool plug in the line between D and F.

After sufficient time had been allowed for crystal formation, suction was applied by opening stopcock M'. The solvent was drawn from the filter-weigh bottle, leaving the precipitated complex on the septum. The solvent was collected in dry-ice trap H or liquid-air trap J.

Since the solution had been removed from D, the stopcock K

could be opened and a stream of air allowed to pass through C, removing any remaining solvent.

K was then closed. The dewar containing dry ice-acetone was

removed from G and pumping continued one or two minutes while allowing the filter-weighing bottle to warm slightly. This was to assure













complete removal of the solvent. The stopcock K was opened, pumping stopped, and G removed from the apparatus. The teflon plugs were inserted, stopcock grease removed from the joints, and the bottle weighed on the analytical balance.

After this, the contents of the bottle could be titrated either for acid or for total oxidizing power. The titrations were performed directly in the filter-weighing bottle, using a micro burette.

The precipitated silver residues were then washed free of

titrating solution. The samples which had been titrated for available acid were treated with a small amount of thiosulfate solution to reduce any AgIO3 to AgI.

Excess standard KCN solution was added and the bottle stoppered and allowed to stand with occasional shaking until all the AgI was dissolved. The solution was then transferred to an erlenmeyer flask and titrated with standard silver nitrate according to the Deniges method.

Experimental conditions and results appear in Table 7. In this table the experimental data relating to the preparation of the complex to be isolated is tabulated, along with the amounts of complex isolated in each case. Values for oxidizing power, acid number, and recovered silver are tabulated. The corresponding theoretical values are also listed for comparison.

Results of the above measurements.--The elusive intermediates in the reaction between (S) and 12 have been isolated and identified. Their formula has been shown to be (QCoO)2AgI. The technique for their separation and analysis has been described. The corresponding











145

complexes for bromine were not isolated.

(GC00)2AgI, though evidently quite stable in solution, is prone to decomposition when removed from the solvent. The complexes were isolated at -780 as white crystals which rapidly took on a yellow-brown cast when warmed to room temperature. Evidently the occurrence of partial or total decomposition of the complex is marked by this color change, since in early experiments when the complex was allowed to warm to room temperature during the latter part of the separation prior to weighing, the silver content was high in comparison to the acid equivalent. In these cases the precipitate had developed the yellow color.

It can then be concluded that the complex exists as the stoichiometrio entity (GCOO)2AgI only in solution or at reduced temperatures.

The fact that the precipitated mixture of (eCoo)2Agi and AgX is stable, retaining its oxidizing paver over long periods of time, and that equivalent mixtures of (S) and X combined in the absence of a solvent produce the observed homogeneous yellow powder which is

stable in the absence of moisture or reducing agents, would indicate that the complex is stabilized by the presence of an additional mole of AgX with the formation of a material of the stoichiometric formula GCOOAgX which resists addition of further X.






TALE 7
PREPARATION, ISOLATION, AND ANALYSIS OF M SOLID COMPLEYLS

PREPARATION ANALYSIS Sol- Weight M.Eq. Oxidiz- M.q. Acid
(s) go M.Eq. 9 .. vent, 'Of ingPor Numer Eq. AgI g. CC. ct I -r *.eo,
8.

C3F7COOAg 2.400 7.47 .9418 7.41 # .14504 .2807 1.362 .2285 .6813 50

CF3COOAg 1.5308 6.926 .8778 6.912 # .1757 .6490 .7622 .3957 .3811 50 C2F5COOAg 1.4233 5.253 .6471 5.095 .52560 .8659 .9126 4613 .4563 C3FTCOOAg 2.9342 9.14o 1.1378 8.959 #0 2125 .622o .64o .398 .3210 50 .125 .622l .u42 .*198 .210 C4F9COoAg 2.5466 6.864 .8674 6.829 65 .1989 .50 .5230 .2600 .2615

C F11COOAg 3.2135 7.633 .9390 7.394 #1 .236o .5320 .5466 .2730 .2733
5 ~651 11


















USES CF THE COMPLZIES IN THE SYNTHESIS OF FILTROCARBON DERIVATIVES


General Discussion

Mixtures of (S) and X have been found to be capable of

halogenating aromatic substances. Haszeldine and Sharpe12 and Henne and Zimer18 have investigated this reaction. Earlier investigators arrived at similar conclusions when investigating the reactions of the silver salts of the ordinary carboxylic acids.

While the present writer cannot, in view of the above-cited findings, agree with the mechanisms advanced by Easzeldine and Henne and their co-workers, it was decided that further investigation into the reaction to clarify more thoroughly all the factors involved should not be undertaken at this time.

The overall equation for the reaction may be written:

OCOQAg + k2 + ArE -- QCOOE + AgX + ArX

The above investigators have demonstrated this to be a highly useful and selective means of obtaining ring-halogenated arouatic compounds.

The reactions involving decarboxylation of (S), either with or without other materials being present, are discussed in the introduction.

When equivalent quantities of (S) and X are mixed in the presence of a solvent and the mixture heated to 120-1500, CO2 is 'V7













evolved; and a mixture of products is formed, consisting of the fluorocarbon mono-iodide, the fluorocarbon carbcxylic acid anhydride, and a small amount of straight-chain fluorocarbon (usually containing twice the number of carbon atoms as the original fluorocarbon radical in the starting material).

When an excess of halogen is present, the reaction proceeds toward the formation of almost 100 percent halide by the previouslycited overall reaction

QCOOAg + + AgX + 9I

Although mechanisms involving the complex (GCOO)2AgX might be postulated for reactions of this type, the present writer feels that nothing would be gained through the suggestion of such mechanisms without the confirmatory evidence which could be gained only through a kinetic study of each process for which mechanisms are advanced.

One immediately is struck by the possibility of substituting other elements or functional groups for the halogen involved in the complex, or substituting other elements or compounds for the second equivalent of halogen whose presence in the reaction mixture makes possible high yields of fluorocarbon halides. The manner in which the decarboxyation would proceed in the presence of these substitute ingredients presents an intriguing problem to the investigator. Much of the following experimental work can be classified under this heading.

The ability of the Simonini complexes to add to olefinic substances with the formation of halo-esters followed by di-esters, as











49

previously described under the work of Prevost,30 suggests an interesting line of investigation. Reasoning by analogy (an admittedly hazardous procedure), one could suggest that a reaction of the following sort might occur

(QCOO)2AgX + -C-C-M- O0CE-CH I + QCOOAg A displacement reaotion of the type

GCOO-CH-CEX + QCOOAg -. OCOO-CHOOCO + AgX

might then occur. If the reactions were carried out using a fluorocarbon olefin, the first reaction might conceivably occur. However, the second step would be precluded by the demonstrated fact that bydrogen-free fluorocarbon esters are not obtainable by a reaction of the latter type.

It was shown by Hauptchein and Von Grosse,15 who attempted to prepare a hydrogen-free fluorocarbon ester by reacting CF 3I with C3 COOAg in a bomb at elevated temperatures, that ester formtion did not occur to any measurable extent. The present writer independently arrived at similar conclusions through the reaction of CF3COOAg with 03F71 in a sealed tube at temperatures up to 2500.











50

Substitution of Pyridine for X in the Complex-Forming Reaction

Discussion:--In view of the unusual capacity of pyridine for

complex formation, it was decided that an experiment involving the substitution of pyridine for I in the reaction of (8) with X would be produotive of interesting results -- perhaps leading to the development of a route to alkforylated pyridines by a path of the type

QCOOAg + Py -oC02 + 9Fy + Ag

Pyridine was found to react with (S) at room temperature and in the absence of a solvent in a 1:1 mole ratio. A solid addendum compound, which was isolated and identified, resulted. The preparation and identification of these substances and their subsequent reactions are described in the experimental part of this section.

Experimental -- A. Preparation, purification, and identification.--The substances of the formula GCOOPyAg were prepared from the silver salts of acetforio acid through valerforic acid. Equimolar quantities of pyridine and (S) are combined under anhydrous conditions at room temperature. Power stirring is employed to mix the reactants thoroughly. An exothermic reaction occurs, which warms the reaction mixture to 45-600, depending upon the amount of material involved. After the heat of reaction has dissipated, white needle or rhombic crystals are formed. These crystals have the above formula and may be used without further purification in the reactions to be described in the "B" part of this section. However, establishment of formula necessitates removal of any excess pyridine or impurities that may have been present in either reactant.











51

It was found that (OCOO)PyAg is soluble in benzene, chloroform, ether, and boiling heptane or hexane. In the case of ether or chloroform, however, secondary complexing involving the solvent occurs. A seccmd mole of pyridine, if added to this system, results in the precipitation of an unstable, hygroscopic complex, whose formula is approximately (GCOO)Py2Ag. Removal of ether results in gradual decomposition of this material, with the loss of pyridine. An oil results, from which GCOOPyAg can be recovered. For example: C2F5COOPyAg was dissolved in ether. The solution was chilled to -780 in crder to effect recrystallization. No precipitation occurred. A second mole of pyridine was added. White cyrstals immediately precipitated, thus indicating that C2F5COOAg probably can react with 2 moles of pyridine to give a substance of the formula (C2F000)Py2Ag. The crystals were separated, and ether was removed in vacuo. A white, highly hygroscopic powder was obtained. The powder had an odor of pyridine and upon analysis was found to contain 26.8 percent silver (theoretical for (02F 5COO)Py2Ag is 25.17 percent). This material,was freed of excess pyridine by washing with anhydrous heptane. The resultant product was found to contain 29.94 percent silver (theoretical for (CF 5COO)PyAg is 30.86 percent).

A second experiment of this type, employing a 2:1 mole ratio of pyridine to (S) was conducted. Silver valerforate, 18.5 g. (.05 mole), and pyridine, 7.9 g. (. 1 mole), were combined as before. A heavy oil resulted. An attempt was made to remove the excess pyridine by extraction with heptane. This was unsuccessful; the oil











52

remined undissolved at room temperature. Upon heating, approimtely half the oil went into solution. When the solution vas cooled to 00, the oil separated unchanged. The heptane vas removae, and the oil diesolved in 100 c. ether. Upon oooling, crystals were obtained. These were washed with anhydrous heptan. and recrystallized from boiling heptane, followed by drying in vacuo. The product, a rather yellow powder, was found to contain 23.49 percent silver (theoretical for (C49CO0)PyAg is 24 percent).

The pure addends were prepared by combining (S) with pyridine in a 1:1 mole ratio and recrystallizing the product from boiling heptane. This solvent was found to be most suitable for the purpose, although the solubility of (QCO0)PyAg in this medium is rather low even at the boiling point. The solubility was found to increase with chain length. Resistance of (OCO0)PyAg to hydrolytic action by atmospheric moisture was found to increase with chain length. The analyses were found to approach more nearly the theoretical value as chain length increased.

Data pertaining to the preparation and analysis of the pure addends appear in Table 8.

TAM 8

PREPARATION AND ANALYSIS C PY1IDMI A S CF THE SIVJER FLUKOCARB" CABOIYLATES

(S) g. mole pyridine mole Product round Theor. Q7,COOAg 10,0 .045 3.5 .045 (FCOO)PYA 35.1. 36.00 C2?,COOAg 2.0 .7 .57 .007 (0275000)PYAg 30.31 30.86 C3;.7CO0Ag 37.1 .11 6.81 *i6 (C397000)PyAg 26.6l42.0 C4?9COAg 37.1 .10 7.9 .10 (01479000)PyAg 2375 24..00











53

B. Reactions.--If the addends are heated,, doecarbazylation

begLns to occur at 1600. Oarbon dioxide and fluorocarbon hydride are evolved. Free silver and pyridine are formed, along with a complex

mixture of compounds containing fluorine. Pyridinium hydrofluoride and a hieh-boiling oil, tentatively identified as a difluorobypyridyl., were isolated from this mixture.

When the addend resulting from the addition of pyridine to C23P00OAg was decarbarylated., the above products were identified as follows: Zthforyl hydride was obtained in the cold trap on the exit line from the reactor. It had a molecular weight of 119 (oalo. 120). Its infrared spectrum contained the characteristic C-F and C-H absorption bands. If water was excluded from the reaction, 02F95 was present in approximately half the total amount available from dearbaolation of (s). If, however, the reactants are not thoroug3j dry, a quantity approaching the theoretical value is obtained.

Pyridinium hydrofluoride, obtained in the distillate of the

reactor reuitue, was purified by recrystallization from ethyl alcohol. It was observed to liberate fluoride ion upon solution in water. Treatment with dilute alkali liberated pyridne. Pyridinium bdrofluoride was prepared by reacting pyridine with hydrofluoric acid and recrystallizing the dried produot from anhydrous ethanol. Both materials melted 158-1620. However, a mixed melting point was depressed to 148-1580.

An oil boiling 300-3040 containing both fluorine and nitrogen was isolated. This material, when dissolved in anhydrous ether and











54

treated with hydrogen chloride, imediately gave a white precipitate. After repeated washings with ether followed by vacuum drying, the precipitate was analyzed for chlorine and titrated for HCi content.

Two samples analyzed 27.14 and 27.25 percent chlorine. (Caloulated values for wonofluorobypyridyl dihydrochlride and difluorobypyridyl dihydrochloride are 28.51 and 26.79 percent, respectively.) One sample analyzed 27.93 percent HCl. (Calculated values for the above compounds are 29.3 and 27.54 percent, respectively.) The hydrochloride was found to contain fluorine. These analytical data indicate that the material was a difluorobypyridyl.

The solid residue remaining in the reactor was thoroughly washed free of organic material with ether and heptane, dried, and analyzed for silver. The silver content was found to be 98.9 percent.

Substantially similar results were obtained from the decarbcxylation of the complex resulting from the addition of pyridine to silver acetforate. If pyridine and (S) are present in a 1:1 mole ratio, deoarboylation does not occur until a temperature of 220-2400 is reached, and considerably more charring and formation of pyridinium hydrofluoride occurs.

Since it was suspected that alkforylaticn of the pyridine

nucleus had occurred to a small extent, it was decided to introduce into the reaction mixture an aromatic substance which would be expected to undergo alkforylation much more readily than pyridine. Experiments were conducted using 2-butyl benzene and meta di-isopropyl benzene. No alkforylation products were isolated.













Results of the above measurements.--Pyridine reacts with (S) in a 1:1 mole ratio to give a crystalline solid, whose analysis indioatcs it to contain 1 mole of pyridine and 1 mole (s). This substance will decarboxylat' at 160-1700, yielding the corresponding fluorocarbon hydride in high yield, along with pyridinium hydrofluoride, mono- and/or di-fluorobypyridyl, and a complex mixture of oils containing fluorine.

This reaction should be investigated more thoroughly, as it shows promise of providing a route to alkforylated pyridines or aromatics.











56

An Improved Method for the Synthesis of Plurocarbon Mono-lodides8

Discussion.--A method of preparing fluorocarbon iodides which consisted of the decarbcuylation of the silver salt of a fluorocarbon oarboxylic acid (8) in the presence of iodine was developed by Simons ard Brice.38 The reaction mixture was kept in suspension in a fluorocarbon diluent by means of mechanical stirring. The process was carried out under strictly anhydrous conditions.

OtherslO,l4,16 prepared the iodides by heating intimate mixtures of (S) and iodine.

The method herein reported consists of bringing the iodine vapors in contact with (S) under controlled conditions of temperature and pressure without a solid or liquid diluent being present. A sharp reaction zone progresses through the mterial until (S) is consumed. The fluorocarbon iodide passes from the reactor and is collected as in the original method. Power stirring, arni prepurification and drying of the iodine, as well as the limitations imposed by the diluent, are eliminated. Maintenance of anhydrous conditions, recovery of silver and control of reaction temperature are facilitated. Higher product yields are obtained.

Two types of reactors are currently in use in these laboratories, each having advantages depending upon the particular situation. Procedures using these reactors are:

PROCEDURE 1: This method employs a reactor which is simple

to construct and provides a quick method of obtaining small quantities











57

of fluorocarbon iodides in the laboratory. It is also adapted to reactions of this type in which the absence of any atmosphere other than the reaction gases is desired. The reactor is a vertical Pyrex tube, closed at the bottom and stoppered at the top. An exit line attached near the top connects it through two cold traps to a vacuum source. A manometer is attached to the exit line. A KOH tube for removal of CO2 is placed between the traps and the vacuum source. Reaction temperature is followed by means of a thermocouple or thermometer extending into the tube through the top.

Erperimental.--The entire apparatus is vacuum dried. Iodine followed by phosphorus pentoxide (P) is poured into the tube. (8) may be either suspended over the iodine and (P) in a steel-wire cage or poured directly into the tube with a layer of glass wool separating

(S) from (P). The apparatus is flushed with dry air. The reactor tube is heated in an oil-bath. The pressure is reduced in order to provide adequate volatilization of iodine at the optimum reaction temperature (13o-16o�). Completion of reaction is noted by cessation of evolution of gases. The bath is heated to 1850 to assure complete conversion of (S). The iodide is collected in the cold traps. All products were fractionated. Iodine was determined by peroxide decomposition and thiosulfate titration. Reaction conditions and yields are in Table 9A.

PROOAMDRE 2: This method is superior in cases in which relatively large amounts of various fluorocarbon iodides must be synthesized in the laboratory. The apparatus is shown in Figure 2.













Operation is at atmospheric pressure. A stream of dry air carries the iodine vapor from the pot to the reaction zone. (8) is supported in the reactor by means of a glass "X" member and a thin layer of glass wool. The reactor and the iodine container are separate units whose temperatures are controlled independently.

Experimental.--The apparatus is assembled. (8) is poured into the reactor from the top. Final drying is effected by passing a current of dry air through the reactor while heating it to 70-800. Excess iodine followed by (P) is placed in the pot. The iodine is heated to sublimation temperature and an extremely gentle air current started. The heat of reaction tends to increase the reactor temperature. The jacket heating is adjusted so as to maintain an optimum reaction temperature of 130-1600. The reaction requires 20-40 minutes, depending on the type and amount of (S). The iodide is collected in cold traps. Purification and identification were carried out as in procedure 1. Reaction coroditions and yields are in Table 9B.

Procedure 2 was used in the preparation of CF7Br. No air current was employed. Reflux action of bromine was provided by an ice-water-cooled condenser on the upper end of the reactor. Reaction temperature was 130-150�. Products were fractionated and analyzed for bromine. A 67 percent yield was obtained.













TABLX 9

IMPROVED SYNT!EIS FOR FUO1OCARBON IODIIES

A. Procedure 1


Heat- Operat- Operating
Starting 9* ing ing Pere., Basket Yield, Yie3d Material g" Time, Temp., Pressure, Support 0.
Min. oCm.


CF3COOAg 75.1 60 1io-15O 400 -500 Not used 47.6 71 C!.COOAg 15.5 80 140-145 450-500 Not used 13.1 93 CT3COOAg 75.4 140 150-160 400-500 Used 52.5 78 C3F7COAg 21.2 115 150-160 400-5o0 Used 16.5 87



B. Procedure 2

Starting Heating Temp., Yield., Yield, Material g" Time, Min. �c g. % CF3COOAg 33.8 21 150-160 22.3 74 CF3cOAg 20o2.1 48 15-16o 161.0 90 CI _CO0A i 2.o 25 130-160 36.5 94 C3F7CO0&g 200.4 45 150-160 175.5 95






Fig.


Reactor


Dry Or


Hedee co I Jacket apertwre
S
'Supor*


P
Iodine











61

Reaction of (s) and the Complex with Phosphorus

Discussion.--Bennet, Brandt, Zmeleus, and Haszeldine2 have described the formation of trimethforyl phosphine and iodomethfaryl phosphines from the reaction of phosphorus with methforyl Lodide in a closed vessel.

It was decided that the decarboxylation of (S) in the presence of red phosphorus might provide a source of trialkforyl phosphines without resorting to the preparation of the iodide and without obtaining the iodoalkforyl phosphines as by-products.

Instead, the reaction yielded the fluorocarbon oarboaylio acid anhydride. The red phosphorus was converted to the yellow allotropic form.

Haszeldine and Sharpe13 have since reported that a mixture of red phosphorus and silver acetforate yielded acetforio anhydride upon thermal decomposition. The results independently obtained in these laboratories are nevertheless reported here, in some detail, as confirmaticn and clarification of the work of the above investigators.

The reaction of red phosphorus with the solid complex formed from the reaction of silver aoetforate with iodine in the absence of any solvent is described.

LXperimental.--Silver acetforate,75 g. (.34 mole), was mixed with red phosphorus, 26 g. (.84 mole). The mixture was placed in a flask to which was attached a gas collection train. An attempt was made to initiate a self-propagating reaction by heating a small spot on the flask with a gas torch. CO2 evolution immediately occurred













but stopped as soon as the flame was removed.

The entire flask was then heated with a heating mantle. Decarboxylation bean at approximately 2000. Heating was continued until 002 evolution ceased. The product was fractionated. The fraction boiling 38.5-400 consisted of 32.5 g. of material. A .5001 g. sample was treated with water. The material reacted violently, liberating acetforio acid which was neutralized by titration of 37.64 oc. of .1266 N. XaOK (correct for acetforic anhydride).

The reactor residue consisted of silver oxide and yellw phosphorus.

Silver aoetforate, 45.4 g. (.2054 mole), was mixed with iodine, 25 g. (.0984 mole), in the absence of a solvent. A homogeneous yellow powder resulted, having the composition CF300OAgI. To this was added a slight excess of red phosphorus. Upon stirring, a homogeneous gray powder was formed. This material spontaneously, and without warning, exploded -- doing considerable damage to both the apparatus and the digaity of the present writer. No products were obtained.

Results of the above eaurements.-Silver aoetforate (and

presumably the higher salts) when decarbcxylated with red phosphorus in the absence of any solvent. gives almost a 100 percent yield of acetforic anhydride. The red phosphorus is converted to the yellow form in the process. The equation for the overall process may be written

2 CI'3 OOAg + (M) P (red) 3j, (r3co)2o + AV + () P (yefla)

The complex mixture of the formula CF3COQAgI reacts spontaneously with red phosphorus.


















RELATED SYNTMCTIC PiROCDUHK8


Preparation of Fluorocarbon Nitriles from (8) and Cyanogen

Discussion.--A novel method for the preparation of fluorocarbon oarbazylic acid nitriles which apparently is analogous to previously describmd fluorocarbon halide syntheses was developed. The method involves the decarbozylation of the silver salt of the fluorocarbon oarbxoylio acid while passing cyanogen ips through or over the reacting mass. In this procedure, use is made of the halogenoid properties of cyanogen. The overall equation for the reaction may be written

QGOOAg + NMCN -)> 0W + Ag= + 002

Inaumich as the process was conducted at high temperature and no intermediates were isolated, the author will not hazard the postulation of any mechanism or the suggestion of an initially-formed reaction product involving (9) and cyanogen.

The fluoroarbon carboxylio acid anhydride was isolated from the reaction product along with a small amount of straight-ohain fluorocarbon which proved to be the dimer of the fluorocarbon radical involved. The equations for these reactions in which d carboxylation occurs without the involvement of cyanogen may be written
2 OCOOA.g - (eoo()2o + A,20 and
2 QC00A -* e + 2 C02 + 2 Ag 63













respectively. The deoarboxoylation of (S) in the absence of other reactive materials has been investigated by Kirshenbaum, Streng, and Hauptsohein.21 Similar products were obtained.

Experimental.--The procedure used in the preparation of

butyrforo nitrile is described as follows: The reaction was carried out in a Pyrex tube heated by a muffle furnace. Silver butyrforate mixed with glass chips was packed into the tube. The temperature was slowly raised while maintaining an atmosphere of cyanogen in the tube. The reaction begmn at approximately 3000. Carbon dioxide was detected by means of a calcium hydroxide scrubber on the exit line. The cyanogen was allowed to stream in at one end while the effluent gases were collected in cold traps connected to the opposite end. The reaction evidently is greatly hindered by the difficulty of providing adequate silver salt-cyanogen contact during the reaction process. The low yields are attributable to this factor, along with the above-mentioned side reactions.

The products were fractionated and identified by physical and chemical properties. A 4.1 s. fraction of clear, inert liquid was obtained, boiling at 55-570 and having a molecular weight of 335 as determined by the Dumas method. (Calc. for C6F14 is 338.) A 6 g. fraction was obtained, boiling at 103-1060. This material reacted violently with water, liberating butyrforic acid. The latter was converted to the amide, which melted at 103-1050 (correct for butforamide). Essentially the same procedure was followed in the preparation of valerforo and caproforo nitriles. In every case the nitriles were










65

hydrolize& to the acid by treatment with H2S04 and converted to the methyl ester, from which the amides were prepared. For oomparieon purposes, the same nitriles were prepared by the conventicnal method involving dehydration of the acid amide.

Two previously unreported nitriles, C419CN and C57FCN, were prepared. Nitrogen analyses on these were performed by Peninsular CheuRe search, Ina.

Experimental data appear in Table 10A. Analytical data appear in Table 103.

Results of the above measurements.--The investigation was

undertaken in order to determine if the previously described method for the preparation of fluorocarbon halides by the deoarboxylation of

(S) in contact with the vapors of X could be extended to the use of omipounds described as "pseudo halogens" or "halogenoids" in place of X. The reaction reported, though hardly a practical means of preparing fluorocarbon oarbcylic acid nitriles, serves to illustrate the "halogenoid" character of oyanogen as related to reactions of this type












TABIZ 10
PREPARATION OF FLUOROCARBON NIIIZS BY TEX DECARBOXYIATION OF SILVER SALTS OF FIUOROCARP" CARBOXYLIC ACIDS IN THE PRESENCE OF CYANOOFN A. Experimental Data

Reactor Nitrile Yield,
(8) g. moles Temp.,�OC Obtained g. moles CF7COOAg 40 .125 340-35 0 5F CI 5.6 .029 23

CF9CGAg 88 .237 350 c47'90N 7.5 .031 13.1

C9511C00A8 85 .203 350 C9'iiCN 6.1 .021 10


B. Analytical Data

Mol. Nitrogen, % Amide Mixed
_itrile Source B.P. Wgt. Calc. Found M.P. M.P.
From (s) -1 - +1 143 103 03770N Literature -1 - +1 145 105

From (8) 28.5 196 5.72 115.63 110 From Amide 28.0 196 i1l From (s) 56.0 244 4.75 4.55 115
C rlFi Fro Amide .0 2 114 114













Synthesis of Fluorocarbon Mono-Iolides from Fluorocarbon Carboxylic Acid Anhydrides

Diecussion.--It was discovered that fluorocarbon mono-iodides can be prepared continuously by passing a vaporized mixture of iodine and fluorocarbon oarboiylic acid anhydride through a hot tube containing silver iodide. In this procedure, the anbydride is made as a part

of the overall process, thus providing a method for going from the fluorocarbon carboxylic acid to the corresponding iodide in one prooedural step. The overall equation for the reaction .my be written

(9COO)20 + I2 2 GI + CO2 + CO

The function of the silver iodide is not clear. However, its beneficial effect is evident upon comparison of the results obtained with and without this material in the reactor tube.

Haszeldine12 prepared methforyl iodide by heating acoetforic anhydride with iodine at 3500 for 12 hours, presumably in a sealed vessel.

Experimental.--The fluorocarbon carboxylic acid is metered into a heated three-neck flask containing phosphorus pentoxide. To one neck of the flask is attached a line through which dry air can be admitted, if necessary, to carry the reactant vapors into the reactor tube. The anhydride is formed by reaction of the acid with phosphorus pentoxide, and passes out through the other neck of the flask into a large tube leading into a furnace. A flask containing iodine and phosphorus pentoxide is also connected to this tube. The flask is heated, causing the iodine to sublime into the tube and mix with the vapors of anhydride











68

pasing into the furnace. The silver iodide, mixed with glass chips, is packed into the heated portion of the tube. The mixed vapors pass through and over the silver iodide. A condenser is provided at the exit end of the tube to prevent the passage of iodine into the collectIng traps. The product, along with some unreacted anhydride*, is collooted in the traps which are appropriately cooled, depending upon the boiling points of the products being formed. Evolution of carbon dioxide is detected by means of a calcium hydroxide scrubber attached to the exit line. Reactor temperatures ranging from 3000 to 400 were employed, althou h at the higher temperatures decomposition begins to occur, as evidenced by the appearance of fluoride ion in the calcium hydroxide scrubber. Experimental data and product yields appear in Table U.

All products were fractionated. Fluorocarbon iodides were

identified from fractionation data and molecular weights. The latter were determined by the Dumas method.

Results of the above measuwements.--The presence of silver iodide in the reactor tube has a definite beneficial effect, both on conversion and yield. The latter term is based on the amount of unrecoverable acid. Lower reactor temperatures can be employed with resultant decrease in the formation of decomposition products. The overall yields are correspondingly improved. Apparently the optimum reactor temperature is around 325 for the materials investigated. Higher conversions might be achieved through the use of a longer reactor tube with specially prepared and supported silver iodide.











69

A reactor designed to re-cycle the unreacted anbdride might also achieve the some end. However, time in which to do further development work on this interesting synthetic procedure was not available.








PREPARATION OF FLUR0CARON


TABIR 11

IDDES FRcm FIDOR ocARBON CARBaryLic ACD Am RIDES


Temp. Iodide Ccrv AnhyYield, eCver- dride
Acid 9. moles A 0 Obtained, moles sion Recov- CaC g. % ered, swied, % moles

CF5COC 38.42 �337 Not 350-700 2 .01 3 --Used

CF5COc 55.18 .4,84 Not 4oo-6oo lo. 4 .053 11 20.0 .294 18 Used

C3F7O0OG 21.53 .1006 Not 355-510 1.8 .0061 6 5.16 .0766 8 Used

CF 3COEC 25.54 .224 Used 300-36c 7.2 .0367 16.3 8.88 .1394 26 CF 3OCK 51.3 .450 Used 200-330 10.3 .0525 11.7 23.0 .2300 23



039TcOC( 16.14 .0754 Used 230-00 9.0 .00k 4o 5.16 .0502 62



















DISCUSSION QF liESUTS


The investigation reported in this dissertation was undertaken in order to expand the knowledge pertaining to the fluorocarbon oarboxylic acid. Special attention was given to the identification and characterization of the intermediates in the reactions between the silver salts of these acid.s and halogen or halogen-like materials.

The author feels that this end was achieved insofar as iodine and bromine are concerned. Chlorine and fluorine were not included in this investigation, as their physical and chemical properties render them unadaptable to the experimental procedures developed and used for the reactions involving iodine and bromine. The author wishes to point out that no extrapolation of the findings reported for iodine aid bromine to the other halogens is either stated or implied.

It was found that when iodine or bromine are combined with (S) in a completely inert solvent at temperatures below those at which decarboxylation can occur, one equivalent of X is consumed for each equivalent of (8) present. It was further demonstrated that any X added past this point is present only in the free state.

Inert fluorocarbons and fluorocarbon derivatives were found to be the only solvents suited to an investigation of this type. The product of the addition of X to (8) under these conditions was found 71











72

to be an addendum compound of the formula (GC0O)2AgX. These substances were characterized in solution in OF solvents by determining the relationship between acid content, oxidizing power, ani silver content. Addend of the formula (GCOO)2AgI were isolated and identified.

The complexes can be isolated for analysis only by crystallization from solution at reduced temperature. They undergo partial or total decompositim if allowed to warm to room temperature in the absence of a solvent. However, the substance (@0)2AgX apparently is stabilized by the presence of another mole of AgX, forming an insoluble material of the empirical formula GCOOAgX, which is relatively stable in the absence of moisture.

(eC0o)2Agj is stable in solution. (@CO0)2AgBr, however,

gradually deccmposes, as evidenced by decrease In oxidizing power with time. Stability of the complexes was also observed to increase with chain length.

Pyridine was found to undergo a complex formation with (S). The complexes can be isolated'and purified by recrystallization from ether, and were found to have the general formula (OCOO)PyAg. The pyridine complex undergoes decarboxylation at 1600 with the formation of the corresponding fluorocarbon hydride, pyridinium hydrofluoride, and a number of other fluorine-containing products. This reaction shows promise of providing a route to alkforylated pyridines or aromatics.

The silver salts of fluorocarbon carboxylic acids can be decarboxylated in the presence of red phosphorus to give the











73

corresponding anhydride in high yields. The complex mixture GCOOAgI explodes on treatment with red phosphorus.

An improved synthesis for fluorocarbon iodides was developed which produces higher yields and eliminates many of the tedious ani time-consuming operations previously inherent in the procedure.

Fluorocarbon oarboxylic acid nitriles can be prepared by decarboxylating (8) in the presence of cyanogen. The reaction apparently analogous to the above iodide preparation in that cyanogen behaves in a halogen-like manner under the reaction conditions.

A process was developed for the preparation of fluorocarbon iodides by a deoarboxylation-decarbonylation of a fluorocarbon carboxylic acid anhydride in the presence of iodine. The yields and conversions are improved markedly by the presence of silver iodide in the reactor.

Through the isolation, characterization, and establishment of the formula of the halogen addends of the silver salts of the silver fluorocarbon carboxylates, the author feels that some light has been shed on an area in which considerable confusion and disagreement existed.

By comparison of the data found in this investigation with

that available on the reactions of the salts of the hydrocarbon carboxylic acids with the halogens, considerable insight into the fundamental differences existing between the fluorocarbon carboxylic acids and their organic counterparts can be gained.

A number of interesting and useful processes have been











74

developed in the course of this investietim. Also, several reactions were discovered which may provide avenues for future research.
















BIBLIOGRAPHY


1. oMnUs, J., J. Chem. Soc., 3755 (1953).
2. Bennet, F. W., Brandt, 0. R. A., Emeleus, H. J., and Haszeldine, R. N., Nature, 166, 225 (1950).
3. Birckenback, L., Goubeau, J., and Berninger, E., Ber., 65, 1339 (1932).
14. Birnbaum, K.,. and Oaier, 3-., Ber., 12. 3270 (1880).
5. Birnbaum, K., and Reinherz, H., Ber., 21, 456 (1882).
6. Bochem ller, W., and Roffian, F. W., An 165 (1935).
7. Cady, G., and Kellogg, K., J. Am. Mtem. So., 75, 2501 (1953).
8. Crawford, G. H. and Simons, J. H., J. Am. Chen. Soc., Ti, 5737 (19531.
9. Carlsohn, Ublier Eine Neue KLasse Von Verbindungen Des Positiv Eintwertigen Jods," Verlag von S. Hirzel, Leipzig, 1932. 10. G scit, s., and Grafinger, G., P 68, 279 (1935). 11. Haszeldr e, B. N., ,. Cha.m. Soc., 584 (1951). 12. Hiszeldine, R. N., and Sharpe, A. G., J. Chm. Soc., 993 (1952). 13. Haszelline, R. N., and Sharpe, A. G., J. Chem. Soc., 4259 (1952). 14. Eszeldine, R. N., and Jander, J., 3. Chem. Soc., o4172 (1953). 15. Hauptschein, M. and Von Grosse, A., J. Am. hai. Soc., D,
2461 (1951. "
16. auptachein, Mo., and Von Grosse, A., J. Am. Chem. Soc., 7.,
h5 (1952). J.









76
17. Name, A. L., and Finnegan, W. G., J. Am. Chem. Soc., 72,
3806 (1950).
18. Henna, A. L., and Zimmer, W. P., J. Am. Chem. So.,
1362 (1951).
19. Herschberg, E. B.,, Helv. Chia. Acta, 17,9 351 (1934). 20. Hunsdiecker, H., and Hunsdiecker, C. L., Bar., 7M, 291 (1942). 21. Kirshenbaum, A. D., Streng, A. G., and Hauptschein, M.,
J, Am. Chem. Soc., 75, 3141 (1953).
22. ZLeixaberg, J., Chem. Revs., 40, 381 (1947). 23. Luttringbausa, A., and Schade, D., Bar., 4, 1565 (1941). 24. Menefree, A., and Cady, G., J. Am. Chem. Soc., 26, 2020 (1954). 25. Oldham, J. W. H., and Ubbelohde, A. R., J. Chem. Soc., 368 (1941). 26. Oldham, J. W. H., 3. Chem. Soc., 100 (1950). 27. Pelgot, Couqt. Rend., ' 9 (1836). 28. Prelog, V., and Seiverth, R., Bar., L, 1769 (1941). 29. Provost, C., R , 1129 (1933). 30. Prevost, c., coupt. Rend., IL' 1661 (1933). 31. Prevost, C., 22W. Rend., 200, 942 (1935). 32. Prevost, C., and Lutz, R., Cqt. Rendl., 1, 2264 (1934). 33. Prevost, C., and Wieusnn, J., Cop.Rn. 20i4, 700 (1937). 34. Prevost, C., and Wiemjn, J., cowt. Rend., 204, 989 (1937). 35. simnini, Monatsch, D, 328 (189). 36. Simonini, )konatsch, 14, 81 (1893). 37. Simms, J. H., et al., J. Electrochem. Soc., No. 2, +7 (1949). 38. Simons, J. H., and Brice, T. J., U.S. Patent 2,554,219,
MaY 21, 1951.









77

39. Swvsz.t, F., Bull. Sc.. Aced. Roy. B.., 12, 721 (1926).
40. Swa'ts, F., Anal. Soc. Pis. 9,uim., 27, 683 (1929). 41. Uschak.v, No ., and Tchisto,, W. 0., Ber 68, 824 (1935). 42. Wielad, H., and Fischer., F. G., Mn., 6 19 (1926). 43. Windbaus, A., and KLenhardt, F.,Ler. ,,581 (192l). 44. Windhaus, A., and Kahardt, F., Ber, 2., 3981 (1922).


















VITA


George Homer Crevford,, Jr., was born in Houston, Texas, o Ootober 6, 1928, the sn of George f. Crawford, Sr., and Augustine Sabayrac Crawford. In 1930 his family took up residence at Baytwn, Texas. In )by, 19116, he oompleted his work at Robert 1. Lee High School; and he attended the summer session of Lee Junior College at Baytown. From September, 19146, to May, 1947, he attended Sam Houston State Teachers College at Huntsville, Texas. He entered aylor Wiversity at Waoo, Texas, in September, 1947. In 3aroh, 1949, he was married to Rvelyn Brasher. He received the B. A. degree in March, 19505 and entered graduate school at Baylor University immediately following graduation. After receiving the M. A. degree in August, 1951, he entered the University of Florida, where he had been granted a pre-dootoral research fellowship under Dr. J. H. Simns. Upon receipt of the Ph. D. degree in June, 1954, he plans to pursue his professional career in industry.












This dissertation was prepared under the direction of the chairman of the candidate's supervisory comittee and has been approved by all members of the committee. It was submitted to the Dean of the College of Arts and Sciences and to the Graduate Council and wan approved as partial fulfillment of the requirements for the degree of Doctor of Philosophy.

June 7, 1954.









Dean, College o Arts and Sciences




Dean, Graduate School SUMNISORY COMITTEE:













/ ,


M gZ-4




Full Text

PAGE 1

THE NATURE AND REACTIONS OF HALOGEN ADDENDS OF THE SILVER FLUOROCARBON CARBOXYLATES By GEORGE H. CRAWFORD A DISSERTATION PRESENTED TO THE GRADUATE COUNCIL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA June, 1954

PAGE 2

ACKNOWLEDGEMENT The author wishes to express his appreciation to Dr. J. H. Simons for the encouragement and advice provided during this investigation. The author also wishes to acknowledge the sponsorship of Minnesota Mining and Manuf ac turing Company , without which this work would not have been possible. ii

PAGE 3

TABU! OF CONTENTS Page LIST OF TABIES . iv LIST OF ILUJSTRATI OHS . vi I. INTRODUCTION . 1 II. PREPARATION, ISOLATION, AND IDENTIFICATION OF THE COMPLEXES . 13 Determination of Extent of Reaction Analysis of Reaction Mixtures Analysis of the Complexes in Solution Isolation of the Solid. Complexes in. USES OF THE COMPLEXES IN THE SYNTHESIS OF FLUOROCARBON DERIVATIVES 47 General Disoussion Substitution of Pyridine for X in the Complex -Forming Reaction An Improved Method for the Synthesis of Fluorocarbon Mono-Iodides Reactions of (s) and the Complex with Phosphorus IV. RELATED SYNTHETIC PROCEDURES 63 Preparation of Fluorocarbon Nitriles from (s) and Cyanogen Synthesis of Fluorocarbon Mono-Iodides from Fluor ooarban Carboxylic Acid Anhydrides V. DISCUSSION OF RESULTS 67 VI. BIBLIOGRAHTY 75 VII. VITA 78 ill

PAGE 4

LIST OF TABLES Table 1 . 2 . 5. k. 5. 6 . 7. 8 . 9. Page Source and Purity of Materials Used in the Investigations Herein Reported 25 Preliminary Experiments Determining the Stoichiometry of the Reaction of Silver Aoetforate vith Iodine or Bromine in Various Solvents 2k The Stoichiometry of the Reaction of Silver Salts of Fluorocarbon Carboxylic Acids with Iodine or Bromine in Fluorocarbon Solvents 26 Measurement of Excess Iodine Present in Solutions of Reaction Mixtures by the Color Comparison Method.... 27 Preparation and Analysis of Reaction Mixtures of Silver Salts of Fluorocarbon Carboxylic Acids and Iodine or Bromine A. Preparation B. Analysis Analysis of the Complexes in Solution in Fluorooarbam Solvents A. Relationship Between Acid Number, Oxidizing Power, and Recoverable Silver B. Comparison of the Stabilities of the Complexes in Solution. Preparation, Isolation, and Analysis of the Solid Complexes 1 ^ UO 41 35 36 Preparation and Analysis of the Pyridine Addends of the Silver Fluorocarbon Carbaxylates Improved Synthesis for Fluorocarbon Iodides A. Procedure 1 B. Procedure 2.. iv

PAGE 5

LIST OP T ABIES — Continued Tal51 ® Page 10 . Preparation of Fluorocarbon Nitriles by the Decarboxylation of Silver Salts of Fluorocarbon Carboxylic Acids in the Presence of Cyanogen A. Experimental Data 66 B. Analytical Data 55 11 . Preparation of Fluorocarbon Iodides from Fluorocarbon Carboxylic Acid Anhydrides 70

PAGE 6

LIST OF ILLUSTRATIONS Figure Page 1 30 2 . 60 X . i

PAGE 7

INTRODUCTION At the current stage in the development of the chemistry of the fluorocarbons and their derivatives, it is highly desirable to enlarge and expand the knowledge of the fluorocarbon carboxylic acids and their related compounds. Toward this end, the specific problem posed to the author and dealt with in the work herein reported is as follows: isolate, establish formula, and characterize as to physical and chemical properties the intermediates in the reactions between the silver salts of the fluorocarbon carboxylic acids and the halogens or halogen-like substances. The fluorocarbon carboxylic acids have already proven themselves to be valuable starting materials in the synthesis of fluorocarbon derivatives. Their acid chlorides and anhydrides have also been adapted to a number of synthetic methods through modifications of procedures familiar to the organic chemist. In addition, several processes have been developed involving decarboxylation of the metallic salts of these acids, either with or without other materials being present. A knowledge of the nature of the halogen addends of the silver fluorocarbon carboxylates is essential to the explanation of the mechanism of certain reactions of this type and is contributory to the understanding of the general reactions of the fluorocarbon 1

PAGE 8

2 carboxylic acids and their derivatives. More important, however, is the value of this knowledge in the prediction of entirely new routes to fluorocarbon derivatives hitherto unavailable through the synthetic approaches used in organic chemistry. The literature contains a number of papers relating to the compounds produced through the interaction of the metallic salts of the organic carboxylic acids with the halogens. There is considerable disagreement as to the nature of the intermediate substances, and many of the conclusions drawn regarding reaction mechanisms appear conjectural. Some reactions of mixtures of the salts of the fluorocarbon carboxylic acids with the halogens have also been investigated. In the development of synthetic procedures involving the use of these mixtures, some investigators have suggested mechanisms apparently based upon some of the earlier conclusions regarding the silver salts of the organic carboxylic acids. The suggested mechanisms involved intermediates which were never isolated or identified and whose existwas purely speculative, in addition, extrapolation from orgmic to fluorocarbon chemistry has been demonstrated to be, at best, hi^ily hazardous, even when based upon completely substantiated organic data. It was evident that a detailed quantitative investigation should be undertaken in order to eliminate the confusion existing in this area. The intermediates in the reaction between the silver salts of the fluorocarbon carboxylic acids and the halogens were found, by methods described in the experimental section of this dissertation, to be addendum compounds of the composition (9C00) 2 AgX where 6 is a

PAGE 9

3 fluorocarbon radical and X is a halogen. The properties found for these substances Berve to demonstrate again both the similarities and the fundamental differences between the conventional organic compounds and the fluorocarbons and their derivatives. Information is provided which is useful in the chemistry of both silver and the halogens. Halogen-like properties of other materials substituted for X in this c cmpound are clarified. The value of these addendum compounds in the synthesis of the derivatives of the fluorocarbons is also demonstrated, in part, in the experimental section of this dissertation. A number of approaches to the problem of characterizing these intermediates might have been suggested. The problem could be attacked through purely theoretical methods based on previously -determined thermodynamic data. This approach was discarded since the data necessary to such an investigation were not available. The molecular structure and configuration of the intermediates might be determined through the use of X-ray diffraction, infrared spectra, and other physical means. This method was also discarded due to the fact that preliminary investigations had indicated that the substances were not adapted to such an approach. A third possibility was the synthetic approach, involving preparation and isolation of the intermediates under rigorously controlled conditions, followed by their characterization through chemical reaction. This method was adopted since both the information in the literature and that gained through preliminary investigation indicated

PAGE 10

that it should he most productive of conclusive results. Historical background . --With the original preparation and investigation of trifluoroaoetic acid by Swarts ,^9 it became evident that here vas a new series of acids whose characteristics showed every promise of proving to be markedly different from those of their organic analogs. However, only since the development of the Simons electrochemical procees 37 have the fluorocarbon carboxylic acids become readily available for research purposes. Before considering the reactions of the silver salts of these acids with the halogens, the earlier work involving the silver salts of the organic carboxylic aoids should be cited. The first reference in the literature is to the reaction between silver benzoate and bromine. Peligot, 2 ? in 1836, combined these materials in benzene solution and obtained meta-bromo benzoic acid as a product. Simonini 35>36 con ,i uc te
PAGE 11

5 that the products of reactions involving these starting materials depend upon five factors, namely: (l) the nature of the silver salt, (2) the temperature of reaction, (3) the presence or absence of other reactive materials, (4) the nature of the solvent, and (5) the equivalent ratio of silver salt to halogen. Kleiriberg states that the reaction appears to fit a definite pattern only when the equivalent ratios of silver salt to halogen are 1:1 or 1:2. According to Oldham and Ubbelohde,^ however, a third category exists, involving an equivalent ratio of 3:4. The above-cited work of Simonini falls into the 1:1 group, as does that of Prevost^9-3^ who, in a series of papers, elaborated upon the earlier work of Simonini. He described the addition of the complex intermediates to olefinic compounds to give the halogen -containing esters followed by di -esters which could then be hydrolized to give glycols. He contended that the complex first added across the double bond, giving the halogen-containing ester and liberating RCOOAg which then underwent an exchange reaction with the halogen in the ester to produce the di-ester, with the liberation of silver halide. He found that the proportion of di -ester formed was directly dependent upon the solubility of the silver salt in the solvent employed. Prevost also investigated the reaction between the complex and butadiene, and concluded that the addition was 1, 2 rather than the 1, 4 exhibited by halogens. He described the iodinating action of the complex formed with silver benzoate and iodine and refers to the "metallic" character of halogens in such complexes. Prevost confines himself almost

PAGE 12

6 entirely to the use of silver benzoate and iodine as starting materials. Herschberg^ and Wieland and Fischer 1 * 2 also investigated these reactions and arrived at similar conclusions. The reactions of the silver salts of the dicarhoxylic acids with the halogens have been investigated. ^>6,10,42,43,44 Both ^ and 1:2 silver salt to halogen equivalent ratios have been employed. Investigations involving a 1:1 ratio were first carried out by Bimbaum and Gaier,^ who reacted the salts of malonio, malic, suooinie, tartario, fumaric, and maleic acids with iodine and obtained the corresponding acids or anhydrides, carbon dioxide, and silver iodide. Wieland and Fischer 1 * 5 obtained lactones along with the acid or anhydride. Similar results were obtained by other investigators . When equimolar quantities of halogen are heated with the silver salts of the dicarboxylic acids carbon dioxide, silver halide, and the corresponding di-halide are obtained. For example, Bochemuller and Hoffman^ obtained 1,4-dibrotnobutane from silver adipate and bromine. This second general classification, that involving an equivalent ratio of 1:2, is represented in the work of Birckenback, Goubeau, and Bemingerp Bochemuller and Hoffman;^ and others 26 According to these investigators, an intermediate substance of the formula RCOOX apparently is formed by the following reaction: RCOOAg + Xq * RCOOX + AgX jpon heating, this material will decarboocylate according to the following equation: RCOOX -*£ *% . RX + C02

PAGE 13

7 The work of Bochemuller and Hoffman can be cited as being typical of the investigations involving the 1:2 equivalent ratio. In the process of chlorinating allyl chloride in acetic acid solution, these investigators obtained a dichloro ester by a process for which they postulated the following mechanism: (1) CH^COCH + Clg *CHjCC^Cl + HC1 (2) CH 5 C0 2 C1 + CH2=CHCH 2 C1 *CHjCOgCHgCHClCBkCl The intermediate acetyl hypochlorite was never isolated, nor were the corresponding hypobromitee or hypoiodites in subsequent investigations by these and other workers. They found that equilibrium (l) could be shifted to the right by the use of a silver salt in preference to the free acid. They found that the reaction mixture would take up another mole of silver salt, with the formation of complexes of the formula (RC00) 2 AgX. These are, of course, the previously described Simonini complexes . Bochemuller and Hoffman^ describe the thermal decomposition of the acyl hypoiodites as occurring by two paths, one giving the alkyl iodide as above and the second giving the ester by the following reaction: 2 RCOOX ft eat v RCOOR + CC^ + Iq Upon reacting bromine with silver benzoate in boiling carbon tetrachloride, these investigators obtained bromobenzene in an 80 percent yield. Luttringhaus and Schaede^ disagreed with these findings, maintaining that the degradation was unsatisfactory. Bochemuller and Hoffman;^ Birckenback, Goubeau, and Berninger p

PAGE 14

8 and TJechakov and Tchistcw^ 1 Investigated the reactions of the acyl hypohalogenites with olefinic substances. Birckenbaok and co-workers^ reacted the filtrate of the reaction between equimolar quantities of silver acetate and iodine at -80° in ether with cyclohexene. He obtained the acetate of 2-iodo1-oyclohexanol. ELeinber^ 2 states that this is undisputed evidence of the presence of CH^COOI in solution. These results would seem to indicate that the ester-producing reaction is due to the presence of ROOOX rather than the Simonini complex as was concluded by Prevost. Carlsohn^ was able to isolate a substance of the formula RC00(Py)l from the reaction of pyridine with solutions resulting from the reactions of silver salts with iodine. Oldham and TJbbelohde 2 ^ characterized a third class of canpounds, the iodine triacyls, which were obtained when an equivalent ratio of 3 silver salt to 4 iodine wbb employed. They postulated the following equation for their production: 3 RCOOAg + 2 Ig ^3 Agl + I(0C0R) 5 These compounds contain tervalent iodine, as opposed to the univalent positive iodine said to be present in both the Simonini oomplex and the acyl hypdhalogenite. They were able to isolate the iodine triaoyls by chilling filtered solutions of the triacyl in oarbcn tetrachloride or petroleum ether and collecting the material which crystallized on a filter septum. They postulated a number of free -radical mechanisms for the thermal decompositions of these substances and present considerable analytical data in support of their findings. It

PAGE 15

9 appears that the reaction products of the iodine triacyls are not substantially different from those of the Simonini complexes or acyl hypohalogenites . Oldham‘S was not able to obtain the corresponding bromine triacyls. He investigated the reactions of bromine with monobasic aliphatic acids, monobasic aromatic acids, monobasio aliphatic monoketo acids, dibasio aliphatic acids, and dibasic aromatic acids. His results were similar to those previously obtained by other investigators. The first investigation into the reactions of the silver salts of the fluorocarbon carboxylic acids was undertaken by Swarts,^ who reacted silver trifluoroacetate with iodine in benzene solution at 60°. He gives the following equation, which is representative of the overall reaotian: 2 CFjCC^Ag + 2 I 2 + 2 C5H5 (CFjCOgJO + H 2 0 + 2 Agl + 2 CgH^I Thus, the reaction mixtures of the silver salts of the fluorocarbon carboxylic aoids with iodine were demonstrated to possess iodine ting power similar to that previously observed in the organic carbaxylio acids. Simons and Brice-^® developed a process for the production of fluorocarbon iodides through the decarboxylation of the silver salt of a fluorocarbon oarboxylio aoid in the presence of exoess iodine in a high -boiling fluorocarbon solvent, and were granted a patent. No attempt was made to postulate a mechanism for this reaction. Henne and Zimmer 1 ® investigated the action of mixtures of

PAGE 16

10 sliver trifluoroacetate .and bromine or iodine on toluene. They found that ring halogens. tl on occurred below 100°, and advanced the following equations for the reaction: CFjCOgAg + 3^ > AgX + CF 5 C0 2 X GF 3 C 0 2 X + CH 3 C 6 H 5 ^ CF 5 C0 2 H + CH 5 C 6 H 4 X In this mechanism, the iodinating entity is the CF3COOX. They state that the positive halogen "hangs" from a polarizable atom attached to one or two strongly -active methforyl groups. Haszeldine and co-workers, 11 * 111 in a series of papers, have reported investigations of the reactions of a number of the salts of the fluorine -containing monocarboxylic acids with the halogens. In the first of this series, Haszeldine 11 describes the decarboxylation of silver trifluoroacetate in the presence of iodine to give trifluoromethyl iodide. He states that the intermediate in the reaction is CF^COOI. This substance, he contends, is formed when iodine is taken up by silver trifluoroacetate in a 2;1 equivalent ratio. In the second of this series, Haszeldine and Sharpe, 1 ^ after obtaining sane results somewhat inconsistent with the above postulate, suggest that an equilibrium exists between the hypohalogenite and a Simcnini-type complex in whioh the iodine and the iodine trifluoroacetate are competing for the silver salt. CFjCOgAg + CF3COOI (CF^CC^Agl If a readily iodinated substance, such as phenol, is added, the iodine trifluoroacetate is removed and the proper amount of silver iodide is

PAGE 17

11 precipitated in accordance with the equation cited under the work of Henne and Zimmer. 1 -® Haszeldine 1 -^ was unable to isolate either iodine trifluoroacetate or the Simonini complex. He used nitrobenzene as a solvent in the above experiments. Haszeldine1 brominated and iodinated a number of aromatic compounds by this method and devoted considerable space to the discussion of ionic and free -radical mechanisms for this and related reactions. In the third paper in this series, Haszeldine and Sharpe 1 -^ describe the decomposition of the silver salts of the fluorohalogenoacetic acids in the presence of the halogens to give the fluorohalogeno alkyl halide in the same manner as with the fluorocarbon carboxylic acid salts. They also describe the preparation of higher fluorocarbon mono-iodides from the higher acids. They discuss the preparation of methforyl halides by the same reaction, using the potassium or sodium salt as an alternative to the silver salt. A number of miscellaneous reactions and attempted reactions are described. In the fourth paper in this series, Haszeldine and Jander 1 ^ report the synthesis of fluorocarbon nitro and nitroso compounds by the decarboxylation of the silver salt of acetforic acid in the presence of nitrosyl chloride. They give the following generalized equation for the reaction;

PAGE 18

12 CP 5 *(CF2)n C 0 2 A 8 AgCl + CF 5 *(CF 2 ) n COONO CFj.CCFgJnC^O CF 3 .(CF 2 ) n N° + CF^ . ( CP 2 ) n N 0 2 + COg, etc. This involves an intermediate vhioh is analogous to the tr if luoroa cetyl hypohalogenite previously postulated. ThiB intermediate, like the others, vas never isolated or characterized. Banus^had previously synthesized fluorocarbon nitro and nitroso compounds by another method. Cady and Kellogg, ^ in an interesting paper, describe the isolation of an unstable substance having the formula CF^COOF from the reaction of acetforio acid with fluorine in a closed glass system. Menefree and Cady 21 * extended this work to the preparation of C^^COOF and CjFyCOCF. The hypafluorites were deteoted by explosions which could be initiated in the reactor. Aoetforyl hypofluorite and propionforyl hypofluorite were identified through the determination of molecular weights and the analysis of decomposition produots. Thus, the existence of substanoes of the formula ©COOF has been established. However, the existence of the analogous hypohalogenites of the general formula ©COOX in the reaction mixtures of silver salt with halogen remains to be proved.

PAGE 19

PREPARATION, ISOLATION, AND IDENTIFICATION OF THE COMPLEXES Determination of Extent of Reaction Discussion , --The stoichiometry of the reaction between the silver salt of a fluorocarbon carboxylic acid, (hereafter designated S) and a halogen (x) is of paramount importance, for it provides an essential clue as to the nature of the product formed at temperatures below that of decarboxylation and in the absence of other reactive materials. Thus, if the reaction leads to the formation of a substance of the formula ©COOX (where © is a fluorocarbon radical) by a path of the type (S) + X 2 v © COOX + AgX one mole of (S) should consume two equivalents of X. If the reaction leads to the formation of the fluorocarbon analog of a Simonini complex by a path of the type 2 (S) + X 2 v (©COO) 2 AgX + AgX one mole of (S) should take up one equivalent of X. If, as Haszeldine subsequently contended, an equilibrium exists of the type (s) + ecooi (ecoo) 2 Agi free iodine should be present during the entire process, its color remaining relatively constant or becoming progressively more intense, depending upon the value of K^. (No values for K were provided.) 13

PAGE 20

14 In order to obtain accurate and reproducible results, a solvent must be employed which can in no way react chemically with the materials under investigation. The solvent should be clear, non viscous, reasonably low-boiling, and capable of dissolving the active intermediate under consideration. In the preliminary investigations, a number of different common organic solvents were tried and subsequently discarded as failing to meet one or more of the above specifications. Erratic, non-reproducible results could usually be attributed to the occurrence of secondary reactions involving the solvent. Among these are oxidative attack upon the solvent by the intermediate, halogenation of the solvent by the intermediate, hydrolysis of the intermediate by the solvent itself or by traces of water in solvents having a hi$i affinity for water, and secondary complex ing between the solvent and (S) or the intermediate. Ether, ohloroform, carbon tetrachloride, and pentane were tested and discarded. Of these, pentane gave the most consistent results, checking closely with those obtained using a solvent consisting of fluorocarbons or their inert derivatives. The results of these preliminary investigations appear in Table 2. Fluorocarbons, fluorocarbon nitriles, and fluorocarbon cyclic oxides were used as solvents. These materials are ideally suited to this investigation because of their inertness to chemical attack, complete lack of color, low viscosity, and the fact that a particular

PAGE 21

15 solvent could be chosen having a boiling point suited to the experiment being performed. The low solubilities of halogens and (S) in these media do not constitute a serious drawback. The rate of reaction of (S) with X is * decreased and becanes partially dependent upon the stirring rate. However, since the investigation is not a study of rate phenomena, this is not a deterrent factor. On the contrary, it is desirable to have the reaction proceed at a reasonable and orderly rate, as contrasted to the violent manner in which it occurs in solvents such as ether, in which both (S) and X are hi^ily soluble. The solvents consisting of fluorocarbons or their derivatives were subjected to rigorous purification and deactivation before being used in experiments. The procedure for this is described in the experimental part of this section. The source and purity of the fluorocarbon carboxylic acids and other materials employed in the following investigations are described in Table 1. Experimental . --Determination of the stoichiometry of the reaction between (S) and X was carried out using the apparatus shown in Figure 1. The silver salts of the fluorocarbon carboxylic acids were prepared by refluxing the acid with excess Ag^O in aqueous solution, followed by filtration and removal of water under vacuum. The salts were purified by repeated recrystallization from anhydrous ether. Final ether removal and drying was effected by warming in vacuo. Since the solubility of X in the fluorocarbon solvent is low,

PAGE 22

16 it is impractical to make up the complexes by addition of a standard solution of X to a slurry of (s). The following techniques were adopted . Iodine (analytical reagent grade) was resublimed through P 2 0 5 and collected on a water-cooled coldfinger in the sublimation flask. The apparatus was transferred to the drybox and the iodine removed from the coldfinger and stored in an appropriate container. Samples of the iodine were transferred in the drybox to wei^ied 5" or k” test tubes. The tubes were then quickly sealed, using a gas oxygen torch. The tube and contents plus the scrap were weighed on the analytical balance and the net amount of iodine per sample determined. Analytical reagent grade bromine was dried by distillation from P2O5, and weighed samples were made up In the same manner as with iodine. The following fluorocarbons and fluorocarbon derivatives (hereafter designated 0F) were used as solvents: #1. R and iso-pentforane mixture (C^F^) -This material was washed with anhydrous ether for removal of fluorocarbon hydrides, refluxed with KOH, refluxed with CF^COOAg + I, treated with sodium thiosulfate, dried over CaClg, and fractionated. The fraction, boiling 29-31°, was taken. The fluorocarbon solvent, thus freed of any impurities which might react with the intermediate, was stored over P 2 0(j in sealed glass tubes. #2. Nonforane mixture (C^q) — This material was treated as above, and distilled. A rather wide fraction, boiling II8-I30 0 ,

PAGE 23

17 was obtained. This was stored over PgO^. #5. Butforyl nitride (C^FgyN) — This material was treated as above and used for a solvent in the fluorocarbon mono-iodide synthesis of Simons and Brice. 58 It was distilled; and the fraction, boiling 165 180 °, consisting of various isomers was obtained. The material was stored over PgO^. #*<. "Fluorochcmioal 0.75" (a cyclic oxide of the formula ^ 8 ^ 16 °) This material was treated in the same manner as the butforyl nitride, and fractionated. The fraction, boiling 100-101.5°, was stored over PgO^. The procedure used for carrying out controlled reactions of (S) with X under absolutely anhydrous conditions and for determining the stoichiometry of that reaction was as follows: The reactor (Figure 1) was thoroughly dried by alternate evacuation and admission of dry air. A sample of (S), weighed to be somewhat (usually in the vicinity of .1 gram) less than that required for a 1:1 equivalent ratio with one of the previously weighed X samples, was carefully poured into the graduated reactor vessel D. This was accomplished by lowering D a few inches, while maintaining a stream of dry air through the system from L and pouring (S) from the weighing bottle directly into D. Great care was taken to avoid getting any (s) on the inside of the Joint. D was replaced and the drying process repeated to remove any water that might have entered the system. The weighing bottle was then reweighed and the net amount of (S) determined.

PAGE 24

18 The vial of X was dropped in through the top while the air stream from the system was maintained. Breakage of the vial could he assured by shooting it in with a sling-shot arrangement attached to the rim of the tube. The solvent was then admitted from B, removing any (S) that mi^it have clung to the sides of D. Magnetic stirring was started, and a reasonable time was allowed for the reaction to go to completion. If all color was removed, more halogen could be added. If, on the other hand, the characteristic red-brown or red-purple color due to free bromine or iodine remained, enough (s) was added to bring the equivalent ratio up to, or perhaps very slightly in excess of, 1:1. Completion of reaction could then be noted by the disappearance of all color in the solution when this point was reached. In every case free halogen remained until additional (s) was added, after which a few minutes' stirring removed all trace of color. The numerical data pertaining to these experiments appear in Table 3. Thus it is demonstrated that bromine and iodine are consumed quantitatively to the 1:1 equivalence point, and that immediately after passing that point the color of free halogen is evident. Next it is desirable to determine if the amount of free halogen remains relatively constant as X is added, due to the establishment of an equilibrium of the type cited under the work of Haszeldine, or whether all the halogen added past the end point is present in the free state. Only one unassailable method for such an experiment appears to exist. That is to bring a mixture of (s) and Xg to the 1:1 equivalence point and then to

PAGE 25

19 add. known quantities of standard X solution past that point. A Bample of the Bame solvent not containing any reactant materials (hereafter referred to as the blank) is then titrated with the same X solution to the point at which its oolor matches that of the above solution containing the reactants. The ratios of added solution to original solvent in the reaction mixture are compared to the corresponding ratios in the blank at the point at which the colors are observed to match. Thus, if all halogen added to the reaction mixture in excess of the 1:1 stoichiometric point is present in the free Btate, the values for these ratios should coincide at each point of identical color intensity . The tabulated values for the above ratios appear in Table 4. The degree to which they coincide as successive amounts of iodine-containing solvent are added is evident. A saturated solution of I 2 in "Fluorochemical 0.75" (solvent #4) was prepared and standardized against thiosulfate. The violet solution had a normality of only .00134. Agl was prepared from A$f0j and KI and dried by warming in racuo in the absence of li#rt. This was to be added to the blank in order to provide identical visual conditions. However, it was discovered that Agl itself is capable of removing small quantities of iodine from solution. The extent of this phenomenon was measured by titrating 9.5074 g. Agl in 10 cc. of solvent with the standard iodine solution. At 65 cc. a noticeable pink color appeared. Thus, the Agl decolorizes I 2 in a 121 to 1 equivalent ratio of Agl to I 2 in the 6ST solvent.

PAGE 26

20 The experiment was repeated using reagent grade nitrobenzene. A .026 N solution of iodine in this solvent was prepared and standardized against thiosulfate. A suspension of *M059 g. Agl in 10 co . pure nitrobenzene decolorized 1^.90 cc. of I 2 solution. Detection of color change is difficult in a colored solvent such as nitrobenzene; hence the I 2 solution was added until the color was obvious. The nitrobenzene solution was then filtered and back-titrated w$th .5 cc . of .1000 IT thiosulfate to a starch end-point. The Agl was found to decolorize I 2 in a to 1 equivalent ratio of AgX to I 2 in nitrobenzene. This phenomenon, though small, is Just another of the complicating factors which can lead to erroneous results, if not properly rec ognized. The experiments were conducted using an amount of Agl in the blank proportional to that which would be produced by the complexfarming reaction. The color-change characteristics of free iodine in 6 F solvents, plus the extreme dilution of the solution used, enabled this investigator to make measurements of high precision with a minimum of hunan error. The results obtained were in remarkably close agreement with theory. Results of the above measurements .--The object of these measurements has been to establish the stoichiometry of the reaction of (S) with X. This was accomplished through the addition of X to (S) in a 6 F solvent and observation of the point in the addition at which free halogen appeared and the manner in which the color intensity increased

PAGE 27

21 past this point. Similar results were obtained, for (s) from silver trifluoroacetate through silver valerforate and for bromine and iodine. It was demonstrated that after the 1:1 equivalence point is reached in the addition of X to (S), free halogen appears in what had been, up to this point, a colorless solution. Furthermore, the deepening of the color as Ig was added past this point progressed linearly with the addition of Ig, paralleling the color intensity deepening of a simultaneously-run blank. It was therefore demonstrated that when the reaction is oarried out at temperatures below that at which decarboxylation can occur, in a completely inert solvent, and under strictly anhydrous conditions, the reaction occurs in a 1:1 equivalent ratio and that any halogen added in excess of this is present only in the free state. The latter measurement was conducted using iodine only. Similar color deepening past the end point was observed with bromine, but the color characteristics of bromine in the solvent used and the high volatility of bromine (resulting in its escaping from solution into the air space above) prevented precision observation of color change. Some cf the investigators have arrived at conclusions as to the extent of the reaction by weighing as AgX the precipitate obtained at various stages during the halogen addition. This involves the erroneous assumption that all the material precipitating during the reaction is silver halide. It will be shewn that the precipitated material is, in reality, a mixture of AgX and (9C00)gAgX. The proportion of the latter in the precipitate is inversely proportional

PAGE 28

22 to Its solubility in the solvent employed. Then, when the mixed precipitate is removed from the reaction mixture by filtration, and presumably washed and dried prior to weighing, the (6COO) 2 AgX is converted to AgX, AgZOj, and free acid by a reaction of the type (0COO) 2 AgX + H 2 0 >2 6C00H + AgOX followed by 3 AgO X > 2 AgX + AgXO^ as will be shown in the following section. The danger of confusion is further increased by the great similarity in appearance between AgX and the mixture AgX + (0C00) 2 AgX. In the preliminary experiments (see Table 2) and in subsequent investigitions, (s) was stirred with I 2 under anhydrous conditions and in the absence of a solvent. A yellow, amorphous powder resulted, having the overall composition ©COOAgl. This material cannot be visually distinguished from Agl. The addition of iodine in excess of the 1:1 equivalent ratio produces the red-purple color of the free halogen. The phenomenon of iodine absorption by silver iodide, though small, is still a complicating factor. It was demonstrated to be over twice as large in nitrobenzene as in "Fluorochemical 0.75."

PAGE 29

23 TABIE 1 SOURCE AND PTJRITY OF MATERIALS USED IN THE INVESTIGATIONS HEREIN REPORTED Compound Source Purity Further Purification Acetforic acid M. M. & M. Co. 95-98$ fractionation Propicnforic acid M. M. & M. Co. 90-95$ fractionation Butyrforic acid M. M. & M. Co. 75-85$ fractionation Valerforic acid M. M. & M. Co. 95-98$ fractionation Caproforic acid M. M. & M. Co. 90-95$ fractionation Bromine Mallinckrodt Anal. R. described on p. 16 Iodine Mallinckrodt Anal. R. described on p. 16 Pyridine Mathieson U.S.P. fractionation and drying Pyridine Eastman Kodak (DPI) Spectral Grade drying Phosphorus (red) Mallinckrodt Teoh. dessication M-di isopropyl benzene Dow Tech. fractionation 2 -butyl benzene Koppere Tech. fractionation Silver oxide Fischer C.P.

PAGE 30

:y eeferdents determining the stoichiometry of reaction of silver acetforate with IODINE OR BROMINE IN VARIOUS SOLVENTS 24 « 0 1 © & •P O « O CQ o 4 W 60 M a 4 M 60 CO 3 © P © •p •p rH • • s © p © rH $ © rH £ pi Ti © CVI Pi g © iH S 4 1 © rH Pi i © H Pi as 4 e § o s g O (3 g 8 g rc\ H o •H O rl O On K) 60 60 o q ^ H O fl pi 83 83 °^-3 O VO 60 o CO 6$ ?«> V 60 2 i m ?4 rH ' 1 CM tt! , 1 CM C . 3 PA Pi cm q • 3 cm q • t o © T o © O © 1 o © o £ o £ o £ o £ lA r© IA P IA ,© VO ,o -4 -P -4 +> vo -P IA -P CM 0 CM q CM 0 CM Pi CM © CM © CM © CM © •p P P -P © © © CQ * ’g © >» -s t.C n !>» •8 • J 1 © 4 § p © •q -p 44 ?I SI s IA H Si sj $3 Vi 83 Vi • p • +3 • H O O © WOO • • o • • ° jj $ IA * S 4 PA PM • VO • Ov • rH • CM VO % i © CM © # & H 1 & CO PA s & K 8 cv» 0 Lz 100 o PA K OJ VO d O rH -4 VO lA O * 8 E Ov vo PA rH CO A CO VO PA IA IA 8 CM CM VO d PA rH OJ CM g I g g 3 2 8 8 8 8 8 8 8 8 O o o o o o o o K*\ A IA JA PA PA PA PA b b b 6 b 8 fe

PAGE 31

25 % § •H 49 g V I CVJ a * P -H *» O X § N ff I 63 I +» o « +> S > CO CT* w 60 i X 60 CO CVJ 10 00 o tM cm c CM « 3 n rH X O © O 5 KN © IfN in VO • rH rH ON rH $ rH £ -=f o H K\ O CM • rA I o KN 6 © +9 c i — ! & o o VO 60 CM S3 o’ t IfN +9 CM ® o KN I §8 PjH On KN in n ITN n oo 9 © +9 o [CM I I o ITN +9 CM © IfN -=)• in KN ON On CM • CM KN KN • O f— I 2 o 0° « +> © rH f O O VO CM 2 £ o IfN -P CM ® * CM © ! I 49 O © * 3 1 I IfN t— KN 5 O CM 2 CJ £~ KN O

PAGE 32

26 K'V a I NOTE: In the second part of each of the six experiments shown, on additional amount of the indicated reactant was added to the reaction mixture .

PAGE 33

27 a § cvjcoot— WHOOO O ACM CM CM O O O O • • • • • «H rH rH rH rH 8 a a-4 on M3 O 1A04 ' • • • • rH rH CM CM IAO O 0 0\0 o co H CM On On O On H o o o o • o CM 8 8 8 O 8 • • • • rH r-l CM CM A A A A A CM CM CM CM CM 3 0 S o o o o o CM CM CM CM CM cr* s CM tt) -4" 21 A 8 6 & 85 * ftrH KN H O O O O O • • • • rl H rl H ir\ On rH A kn On 4 O iA04 • • • • H rH CM CM Q A A Q On -4 CO CO ON O O On ON rH rH W CO 9 o o o CM §S w Jo A I 8 8 88 8 • • • • rH rH CM CM A A A A A CM CM CM CM CM a a a a c E s s •g £ o o o o o A CM CM CM CM a* W A CM m to o' H A CM A 00 H -4" CO CO O o 8 8 -8 §s 1A CO rH rH ON ON O O • • • rH rH A A C— CO H CM OMAO A H CM CM i A ' 00 Q A A i |A CM A i On O O O ON rlHH O O o CM 88888 • • • • rH rH CM CM A A A A A CM CM CM CM CM C d C C .. rH H A u £ ON A 8 & fc 1.00

PAGE 34

28 o mcy o h O ITvCO On On o o o o o O On On On On P3IK • • • • • * « • < • r-H rH rH r—j r—i rH % PP Q O • o ir\ rr\ o cm o tot-t-co tT\ Oc-4 On-? 3 -2 t> rH ir\ o tr\ o ir\ o H H CM CVJ r— < l — 1 CCJ K P CQ oi o o g Q Q a Q K\ O lAr-l J O H H 1A KC 0 04 Oco 3 oodo\o O Oc O O On CQ H H rl rH rH rH rH • o • • o • o to o CD to •p § o ft °. ft § -p • o o 6 OJ OJ o *s *N o § ft § ft H o o o !s 3-2 > H ft 8 ft o ft • • • • • ft 8 ft 8 ft • • • • • 03 *H o H rH OJ OJ H rH OJ OJ K P a • 'd O r5 5 lT\ ITv IfMfMTN tr\ ir\ ir\ itmtc o o CQ 3 OJ OJ OJ OJ CVJ OJ OJ OJ CM OJ CO S3 c d a d c a c c a a •H ^ © 3 a a •’a e •H *H «H »rH *H si e S B lf> o o o o AOOOO OJ OJ OJ OJ -4 OJ OJ CU OJ o 4 o 4 O 4 O 4 w w W M s s s ss ft 8 » a CO CO CO CO w J • • to to m to ca O «P o Oc H o s OJ -=f o OJ Oc CQ Oc w CO O ft 8 £ OJ 0o • • • O o • • • * ft w ft *N CsO $ g 8 O o o rH On rH Ee, •» ^ -4 OJ fa O M ® (V | ^ *\ cT hF &

PAGE 35

29 KEY TO FIGURE I A. Drying tube B. Burette C. Manometer #1 C ' . Jfenoraeter #2 D. Graduated reactor vessel E. Magnetic stirrer F. Stopcock G. Filter -weighing bottle H. Trap I. Saran flex Joint #1 I'. Saran flex Joint S2 J. Trap K. Stopcock L. Dry -air supply M. Vacuum M' . Vacuum N. 5 -liter gas bulb

PAGE 36

Fig. I

PAGE 37

31 Analysis of Reaction Mixtures Discussion . —Further information as to the nature of the halogen addends of the silver fluorocarbon oarboxylates can be obtained through the analysis of the reaction mixtures far total oxidizing power and recoverable acid. A comparison is made between the total available oxidizing power far the complexes farmed with silver acetf orate through silver caproforate and bromine or iodine. From these data important conclusions as to the formulae and properties of the complexes may be drawn. Recoverable acid is important from the standpoint of determining whether or not all the (S) is consumed in the reaction forming the readily hydrolizable complex. Also, it serves to determine the extent of Bide reactions such as decarboxylation, which could eliminate acid from the system. The extent of the occurrence of the decarboxylation reaction, if any, was also checked by the observation of any pressure changes occurring in the system during the initial reaction and by measuring the weight increase of an ascarite tube attached to the system, as shown at M in Figure 1. No appreciable pressure rise or weight change could be detected for iodine, althou^i the bromine reaction did give some evidence of gas evolution. In a few cases, recoverable silver was determined as a check on the analytical method to be used in subsequent determinati ns of the silver content of the purified solutions and isolated complexes, in which this factor is highly critical.

PAGE 38

32 Experimental. --The reactions were conducted as before, using tbe apparatus in Figure 1 and a 9F solvent. Total oxidizing power was determined by adding KI to the colorless reaction mixture, shaking until the reaction appeared complete, adding water, and immediately titrating the liberated' iodine with standard thiosulfate. Liberated acid was determined on separate reaction mixtures by treatment with water, followed by titration with standard base to a phenolpthalein end point. In other cases, a considerable excess of (S) was employed. The acidic solution resulting upon the addition of water was distilled into a closed receiver. More water was added and the process repeated until all the free acid was obtained. This was then titrated to a phenolpthalein end point with standard base. Recovered silver was determined by the Deniges cyanide method. Experimental data and analytical results appear in Table 5. Results of the above measurements . --The results obtained lead to the following conclusions: (1) In the case of both bromine and iodine, one equivalent of oxidizing power is available for each equivalent of halogen consumed. (2) The total oxidizing power is somewhat low for the reaction mixture of silver acetf orate and X, but closely approaches the theoretical value as the molecular weight increases. (3) The reaction time for (s) with X appears to increase with molecular weight of (s) and in going from Br 2 to . (U) When the total acid, both liberated and present as the

PAGE 39

33 silver salt, was titrated in the reaction mixture, an equivalent of acid was found for every equivalent of (s) originally present. In cases in which the original excess of (s) was large, AgCE was precipitated due to the reaction of the excess Ag ion in solution with the titrating base. However, when the liberated acid was distilled from the mixture in order to separate it from the unreacted (s), one equivalent of acid was found in the distillate for each equivalent of X originally present in the reaction mixture. Thus, the acid liberated from the reaction mixture upon hydrolysis with water is independent of the amount of (s) present in excess of the 1:1 equivalent ratio. (5) In several instances the precipitate obtained upon treatment of the reaction mixtures with water was separated, washed, and vacuum dried. This precipitate was observed to possess oxidizing power when treated with Id solution. The precipitate was presumed to contain AglO^ . (6) In every case the precipitate obtained did not seem to be as subject to decomposition by light as is AgX obtained by more conventional methods. Samples of the precipitate obtained after treatment of the reaction mixture were washed and vacuum dried. These were then analyzed according to the Deniges cyanide method and were found in every case to be almost pure AgX. From the above findings, the equation for the reaction of the complex with id may be written (9C00) 2 A gX + 2 id AgX + 2 9C00K + Ig

PAGE 40

3b while the reaction with water may he written as before, i. e., (9C00) 2 AgX + H 2 0 wAgOX + 2 GCOO0 3 AgOX y2 A£X + AgXOj

PAGE 41

35 ir\ 3 I (!) U Pi

PAGE 42

36 a

PAGE 43

37 Analysis of the Complexes in Solution Discussion . — In the following series of experiments, clear solutions of the complex in the GfF solvent are analyzed for the same quantities determined for the entire reaction mixtures in the previous section. The complicating factors due to precipitated Agl, unreacted (S), and undissolved complex are eliminated. The acid number, total oxidizing power, and recoverable silver are obtained as functions of the amount of complex present in solution. Since the complexes are only partially soluble in the 6F solvent, the greatest significance will be in the comparative values of the three measured quantities. After the samples of solution are removed, the residual solution and precipitated mixture can be titrated and the summation of all the quantities compared to the theoretical totals. If the formula for the complex is (9C00) 2 AgX, and if its reactions with KI and HgiO are represented by the previously -cited equations, then the total oxidizing power in equivalents should equal the acid number, while the recoverable silver should be one half that value . Also included in these experiments are two designed to determine the relative stability of solutions of the complexes containing bromine as compared to those containing iodine. Observations of the behavior of the complexes of br ranine have indicated that they are less soluble in the solvents employed, and that the solutions are less stable, losing their measurable activities

PAGE 44

38 if allowed to stand for any period of time. Total oxidizing power was chosen as the parameter for the measurement of the stability of (©C00) 2 AgBr in solution, since this quality has been demonstrated to be reasonably dependable and independent of excess (s). Experimental . — The complexes were prepared as before, using the apparatus in Figure 1 and several of the different fluorocarbon solvents. As soon as the reaction was complete, the precipitate was allowed to settle until no particles remained suspended in the clear supernatant liquid. A special pipette was constructed with a stem sufficiently long to reaoh into the apparatus from the top and remove 10 cc. samples. A tiny chamber containing calcium chloride was sealed into the stem to eliminate the possibility of contamination of the solution by the operator. The lower end was further constricted to prevent loss of the dense, non -viscous solution during transfer. The pipette was dried by passing a stream of dry air through it. The top was removed from the reactor, while maintaining a stream of dry air from L. The pipette was partially inserted and allowed to remain suspended in the air stream for a few minutes to remove water from its outside surface. It was then lowered into the supernatant • liquid and the sample obtained. The pipette and sample were then removed, the sample rapidly transferred to an erlenmeyer flask and titrated immediately. This prooess was repeated until the solution level was too low to permit removal of samples without the danger of

PAGE 45

39 their being contaminated with precipitated material from the bottom of the reactor. In the last two experiments a measured time was allowed between removal of samples for titration. The reactor vessel D was then removed and the remaining material titrated directly in the vessel. Experimental and analytical data appear in Table 6. In Part A of this table, values for oxidizing power, liberated acid, and recovered silver for samples removed from reaction mixtures as described above are tabulated. In Part B, values of total oxidizing power for samples removed from reaction mixtures at measured time intervals are tabulated. The time variation of this value for samples of solutions containing (C^COO^Agl and (C 2 F 5 C00) 2 AgBr may be compared. Results of the above measurements .--The tabulated experimental data and results clearly indicate that the theory cited on page 33 is accurate and that the complexes in solutions separated from precipitated material react with water and KI in accordance with the suggested equations . The complexes formed with iodine are evidently quite stable in solution, since no marked decrease was noted in the measured quantities during the period of measurement. On the other hand, the solutions of complexes of bromine rapidly lose their oxidizing power an starring. Furthermore, their initial solubilities in the solvent are somewhat lower than those of the iodine complexes. For these and other reasons, iodine was used more frequently than bromine in the foregoing and following measurements.

PAGE 46

'IS OF THE COMPLEXES IN SOLUTION IN FLUOROCARBON SOLVENTS Itf) In obtaining totals for oxidizing power and recoverable silver, appropriate values substituted for those not determined.

PAGE 47

41

PAGE 48

ks Isolation of the Solid Complexes Discussion . — The problem of isolating the solid complexes from their solutions was not easily solved. It was necessary to obtain solutions absolutely free of precipitated Agl and ( SCOO^ Agl, and to devise a method for separating the complex from the solvent without decomposing it in the process. The entire operation would, of course, have to be carried out under strictly anhydrous conditions. A number of attempts to remove the solvent from the complex by distillation were made. This invariably resulted in the decomposition of the complex, even when the distillations were carried out under vacuum. It was discovered that when dear solutions of the complex in a low -boiling 6F solvent were chilled to -78°, white crystals precipitated. These were presumed to be the complex, and the apparatus (Figure 1) was adapted to effect their separation by this method. The complexes, thus removed by crystallization and filtration, were weighed and analyzed for total oxidizing power, acid equivalent, and recoverable silver. Experimental . --The entire apparatus was very carefully dried by alternate evacuation and admission of dry air. The complexes were prepared by the previously -described method, employing as a solvent the low -boiling fluorocarbon mixture (#1) . During the preparation of the complex, carried out in the flask designated D in Figure 1, the remainder of the system was subjected to more rigorous vacuum drying, with special attention being paid to the filter -weighing bottle G.

PAGE 49

^3 The latter was specially constructed, using a "fine" grade sintered-glass disc as the filter. It was designed to he sufficiently light for use as a weighing bottle (weight 55.4701 g.) and sufficiently compact to fit inside a 1-liter dewar; yet it had a capacity of 42 cc., which enabled it to hold enough solution to assure that a measurable amount of complex could be obtained. It wa8 connected to the system by means of 10/30 standard taper joints, as shown in the diagram. Teflon plugs were provided for insertion in the inside of the male joints on the bottle during the weighing process. After the reaction had reached completion, the precipitate was allowed to settle. G was chilled in dry ice-acetone. Stopcock F was then opened and the solution allowed to pass into G. Trapped air in G was vented by opening stopcock K. Any suspended particles were removed from the solution by a glass-wool plug in the line between D and F. After sufficient time had been allowed for crystal formation, suction was applied by opening stopcock M'. The solvent was drawn from the filter -weigh bottle, leaving the precipitated complex on the septum. The solvent was collected in dry -ice trap H or liquid-air trap J. Since the solution had been removed from D, the stopcock K could be opened and a stream of air allowed to pass through G, removing any remaining solvent. K was then closed. The dewar containing dry ice -acetone was removed from G and pumping continued one or two minutes while allowing the filter-weighing bottle to warm slightly. This was to assure

PAGE 50

kk complete removal of the solvent. The stopcock K was opened, pumping stopped, and G removed from the apparatus. The teflon plugs were inserted, stopcock grease removed from the Joints, and the bottle veiled on the analytical balance. After this, the contents of the bottle could be titrated either for acid or for total oxidizing power. The titrations were performed directly in the filter -weighing bottle, using a micro burette. The precipitated silver residues were then washed free of titrating solution. The samples which had been titrated for available acid were treated with a small amount of thiosulfate solution to reduce any AglOj to Agl. Excess standard KCN solution was added and the bottle stoppered and allcwed to stand with occasional shaking until all the Agl was dissolved. The solution was then transferred to an erlenmeyer flask and titrated with standard silver nitrate according to the Deniges method. Experimental conditions and results appear in Table 7. In this table the experimental data relating to the preparation of the complex to be isolated is tabulated, along with the amounts of complex isolated in each case. Values for oxidizing power, acid number, and recovered silver are tabulated. The corresponding theoretical values are also listed for comparison. Results of the above measurements . --The elusive intermediates in the reaction between (s) and I 2 have been isolated and identified. Their formula has been shown to be (OCOCOgAgl. The technique for their separation and analysis has been described. The corresponding

PAGE 51

li5 complexes for bromine were not isolated. ( GC 00 ) 2 Agl, though evidently quite stable in solution, is prone to decomposition when removed' from the solvent. The complexes were isolated at -78° as white crystals which rapidly took on a yellow-brown cast when warmed to room temperature. Evidently the occurrence of partial or total decomposition of the complex is marked by this color change, since in early experiments when the complex was allowed to warm to room temperature during the latter part of the separation prior to weighing, ! the silver content was high in comparison to the acid equivalent. In these cases the precipitate had developed the yellow color. It can then be concluded that the complex exists as the stoichiometric entity (9C00)2AgI only in solution or at reduced temperatures . The fact that the precipitated mixture of (QCOO^AgX and AgX is stable, retaining its oxidizing power over long periods of time, and that equivalent mixtures of (S) and X combined in the absence of a solvent produce the observed homogeneous yellow powder which is stable in the absence of moisture or reducing agents, would indicate that the complex is stabilized by the presence of an additional mole of AgX with the formation of a material of the stoichiometric formula GCOOAgX which resists addition of further X.

PAGE 53

uses car the complexes in the synthesis of fluorocarbon derivatives General Discussion Mixtures of (S) and X have been found to be capable of halogenating aromatic substances. Haszeldine and Sharpe 12 and Henne 1 P) and Zimmer1 have investigated this reaction. Earlier investigators arrived at similar conclusions when investigating the reactions of the silver salts of the ordinary carboxylic acids. While the present writer cannot, in view of the above-cited findings, agree with the mechanisms advanoed by Haszeldine and Henne and their co-workers, it was decided that further investigation into the reaction to clarify more thoroughly all the factors involved should not be undertaken at this time. The overall equation for the reaction may be written; ©COOAg + Xg + ArH s* ©COOH + AgX + ArX The above investigators have demonstrated this to be a highly useful and selective means of obtaining ring-halogenated aromatic c expounds. The reactions involving decarboxylation of (S), either with or without other materials being present, are discussed in the introduction. When equivalent quantities of (s) and X are mixed in the presence of a solvent and the mixture heated to 120-150°, COg is bl

PAGE 54

48 evolved; and a mixture of products is formed, consisting of the fluorocarbon mono-iodide, the fluorocarbon carboxylic acid anhydride, and a small amount of straight-chain fluorocarbon (usually containing twice the number of carbon atoms as the original fluorocarbon radical in the starting material). When an excess of halogen is present, the reaction proceeds toward the formation of almost 100 percent halide by the previously cited overall reaction ecooAg + x 2 v co 2 + Agx + ei Although mechanisms involving the complex (6C00) 2 AgX might be postulated for reactions of this type, the present writer feels that nothing would be gained throu^i the suggestion of such mechanisms without the confirmatory evidence which could be gained only through a kinetic study of each process for which mechanisms are advanced. One immediately is struck by the possibility of substituting other elements or functional groups for the halogen involved in the complex, or substituting other elements or compounds for the second equivalent of halogen whose presence in the reaction mixture makes possible high yields of fluorocarbon halides. The manner in which the decarboxylation would proceed in the presence of these substitute ingredients presents an intriguing problem to the investigator. Much of the following experimental work can be classified under this heading. The ability of the Simonini complexes to add to olefinic substances with the formation of halo-esters followed by di -esters, as

PAGE 55

49 previously described under the work of Provost, suggests an interesting line of investigation. Reasoning by analogy (an admittedly hazardous procedure), one oould suggest that a reaction of the following sort might occur (eC00) 2 AgX + -CH=CH >9C00-CH -CHX + ©COOAg A displacement reaction of the type 9C00-CH-CHX + ©COOAg > ©COO-CH-CHOOCG + AgX might then occur. If the reactions were carried out using a fluorocarbon olefin, the first reaction might conceivably occur. However, the second step would be precluded by the demonstrated fact that hydrogen-free fluorocarbon esters are not obtainable by a reaction of the latter type. It was shown by Hauptschein and Von Grosse, 1 ' who attempted CF COQAg in a bcxnb at elevated tempera tures, that ester formation 5 did not occur to any measurable extent. The present writer independently arrived at similar conclusions through the reaction of CF^COOAg with C^Fyl in a sealed tube at temperatures up to 250°.

PAGE 56

50 Substitution of Pyridine for X in the Complex -Forming Reaction Discussion : — In view of the unusual capacity of pyridine for complex formation, it was decided that an experiment involving the substitution of pyridine for X in the reaction of (s) with X would be productive of interesting results — perhaps leading to the development of a route to alkf arylated pjoridines by a path of the type eCOOAg + Py >. COg + G Py + Ag Pyridine was found to react with (s) at room temperature and in the absence of a solvent in a 1:1 mole ratio. A solid addendum 4 compound, which was isolated and Identified, resulted. The preparation and identification of these substances and their subsequent reactions are described in the experimental part of this section. Experimental -A. Preparation, purification, and identification . --The substances of the formula GCOORyAg were prepared from the silver salts of acetforic acid through valerforic acid. Equimolar quantities of pyridine and (S) are combined under anhydrous conditions at room temperature. Power stirring is employed to mix the reactants thoroughly. An exothermic reaction occurs, which warms the reaction mixture to 45-60°, depending upon the amount of material involved. After the heat of reaction has dissipated, white needle or rhombic crystals are formed. These crystals have the above formula and may be used without further purification in the reactions to be described in the "B" part of this section. However, establishment of formula necessitates removal of any excess pyridine or impurities that may have been present in either reactant.

PAGE 57

51 It was found that (9C00)PyAg is soluble in benzene, chloroform, ether, and boiling heptane or hexane. In the case of ether or chloroform, however, secondary complexing involving the solvent occurs A second mole of pyridine, if added to this system, results in the pre^ cipitation of an unstable, hygroscopic complex, whose formula is approximately (9C00)Py2Ag. Removal of ether results in gradual decomposition of this material, with the Iosb of pyridine. An oil results, from which ©COOPyAg can be recovered. For example: C2F^C00PyAg was dissolved in ether. The solution was chilled to -78° in order to effect recrystallization. No precipitation occurred. A second mole of pyridine was added. White cyrstals immediately precipitated, thus indicating that C2F^C00Ag probably can react with 2 moles of pyridine to give a substance of the formula (CgF^COOjlfrgAg. The crystals were separated, and ether was removed in vacuo . A white, highly hygroscopic powder was obtained. The powder had an odor of pyridine and upon analysis was found to contain 26.8 percent silver (theoretical for ( ^Fj-COO ) Ry 2Ag is 25.17 percent). This material, was freed of excess pyridine by washing with anhydrous heptane. The resultant product was found to contain 29-9^ percent silver (theoretical for (CgF 5 C 00)£yAg is 50.86 percent). A second experiment of this type, employing a 2 : 1 mole ratio of pyridine to (s) was conducted. Silver valerforate, 18.5 g. ( .05 mole), and pyridine, 7.9 g. (. 1 mole), were combined as before. A heavy oil resulted. An attempt was made to remove the excess pyridine by extraction with heptane. This was unsuccessful; the oil

PAGE 58

52 remained undissolved at room temperature. Upon heating, approximately half the oil went into solution. When the solution was oooled to 0°, the oil separated unchanged. The heptane was removed, M the oil dissolved in 100 cc. ether. Upon cooling, crystals were obtained. These were washed with anhydrous heptane and recrystallized from boiling heptane, followed by drying in vacuo . The product, a rather yellow powder, was found to contain 23.49 percent silver (theoretical for (C^COOjPyAg is 24 percent). The pure addends were prepared by combining (s) with pyridine in a 1:1 mole ratio and recrystallizing the product from boiling heptane. This solvent was found to be most suitable for the purpose, although the solubility of (9C00)l^Ag in this medium is rather low even at the boiling point. The solubility was found to increase with chain length. Resistance of (ecOO)pyAg to hydrolytic action by atmospheric moisture was found to increase with chain length. The analyses were found to approach more nearly the theoretical value as chain length increased. Data pertaining to the preparation and analysis of the pure addends appear in Table 8. TABLE 8 PREPARATION AND ANALYSIS OF PYRIDINE ADDENDS CF THE SILVER FLUOROCARBON CARBOXYTATES (S) g. mole pyridine mole Product "T*Xg Found % Ag 1 The or. CFjCOQAg 16.0 “o45 $.5 ir\ do (CF^COO)PyAg 35.64 36.00 CgF^COOAg 6.0 ".ooY “ 37 “ (C 2 F 5 COO)PrAg 30.31 w C^FyCOOAg 37.1 .11 .086 (C 3 F 7 COO)PyAg 66.54 67.00 Cl % 1 37.1 .10 ~T7T ,lo (C^COOjFYAg '“ 25:75 '24.00

PAGE 59

55 B. Reactions . --If the addends are heated, decarboxylation begins to occur at 160°. Carbon dioxide and fluorocarbon hydride are evolved. Free silver and pyridine are formed, along with a complex mixture of compounds containing fluorine. Pyridinium hydrofluoride and a hi^i-boiling oil, tentatively identified as a difluorobypyridyl, were isolated from this mixture. When the addend resulting from the addition of pyridine to CgFjjCOQAg was decarboxylated, the above products were Identified as follows: Ethforyl hydride was obtained in the cold trap on the exit line from the reactor. It had a molecular weight of 119 (calo. 120). Its infrared spectrum contained the characteristic C-F and C-H absorption bands. If water was excluded from the reaction, CgF^H was present in approximately half the total amount available from decarboxylation Qf (S). If, however, the reaotants are not thoroughly dry, a quantity approaching the theoretical value is obtained. Pyridinium hydrofluoride, obtained in the distillate of the reactor residue, was purified by recrystallizaticn from ethyl alcohol. It was observed to liberate fluoride ion upon solution in water. Treatment with dilute alkali liberated pyridine. Pyridinium hydrofluoride was prepared by reacting pyridine with hydrofluoric acid and recrystallizing the dried product from anhydrous ethanol. Both materials melted 158-162°. However, a mixed melting point was depressed to 146-158°. An oil boiling 500-504° containing both fluorine and nitrogen was isolated. This material, when dissolved in anhydrous ether and

PAGE 60

treated with hydrogen chloride, immediately gave a white precipitate. After repeated washings with ether followed by vacuum drying, the precipitate was analyzed for chlorine and titrated for HC1 oontent. Two samples analyzed 27.14 and 27.25 percent chlorine. (Calculated values for non of lu or oby py r idyl dihydrochloride and difluorobypyridyl dihydrochloride are ^ 8.51 and 26.79 percent, respectively.) One sample analyzed 27.93 percent HC1. (Calculated values for the above compounds are 29.3 and 27.54 percent, respectively.) The hydrochloride was found to contain fluorine. These analytical data indicate that the material was a difluorobypyridyl. The solid residue remaining in the reactor was thoroughly washed free of organic material with ether and heptane, dried, and analyzed for silver. The silver content was found to be 98.9 percent. Substantially similar results were obtained from the decarboxylation of the complex resulting from the addition of pyridine to silver acetf orate. If pyridine and (s) are present in a 1:1 mole ratio, decarboxylation does not occur until a temperature of 220-240° is reached, and considerably more charring and formation of pyridinium hydrofluoride occurs. Sinoe it was suspected that alkf orylatian of the pyridine nucleus had occurred to a small extent, it was decided to introduce into the reaction mixture an aromatic substance which would be expected to undergo alkforylation much more readily than pyridine. Experiments were conducted using 2 -butyl benzene and meta di-isopropyl benzene. Eo alkforylation products were isolated.

PAGE 61

55 Results of the above measurements . --Pyridine reacts with (s) in a 1;1 mole ratio to give a crystalline solid, whose analysis indicates it to contain 1 mole of pyridine and 1 mole (S). This substance will decarboxylate at 160-170°, yielding the corresponding fluorocarbon hydride in high yield, along with pyridinium hydrofluoride, monoand/or di -fluorobypyridyl, and a complex mixture of oils containing fluorine. This reaction should be investigated more thoroughly, as it shews promise of providing a route to alkforylated pyridines or aromatics.

PAGE 62

56 An Improved Method for the Synthesis of Fluorocarbon Mono -Iodides^ Discussion . --A method of preparing fluorocarbon iodides which consisted of the decarboxylation of the silver salt of a fluorocarbon carboxylic acid (s) in the presence of iodine was developed by Simons 38 and Brice. The reaction mixture was kept in suspension in a fluorocarbon diluent by means of mechanical stirring. The process was carried out under strictly anhydrous conditions. Other s 10 > 11; ,l6 prepared the iodides by heating intimate mixtures of (S) and iodine. The method herein reported consists of bringing the iodine vapors in contact with (s) under controlled conditions of temperature and pressure without a solid or liquid diluent being present. A sharp reaction zone progresses through the material until (S) is consumed. The fluorocarbon iodide passes from the reactor and is collected as in the original method. Power stirring, and prepurification and drying of the iodine, as well as the limitations imposed by the diluent, are eliminated. Maintenance of anhydrous conditions, recovery of silver and control of reaction temperature are facilitated. Higher product yields are obtained. Two types of reactors are currently in use in these laboratories, each having advantages depending upon the particular situation. Procedures using these reactors are; PROCEDURE 1: This method employs a reactor which is simple to construct and provides a quick method of obtaining small quantities

PAGE 63

57 of fluorocarbon Iodides in the laboratory. It is also adapted to reactions of this type in which the absence of any atmosphere other than the reaction gases is desired. The reactor is a vertical Pyrex tube, closed at the bottom and stoppered at the top. An exit line attached near the top connects it through two cold traps to a vacuum souroe. A manometer is attached to the exit line. A KOH tube for removal of COg is placed between the traps and the vacuum source. Reaction temperature is followed by means of a thermocouple or thermometer extending into the tube through the top. Experimental . —The entire apparatus is vacuum dried. Iodine followed by phosphorus pentoxide (P) is poured into the tube. (S) may be either suspended over the iodine and (P) in a steel-wire cage or poured directly into the tube with a layer of glass wool separating (s) from (P). The apparatus is flushed with dry air. The reactor tube is heated in an oil-bath. The pressure is reduced in order to provide adequate volatilization of iodine at the optimum reaction temperature (13O-I6O 0 ). Completion of reaction is noted by cessation of evolution of gases. The bath is heated to 185° to assure complete conversion of (s). The iodide is collected in the cold traps. All products were fractionated. Iodine was determined by peroxide decomposition and thiosulfate titration. Reaction conditions and yields are in Table 9A. PROCEDURE 2: This method is superior in cases in which relatively large amounts of various fluorocarbon iodides must be synthesized in the laboratory. The apparatus is shown in Figure 2.

PAGE 64

58 Operation is at atmospheric pressure. A stream of dry air carries the iodine vapor from the pot to the reaction zone, (s) is supported in the reactor by means of a glass "X" member and a thin layer of glass wool. The reactor and the iodine container are separate units whose temperatures are controlled independently. Experimental . --The apparatus is assembled, (s) is poured into the reactor from the top. Final drying is effected by passing a current of dry air through the reactor while heating it to 70-80°. Excess iodine followed by (p) is placed in the pot. The iodine is heated to sublimation temperature and an extremely gentle air current started. The heat of reaction tends to increase the reactor temperature. The Jacket heating is adjusted so as to maintain an optimum reaction temperature of I 3 O-I 6 O 0 . The reaction requires 20-k0 minutes, depending on the type and amount of (s). The iodide is collected in cold traps. Purification and identification were carried out as in procedure 1. Reaction conditions and yields are in Table 9B. Procedure 2 was used in the preparation of CjFjBr. No air current was employed. Reflux action of bromine was provided by an ice -water -cooled condenser on the upper end of the reactor. Reaction temperature was 130-150 0 . Products were fractionated and analyzed for bromine. A 67 percent yield was obtained. f

PAGE 65

59 TABLE 9 IMPROVED SYNTHESIS FOR FLUOROCARBON IODIDES A. Procedure 1 Starting Material g. Heating Time, Min. Operating Temp . , °C Operating Pressure, mo.. Basket Support Yield, g. Yield, i CFjCOQAg 75.1 60 140-150 1+00-500 Not used 47.6 71 C^COOAg 15.5 80 11+0-145 450 -500 Not used 15.1 95 CF^COQAg 75.4 l4o 150-160 400-500 Used 52.5 78 C^COOAg 21.2 115 150-160 400-500 Used 16.5 87 B. Procedure 2 Starting Material g. Heating Time, Min. Temp. , °C Yield, g. Yield, * CFjCOQAg 55.3 21 150-160 22.3 74 CF^COOAg 202.1 48 150-160 161.0 90 C^F^COOAg 42.0 25 150-160 56.5 94 C^F^COOAg 200.4 45 150-160 175.5 95

PAGE 66

60 Fig. 2 To •fro Heater coi I Reactor Support Pot" P Iodine

PAGE 67

61 Beaction of (S) and the Complex with PhosphoruB Discussion . — Bennet, Brandt, Emeleus, and Haszeldine^ have described the formation of trimethf oryl phosphine arai iodomethf aryl phosphines from the reaction of phosphorus with methforyl iodide in a closed vessel. It was decided that the decarboxylation of (s) in the presence of red phosphorus might provide a source of trialkforyl phosphines without resorting to the preparation of the iodide and without obtaining the iodoalkforyl phosphines as by-products. Instead, the reaction yielded the fluorocarbon carboxylic acid anhydride. The red phosphorus was converted to the yellow allotropic form. Haszeldine and Sharped have Bince reported that a mixture of red phosphorus and silver aoetforate yielded acetforic anhydride upon thermal decomposition. The results independently obtained in these laboratories arc nevertheless reported here, in some detail, as confirmation and clarification of the work of the above investigators. The reaction of red phosphorus with the solid complex formed from the reaction of silver aoetforate with iodine in the absence of any solvent is described. Experimental . --Silver aoetforate, 75 g. (.5^ mole), was mixed with red phosphorus, 26 g. (.81)mole) . The mixture was placed in a flask to which was attached a gas collection train. An attempt was made to initiate a self -propagating reaction by heating a small spot on the flask with a gas torch. CO2 evolution immediately occurred

PAGE 68

62 tut stopped, as soon as the flame was removed.. The entire flask was then heated, with a heating mantle. Decarboxylation began at approximately 200°. Heating was continued, until COg evolution ceased. The product was fractionated. The fraction boiling 38 .5-hO° consisted of 32.5 g. of material. A .5001 g. sample was treated with water. The material reacted violently, liberating acetforic acid which was neutralized by titration of 37.6k 00. of .1266 N. NaOH (correct for acetforic anhydride). The reactor residue consisted of silver oxide and yellow phosphorus . Silver aoetf orate, k5.k g. (,205k mole), was mixed with iodine, 25 g. (.098k mole), in the absence of a solvent. A homogeneous yellow powder resulted, having the composition CF^COOAgl. To this was added a slight excess of red phosphorus. Upon Btirring, a homogeneous gray powder was formed. This material spontaneously, and without warning, exploded -doing considerable damage to both the apparatus and the dignity of the present writer. Ho products were obtained. Results of the above measurements . — Silver acetf orate (and presumably the higher salts) when decarboxylated with red phosphorus in the absence of any solvent, gives almost a 100 percent yield of aoetforic anhydride. The red phosphorus is converted to the yellow form in the process. The equation for the overall process may be written 2 CF^COOAg + (X) P (red) ^(CF^CO^O + A^0 + (x) P (yellow) The complex mixture of the formula CFjCOOAgI reacts spontaneously with red phosphorus.

PAGE 69

RELATED SYNTHETIC PROCEDURES Preparation of Fluorocarbon Nitriles from (s) and Cyanogen Discussion . — A novel method for the preparation of fluorocarbon carboxylic acid nitriles which apparently is analogous to previously described fluorocarbon halide syntheses was developed. The method Involves the decarboxylation of the silver salt of the fluorocarbon carboxylic acid while passing cyanogen gas through or over the reacting mass. In this procedure, use is made of the halogenoid properties of cyanogen. The overall equation for the reaction may be written QCOOAg + RCCN > ©CN + AgCN + COg Inasmuch as the process was conducted at high temperature and no intermediates were isolated, the author will not hazard the postulation of any mechanism or the suggestion of an initially -formed reaction product involving (s) and cyanogen. The fluorocarbon carboxylic acid anhydride was isolated from the reaction produot along with a small amount of straight -chain fluorocarbon which proved to be the dimer of the fluorocarbon radical involved. The equations for these reactions in which decarboxylation occurs without the involvement of cyanogen may be written 2 QCOOAg >-(9C00) 2 0 + Agr,0 and 2 QCOOAg > oe + 2 CO 2 + 2 Ag

PAGE 70

64 respectively. The decarboxylation of (S) in the absence of other reactive materials has been investigated by Kirshenbaum, Streng, and Hauptsche in . Similar products were obtained. Experimental . --The procedure used in the preparation of butyrforo nitrile is described as follows: The reaction was carried out in a Pyrex tube heated by a muffle furnace. Silver butyrf orate mixed with glass chips was packed into the tube. The temperature was slowly raised while maintaining an atmosphere of cyanogen in the tube. The reaction began at approximately 300°. Carbon dioxide was detected by means of a calcium hydroxide scrubber on the exit line. The cyanogen was allowed to stream in at one end while the effluent gases were collected in cold traps connected to the opposite end. The reaction evidently is greatly hindered by the difficulty of providing adequate silver salt-cyanogen contact during the reaction process. The low yields are attributable to this factor, along with the above-mentioned side reactions. The products were fractionated and identified by physical and chemical properties. A 4.1 g. fraction of clear, inert liquid was obtained, boiling at 55-57° and having a molecular weight of 335 as determined by the Dumas method. (Calc, for is 338.) A 6 g. fraction was obtained, boiling at 103-106°. This material reacted violently with water, liberating butyrforic acid. The latter was converted to the amide, which melted at 103-105° (correct for butforamide). Essentially the same procedure was followed in the preparation of valerforo and caproforo nitriles. In every case the nitriles were

PAGE 71

65 hydrolized to the acid by treatment with HgSO^ and converted to the methyl ester, from which the amides were prepared. For comparison purposes, the same nitriles were prepared by the conventional method involving dehydration of the acid amide. Two previously unreported nitriles, C 1 F 9 CN and C 5 F 3 JCN, were prepared. Nitrogen analyses on these were performed by Peninsular ChemRe search, Inc. Experimental data appear in Table 10A. Analytical data appear in Table 10B. Results of the above measurements . --The investigation was undertaken in order to determine if the previously described method for the preparation of fluorocarbon halides by the decarboxylation of (S) in contact with the vapors of X could be extended to the use of compounds described as "pseudo halogens" or "halogenoids” in place of X. The reaction reported, though hardly a practical means of preparing fluorocarbon carboxylic acid nitriles, serves to illustrate the "halogenoid" character of cyanogen as related to reactions of this type.

PAGE 72

66 TABLE 10 PREPARATION OF FLUOROCARBON NITRILES BY THE DECARBOXYLATION OF SILVER SALTS OF FLUOROCARBON CARBOXYLIC ACIDS IN THE PRESENCE CF CYANOGEN A. Experimental Data (S) gmoles Reactor Temp . , °C Nitrile Obtained g. moles Yield, i CjF^COOAg hO .125 5ltO-550 c 5 f t cn 5.6 .029 23 C gC OOAg 88 • 237 350 Ci^CN 7.5 .031 13.1 CkJF ^COOAg 85 .205 350 C 5 F 11 CN 6.1 .021 10 B. Analytical Data Nitrile Source B.P. Mol. Nitrogen, # Amide Mixed Wgt. Calc. Found M.P. M.P. c 5 f t cn From (s) -1 — +1 1^3 105 Literature -1 — +1 145 105 C^F g CN From (S) 28.5 196 5.72 5.65 110 110 From Amide 28.0 196 111 c 5 f 1 i cti From (S) 56.0 244 4.75 4.55 115 114 From Amide 56.0 245 114

PAGE 73

67 Synthesis of Fluorocarbon Mono-IocLides from Fluorocarbon Carboxylic Acid Anhydrides Discussion . —It was discovered that fluorocarbon mono-iodides can be prepared continuously by passing a vaporized mixture of iodine and fluorocarbon carboxylic acid anhydride through a hot tube containing silver iodide. In this procedure, the anhydride is made as a part of the overall process, thus providing a method for going from the fluorocarbon carboxylic acid to the corresponding iodide in one procedural step. The overall equation for the reaction may be written (ecoo) 2 o + i 2 > 2 ©i + co 2 + co The function of the silver iodide is not clear. However, its beneficial effect is evident upon comparison of the results obtained with and without this material in the reactor tube. 12 Haszeldine prepared methforyl iodide by heating acetforic anhydride with iodine at 350 ° for 12 hours, presumably in a sealed vessel. Experimental . — The fluorocarbon carboxylic acid is metered into a heated three-neck flask containing phosphorus pentoxide. To one neck of the flask is attached a line through which dry air can be admitted, if necessary, to carry the reactant vapors into the reactor tube. The anhydride is formed by reaction of the acid with phosphorus pentoxide, and passes out through the other neck of the flask into a large tube leading into a furnace. A flask containing iodine and phosphorus pentoxide is also connected to this tube. The flask is heated, causing the iodine to sublime into the tube and mix with the vapors of anhydride

PAGE 74

68 passing into the furnace. The silver iodide, mixed with glass chips, is packed into the heated portion of the tube. The mixed vapors pass through and over the silver iodide. A condenser is provided at the exit end of the tube to prevent the passage of iodine into the collecting traps. The product, along with some unreacted anhydride, is collected in the traps which are appropriately cooled, depending upon the boiling points of the products being formed. Evolution of carbon dioxide is detected by means of a calcium hydroxide scrubber attached to the exit line. Reactor temperatures ranging from 300° to hOO° were employed, although at the higher temperatures decomposition begins to occur, as evidenced by the appearance of fluoride ion in the calcium hydroxide scrubber. Experimental data and product yields appear in Table 11. All products were fractionated. Fluorocarbon iodides were identified from fractionation data and molecular weights. The latter were determined by the Dumas method. Results of the above measurements . --The presence of silver iodide in the reactor tube has a definite beneficial effect, both on conversion and yield. The latter term is based on the amount of unrecoverable acid. Lower reactor temperatures can be employed with resultant decrease in the formation of decomposition products. The overall yields are correspondingly improved. Apparently the optimum reactor temperature is around 325 ° for the materials investigated. Higher conversions might be achieved through the use of a longer reactor tube with specially prepared and supported silver iodide.

PAGE 75

69 A reactor designed to re-cycle the unreacted anhydride might also achieve the same end. However, time in which to do further development work on this interesting synthetic procedure was not available.

PAGE 76

70 a i r\ •d rH © i 00 CO M3 m CVJ •H H i iH CM CM VO MO -3t o CVJ . , •cJ ® t • •4 1 MO ov o 8 •d » © © Q\ cm in Ti S H i CM o rH CM o o o 3 o •* VD CO VO b’d ° 'd * i o rH CO o rH ,© h O © *30 i • • • d fj © o-^s. K"\ iH MO VO i — i s g 3 H 1 — 1 rH o © H t in -d" © K\ VO VO CM O rH rH in 8 KV in m O O o O o o S • • • • • •N a—J © Td •d cj •H d t •d +5 © +3 © +3 © © 0 o © O CO o © © © CQ K t© t=> £> £> © V0 -4 © t•Jt o -=f -4 lA i — I o m in 3 9 CM CM t— O S • • • cc rH in in -3" in KA -A. — r rH • • • • • • • w 00 tf\ H ITS rH VO K"\ w 8 lT\ CM OJ in H »d o a n Q O •H o o CO o o CO o o o tO o t— to K> KA m pH 15 15 m o b 15 m CO

PAGE 77

DISCUSSION OF RESULTS The investigation reported in this dissertation was undertaken in order to expand the knowledge pertaining to the fluorocarbon carboxylic acids. Special attention was given to the identification and characterization of the intermediates in the reactions between the silver salts of these acids and halogen or halogen-like materials. The author feels that this end was achieved insofar as iodine and bromine are concerned. Chlorine and fluorine were not included in this investigation, as their physical and chemical properties render them unadaptable to the experimental procedures developed and used for the reactions involving iodine and bromine. The author wishes to point out that no extrapolation of the findings reported for iodine and bromine to the other halogens is either stated or implied. It was found that when iodine or bromine are combined with (s) in a completely inert solvent at temperatures below those at which decarboxylation can occur, one equivalent of X is consumed for each equivalent of (s) present. It was further demonstrated that any X added past this point is present only in the free state. Inert fluorocarbons and fluorocarbon derivatives were found to be the only solvents suited to an investigation of this -type. The product of the addition of X to (S) under these conditions was found 71

PAGE 78

72 to be an addendum compound of the formula (9C00) 2 AgX. These substances were characterized in solution in OF solvents by determining the relationship between acid content, oxidizing power, and silver content. Addends of the formula (9C00) 2 AgI were isolated and identified. The complexes can be isolated for analysis only by crystallization from solution at reduced temperature. They undergo partial or total decomposition if allowed to warm to room temperature in the absence of a solvent. However, the substance (eC00) 2 AgX apparently is stabilized by the presence of another mole of AgX, forming an insoluble material of the empirical formula GCOQAgX, which is relatively stable in the absence of moisture. (0COO) 2 AgI is stable in solution. (6C00) 2 AgBr, however, gradually decomposes, as evidenced by decrease in oxidizing power with time. Stability of the complexes was also observed to increase with chain length. Pyridine was found to undergo a complex formation with (s). The complexes can be isolated and purified by recrystallization from ether, and were found to have the general formula (eCOO)PyAg. The pyridine complex undergoes decarboxylation at 1 60° with the formation of the corresponding fluorocarbon hydride, pyridinium hydrofluoride, and a number of other fluorine -containing products. This reaction shows promise of providing a route to alkforylated pyridines or aromatics. The silver salts of fluorocarbon carboxylic acids can be decarboxylated in the presence of red phosphorus to give the

PAGE 79

I 73 corresponding anhydride in high yields. The complex mixture GCOOAgI explodes an treatment with red phosphorus. An improved synthesis far fluorocarbon iodides was developed which produces higher yields and eliminates many of the tedious and time-consuming operations previously inherent in the procedure. Fluorocarbon carboxylic acid nitriles can be prepared by decarboxylating (s) in the presence of cyanogen. The reaction apparently analogous to the above iodide preparation in that cyanogen behaves in a halogen-like manner under the reaction conditions. A prooess was developed far the preparation of fluorocarbon iodides by a decarb axy la ti on -decarbony la ti on of a fluorocarbon carboxylic acid anhydride in the presence of iodine. The yields and conx versions are improved markedly by the presence of silver iodide in the reactor. Through the isolation,, characterization, and establishment of the formula of the halogen addends of the silver salts of the silver fluorocarbon carboxylatee, the author feels that some light has been shed on an area in which considerable confusion and disagreement existed. By comparison of the data found in this investigation with that available an the reactions of the salts of the hydrocarbon carboxylic acids with the halogens, considerable insight into the fundamental differences existing between the fluorocarbon carboxylic acids and their organic counterparts can be gained. A number of interesting and useful processes have been

PAGE 80

lh developed in the course of this investigation. Also, several reactions were discovered which may provide avenues for future research.

PAGE 81

BIBLIOGRAPHY 1. Banus, J., J. Chem. Soc. , 3755 (1953). 2. Bennet, F. W. , Brandt, G. R. A., Emeleus, H. J., an d Haszeldine, R. N., Nature , 166 , 225 (1950). 3. Birckeriback, L., Goubeau, J., and Berninger, E., Ber., 65 , 1359 (1932). ~ 4. Birnbaum, K., and Gaier, J., Ber., 1%, 1270 ( 1880 ). 5. Birnbaum, K. , and Reinherz, H., Ber. , 1£, 456 (l88e). 6 . Bochemuller , W., and Hoffman, F. W., Ann. , 519 , 165 (1935). 7. Cady, G., and Kellogg, K. , J. Am. Chem. Soc. , 75, 2501 (1953). 8 . Crawford, G. H., and Simons, J. H., J. Am. Chem. Soc., 75, 5737 (1953). 9. Carlsohn, : Uber Eine Neue KLasse Von Verbindungen Des Positiv Eintwertigen Jods," Verlag von S. Hirzel, Leipzig, 1932. 10. Goldschmidt f S., and Grafinger, G., Ber. , 68 , 279 (1935). 11. Haszeldine , R. N., J. Chem. Soc. , 584 (1951). 12. Haszeldine, R. N., and Sharpe, A. G., J. Chem. Soc. , 993 (1952). 13. Haszeldine, R. N., and Sharpe, A. G., J. Chem. Soc. , 4259 (1952). 14. Haszeldine, R. U., and Jander, J., J. Cliem. Soc. , 4172 (1953) • 15. Hauptschein, M., and Von Grosse, A., J. Am. Chem. Soc., 73, 2461 (1951). 16. Hauptschein, M. , and Von Grosse, A., J. Am. Chem. Soc., 74, 4454 (1952). 75

PAGE 82

76 17. Henne, A. L., and Finnegan, W. G., J. Am. Chem. Soc.. 72, 580o (1950). 18. Henne, A. L., and Zimmer, W. F., J. Am. Chem. Soc., 73, 1362 (1951). 19. Herschberg, E. B., Helv. Chim. Acta , 17, 351 (1934). 20. Hunsdiecker, H., and Hunsdiecker, C. L., Ber., ][5, 291 (1942). 21. Kirshenbaum, A. D., Streng, A. G., and Hauptsckein, M. , J. Am. Chem. Soc. , 75, 3l4l (1953)* 22. KLeiriberg, J., Chem. Revs, , 40 , 381 (1947). 23. Luttringhaus, A., and Schade, D. , Ber. , j4, 1565 (l94l). 24. Menefree, A., and Cady, G., J. Am. Chem. Soc. , 26, 2020 (1954). 25. Oldham, J. W. H. , and Ubbelohde, A. R., J. Chem. Soc. , 368 (l94l). 26. Oldham, J. W. H., J. Chen. Soc. , 100 (195O) . 27. Peligot, Coopt, Rend. , 3, 9 (1836). 28. Prelog, V., and Seiverth, R., Ber. , 74 , 1769 (1941). 29. Prevost, C., Coopt. Rend. , 196 , 1129 (1955). 50. Prevost, C., Conrpt. Rend. , 197 , l66l (1933). 31. Prevost, C., Compt. Rend. , 200 , 942 (1935). 32. Prevost, C., and Lutz, R., Coopt. Rend. , 198, 2264 (1934). 33. Prevost, C., and Wiemann, J., Compt. Rend, , 204 , 700 (1937). 34. Prevost, C., and Wiemann, J., Compt. Rend. , 204 , 989 (1937). 35. Simonini, Monatsch , 13 , 328 (1892). 36. Simonini, Monatsch , 14, 8l (1893). 37. Simons, J. H., et al., J. Electrochem, Soc,, 22 Wo. 2, 47 (1949). 38. Simons, J. H., and Brice, T. J., U.S. Patent 2,554,219, May 21, 1951.

PAGE 83

77 39 . 40 . 41. 42 . * 3 . 44 . Svarts, F, , Bull. Sci. Acad. Roy, Belg. , 12 , 721 (1926). Swarts, F., Anal. Soc. Fis. Quim. , 27, 683 (1929). Uschakov, M. I., and Tchistcw, W. 0., Ber . , 68 , 824 (1935)* Wieland, H., and Fischer, F. G., Ann. , 446 , 49 (1926). Windhaus, A., and Klanhardt, F., Ber. , 581 (1921). Windhaus, A., and Klaniiardt, F., Ber. , 33 > 398l (1922).

PAGE 84

VITA George Homer Crawford, Jr., was "born in Houston, Texas, cm October 6, 1928, the sen of George H. Crawford, Sr., and Augustine Sabayrac Crawford. In 1930 his family took up residence at Baytown, Texas. In May, 1946, he completed his work at Robert E. Lee Hig£i School; and he attended the summer session of Lee Junior College at Baytown. From September, 19 46, to May, 19U7, he attended Sam Houston State Teachers College at Huntsville, Texas. He entered Baylor University at Waco, Texas, in September, 1947. In March, 1949, he was married to Evelyn Brasher. He received the B. A. degree in March, 1950, and entered graduate school at Baylor University immediately following graduation. After receiving the M. A. degree in August, 1951, he entered the university of Florida, where he had been granted a pre-doctoral research fellowship under Dr. J. H. Simons. Upon receipt of the Ph. D. degree in June, 1954, he plans to pursue his professional career in industry. 78

PAGE 85

This dissertation was prepared under the direction of the chairman of the candidate's supervisory committee and has been approved by all members of the committee. It was submitted to the Dean of the College of Arts and Sciences and to the Graduate Council and was approved as partial fulfillment of the requirements for the degree of Doctor of Philosophy. June 7> 195^Dean, Graduate School SUPERVISORY COMMITTEE: 7777 ‘-7 7