AN ABSTRACT OF THE THESIS OF
George Vance Gritton for the M. S. (Name of student) (Degree) in Chemistry presented on p^,--.,, •;•?. (Major) (Date)
Title: STUDIES RELATING TO THE PREPARATION AND
PROPERTIES OF PERFLUOROALKYL SUBSTITUTED
BORANES
Abstract approved: Redacted for privacy Theran D. Parsons
The structures of the boron hydrides are surveyed. Aspects of several theories of electron deficient bonding are discussed.
Mechanisms for the disproportionation of unsymmetrical boranes are reviewed and the chemical consequences of the ability or lack of ability of the CF - group to participate in three-center bonds are discussed.
Possible approaches to the formation of CF-B bonds, including successful and unsuccessful attempts to synthesize such bonds, are examined.
Successful synthesis of a new compound, CF B(CH_) , has been accomplished. The preparation reactions are: 1. 2CF-I + Hg +Cd ACd • (CF,)_Hg +Cdl 3 Amalg. 3 2 2 2. BBr3 + 2(CH3)3B -3(CH3)2BBr 3. (CF3)2Hg + 2(CH3)2BBr -2 CF^fCH^ + Hg Br2 The experimental molecular weight of CF~B(CH_)~ has been determined to be 112. The compound is monomeric in the vapor phase.
CF B(CHJ is unstable under the conditions of its synthesis.
The major identifiable product of the decomposition was determined to be (CH3)_BF, There is evidence that, in addition, a polymer of
CF_ is formed.
The conclusion is drawn that the CF - group, in the physical and molecular environment involved, is unable to participate in the formation of three-center bonds. Studies Relating to the Preparation and Properties of Perfluoroalkyl Substituted Boranes
by
George Vance Gritton
A THESIS
submitted to
Oregon State University
in partial fulfillment of the requirements for the degree of
Master of Science
June 196 7 APPROVED:
Redacted for privacy
Redacted for privacy
Chairman of Department of Chemistry
Redacted for privacy
Dean of Graduate School &
Date thesis is presented ^"jJUugULu. °?y / 7
Typed by Opal Grossnicklaus for George Vance Gritton ACKNOWLEDGMENT
The author wishes to express his sincere appreciation of the guidance, encouragement, and patience of Dr. Theran D.
Parsons throughout the course of this investigation.
The material aid of the National Science Foundation for a portion of this work is acknowledged with appreciation. TABLE OF CONTENTS
I. INTRODUCTION 1
II. EXPERIMENTAL 17
Apparatus and Equipment 17 Reagents 19 Preparations and Reactions 24
III. DISCUSSION AND CONCLUSIONS 48
IV. SUMMARY 55
BIBLIOGRAPHY 56 LIST OF FIGURES
Figure Page
1 Typical reaction or storage vessel. 17
2 Vapor pressure vs. l/T for CF BMe 32
3 IR spectrum of one fraction of impure Me B from the decomposition of CF BMe . 34
4 IR spectrum of a second fraction of impure Me B from the decomposition of CF BMe . 35
5 IR spectrum of Me BF from the decomposition of CF3BMe2. Z 37
6 Vapor pressure vs. l/T of Me BF from the decomposition of CF BMe . 38
7 IR spectrum of the highly volatile fraction from the decomposition of CF BMe . 39
8 Vapor pressure vs. l/T of an unknown substance resulting from the reaction of CF I with Me N. 41
9 Vapor pressure vs. l/T of an unknown substance resulting from the reaction of CF I with pyrolized LiBH4. 45
10 IR spectrum of an unknown substance resulting from the reaction of CF I with pyrolyzed LiBH . 46 LIST OF TABLES
Table Page
1. Constant temperature baths 20
2. Vapor pressure thermometers 20 STUDIES RELATING TO THE PREPARATION AND PROPERTIES OF PERFLUOROALKYL SUBSTITUTED BORANES
I. INTRODUCTION
Until recently, interest in boron hydride chemistry was primar ily academic and was mainly concerned with the failure of the boron hydrides and some of their derivatives, as well as certain analogous compounds of other metals such as aluminum, gallium, and platinum, to conform with the rules which relate chemical composition with classical valence theory. Within the last 15 years or so, however, the boron hydrides and their derivatives have assumed great practical importance. This surge of interest has been the result of the use of certain boron compounds as high energy fuels, of others as fuel addi tives, and of still others as powerful and selective reducing agents in organic chemistry.
Because these compounds depart from classical valence theory, it is apparent that knowledge of their structure and bonding is of prime importance in applications such as those above and in the potential discovery of new applications. It is also possible that a knowledge of the bonding of these compounds may be of help in elucidating the nature of the bonding in other troublesome areas, such as the bonding in intermetallic compounds. For these reasons, it is interesting to trace the development of the theory of the structure and bonding of the boron hydrides and their derivatives.
A series of boron hydrides, or boranes, was first character ized by Alfred Stock (46) and his collaborators in the period from
1922 to 1936. His pioneer work in the field stimulated others, not ably H. I. Schlesinger, to enter this area of inorganic chemistry around 1930. Schlesinger and his co-workers developed new and improved methods of production of the boron hydrides, characterized many new compounds related to them, and made numerous structural studies. By 1942 boron hydride chemistry was considered of suffi cient importance to warrant an extensive review by Schlesinger and
Burg (42).
The hydrides of boron and their related compounds are in a class characterized by electron-deficient bonding. This term is applied to those compounds which contain insufficient valence elec trons to permit all of the adjacent atoms to be held together by simple two-atom electron-pair bonds. Other examples of compounds of this typeare Al (CHJ,, Al CI,, and [ (CH3) Pt] . In a wider sense, the term "electron-deficient" is used to designate those compounds in which the number of available orbitals of approximately equal energy is greater than the number of normally available electron pairs. This definition allows the inclusion of compounds such as
(CH-KB and BF., in which the p orbital of the valence shell of boron is nominally vacant. Diborane, B_H/? may be considered as an example of such 2 6 compounds. Each boron atom can contribute three bonding electrons and each hydrogen atom, one. There is available for bonding, then, a total of 12 electrons, or six electron pairs. This is not sufficient to bond together the eight atoms with "normal" two-electron bonds.
Soon after the determination of the molecular formula of diborane by Stock, several structural theories were proposed (42).
The first, however, that met with any degree of success was that of the one-electron bond proposed by Sidgwick (43, p. 103). This was later expanded by Pauling (29, p. 23 9-244) to include other resonance structures, including a no-electron bond. Bauer's interpretation of electron-diffraction data (4, 5), which precluded bridged structures such as Dilthey (10) had earlier proposed, strongly supported the essentially ethane-like structures upon which these bonding theories were based.
As further experimental data were recorded, these structures were abandoned. The compounds were found to be diamagnetic, a finding which was at variance with any structure postulating a one- electron bond. Infra-red spectroscopy studies by Price (32, 33) were instrumental in establishing a bridged structure to be the correct configuration of diborane. More precise electron diffraction studies by Hedberg (18) aided in definitely establishing this structure, and in removing the objections raised by the previous work of Bauer. 4
An ethylene-type structure with two protons buried in a double
bond connecting the two boron atoms was proposed by Pitzer (31),
It was felt, however, that proposals concerning the bonding of dibor
ane should be applicable to compounds such as the dimers of alumi
num chloride and trimethyl aluminum as well, and here Pitzer's
concept was not well received, for "methylated double bonds" seemed
quite strange, while "halogenated double bonds" strained credulity.
In 194 7 Rundle published a paper on "half-bonds" applied to
electron deficient bonding (34). The proposed half-bonds used one
orbital and one electron pair to form two different bonds. His work
was expanded and continued in two subsequent papers (35, 36). This
quantum-mechanical treatment led to the very important conclusion
that "normal" structures which leave one or more low-energy orbitals
vacant are unstable with respect to electron-deficient bonding with half-bonds in which all of the low-energy orbitals are used. In his view, the tendency to use all low-energy orbitals is the underlying
principle of electron-deficient bonding. In 1954 Eberhardt, Crawford, and Lipscomb (11, 22) presented a similar treatment, though in greater detail, of the boron hydrides and coined the term
The pitfalls of boron hydride structure theory are entertain ingly pointed out in this quotation from Eberhardt et al. ; "We have even ventured a few predictions, knowing that if we must join the ranks of boron-hydride predictors later proved wrong, we shall be in the best of company. " "three-center bond. " They pointed out that in a normal covalent
bond two atoms use two orbitals, one centered on each atom, which
interact to form one bonding and one antibonding orbital. If two
electrons are present, they will fill the bonding orbital to produce a
normal covalent bond. Similarly, three atoms may provide three
orbitals which combine to form one bonding and two antibonding
orbitals; two electrons may fill this bonding orbital to form a three-
center bond. Again it is pointed out that for a given atom the forma
tion of a three-center bond which uses all available orbitals results
in a lower energy state, and hence is a more stable system, than
does the formation of normal two-center bonds which leave a vacant
orbital.
Three elementary "selection rules" are given by which the type
of bonding to be expected may be determined. For a situation in
which n orbitals and m electrons are available, there are three
possible cases:
1) When m > n, the excess electrons are found as "lone
pairs" (unshared pairs).
2) When m = n, normal bonds are formed.
3) When m < n, three-center bonds are formed.
A later paper by Rundle (3 7) gives an excellent summary of
the present status of the three-center bond and touches on four-
center bonds and those of higher order. The following pair of 6 diagrams taken from this paper illustrate the similarities of normal and three-center bonds.
1 N/7(W X-^w Q-B Q- \F2 B 2^A^C)Q Vb= Q+B Q+ VTb VT In this paper, Rundle also states that the halogen bridge in compounds such as Al CI, is not a three-center bond, but instead consists of two normal covalent bonds with the bridge halogen con tributing a usually unshared pair to form the second bond. This type of bonding is contrasted with the three-center bond bridge in A1_(CH„), in the following pair of diagrams: c 3 b Me Me A study of the literature cited leads to the definite conclusion that a compound with a vacant orbital in the valence shell will exhibit no appreciable degree of stability. (Perhaps one could say that "Nature abhors an unfilled orbital. ") Compounds such as the boron halides and the alkyl boranes, which exist as monomers, must be stabilized by some mechanism which makes an extra, electron pair, donated by one of the substituent groups, available to the p orbital of the boron atom. In the case of the boron halides, the primary B-X bonds are 2 presumably "cr-type" sp hybrids. The remaining p orbital of boron perpendicular to the molecular plane can then accept an electron pair from orbitals of similar symmetry of the X atoms (23, p. 147). The extra orbital of boron is thus satisfied, and the molecule is stabilized via an intramolecular dative bond or "back coordination. " Hyperconjugation (30, p. 308-309) is a second method of stabil ization and presumably is the mechanism by which the alkyl boranes are stabilized. The double bond character imparted to the boron - carbon bond by this mechanism allows the p orbital of boron to be satisfied, and thus the molecule is stabilized. A third way in which a boron atom with three existing normal bonds may be stabilized is by coordination with an external electron donating group. Thus, complexes with ethers and amines are com mon, The methods of stabilization of electron-deficient compounds may thus be classified into three main categories: the three-center bond, internal stabilization due to electron donation from some substituent group, and external stabilization by some donor group. 8 More than one method of stabilization may be present in a compound. Although it has previously been mentioned that the prin cipal method of stabilization of the alkyl boranes is the intramolecu lar mechanism of hyperconjugation, the disproportionation of many unsymmetrically substituted boranes indicates that the formation of three-center bonds is also available to such compounds as a method of stabilization. Parsons and Ritter (27), and simultaneously Long and Dollimore (24), proposed a mechanism for disproportionation which requires the formation of a bridge structure in an intermedi ate compound. An electronic shift follows, giving products which are more symmetrical than the original borane. In the following illustra tion of this mechanism, R and R' represent different halogens, hydro gen, or small alkyl groups, and the small arrows indicate the elec tronic shifts: R R R RR , \ \ /—\ / \ 2 B- R *- BB >- B—R / / vv \ / R R R RR / / R R \ / + B >etc. I R The over-all reaction tends to proceed in such a manner that the resultant products are symmetrical. However, each reaction should be considered reversible, for the unsymmetrical members of a series often may be prepared by reaction between the symmetrical members. For example, as described in the experi mental section of this thesis, (CH.) BBr may be prepared by heating together (CH )B and BBr in 2:1 ratio. (CH,) BBr may then be isolated from the resulting mixture. At the present time several factors are known which influence the bonding of electron-deficient compounds. A brief review of these factors is instructive. Both in the case of stable dimers and in the case of the bridged transition states in the disproportionation reactions of unsymmetrical compounds, steric factors due to the size and nature of the substitu ent groups have an important bearing on the relative ease of forma tion of a bridge bond. This is well illustrated by comparing the sta bility toward disproportionation within a series of unsymmetrically substituted boranes. For example, dimethylchloroborane and di- methylethylborane are unstable toward disproportionation (27), while di-n-butylchloroborane and di-n-butylmethylborane are both stable toward disproportionation up to 110° C (1). Thus, as the size of the substituent group is increased, bridge bonds are less easily formed and the borane becomes stable toward disproportionation. The acidity of the metal in its molecular environment, in the sense proposed by G. N. Lewis (26, p. 326-329), is another impor tant determinant for the relative strength of a possible bridge bond. The Lewis acidity is dependent upon both the inherent electron 10 deficiency of the metal and the extent to which this deficiency is satisfied intramolecularly by the substituent groups. For a given set of substituents the electron deficiency, and hence the Lewis acidity, will be dependent only upon the central metal atoms. When the methyl substituted compounds of the Group III metals are exam ined, it is noted that "trimethyl aluminum" is a dimer in the gaseous state, while the trimethyl derivatives of boron, gallium, indium, and thallium are monomers (26, p. 743-744). Inspection of the elector- negativity of these metals shows that aluminum is the least electro negative (30, p. 93). Thus it is indicated that a metal with a low electronegativity is better able to form a three-center bond than is one with a high electronegativity. When the same metal is used to form different compounds, the ability of the substituent group to satisfy the electron deficiency be comes the principal factor. Thus, at room temperature chloro- dimethylborane is very unstable toward disproportionation while bromodimethylborane, although it does disproportionate, is suffi ciently stable that its physical properties may be determined. Bromine, because of its lower electronegativity (30, p. 93), is better able to satisfy the electron deficiency of the boron atom by a more favorable distribution of the electrons in the sigma bond, and hence the bromo- compound is less likely to form a bridge bond than is the chloro- compound. 11 Another factor which greatly influences bonding in halogen substituted boranes is the degree of bonding between the halogen and the boron. On the basis of the discussion of the previous paragraph, one might predict that fluorodimethylborane would be extremely unstable toward disproportionation, whereas it is actually very stable (9). This stability may be explained by the fact that in the case of the smaller halogen, fluorine, the relatively large region of p-orbital overlap allows the formation of a stronger dative bond to the boron atom. The double-bond character imparted to the B-X bond by the formation of a dative bond should lead to a shortening of the bond. The observed bond lengths in BX molecules correspond to 33% double-bond character in BF,, 22% in BC1,, 15% in BBr and 6% in BI (30, p. 318). (Thirty-three percent is, of course, the maximum possible for resonance structures in which one double bond resonates among the three B-X bonds. ) When BF, is complexed with NH , the p orbital of boron is no longer available for formation of a dative bond. In this compound, BF • NH,, the B-F bonds are found to be in good agreement with the calculated single bond distances (30, p. 319). The energy required to dissociate a B-X bond in BX is given as 15^ . 8 Kcal/mole for BF 108. 5 for BCL, and 90. 2 for BBr (44). The Lewis acid strength of the boron halides, at least with many Lewis bases, is in the order BBr, > BC1 > BF (8). This is 12 in agreement with the principle that the tt bonding would increase along the series BBr , BC1,, BF , thereby reducing the accep tor power of boron by successive amounts (8). The principles discussed in the previous two paragraphs tend to oppose each other. It appears that one factor operates to stabil ize the fluoro- substituted boranes, that the other operates to stabil ize the bromo-substituted boranes, and that neither factor predom inates sufficiently to stabilize the chloro- substituted boranes. Calculations by Rundle (36) indicate that the formation of a three-center bond is energetically favored when the three atoms involved are of about the same electronegativity. While the various factors discussed above are of great impor tance, a number of as-yet-unexplained anomalies exist. For exam- 13 pie, there is evidence, derived by mass spectral studies of C tagged trimethylborane, that methyl groups are not exchanged between molecules of trimethylborane in either the liquid or the gaseous state (47), while it would be expected that a mechanism similar to that proposed for disproportionation would make the compound readily susceptible to exchange. Also, the chlorofluoro- derivatives of boron disproportionate readily, while the stability of the methyl- fluoroboranes might lead one to predict that they would be stable toward disproportionation. One must conclude that the factors which influence the bonding 13 of electron-deficient compounds, and the degree of this influence, are as yet not known with any great degree of certainty. It is to this end that much of the present day research with boranes is ultimately directed. The investigations center around studies con cerning the existence, stability, mode of decomposition, and degree of association of variously substituted boranes. The need for assembling more data about the substituted bor anes is apparent. It is felt that the perfluoromethyl group, CF,-, should be an especially interesting substituent. Because this group is dimensionally similar to a methyl group, steric effects should be negligible in a comparison of perfluoromethyl substituted boranes with their methyl analogs. The highly electrophillic nature of the perfluoromethyl group, which has an electronegativity between that of fluorine and that of chlorine (12), makes it in this respect resem ble fluorine more closely than it does the methyl radical. One would not expect the CF - group to make electrons available to satisfy the electron deficiency of boron via hyperconjugation; thus, the total electron contribution of the group as a substituent should be only that of the normal sigma bond, unless it can enter into the formation of a three-center bond. If the CF - group can participate in the formation of three- center bonds, boranes with this substituent should be more prone to exist as dimers than should the corresponding methyl-substituted 14 boranes. Also, if three-center bond formation is favored, unsym metrical boranes with perfluoromethyl substituents should be unstable toward disproportionation. If, because of the wide difference between the electronegativity of the perfluoromethyl group and boron, this group can not participate with boron in the formation of three-center bonds, boranes with this substituent should be stable toward dispro portionation and should exist as monomers. The considerations given above have prompted the present investigation of the means of preparation of, and the properties of, perfluoromethyldimethylborane. Because of the electrophillic nature of the perfluoromethyl group, its derivatives of the Group III elements are difficult to pre pare. An unsuccessful attempt to prepare such derivatives of boron has been reported by Bartocha, Graham, and Stone (3), who treated diborane with C?F_, C_F H, C_F H , and C_FH,. They were unable to isolate any perfluoroalkyl derivatives of boron. The products from the reactions consisted of BF,, triethylborane, and mixtures of ethylfluoroboranes. A specific attempt to prepare perfluoromethyl boron compounds was reported by Lagowski and Thompson (20), who treated (CF ) Hg with boron halides and with substituted boron halides. No experi mental details are given, save that the reaction with BC1, does not proceed below 190°C, and that BF is the main product above this 15 temperature. The reactions yielded mainly BF^ with some perflu oromethyl halides. They state that infrared studies on the system ^2F5^2^S " BCI3 indicate that the reaction does not involve direct fluorination. They also state that several observations, the nature of which is not specified, indicate that B-(perfluoroalkyl) bonds are initially formed and that the resulting compounds are unstable with respect to BF?. This conjecture is apparently based on work by Goubeau and Rohwedder (17) with fluoromethyl boron difluoride, CH„FBF.,. This latter paper reports that this compound is a dimer and is unstable toward BF,. The mechanism of dimerization, as indicated by infrared studies, is said not to involve a three-center bond, but to involve association which takes place between a fluorine bonded to a carbon atom and the tervalent boron of a second molecule: F F\ ? 1 J3-C-F B-F FH CH2F Preparation of the compound CF BF has been reported by Parsons, Baker, Burg, and Juvinall (28, 2). Parsons and Baker prepared the compound by reacting CF I with KB(n-C.HQ), in ether solution and subsequently treating the filtered solution with BF . The intermediate compound, CF B(n-C .H ), was not isolated. Burg and Juvinall prepared perfluoromethyldifluoroborane by a different method. However, the product was not isolated as such, but was handled as a complex of dimethyl ether. Their 16 evidence for the compound was not as clear-cut as was that of Parsons and Baker. This is the only reported preparation of a perfluoroalkyl bor ane. The compound was found to be a monomer and was stable at room temperature when carefully kept from contact with air, mois ture, and organic materials. When catalyzed by a trace of air or by certain organic materials, the substance decomposed, yielding BF, and a white solid which was presumed to be a (CF ) polymer. With this compound, it is not possible to ascertain if decompo sition proceeded via a bridged compound or by direct splitting out of (CF ) groups, as the products would be the same in either case. Of the possible methods of preparation of perfluoromethyl- dimethylborane, it was decided to investigate primarily reactions of the Grignard reagent, CF,MgI, and the mercurials CF Hgl and (CF,) Hg with dimethylbromoborane. The mercurials were chosen in spite of their apparent failure to give perfluoromethyl compounds as reported by Lagowski and Thompson (20). The experimental information reported by these workers was very scanty, and it was felt that these compounds might still prove useful under the right experimental conditions. During the course of the investigation, other reactions related to the major problem were cursorily investigated and will be reported in the section devoted to discussion and conclusions. 17 II. EXPERIMENTAL Apparatus and Equipment The major piece of equipment used in this investigation was a vacuum apparatus, or "line, " similar to those described by Stock (46, p. 173-205) and Sanderson (38) for the handling of volatile com pounds. Although a line with mercury float valves, such as described in the references given, was used for a portion of the work, most of the experimentation was carried out using a similar line which was equipped with conventional high-vacuum stopcocks. Unless otherwise noted, all reactions and experiments were performed in this vacuum line or in sealed reaction vessels consist ing of glass bulbs of appropriate size, from about 25 milliliters to approximately two liters, to which were sealed two or more break seals. Standard taper joints were then sealed to the break seal tubes, and the reaction vessel thus attached to the line. After evacuation of this connection, the vessel could be opened to the line by breaking the break seal with a glass-enclosed iron hammer which had previ ously been placed in the connecting tube above the break seal. The connecting tube between the break seal and the vessel itself was usually constricted so that it might easily be sealed with a hand torch after the reactants had been introduced. A typical reaction vessel is illustrated in Figure 1. FIGURE I Reaction / Storage Vessel ?i Pi NOT TO SCALE 19 Storage of volatile compounds was in glass bulbs closed either with break seals or with conventional high-vacuum stopcocks. Special apparatus for reactions and for purification of reagents were constructed on occasion and are described in connection with the experiments where they were used. Low temperature baths for fractional condensation and for the measurement, of vapor pressures consisted of liquid nitrogen, dry ice - acetone mixtures, and solid - liquid equilibrium mixtures, or "slushes, " of appropriate solvents. The baths commonly used, with their nominal temperatures, are given in Table 1. When used for fractional condensations, the baths were usually assumed to be at their nominal temperatures; when used for vapor pressure measure ments, these temperatures were more precisely determined by the use of vapor pressure thermometers. The composition and temper ature ranges of the four thermometers used are given in Table 2. A Perkin-Elmer Model 21 Infrared Spectrophotometer and a Beckmann Model IR-7 Infrared Spectrophotometer were used to obtain infrared spectra. Gaseous samples were contained in a ten- centimeter gas cell with polished sodium chloride windows. Reagents The common laboratory chemicals employed were stock reagent grade materials, used without further purification. 20 Table 1. Constant temperature baths Composition Nominal temperature ° C Water - ice 0 Carbon tetrachloride slush -23. 0 Bromobenzene slush -30. 7 Chlorobenzene slush -45. 1 Chloroform slush -63. 0 Dry ice - acetone -80 Toluene slush -95.0 Carbon disulfide slush -111. 8 Ethyl bromide slush -119.0 Methylcyclohexane slush -126. 7 Pentane slush (mixture of isomers) -140 to -150 Isopentane slush -160.0 Liquid nitrogen -196.0 Table 2. Vapor pressure thermometers Composition Usable temperature range ° C Sulfur dioxide -10 to -57 Ammonia -33 to -77 Carbon dioxide -78 to -110 Ethylene 103 to -150 21 Volatile reagents were purified by fractional condensation. In many cases, the fractionation was considered to yield sufficiently pure materials, and no further check on the purity was made. When deemed necessary, the purity was checked by vapor pressure meas urements. Two methods were employed. In the first, the vapor pressure at a measured temperature was compared with literature values. In the second, the vapor was allowed to expand into a larger volume after an initial measurement of the vapor pressure. A change in vapor pressure would serve to indicate a mixture, while a con stant vapor pressure would serve to indicate a pure substance, since the possibility of an azeotropic mixture is considered remote. In those cases where standards of adequate purity were available, infrared spectra were also used as checks on purity. Special reagents and reagents receiving special purification techniques are given below. Boron Trifluoride, Hydrogen Chloride, and Trimethylamine: Tank reagents from Matheson Co. were purified by fractional con densation. Nitrogen: Tank nitrogen (National Cylinder Gas Co. ) was dried by passing through a CaCl? - BaO - P-,0 column and used to pro vide an inert atmosphere where needed. Tetrahydropyran (THP): Eastman commercial grade THP was distilled, using an Oldershaw column, with the fraction that distilled 22 over between 82 °C and 87° C being retained. The entire batch was used as a solvent for the preparation of a dilute solution of ethyl magnesium iodide. The THP was then distilled as needed from this "keeping" Grignard. Magnesium: For the attempted preparation of CF Mgl a sample of high-purity sublimed magnesium crystals was kindly furnished gratis by Dow Chemical Company, Midland, Michigan. The purity was stated to be in excess of 99. 9%. Boron Trifluoride - di-n-butyl Ether Complex: The complex was prepared by bubbling tank BF into reagent grade di-n-butyl ether which had been dried with metallic sodium. Trimethylborane: Trimethylborane was prepared as suggested by Brown (7). Boron trifluoride - di-n-butyl ether complex, in dibutyl ether solution, was added dropwise to freshly prepared methyl magnesium iodide reagent in a nitrogen atmosphere. A stream of dry nitrogen was maintained to sweep the product into the line through a reflux condenser. The borane was purified by frac tional condensation from -80° C through -112° through -140° into -196° . The -140° fraction was transferred to a storage bulb built into the line and kept at -80° until used. Diborane: Diborane was prepared after the method of Schlesinger and Brown (41), using BC1 and LiBH . The diborane was purified before use by fractional condensation through -140° 23 into -180° . Boron Tribromide: Boron tribromide was prepared after the method of Gamble (15). Prior to use, the reagent was purified by fractional condensation through -30° into -80°. Dimethylbromoborane: Dimethylbromoborane was prepared by a slight modification of the method used by Schabacher and Goubeau (39). A two liter reaction vessel was charged with 50. 9 millimoles of trimethylborane, 25. 7 millimoles of boron tribromide, and 0.4 millimoles of diborane as a catalyst. The vessel was then sealed, allowed to warm to room temperature and subsequently heated for 31 hours on a steam bath. The sealed vessel, now containing a mix ture of the various methyl - bromo - boranes, was stored until a sample of Me BBr was needed. At that time, the reaction vessel was attached to the line through a break seal connecting tube and the con tents separated by fractional condensation. Me B passes through a -80 °C trap rapidly, while Me_BBr passes through very slowly. MeBBr and BBr are completely stopped by a -80° trap. The dimethylbromoborane was isolated, therefore, by allowing the con tents of the reaction vessel to pass through -80° into -196° for periods of time ranging from half an hour to two hours. The -196° trap contained all of the Me B and some Me,BBr, while the -80° trap contained all of the MeBBr , all of the BBr , and some of the Me,BBr. A second, separate -196° fraction, which consisted 24 almost entirely of Me.BBr, was then collected from the mixture which remained in the -80° trap. The process was repeated on this fraction until vapor pressure measurements gave satisfactory agreement with literature values. The components which had been removed by fractionation, as well as the Me BBr in excess beyond that which was needed for the experiment in progress, were returned to the storage bulb, which was then resealed for storage. Trifluoroiodomethane: Trifluoroiodomethane was prepared by the method of Henne and Finnegan (19). The CF I was purified by fractional condensation through -95 ° C, through -112°, through -127°, into -196° with the -127° cut being retained. Preparations and Reactions Reaction of CF I with Magnesium: The method followed was essentially that of Haszeldine (13), who has reported the formation of the Grignard reagent CF Mgl, although in very low yields. Di-n-butyl ether was distilled from CaH , using an Oldershaw column. The sample of sublimed magnesium crystals furnished by Dow Chemical Company was drilled while immersed in this purified ether in order to provide uncontaminated high-purity magnesium turnings. About 100 milliliters of THP were distilled from the "keeping" Grignard into a 200 milliliter three-necked flask. The flask was 25 connected via a reflux condenser to the vacuum line. One gram of iodine crystals and 1.30 grams of the special magnesium turnings were added to the flask. The flask was then fitted with a stirrer and flushed with dry nitrogen. The iodine and magnesium turnings were stirred together for about 15 minutes. A solution of 9. 69 millimoles of CF,I (1.88 gm. ) in purified THP was added dropwise. The con tents of the flask were then stirred for 11 hours under gentle reflux (b. p. of THP is 81 °C). No apparent reaction had occurred at the end of this time. Water was added to the mixture and allowed to remain for six hours. The contents of the reaction flask were then fractionated through -80° into -180° . A small amount of material which appeared to be CF,H was obtained, but this accounted for only about three percent of the original CF I. The remainder had apparently under gone no reaction. Because of this very low yield, in spite of rigor ous precautions, and because the mercurials were yet to be investi gated, no further attempts to prepare this Grignard reagent were made. Preparation of Trifluoromethylmercuric Iodide: The method of preparation of this mercurial was suggested by that of Emeleus and Haszeldine (13). A cylindrical pyrex pot of about 1. 3 liters capacity was fitted with a large standard taper joint at the top and with a side arm carrying a stopcock and a small standard taper joint. A Hannova 26 water-cooled mercury arc ultraviolet irradiator was fitted into the pot at the large top joint. Fifty milliliters of Hg and 38. 6 millimoles of CF I were placed in the pot and irradiated for 30 hours. The appearance of considerable quantities of iodides of mercury gave evidence that some reaction had occurred, but only a very little CF Hgl could be recovered from the contents of the pot. The presence of con siderable quantities of non-condensable gases in the pot after irradi ation indicated that the greased joint did not hold up well at the tem peratures involved. A second run was irradiated for seven days, but yielded similar results. Preparation of bis(trifluoromethyl) mercury: This mercurial was prepared by a modification of the method suggested by Emeleus and Haszeldine (14). Trifluoroiodomethane was shaken with dilute cadmium amalgam (about 15 grams of Cd per 100 milliliters of Hg) at room temperature and protected from light. The first runs were made with the CF I pressure below one atmosphere. A 500 milliliter round-bottom flask was equipped with an inner 24/40 standard taper joint. This flask could be capped with a matching outer joint sealed to a stopcock which in turn was sealed to an inner 14/35 joint, by which the assembly could be attached to the vacuum line. In a typical run, 16 milliliters of the amalgam and two grams 27 (10 millimoles) of CF I were frozen into the evacuated flask. The flask and its contents were placed in a mechanical shaker and covered with a cardboard box to protect them from light. The flask was shaken for about 120 hours. At the end of this time the flask con tained, in addition to the amalgam, a considerable quantity of gray powder. The flask was connected to the line and the volatile contents removed. This volatile material consisted mainly of unreacted CF I together with a smaller quantity of a more volatile component. In most runs, about 75 percent of the trifluoroiodomethane was con sumed. (One run yielded a rather large quantity of the more vola tile component - about 13 millimoles yeild from about 20 millimoles of CF,I used as starting material. This compound was not identi fied but was found to be neither CF, norC^F,. A vapor pressure - 4 2 6 f f temperature curve and an infrared spectrum were obtained, but characterization was not pursued further. ) The non-volatile contents of the flask were extracted with dried diethyl ether. The extract was filtered, and the ether removed by vacuum evaporation at 0°C. The residue was sublimed at room temperature onto a cold finger cooled with dry ice - acetone mixture. Several problems were encountered. It was difficult to prevent the extract from becoming contaminated with grease from the joint of the reaction flask, and this contamination prevented the proper 28 evaporation of the extract. The removal of the ether was accom panied by a considerable loss of product. It was difficult to remove the last traces of ether from the residue, and these traces of ether then interfered with proper sublimation. In spite of these difficulties, a small amount of the mercurial was obtained. Later runs were made using a higher pressure of CF,I and afforded much better results. Sealed reaction tubes equipped with side arms closed with break seals were used for reaction vessels. These tubes had volumes of approximately 50 milliliters. The quan tity of CF,I used was sufficient to assure that some of this compound would be in the liquid phase at the start of the reaction. Extrapola tion of the vapor pressure data previously obtained for CF I indi cated that the pressure exerted by the trifluoroiodomethane at room temperature was five to six atmospheres. In a typical high-pressure run, 11 grams of cadmium was dis solved in 25 milliliters of mercury and the amalgam placed in a reac tion tube. The tube was attached to the line and evacuated. Fifty millimoles (9. 8 grams) of CF,I was then frozen into it. The tube was sealed off, allowed to warm to room temperature, and placed in the shaker. A large cardboard box was used to cover the entire shaker to protect the tube from light and also to provide some meas ure of protection in case the tube failed to withstand the pressure. 29 2 Shaking was continued for several weeks. After being shaken, the tube contained a considerable quantity of white crystalline material besides the familiar gray powder and the excess amalgam. No liquid CF I remained. The tube was con nected to the line and opened. A trace of non-condensible gas and 0. 8 millimoles of unreacted CF I were present. The contents of the tube were extracted several times with dried diethyl ether. The extract was placed in a bulb fitted with a 19/38 inner joint at the neck. To this was fitted a "U" tube through a stop cock. The free end of the "U" tube was connected to the line via a 19/38 inner joint. This assembly was then repeatedly flushed with dry nitrogen. An ice - salt bath was placed around the "U" tube, the system was evacuated, and the ether was slowly removed to the line. The ice - salt bath effectively condensed all of the desired product while allowing passage of the ether. A stopcock between the "U" tube and the bulb proved necessary to prevent too great an accumulation of ether in the "U" tube, as at the temperature of the ice - salt bath the ether is removed rather slowly. After complete removal of the ether, the bulb containing the residue was connected to a sublimer. This piece of apparatus con sisted of two 19/38 outer joints sealed together in the form of a 2 In practice, shaking was continued on each run until the con tents were needed for an experiment. The shortest time any run was shaken was six days. 30 flattened, inverted, "U" tube. From the center of the inverted "U" a vertical side arm with a stopcock projected and terminated in a 19/38 inner joint. The bulb containing the residue was connected to one end of the sublimer, and a receiving tube was connected to the other end. The receiving tube consisted simply of a 19/38 inner joint sealed at the end. The assembled sublimer was evacuated and the receiver im mersed in a dry ice - acetone bath. Sublimation proceeded at room temperature for several days. A good crop of white crystals col lected in the receiver. Later, a second and a third crop were obtained. Reaction of Dimethylbromoborane with Bis(trifluoromethyl)mer- cury: A typical experiment will be described. The receiving tube from the sublimer containing an unmeasured quantity of the sublimed mercurial was placed on the line, repeatedly flushed with dry nitro gen, and evacuated. The flushing and evacuation were performed rapidly to avoid excessive loss of the mercurial. Dimethylbromo borane was withdrawn from the storage bulb and purified as described earlier. A 2. 31 millimole portion of the borane was frozen into the tube containing the mercurial. A -80° C bath was placed around the tube and left in position for 1 7 hours. There was no visible change in the contents of the tube at the end of this period. The -80° bath was 31 replaced with a -40° bath. After an hour and a half, the appearance of the contents of the reaction tube had altered, indicating some reac tion had occurred. The tube was held at -40° for a total of ten hours, The tube was then maintained at -80° until further work could be done. The volatile contents of the reaction tube were removed to the fractionation train of the line and fractionated from -80° C through -95°, -112°, -127° into -196° . Minor amounts of material were found in the -95° and -127° fractions, 0.5 millimoles in the -80° fraction, 1.5 millimoles in the -196° fraction, and 0. 2 millimoles in the -112° fraction. The -80° fraction contained any unreacted dimethylbromobor ane as well as any low-volatility products from side reactions and from decomposition of the expected trifluoromethyldimethylborane. High-volatility products from side reactions and from decomposi tion were contained in the -196° fraction. It was expected that the anticipated product, CF.Me.B, would be collected in the intermedi ate fractions. A vapor pressure curve was obtained for the material in the -112° fraction (Figure 2). The molecular weight of this fraction calculated from vapor density measurements was 112. The the oretical molecular weight for trifluoromethyldimethylborane is 109.9. 1000- 900- FIGURE 800- 700- Vapor Pressure Vs. ^~~ 600- CF36Me2 500- 400- 300-J 200- Pv, mm 100- 90- 60- 70- 60- 50- 40- 30- 20- 10- 3.2 3,6 4.0 4.4 4.8 5.2 5.6 6.0 1000/ Temp.(°K) 33 After obtaining the molecular weight, of necessity a room tem perature operation, the vapor pressure was rechecked. The values obtained were considerably different from those originally obtained, indicating a decrease of the purity of the material. Upon refraction- ation, essentially all of the volatile material passed through -127°C into -196°. This material was combined with the remaining frac tions from the original separation, exclusive of the -80° fraction, which consisted of unreacted Me BBr. The mixture was stored and later characterized. Characterization of the Decomposition Products of Perfluoro- methyldimethylborane: The total quantity of the decomposition mix ture reported above was 1. 7 millimoles. The mixture was separated by fractional condensation into a fraction of low volatility (fraction 1, stopped at -127°), a fraction of intermediate volatility (fraction2, stopped at -145°), and a fraction of high volatility (fraction 3, stopped at -196°). Fraction 1 accounted for approximately 19% of the mix ture, fraction 2 for 74%, and fraction 3 for 7%. Fraction 1 was further separated into two fractions by frac tional condensation. The separation was not at all clean. The infra red spectrum of each fraction was obtained (Figures 3 and 4). The two spectra are similar, but by no means identical. The structure common to both spectra is in good agreement with the published spec trum of trimethylborane (21, 45). Vapor pressure measurements indicated SAMPLE= One fraction of impure Me3 B from the decomposition of rl6 15 10cm Cell P = Not recorded Reference = Air Beckman IR7 FIGURE 3 SAMPLE = A second fraction of impure Me3 B from the decomposition of Me2 BCF3 MICRONS 3 3.5 / 100 8 2 13 14 15 16 1 1 90 70 %T • 50 — 30 — 10 3B 36 32 28 V16 15 10cm Cel P = Not recorded Reference =Air Beckman IR7 36 low purity for both fractions. The small quantities of these frac tions, coupled with their evident impurity, precluded the determina tion of molecular weights. Fraction 2, comprising most of the original mixture, was shown to be dimethylfluoroborane, Me BF. The infrared spectrum ob tained from this fraction (Figure 5), after it had been further puri fied by fractional condensation, is in excellent agreement with the spectrum of Me^BF reported by Becher (6). The vapor pressure curve obtained from this fraction (Figure 6) agrees very well with that reported for Me BF by Burg (9). The molecular weight calcu lated from vapor density measurements was 62. 3. The theoretical molecular weight of Me BF is 59. 9. Fraction 3 was not identified. Vapor density measurements showed low purity. The infrared spectrum of this fraction (Figure 7) showed little agreement with published spectra of methyl fluoro boranes, fluoromethanes, or related compounds. There was no indication of any trace of either BF or MeBF in any of the fractions. Reaction of CF I with BF in the Presence of Hg: An attempt was made to prepare CF BF by the direct reaction of CF.,1 with BF in the presence of mercury. In a typical run, 2. 24 millimoles of CF_I and 5. 20 millimoles of BF were frozen into a reaction tube containing approximately 20 milliliters of mercury. The reaction SAMPLE = Me2 BF from decomposition of Me2 BCF3 32 28 100 CM 10cm Cell A= 29% Screen P= 19.5 mm B= 47% Screen Reference = A ir Beckman IR7 FIGURE 5 1000 FIGURE 6 Vapor Pressure Vs. j000 500:- EMP. Mea BF O = Experimenta ^ s Values from Burg (9) 100 50 Pv, mm 10- 4.50 4.75 500 5.25 5.50 5.75 600 1000/Temp. (°K) 6.25 SAMPLE = Most volatile fraction from Me2 BCF3 decomposition 100 90 38 36 32 28 V|6 100 CM 10cm Cell P= 29.7 Reference = Air Beckman IR7 FIGURE 7 40 tube was similar to the tubes used to prepare bis(trifluoromethyl)mer- cury. The tube was allowed to warm, to room temperature and was shaken in the absence of light for several days. At the end of this period, there was little visual evidence of reaction. The tube was then exposed to room light and shaking was continued. Some reac tion occurred, as evidenced by the formation of an orange coating of mercury iodides on the inside of the tube. The tube was attached to the line and the volatile contents were fractionated through -127° C into -196°. The -196° fraction consist ed mainly of BF, with a little CF I. Approximately 90% (4. 6 milli moles) of the original BF was recovered unchanged. The -127° fraction was treated with trimethylamine. The reaction produced a white solid and some gaseous material. The white solid did not react with HC1 and was discarded. Repeated fractionation of the gaseous material yielded a small amount of CF I and a substance which collected at - 70 ° . A vapor pressure curve (Figure 8) was obtained. The molecular weight calculated from vapor density measurements was 131. 6. The above material was treated with HC1, resulting in the formation of a solid and a gas. The solid was discarded. The vapor pressure of the gas checked with that of the original material. An infrared spectrum was obtained. The material was refractionated after the infrared analysis, and a second component, of greater I000-, FIGURE 8 Vapor Pressure Vs. ^^ 500-- Product of the reaction between CF3 I Me3 N too Pv, mm 10 3.0 32 3.4 3.6 3.8 4.0 42 4.4 1000/Temp. (°K) 42 volatility, was now found to be present. The molecular weight of this new component calculated from vapor density measurements was 144. 2. Pyrolysis of Lithium Borohydride: Schlesinger and Brown (40) have reported that LiBH , upon being heated to 275 - 280 ° C, loses about 50% of its hydrogen. A preliminary investigation of the pyroly sis and of the reaction of CF I with the pyrolysis product was car ried out. A bulb with a long, curved neck terminating in a standard taper joint was constructed. A small open-coil furnace controlled by a Powerstat was used to heat the bulb. A chromel-alumel thermo couple attached to the side of the bulb was used to obtain a rather rough indication of the temperature of the contents of the bulb. The temperature of the heating coil, and hence the maximum temperature to which the contents of the bulb could be exposed, was monitored by a thermometer resting upon the coil. The bulb, the lower part of the neck, the thermocouple and the thermometer were fitted into an alundum shield to provide more even heating. The evolved hydrogen was measured by forcing it into a calibrated volume with a Toepler pump. The lithium borohydride was stock material from Metal Hydrides, Inc. , and was stated to be 92% pure. In a typical run, 6.5 7 millimoles of LiBH was weighed out and loaded into the bulb, which was then attached to the line and 43 evacuated. The furnace, shield, thermocouple, and thermometer were placed in position and the temperature of the bulb was raised in a stepwise manner. The pattern of hydrogen evolution was distinctive. When the sample was heated to some particular temperature, hydrogen was evolved rapidly at first. The rate of evolution slowly decreased as the sample was maintained at this temperature. When little hydrogen was evolving, the temperature was raised. Again the hydrogen evo lution was initially rapid, and again the rate slowly decreased. This behavior was observed each time the temperature was increased. When the temperature had been raised to 200 ° C, 3. 77 milli moles of hydrogen had been evolved. This is well over 50% of the total amount that was expected to have been evolved at 275 ° C. When the temperature had been raised to 275 ° , 8. 63 millimoles, or 65. 6% of the total hydrogen of the sample, had been evolved. The residue after pyrolysis was inhomogenous in appearance and consisted of a brownish material with a burned appearance. A dark gray sublim ate appeared on the neck of the vessel up to the top of the alundum shield, and a black film appeared on the neck about three inches above the top of the shield. CF„I was introduced and allowed to remain in contact with the residue for several hours at room temperature. The volatile mater ial was then withdrawn and checked for quantity and vapor pressure. 44 No change in vapor pressure or quantity was observed. The CF I was then returned to the pot and the temperature increased. After several hours at the new temperature, it was again removed and checked. This process was repeated several times at gradually increased temperatures. The vapor pressure of the volatile mater ial started to increase after the mixture had been cooked for five hours at a coil temperature of 105° C. The process of slowly chang ing the temperature and checking the volatile material was continued. After heating the mixture for two hours at a coil temperature of 230° (corresponding to a pot temperature of approximately 130°), the vapor pressure had risen to 24 mm. at -95. 1 ° compared to a vapor pressure for CF I at the same temperature of 9. 7 mm. The volatile material was then fractionated through -127° into -196°, and the -127° fraction was returned to the pot. This process was repeated several times at this temperature. The recovery of volatile mater ial was about 90%, based on the quantity of CF I originally intro duced. The accumulated -196° fractions were combined and subjected to extensive fractional condensation. The separations were not at all clear cut. One fraction of reasonable purity was obtained. The vapor pressure curve (Figure 9) and the infrared spectrum (Figure 10) of the material were obtained. Vapor density measurements were taken which indicated a molecular weight of approximately 50. FIGURE Vapor Pressure Vs -i0Q0- TEMP. Product of the reaction between CF3 1 and pyrolyzed Li B H4 5-75 6.00 6.25 7.00 1000 / Temp. (°K) SAM PL E = Product of the reaction between CF3 I ond pyrolyzed Li BH4 MICRONS 100 4 5 6 7 '0 ',' 12 13 14 a^n 90 1 I I TT"1 !C0 10cm Cell P=45mm £ 6mm Reference = Air Perkin - Elmer FIGURE 10 47 This value may be in error by 10%. A second fraction was obtained which appeared to be impure BF3. For another run, the borohydride was handled in a different manner. A plastic bag was filled with dry nitrogen and all handling of the LiBH was carried out in this protective atmosphere. The evolution of hydrogen followed the same pattern as that of the previ ous run, but the pyrolysis residue was a dirty white with only a trace of brown. Since this area of investigation was not directly related to the major problem, and since the low molecular weight of the volatile material obtained from the reaction of the pyrolysis residue with CF_I precluded any CF. compound, the investigation of this system was not pursued further. 48 III. DISCUSSION AND CONCLUSIONS The first method to be investigated for the preparation of trifluoromethyldimethylborane was a synthesis utilizing the Grignard reagent CF Mgl. It was proposed that this reagent would react with dimethylbromoborane according to the equation CF-Mgl + Me BBr -* Me_BCF + MgBrl. The first step in the research was thus the attempt to prepare the Grignard reagent itself. Although great care was exercised in the selection and handling of the reactants, attempts to prepare the Grignard reagent met with little success. Because of the extremely low yields of this starting material, it was decided that other routes to the preparation of the borane should be investigated before attempts were made to prepare the Grignard reagent on a scale large enough to be of value. It was then proposed that the mercurials of CF Hgl and (CF ) Hg would react with dimethylbromoborane in a manner simi lar to that proposed for the Grignard reagent. The first mercurial to be investigated was trifluoromethyl- mercuric iodide. Attempts were made to prepare this compound by the batch irradiation of CF.,1 and mercury. The yields of this reagent were also very low. It became apparent that the ultraviolet radiation caused the mercurial to decompose almost as fast as it was formed. To prepare trifluoromethylmercuric iodide in quantity 49 would require a system in which the product could be continuously removed from the influence of the ultraviolet light. As the approach using bis(trifluoromethyl)mercury seemed as promising as that using trifluoromethylmercuric iodide, further work on CF Hgl was aban doned, although a fairly simple closed system in which mercury vapor and CF I could be continuously mixed and irradiated and in which the product could be continuously removed was designed. Bis(trifluoromethyl)mercury was prepared by allowing CF I to react with cadmium amalgam in the absence of light. The first trials, conducted with low pressures of CF I, gave low yields but indicated the inherent feasibility of the method. When the pressure of CF,I was increased to about five atmospheres, good yields of the mercurial were obtained. Improvements in methods of separation and purification allowed the production of considerable quantities of (CF3)2Hg. Trifluoromethyldimethylborane was then prepared by the reac tion of dimethylbromoborane with bis(trifluoromethyl)mercury at low temperatures according to the equation: (CF ) Hg + 2Me_BBr -* 2Me2BCF3 HgBr^ The borane thus obtained yielded an experimental molecular weight of 112 , The calculated molecular weight for trifluoromethyl dimethylborane is 109. 9. The logarithm of the vapor pressure over the range of 15 mm. to 350 mm. is shown plotted versus the 50 reciprocal of the temperature in Figure 2. Extrapolation of the straight central portion of the curve indicates a boiling point of approximately 47° C. Trifluoromethyldimethylborane was found to be very unstable. At temperatures sufficiently high to allow the starting materials to react (ca. -40°), the resulting product was extensively decomposed. In the best trial, only 0. 2 millimoles of product and 1. 5 millimoles of decomposition products were isolated from the mixture which resulted from the reaction of 1. 8 millimoles of Me BBr. While the measurement of pressure and volume in preparation for the determin ation of the molecular weight could be carried out with sufficient rapidity to produce an homogenous sample, the material in the molecular weight bulb after the weighing operation was essentially all decomposition product. Because of this instability, the vapor pressure data must be considered to be of very limited accuracy. No infrared data could be obtained from this short-lived borane. The fact that the major decomposition product of trifluoro methyldimethylborane is Me BF allows certain inferences to be drawn concerning the mechanism of decomposition and concerning the ability of the CF - group to participate in three-center bonding. The ability of the methyl group to participate in three-center bonds is well established. If the perfluoromethyl group could also partici pate in such bonds, then disproportionation of 51 trifluoromethyldimethylborane would be expected to occur via the formation of transition bridged structures as indicated: Me CF, CF, \ / ^ / 3 2 Me BCF -»- B} 4/b' —•»• Me B + MeB(CF ) Me Me Me ' »-etc. The decomposition of the disproportionation products of such transi tion structures would of necessity yield MeBF or BF as compo nents of the decomposition mixture. As no trace of either of these compounds is actually produced in the decomposition of trifluoro methyldimethylborane, the conclusion may be drawn that the CF - group is unable to participate in three-center bonds, at least in the present molecular environment and at the temperature at which de composition of the borane occurs. The decomposition of trifluoromethyldimethylborane to yield Me^BF could conceivably proceed by either an intramolecular proc ess involving the expulsion of a • CF • fragment or by an inter- molecular process involving association between two or more mol ecules of the borane followed by an electronic shift. The intramolecular process is supported to some degree by Mahler's production of • CF • fragments by the pyrolysis of (CF3)2PF2(25). The intermolecular process, in turn, is supported by the previously cited work of Goubeau and Rohwedder (17). The com pound studied by these investigators, F BCH.F, was found to be 52 a dimer and to have the structure: Fv HF \ )B-C-FI B-CFH0I / I I 2 FH F This compound was unstable, forming BF, upon decomposition. While trifluoromethyldimethylborane was found to be monomeric, the formation of a transitional dimer of the form Me F ,Me \ ;B-C-F—I B—Me/ Me/ FI \ CF could account for the observed production of Me BF as the product of its decomposition. While this latter mode is believed to be the basic mechanism of the decomposition, the detailed path of the reaction, for example just how the fluorocarbon groups are split off, is not known. How ever, the appearance of a white deposit on the reaction and storage bulbs and the absence of any fluorocarbons in the decomposition mix ture strongly indicate that a fluorocarbon polymer is formed. This series of experiments utilized almost exclusively a vacuum system employing conventional stopcocks. There is a dis tinct possibility that in a system employing mercury float valves, 53 or some other means of preventing all contact with organic mater ials, the trifluoromethyldimethylborane would be found to be much more stable. Parsons and Baker (28) found the rather similar com pound CF3BF2 to be quite stable in a system isolated from organic materials but very unstable when in contact with certain organic materials. In the course of this investigation, certain phenomena were observed which might well be worthy of further study. One of these was the apparent reaction of CF I with Me N to form a stable gas eous compound. Another was the appearance of some foreign mater ial in Me BF when this compound was allowed to stand in glass ves sels closed with a stopcock. The appearance of this contaminant was clearly evidenced by the appearance of a very strong band at 723 cm in the infrared spectrum of Me BF. It could be removed by careful fractionation through -145 °C into -196°. After a period of a few days, the Me BF thus purified would again show evidence of the same material. It is possible that this substance is SiF pro- 4 ^ duced by the reaction of the borane with glass. It is interesting to note that this material was not present in the original mixture of decomposition products of trifluoromethyldimethylborane. This would lead one to believe that the substance formed reacted in turn with other components of the mixture, and may be the source of the unidentified, highly volatile fraction obtained from the 54 decomposition mixture. As a result of preliminary investigations carried out in the course of this study, certain other areas appear worthy of additional research. While the direct reaction of CF I with BF in the pres ence of mercury failed, it is felt that this system should be more carefully investigated. Substition of cadmium amalgam for mercury should prove interesting, as should substitution of other boron halides and alkyl boron halides for BF . The reaction of CF,I with the pyrolysis product of LiBH also failed to give results. The trial, however, was conducted at low pressure. Perhaps reaction at higher pressures or in the pres ence of a solvent might prove of value. The nature of the pyrolysis product itself should be interesting. A flow system might be designed which would allow the more efficient production of Me BCF . The infrared spectrum of the solid borane should greatly assist in identifying the compound more positively. A low-temperature cell would be needed to obtain such a spectrum. 55 SUMMARY The unstable compound trifluoromethyldimethylborane was prepared by the reaction of dimethylbromoborane with bis(trifluoro- methyl)mercury. This borane decomposes rapidly to yield dimethyl- fluoroborane and a fluorocarbon polymer. This evidence indicates that the trifluoromethyl group cannot participate in forming three- center bonds at low temperatures. 56 BIBLIOGRAPHY 1. Auten, Robert W. and Charles A. Kraus. Studies relating to boron. V. Chemistry of the dibutylboron group. Journal of the American Chemical Society 74:3398-3401. 1952. 2. Baker, Everett Duane. Preparation and properties of trifluoro- methyl-boranes. Master's thesis. Corvallis, Oregon State College, 1951. 36 numb, leaves. 3. Bartocha, Bodo, W.A.G. Graham and F. G. A. Stone. Hydride- transfer reactions of borine--I. Synthesis of boron-carbon bonds from the fluoroethylenes. Journal of Inorganic and Nuclear Chemistry 6:1 19-129. 1958. 4. Bauer, S.H. The structure of diborane. Journal of the American Chemical Society 59:1096-1103. 1937. 5. Structures and physical properties of the hydrides of boron and their derivatives. Chemical Reviews 31:43-75. 1942. 6. Becher, H. J. Die Schwingungsspektren der Methylborfluoride. Zeitschrift fur anorganische und allgemeine Chemie 291:151-159. 1957. 7. Brown, Herbert C. Studies in stereochemistry. V. The effect of F-strain on the relative base strengths of ammonia and tri- methylamine. Journal of the American Chemical Society 67: 374-378. 1945. 8. Brown, Herbert C. and Robert R. Holmes. The heats of reac tion of pyridine and nitrobenzene with boron trifluoride, trichlor ide and tribromide; the relative acceptor properties of the boron halides. Journal of the American Chemical Society 78:21 73-21 76. 1956. 9. Burg, Anton B. Methylboric acid and its anhydride; the methyl- boron fluorides. Journal of the American Chemical Society 62: 2228-2234. 1940. 10. Dilthey, Walter. Uber die Konstitution des Wass ers. Zeitschrift fur angewandte Chemie 34:596. 1921. 5 7 11. Eberhardt, W.H. , Bryce Crawford, Jr. and William N. Lipscomb. The valence structure of the boron hydrides. Journal of Chemical Physics 22:989-1001. 1954. 12. Emeleus, H.J. Organometallic compounds containing the fluorocarbon radical. Journal of the Chemical Society, 1954, p. 2979-2986. 13. Emeleus, H.J. and R. N. Haszeldine. Organometallic fluorine compounds. Part I. The synthesis of trifluoromethyl and pentafluoromethyl mercurials. Journal of the Chemical Society, 1949, p. 2948-2953. 14- Organometallic fluorine compounds. Part II. The synthesis of bistrifluoromethylmercury. Journal of the Chemical Society, 1949, p. 2953-2956. 15. Gamble, E. Lee, Paul Gilmont and John F. Stiff. The reaction of boron fluoride with aluminum chloride or bromide. Journal of the American Chemical Society 62:125 7-1258. 1940. 16. Glasstone, Samuel. Textbook of physical chemistry, 2d ed. Lancaster, Pennsylvania, Lancaster Press, 1946. 1320 p. 17. Goubeau, Josef and K. H. Rohwedder. Die Reaktion von Diazomethan mit Bortriflorid in der Gasphase. Justus Leibigs Annalen der Chemie 604:168-178. 1957. 18. Hedberg, Kenneth and Verner Schomaker. A reinvestigation of the structures of diborane and ethane by electron diffraction. Journal of the American Chemical Society 73:148 2-148 7. 1951. 19. Henne, Albert L. and William G. Finnegan. The degradation of silver trifluoroacetate to trifluoroidomethane. Journal of the American Chemical Society 72:3806-380 7. 1950. 20. Lagowski, J.J. and P. G. Thompson. Perfluoroalkyl boron compounds. Proceedings of the Chemical Society, 1959, p. 301-302. 21. Lehman, W. J. , CO. Wilson, Jr. and I. Shapiro. Character istic frequencies of trimethylborane. Journal of Chemical Physics 28:777-784. 1958. 58 22. Lipscomb, William N. Structures of the boron hydrides. Journal of Chemical Physics 22:985-988. 1954. 23. Recent studies of the boron hydrides. Advances in Inorganic Chemistry and Radiochemistry 1:117- 156. 1959. 24. Long, L.H. and D. Dollimore. The chemistry of boron. IV. The boron chlorobromides. Journal of the Chemical Society, 1954, p. 4457-4461. 25. Mahler, Walter. Double addition of a carbene to an acetylene. Journal of the American Chemical Society 84:4600-4601. 1962. 26. Moeller, Therald. Inorganic chemistry. New York, John Wiley and Sons, 1952. 966 p. 27. Parsons, Theran D. andD.M. Ritter. Alkenylmethylborines. Journal of the American Chemical Society 76:1710. 1954. 28. Parsons, Theran D. et al. A trifluoromethylboron compound, CF-BF2. Journal of the American Chemical Society 83:250- 251. 1961. 29. Pauling, Linus. The nature of the chemical bond. Ithaca, Cornell University Press, 1939. 429 p. 30. The nature of the chemical bond. 3d ed. Ithaca, Cornell University Press, 1960. 644 p. 31. Pitzer, KennethS. Electron deficient molecules. I. The principles of hydroboron structures. Journal of the American Chemical Society 67:1126-1132. 1945. 32. Price, W. C. The structure of diborane. Journal of Chemical Physics 15:614. 1947. 33. The absorption spectrum of diborane. Journal of Chemical Physics 16:894-902. 1948. 34. Rundle, R. E. Electron deficient compounds. Journal of the American Chemical Society 69:1327-1331. 1947. 35. _^ Electron deficient compounds. III. The bridge structure for B2H^. Journal of the American Chemical Society 69:2075-2076. 1947. 59 36. Rundle, R. E. Electron deficient compounds. II. Relative energies of "half-bonds. " Journal of Chemical Physics 17: 671-675. 1949. 3 7. Electron deficient compounds. Journal of Physical Chemistry 61:45-50. 1957. 38. Sanderson, Robert Thomas. Vacuum manipulation of volatile compounds. New York, Wiley, 1948. 162 p. 39. Schabacher, W. and J. Goubeau. Darstellung und Schwingung- spektren der Methylborbromide. Zeitschrift fur anorganische und allgemeine Chemie 294:183-190. 1958. 40. Schlesinger, H.I. and Herbert C. Brown. Metallo borohydrides. III. Lithium borohydride. Journal of the American Chemical Society 62:3429-3435. 1940. 41. New developments in the chemistry of diborane and the borohydrides. I. General summary. Journal of the American Chemical Society 75:186-190. 1953. 42. Schlesinger, H.I. and Anton B. Burg. Recent developments in the chemistry of the boron hydrides. Chemical Reviews 31:1- 41. 1942. 43. Sidgwick, Nevil V. The electronic theory of valence. London, Oxford University Press, 1927. 310 p. 44. Skinner, H.A. and N. B. Smith. The heat of hydrolysis of boron tribromide. Transactions of the Faraday Society 51:19-22. 1955. 45. Stewart, James E. Infrared absorption spectrum of trimethyl borane. Journal of Research of the National Bureau of Stand ards 56:337-341. 1956. 46. Stock, Alfred. Hydrides of boron and silicon. Ithaca, Cornell University Press, 1933. 250 p. 47. Tollins, B.C., Riley Schaeffer and H.J. Svec. Mass spectral studies of trimethyl boron. Journal of Inorganic and Nuclear Chemistry 4:273-278. 1957.