INFORMATION to USERS Universiw M Icrxjfilm S Internationéil

INFORMATION TO USERS

This reproduction was made from a copy of a document sent to us for microfilming. While the most advanced technology has been used to photograph and reproduce this document, the quality of the reproduction is heavily dependent upon the quality of the material submitted.

The following explanation of techniques is provided to help clarify markings or notations which may appear on this reproduction.

1. The sign or “target” for pages apparently lacking from the document photographed is “Missing Page(s)” . If it was possible to obtain the missing page(s) or section, they are spliced into the film along with adjacent pages. This may have necessitated cutting through an image and duplicating adjacent pages to assure complete continuity.

2. When an image on the film is obliterated with a round black mark, it is an indication of either blurred copy because of movement during exposure, duplicate copy, or copyrighted materials that should not have been filmed. For blurred pages, a good image of the page can be found in the adjacent frame. If copyrighted materials were deleted, a target note will appear listing the pages in the adjacent frame.

3. When a map, drawing or chart, etc., is part of the material being photographed, a definite method of “sectioning” the material has been followed. It is customary to begin filming at the upper left hand comer of a large sheet and to continue from left to right in equal sections with small overlaps. I f necessary, sectioning is continued again—beginning below the first row and continuing on until complete.

4. For illustrations that cannot be satisfactorily reproduced by xerographic means, photographic prints can be purchased at additional cost and inserted into your xerographic copy. These prints are available upon request from the Dissertations Customer Services Department.

5. Some pages in any document may have indistinct print. In all cases the best available copy has been filmed.

U niversiW Micrxjfilms Internationéil 300 N. Zeeb Road Ann Arbor, Ml 48106

8311779

Miller, Mark Allen

LOW VALENT ALUMINUM COMPOUNDS: REACTIONS AND CHEMISTRY OF TETRAISOBUTYLDIALUMINUM

The Ohio State University Ph.D. 1983

University Microfilms I nternationsil 300 n. zeeb R W . Ann Arbor, M I 48106

LOW VALENT ALUMINUM COMPOUNDS:

REACTIONS AND CHEMISTRY OF TETRAISOBUTYLDIALUMINUM

DISSERTATION

Presented in P artial F u lfillm en t of the Requirements for

the Degree Doctor of Philosophy in the Graduate

School of The Ohio State University

By

Mark Allen Miller, B.A., M.S.

*****

The Ohio State University

1982

Reading Committee: Approved By

Dr. Eugene P. Schram

Dr. Sheldon Shore

Dr. Andrew Wojcicki Adviser Department o f Chemistry TO

MOM

11 ACKNOWLEDGMENTS

I would lik e to thank my adviser. Dr. E. P. Schram, fo r his advice and patience throughout my tenure at The Ohio State University. Special thanks to my colleagues. Bill and Drew.

I would like to thank my typist, LaDonna Northern, for a splendid job under difficult conditions.

i n VITA

February 7, 1953 ...... Born - M assillon, Ohio

1975 ...... B .A ., The College o f Wooster, Wooster, Ohio

1978 ...... M .S ., The Ohio State U n iv ersity, Columbus, Ohio

1975-1981 ...... Teaching Associate, The Ohio State Uni­ v e rs ity , Columbus, Ohio

1975-1982 Research Associate, The Ohio State Uni­ v e rs ity , Columbus, Ohio

FIELDS OF STUDY

Major Field: Chemistry

IV TABLE OF CONTENTS

Page

DEDICATION...... ü

ADKNOWLEDGMENTS...... i i i

VITA ...... iv

LIST OF TABLES...... v ii

LIST OF FIGURES...... v i i i

INTRODUCTION ...... 1

EXPERIMENTAL...... 19

Equipment ...... 19 Analyses...... 29 Reagents ...... 30

Purchased Reagents ...... 30 Prepared Reagents ...... - 33

Boron Compounds...... 33 Aluminum Compounds...... 34 Pentacarbonylmanganese Hydride ...... 38 T ris(triphenylphosph1ne) pi a t 1num(0 ) ...... 39

Reactions of Tetralsobutyldlaluminum ...... 40

With Protonic Reagents ...... 40

With HCl...... 40 With Dimethyl amine ...... 42 With Ethyl ami n e ...... 43 With Dimethyl amine Adduct of D1Isobutyl- dimethyl ami noaluminum ...... 44 With Toluene-3,4-D1 th io l ...... 45

Reactions w ith Aluminum and Boron Compounds ...... 46

With T ri methyl aluminum ...... 46 Attempted Preparation of Lewis Base Adducts of Tetra 1 s obutyl d 1 al uml num...... 47 With Boron Trichloride ...... 48 V TABLE OF CONTENTS (c o n t.)

Page

With Trimethoxyboron ...... 49 With T ri s (dimethyl ami no) boron...... 52

Reactions w ith Transition Metal Compounds ...... 54

With Dimanganese Decacarbonyl ...... 54 With Pentacarbonylmanganese Hydride ...... 55 With Tris(triphenylphosphine)platinum(O)...... 56

RESULTS AND DISCUSSION...... 58

Introduction...... 58

Reactions with Protonic Reagents ...... 63

With H C l...... 64 With Dimethyl ami n e ...... 73 With Ethyl ami ne ...... 86 With Dimethyl amine Adduct o f Diisobutyldim ethyl- aminoaluminum...... 87 With Toluene-3,4-Dithiol ...... 89

Reactions w ith Aluminum and Boron Compounds ...... 96

With Trimethyl aluminum ...... 96 Attempts to Prepare Lewis Base Adducts o f T etrai sobutyl di al uminum...... T09 ' With Boron Trichloride ...... 115 With Trim ethoxyboron ...... 118 With Tris(dimethylamino)boron...... 142

Reactions with Transition Metal Compounds ...... 155

• With Dimanganese Decacarbonyl ...... 155 With Pentacarbonylmanganese Hydride ...... 156 With Tris(triphenylphosphine)platinum (0)...... 161

SUMMARY...... 167

APPENDIX...... 168

LIST OF REFERENCES...... 188

VI LIST OF TABLES

Table Page

1. Infrared Frequencies for Trimethoxyboron...... 170

2. Infrared Frequencies for Tris(dimethylamino)boron ...... 172

3. Infrared Frequencies for Tetrai sobutyl dial uminum ...... 175

4. Infrared Frequencies for Diisobutylmethoxyaluminum ...... 177

5. Infrared Frequencies for Di i sobutyldi methylami noalumi num. . . 179

6 . In frared Frequencies fo r Pentacarbonylmanganese Hydride . . . 181

7. Infrared Frequencies for Tris(triphenylphosphine)platinum . . 183

8 . Infrared Frequencies for Products of Treatment of Tetraiso­ butyl dial uminum with HCl ... 68

9. Dihydrogen Evolution from Tetraisobutyldialuminum and Dimethylamine at Room Temperature ...... 75

10. Dihydrogen Evolution from Tetrai sobutyldialumi num and Dimethyl ami ne at 4 3 °C ...... 76

11. Dihydrogen Evolution from Tetra i sobutyldi alumi n um and Ethyl ami ne ...... 38

12. Infrared Frequencies for Products of Treatment of Tetra- isobutyldial uminum with Toiuene-3,4-Dithiol ...... 94

13. Infrared Frequencies for Products of Treatment of Tetra- isobutyldialuminum with Trimethoxyboron...... 134

14. Infrared Frequencies for Products of Treatment of Tetra- i sobutyl dial umi num with Tris (dimethyl ami no) b o ro n...... 149

15. Infrared Frequencies for Solid from Treatment of Tetra- isobutyldial uminum w ith Pentacarbonylmanganese Hydride. . . . 158

16. Infrared Frequencies for Solid from Treatment of Tetra- isobutyldial umi num with Tri s(tr i phenylphosphi ne)p ia ti- num. A fter Exposure to A ir ...... 165 v ii LIST OF TABLES (cont.)

Table Page

17. Chemical S h ift fo r Various Aluminum and Dialuminum Compounds...... 61

18. Infrared Frequencies for Tri isobutyl aluminum ...... iss

19. Infrared Frequencies for Trimethyl aluminum ...... • • • • 100

v n i LIST OF FIGURES

Figure Page

1. Bi s ( di methyl ami no) t r i methyl t r i al umi num, Al 3(CH3)3[N(CH3)2]2...... ^

2. Bis(2,3-dimethylaluminacyclopropane), AI2C8 H1 6 ...... 2

3. General Form of Tetrasubstituted Dialuminum Compounds, Al2^2^2* 3

4. Hexai sobutyl di al uminum Dianion, Al2 (Ct*Hg)6 . " ^ ...... 3

5. Dicobaltoctacarbonyl Aluminumtribromide Adduct, Co2(C0)g »AlBr3...... 11

6. Dicyclopentadienyltungstendihydride Trimethylaluminum Adduct, (C5H5)2WH2*A1(CH3)3...... 11

7. Cyclopentadi en yltri phenylphosphinodi ethylalumi nomolybdenum- carbonyl, (C5H5)[P(CgHg)3](C 0 )MoAl(€2^15)2...... 13

8 . Dimethylaminoaluminotricarbonyl iron Dibromide, (CG)3Br2FeAlN(CH3)2 ...... 13

9. Vacuum Line, Front Side ...... 20

10. Vacuum Line, Back S id e ...... 21

11. Standard Reaction B ulb ...... 24

12. Filtration Apparatus...... 24

13. Toepler System ...... 26

14. In frared Spectrum of Trimethoxyboron, Gas Phase ...... 169

15. In fra re d Spectrum of T r is (dimethylamino)boron. Gas Phase . . .171

16. Proton NMR Spectrum o f T etrai sobutyl dia l umi num ...... 173

17. In frared Spectrum of T e tra i sobutyl dia l umi num. O i l ...... 174 ix LIST OF FIGURES (cont.)

Fi gure Page

18. Proton NMR Spectrum of Diisobutylmethoxyal umi num ...... igQ

19. In frared Spectrum o f Diisobutylmethoxyaluminum. Neat Liquid ...... 176

20. Proton NMR Spectrum o f Diisobutyl dimethylaminoaluminum .... 80

21. In frared Spectrum of Diisobutyl dimethyl aminoaluminum. Neat Liquid ...... 178

22. In frared Spectrum of Pentacarbonylmanganese Hydride ...... 180

23. In frared Spectrum o f T ris ( t r i phenyl phosphi ne)platinum...... 182

24. Proton NMR spectrum o f Reaction Products from Treatment of Tetraisobutyldialuminum with H C l...... 66

25. In frared Spectrum o f Products from Treatment o f T etra­ isobutyldialuminum with HCl...... 67

26. Proton NMR Spectrum o f Diisobutyl dimethyl aminoaluminum Dimethyl ami ne A dduct ...... 81

27. In frared Spectrum o f Solid Products from Treatment of Tetraisobutyldialuminum with Toluene-3,4-Dithiol ...... 93

28. Proton NMR Spectrum o f Solution from Treatment o f Tetra­ isobutyldialuminum with Toiuene-3,4-Dithiol ...... 95

29. Proton NMR Spectrum o f Product from Treatment o f Tetra­ isobutyldialuminum with Trimethyl aluminum...... 101

30. Proton NMR Spectrum of Product from Treatment o f Tetra­ isobutyldialuminum with 4-PiCO l i n e ...... 111

31. Proton NMR Spectrum o f Product from Treatment o f Tetra­ isobutyldialuminum with Boron Trichloride...... 116

32. Proton NMR Spectrum o f V o la tile Materials Removed from Reaction of Tetraisobutyldialuminum and Trimethoxy­ boron...... 126 X LIST OF FIGURES (c o n t.)

Figure Page

33. Infrared Spectrum o f Product from Treatment of Tetra­ isobutyldialuminum with Trimethoxyboron, Neat Oil...... 133 34. Proton NMR Spectrum of Product from Treatment of Tetrai sobutyldialumi num with Trimethoxyboron, Fully Extracted ...... 144

35. Proton NMR Spectrum of I ...... 132

36. Proton NMR Spectrum o f V o la tile M aterials Removed from Reaction of Tetraisobutyldialuminum and Tris(dimethyl- ami no) boron...... 146

37. Proton NMR Spectrum o f Products from Treatment of T etra­ isobutyldialuminum with Tris(dimethylamino)boron in Cyclopentane ...... 152

38. Infrared Spectrum o f Products from Treatment of Tetra­ isobutyldialuminum with Tris(dimethylami no)boron. Neat O il...... 148

39. Proton NMR Spectrum o f Products from Treatment of Tetra­ isobutyldialuminum with Excess Tris(dimethylamino)boron. . . .1 5 0

40. Infrared S;;;ectrum of Solid from Treatment of Tetraiso­ butyldialuminum w ith Pentacarbonylmanganese Hydride ...... 157

41. Proton NMR Spectrum o f Solution from Treatment of T etraiso ­ butyldialuminum w ith Pentacarbonylmanganese Hydride...... 1 5 0

42. Proton NMR Spectrum o f T ri isobutyl aluminum...... 184

43. Infrared Spectrum o f T ri isobutyl aluminum. Neat Liquid ...... 135

44. Proton NMR Spectrum o f Diisobutylaluminum C h lo rid e ...... igy

45. Dihydrogen Evolution from Treatment o f Tetraisobutyl­ dialuminum with Dimethyl amine ...... 77

46. Proton NMR Spectrum o f a Mixture o f Tetraisobutyldialum i­ num and Dimethyl ami ne ...... 79 xi LIST OF FIGURES (co n t.)

Figure Page

47. Infrared Spectrum of Products from Treatment of Tetra­ isobutyldialuminum with Trimethylamine...... 99

48. Infrared Spectrum of Product of Reaction of Tetraisobutyl- dialuminum w ith Ethyl amine ...... • * *

49. Low Temperature NMR Spectrum of Products from Treatment of Tetraisobutyldialuminum with Teimethyl- aluminum...... 103

50. Reaction Schemes for the Interaction of Tetraisobutyl­ dialuminum and Trimethylaluminum ...... 106

51. Infrared Spectrum of Insoluble Material from Treatment of Tetraisobutyldialuminum with Trimethoxyboron...... 128

52. Proton NMR Spectrum o f Diisobutylmethoxyal uminum a fte r Three Months ...... 121

53. Proton NMR Spectra of a Mixture of Tetraisobutyldialumi­ num and Trimethoxyboron as a Function of Time...... 123

54. Proton NMR Spectra of a Mixture of Triisobutyl aluminum and Trimethoxyboron as a Function of T im e...... 124

55. Proposed Structure of 1 ...... 137

56. Proton NMR Spectrum o f Product from Treatment o f Tetra­ isobutyldialuminum with T ri s (tri phenylphosphine)piati- num...... 164

57. Infrared Spectrum of Solid from Treatment of Tetra­ isobutyldialuminum with Tris(triphenylphosphine)plati­ num a fte r Exposure to A ir. . : ...... 164

58. Proton NMR Spectrum o f Product from Treatment of Tetraisobutyldialuminum with Trimethoxyboron in Cyclopentane ...... 143

x ii INTRODUCTION

Since the synthesis of the firs t catenated aluminum compound bis-

(dimethyl ami no) trim ethyl t r i aluminum, Al 3( 0113) 3 j^N(CH3) 2j 2 (Figure 1), in

1969 by Schram and co-workers by reduction of trimethylaluminum with tetrakis(dimethylamino)diboron , 1 there has been little work published concerning the successful synthesis of compounds containing direct aluminum-aluminum bonds.

In 1975 Skell and McGlinchey postulated the presence of such a moiety as a product of the co-condensation of aluminum atoms and 2- butene.2 This product gave dideuterium upon reaction with deuterium oxide and the mcsety responsible for this behavior could be titrated out with iodine, taking up the latter in a 1:1 mole ratio with the alumi­ num compound. One possibility that was suggested to satisfy the experi­ mental data was bis( 2 ,3-dimethylaluminacyclopropane) [Figure 2 ]. However, no compound was actually isolated and characterized from this reaction; the "purification would be fraught with experimental difficulties."

In 1976 Hoberg and Krause published examples o f another type of catenated aluminum compound.^ Potassium metal was employed to reduce cer­ tain organoaluminum chlorides in non-polar solvents and produce species with the formula X(R)A1-A1(R)X (Figure 3), where R=isobutyl and X=isobutyl, dimethyl amino, or isobutoxy.

The tetraisobutyl compound was characterized by reaction with deutero- methanol, which produced dideuterium and deuteroisobutane in a 1:4 ratio.

1 ^ ^ 2

Al Al / Me *

NME'-

Figure 1

Bi s(di methyl ami no) t r i methyltri a1umi num, A l3(CH3)3 N(CH3);

H ÿ x /CH3 HC

/ \ CH.

Figure 2

Bis(2 ,3-dimethylaluminacyclopropane), Al2CgHi6 , R R

A l — kL y \ X X

Figure 3

General Form of Tetrasubstituted Dialuminum Compounds, A I 2R2X2 .

iB u ^ iBu

iBu— Al—Al— iBu

iBu'^ '^iBu

Figure 4

Hexaisobutyldialuminum Dianion, AlgfC^Hgj^."^ Magnetic studies showed i t to be diamagnetic. The presence of the cova­

lent aluminum-aluminum bond was accompanied by an intense dark red-brown color.

In 1978 this work was furthered to include the reduction of triiso­ butyl alumi num with potassium.** From a prolonged reaction in a non-polar solvent came a product characterized as K 2Al2iBug, the anion having an ethane-like structure (Figure 4). The characterization was performed by molecular weight determination, solvolysis with a deuterated alcohol to afford a ratio of deuterated isobutane: dideuterium of 6 : 1 , and proton nuclear magnetic resonance studies that showed the chemical s h ift of the alpha protons of the isobutyl group shifted upfield with regard to that found in the startin g m a te ria l. The NMR spectrum also showed the absorptions to be invariant upon lowering the temperature to -70°C, in­ dicating the absence of bridging alkyl groups.

The compounds f ir s t synthesized by Hoberg and Krause, i . e . , R(X)A1-

A1(R)X, immediately bring to mind s im ila r boron compounds that have been known fo r some time. Di boron te tra c h lo rid e , B 2Cljt, was the firs t such compound identified,^ prepared by subjecting boron trichloride to elec­ tr ic discharge. This compound is unstable above 0°C and undergoes dis­ proportionation to boron trichloride and a series of cluster compounds with the general formula (BCl)^, where x is 4,8,9® “® and 612^^11

The stability of the boron-boron moiety may be increased by substi­ tuting various ligands for the chloride that are better pi donors, such as -F, -OCH3 , or -N(CH3) 2 . Indeed, tetraki s(di methyl ami no)di boron,

B2 1^ ( 083 ) 2^ 4 , may be d is tille d under reduced pressure a t temperatures up to 200°C with no decomposition o c c u r r in g .T h e o r e tic a l calculations a ttrib u te this increased s ta b ility to the "back bonding" a b ility of the

various ligands. ^i Electron density is delocalized from the ligand to the

boron, u tiliz in g the o rb ital of the la tte r to form a pi o rb ita l.

Mixed diboron compounds involving methyl ligands were synthesized by

Schick in 1966^2 and later by Schram in 1969.^ Attempts to produce the tetramethyl derivative were unsuccessful, resulting in the production of trimethyl boron and intractable p o l y m e r s . The firs t successful synthe­ sis of a tetraalkyl derivative was achieved by Noth in 1980. Tris(t- butyl)methyldiboron was prepared from tetramethoxydiboron, t-butyllithium , and methyl lithiu m . I'*

The interesting chemistry afforded by the diboron compounds might well be extended to include analogous aluminum compounds. The fir s t catenated aluminum compound, mentioned above, and a unique boron com­ pound, B6^u-N(CH3) 2j i 2Al6 ( 083 ) 12» were prepared by the controlled de- stabilization of the boron-boron bond in 82 j^N(CH3) 2jwith trimethyl­ aluminum. The tetrai sobutyldialumi num may well be an excellent starting material for investigations into similar areas involving all-aluminum systems. I t appears that the known dialuminum species have sim ilar stabilities as closely related diboron species. Noth reports that the tetraalkyldiboron compound is stable to at least 50°C and Hoberg gives a temperature of approximately 80°C for the decomposition of tetraisobutyl­ dialuminum.

Attempts a t alkyl exchange, using triethylaluminum with te tra iso ­ butyldialuminum, have been r e p o r t e d . ^ No stoichiometries of the reaction were given but mixing solutions of the two reagents resulted in the depo­ sition of m etallic aluminum. The actual course fo r such a reaction can 6 not be determined unequivocally, but it seems likely that alkyl exchange will occur. This may be followed by disproportionation involving a 1,2 shift of an alkyl group to produce A l(III) species and "AIR". The latter should be extremely reactive and would be the source of m etallic aluminum.

Decomposition via a purely intermolecular process has been suggested fo r similar diboron species, but insufficient evidence is available to favor either route. This ambiguity notwithstanding, the destabilization of the tetraisobutyldialuminum is not an easily controlled reaction and disproportionation cannot be prevented or used to advantage in producing new catenated species.

I t would seem, then, that increased s ta b ility of the aluminum- al umi num moiety is needed. Analogous to the diboron compounds already mentioned, this might be accomplished by exchange of the isobutyl ligands by ligands possessing much better pi back bonding ca p ab ilitie s . These could include halide, alkoxy, or alkyl ami no groups.

There are two basic ways to induce ligand exchange in al kyl-aluminum compounds. First is the reaction with the appropriate protonic reagent according to the general scheme:

AIR3 + nHX AlRa-nXn + nHR (1) where R is alkyl, X is halide, alkoxy, or alkyl ami no, and nisi, 2, or

3. (Note: Although most aluminum compounds e x is t as dimers or higher order aggregrates, they are shown here as monomers for the sake of sim­ plicity.) For the specific case we wish to investigate, the scheme be­ comes :

AI2R4 + nHX -> AlgR^-nXn + nHR (2)

This type of reaction is well known fo r triv a le n t aluminum compounds 7

and by controlling the reaction stoichiometry excellent yields and purity may be obtained.Reaction conditions vary widely depending on the protonic reagent involved. With haloacids, the reaction proceeds at temperatures as low as -78°C, while dialkyl amines generally require ele­ vated temperatures, up to 100°C, to afford the desired products.

A possible complication to this method of ligand exchange is fore­ shadowed by the work of Hoberg cited earlier. In the characterization of the dialuminum compounds, deuterated methanol was reacted with the com­ pound as a means of determining the presence of aluminum-alumi num bonds, oxidation of which by this reagent produces dideuterium. The reagent also causes cleavage of the aluminum carbon bonds in the compound. This la tte r reaction is the one of in tere s t here and i t remains to be seen which bond is the more reactive of the two. Experimental conditions are not given by Hoberg, but in such reactions, vigorous conditions and an excess o f the oxidizing reagent are normally employed to insure complete reaction. Therefore, carefully controlled reaction conditions and stoichiometries may result in finding a situation in which the desired reaction would take place preferentially and to a much greater extent than the oxidation of the aluminum-aluminum bond.

This would be the method o f choice i f such problems can be e lim i­ nated. The isolation of products would be greatly simplified by the fact that the unwanted by-product of the reaction would be a volatile gas, isobutane, and could be easily removed from the reaction environment.

Monitoring of the reaction progress would be facilitated also.

If this scheme does not prove viable, a second type of reaction is available to induce ligand exchange; redistribution reactions involving 8

the appropriate aluminum or boron compound.j^ese can be repre­

sented generally fo r triv a le n t aluminum compounds as:

AIR3 + MX3 AlR3_nXn + MX3_nRn (3)

where M is B or Al. For the specific case we are interested in, the equa­

tion becomes:

A l2R4 MX3 A l2R4-n^n MX3-nRn (4)

Actual products for the reactions depend on specific reagents and

stoichiometries. When the metal is aluminum, the reaction is usually a

simple scrambling of ligands to afford the most stable bridged species,

such as:

2A1(CH3)3 + A IC I3 ^ 3A1C1(CH3)2 (5)

A1(CH3)3 + 2A1C13 -> 3Al(CH3)Cl2 ( 6 )

Here, with chloride as the bridging ligand, a statistical distribution of available ligands about the aluminum atoms results.

With th is in mind, Hoberg's reaction of tetraisobutyldialuminum with

trie th y l aluminum may have been forced to unstable products by an excess of the ethyl species. Once again, this highlights the need for carefully controlled reaction conditions and stoichiometries.

As stated, boron compounds can also be used to e ffe c t a ligand ex­ change. The principles of the reaction are the same but the results be­ come more complicated by the differences in the chemistry of aluminum and boron. The bond strengths o f the various metal ligand systems that may be investigated as usable reagents d iffe r from aluminum to boron. Thus, the products are not always governed by the s ta tis tic a l scrambling of ligands. An example of this is given by the reaction of trimethyl boron and tris(dimethylamino)aluminum dimer, in which the product reflects the 9

fact that only one of the methyl groups o f the boron can be replaced by exchange with the aluminum compound, no matter what the ra tio of react­ ants is.

Fortunately, for this work it seems that the alkyl boron species are preferred products in this type of reaction. That is, the favored products are those which involve a transfer of methoxy or dimethyl ami no group from boron to aluminum and the subsequent replacement o f th a t group by an alkyl group from the aluminum species.

I f the re d is trib u tio n scheme should prove to be the most lik e ly to succeed, other factors must be considered in addition to those mentioned above. The most obvious is that the reaction products may be very d i f f i ­ cult to separate and afford a pure sample of any new species generated.

The tetraki s(dimethyl ami no)- and tetramethoxydiboron compounds are both liquids at room temperature which do not exhibit a significant vapor pressure until heated well above room tem perature.The aluminum counterparts of these compounds would be expected to show sim ilar prop­ erties, but their stabilities may not be as great. Distillation of the reaction mixture to separate components and purifying a desired product may not be possible in this case, as is the norm with diboron analogues.

Alternately, then, it may be possible to remove only the unwanted by-products from the reaction mixture. Here, the reagents of choice would be the boron compounds because o f th e ir higher vapor pressure than corresponding aluminum compounds. For example, trim ethyl boron is a gas a t room temperature, whereas trim ethyl aluminum is a liquid with a vapor pressure of only about 15 torr. (Note: This phenomenon is largely due to the penchant fo r aluminum compounds to exist in dimeric or higher 10 aggregate form, thus more than doubling the molecular weight of an alumi­ num compound as compared to its boron analogue.)

It would appear that solubility differences hold little promise as a means of separation. The dialuminum compounds that are known tend to form oils rather than crystals and are soluble in all common aliphatic and aromatic solvents.One might expect there to be slight, if any, s o lu b ility difference between such sim ilar compounds as triis o b u ty l aluminum and tetraisobutyldialuminum. Possibly, Lewis base adducts of the species may show more of a difference in solubility characteristics.

Extending th is la tte r idea fu rth e r, one might also tend toward the use of boron compounds as the reagent for redistribution reactions.

Aluminum is a stronger Lewis acid than boron and prefers "hard" type bases to a much greater extent than boron.This difference may be the basis for a separation involving formation of adducts of the aluminum species, which w ill increase their stability while only forming weak ad­ ducts to the boron by-products. The la tte r should be easily associated and pumped away under reduced pressure.

Another area of interest that might be available through the use of catenated aluminum compounds is that of direct aluminum-transition metal bonds. A fairly large body of work is present concerning the reactions of aluminum species and transition metal complexes, most notably carbonyl and pi-cyclopentadiene complexes.

There are three basic types of interaction in these systems. First, of little interest here but worth mentioning, is the interaction of alu­ minum compounds and metal carbonyls via the bridging CO ligands of the complexes. Such a reaction is ty p ifie d by the product obtained when n

0 0C \ A / C O OC— Co— Co—— CO o e ^ CO 0

A lB r 3

Figure 5

Dicobaltoctacarbonyl Aluminumtribromide Adduct, CogfCOOg-AIBrg,

Cp H W »Al (CHt )-z Cp ^ H

Figure 6

Di eye1opentadi eny1tungstendi hydri de T ri methyl a1uminum Adduet, (CgHgizWHz-AlfCHgig. 12

Co2(C0)e is treated with aluminum bromide.^3 The product has the formula

Co2(C0 ) 8 *A l(B r)3 and, on the basis of infrared spectrocopy, it was deter­ mined that the oxygen atom of the bridging carbonyl acts as a Lewis donor to the aluminum and results in formation of an adduct with the form shown in Figure 5. The stretching frequency of the bridging CO ligand shows a marked decrease upon coordination of the aluminum species. The preference for this site, rather than the terminal site for adduct formation, is ex­ plained by the increase in basicity of the oxygen of the former afforded by bridging two metal centers. Other examples of this type of coordina­ tion are found with a molybdenumand nickel carbonyl complexes.^"’

The second type of interaction is again a coordination situation, but this time involving a direct metal-aluminum interaction. The first example of such a compound was prepared from dicyclopentadienyltung­ stendi hydride and aluminum trim ethyl.26 x-ray diffraction studies showed the species to have a structure which can be represented in Figure 6.

The rhenium analogue of this species was also prepared.

Further work with this type of system included thermolysis, which resulted in the evolution of methane gas. This phenomenon could be re­ lated to the reaction of a dialkylamine and aluminum trimethyl, in which the in itial reaction at room temperature is adduct formation. Upon heat­ ing to ca. 100°C, methane is formed and a covalent aluminum-nitrogen bond is formed. However, in the case of these metal-aluminum species no stable covalent compounds could be identified or isolated.

This leads us to the final type of interaction (the most germane to the discussion here), formation of direct metal-aluminum covalent bonds.

Such species have been produced in the reaction of aluminum alkyls and 13

CO OC X, / /C y H g Mo— AL

(IP ^ I CoHi'2 ^ 5 Cp

Figure 7

Cyclopentadi enyl t r i phenylphosphi nodi ethylalumi no- molybdenumcarbonyl, (C5H5) PfCGHgjsjfCOjMoAlfCzHgjz.

oc Br B p CO \ / oc— Fe— Al Al- -F e — CO / \ B r CO “ ■> X > .

Figure 8

Dimethyl aminoàl uminotricarbonyl iron Dibromide, (C0)3Br2FeAlN(CH3)2. 14

cyclopentadienylmolybdenumcarbonylhydrides (Figure 7).2? Other molybdenuni-

aluminum compounds posses aluminum atoms which occupy bridging positions

between molybdenum a t o m s .28 ,29

Another bridging interaction is seen in the product of the reaction

of aluminum trichloride and sodium tetracarbonylcobaltate(I). Here,

the aluminum is thought to cap a tricobalt cluster.

A d ire c t metal-aluminum covalent interaction is also found in the

dimeric dimethyl ami noaluminotricarbonyl iron dibromide (Figure 8). This

compound is the product of the reaction of dimethyl aminoaluminum dibro­

mide and triirondodecacarbonyl and has aluminum in the formal oxidation

state of +1.81

A polar solvent, such as THF, breaks up the intermolecular associa­

tion by coordination to the aluminum atom, but s till retains the aluminum

in low oxidation state and the iron-aluminum bond remains in ta c t.

The dialuminum compounds studied here might provide a startin g point

for investigations into the preparation of new transition metal-aluminum

species. Two basic types of reaction should be studied. F irs t would be

the reaction between a hydridic species and the dialuminum compound.

This would be sim ilar to the reactions mentioned fo r protonic reagents.

Pentacarbonylmanganesehydride should prove to be a suitable starting material for such reactions. All five CO groups are bound identically to the metal and the ligands are all of the terminal type, as shown by

X-ray crystallographic studies.82 This is the most stable of the known carbonyl metal hydrides of the firs t row transition metals, decomposing only above 25°C. 15

It is known that this hydride will react with trivalent aluminum alkyls or alkyl hydrides to evolve the appropriate gaseous by-product, i . e . , H 2 or RH. In these reactions, however, no stable compound contain­ ing an aluminum-manganese moiety was isolated from the reaction environ­ ment.33 Whether this is due to the actual mechanism of the reaction or to the fact that the desired product was just not stable under the con­ ditions employed to isolate it cannot be determined unequivocally. The possibility exists that a dialuminum species may offer a different path­ way for reaction to occur or may prove to result in a more stable prod­ uct.

Recalling the discussion involving protonic reagents, above, it is realized th at there are two possible paths for reaction available. One is cleavage of the aluminum-aluminum bond by the hydridic species to eventually evolve hydrogen and form metal-aluminum bonds. A reaction se­ quence may take the form of:

AI2R4 + HMn(C0)5 -> R2Al-Mn(C0 )s + R2AIH (7)

R2AIH + HMn(CO)s 4. R2Al-Mn(C0)5 + H2 ( 8 )

Another possible path could be the replacement of an isobutyl group by the metal carbonyl moiety accompanied by the production of isobutane.

This reaction scheme is illustrated as:

AI2R4 + HMn(C0)5 ^ R2Al-Al(R)Mn(C0)5 + RH (9)

This reaction scheme could be considerably more complicated than the first discussed. Multiple substitution would be possible and a variety of products, some of them isomeric forms, could result from such a reac­ tio n. Whether the aluminum-aluminum bond would remain in ta c t is another question th at this scheme would raise. 16

Of course, as postulated for the reaction of protonic reagents ear­

lier, there is the likelihood that these two reactions will be competi­

tive, giving a large variety of products that may prove very difficult

to isolate. Knowledge gained from the reactions of the protonic reagents

should be helpful in establishing reaction conditions and stoichiometries

fo r the tran sition metal case.

Characterization of any desired species formed should not be d iffi­

cult if isolation of pure compounds can be effected. Hydrolysis of the

metal-aluminum bond should regenerate the starting metal hydride and pro­

duce an AlOH m o i e t y , both easily identified.

Another type o f reaction with a tran sitio n metal compound that can

be envisioned is th a t o f the dialuminum and a saturated zero valent metal compound, i.e ., a d^° system. Specifically, tris(triphenylphosphine)- platinum(O) could be reacted with dialuminum species in an effort to pro­ duce yet unknown aluminum-platinum bonded species.

Upon solvolysis, the platinum complex is thought to dissociate one triphenylphosphine ligand to give "(CgHqJgP-Pt-PfCgHsis" in low concen­ trations, this species being the reactive reagent in such cases.35 if this complex were to react with the dialuminum compound, one might expect the initial interaction to be overlap of d and p orbitals of the metal and aluminum, respectively. The metal would be acting as a Lewis base in supplying electron density to the vacant aluminum orbital.

Steric hinderance may play a part in this initial interaction, as the diphosphine complex is thought to be lin ear and the aluminum-aluminum moiety is well shielded by the bulky isobutyl groups. For sig n ifican t interaction to take place, a rearrangement might be necessary. 17

Two types of covalent interaction can be envisioned, one in which the aluminum-aluminum bond is maintained and the other in which i t is cleaved. The former would re s u lt in a three-membered ring structure in which all of the electrons for the formation of platinum-aluminum bonds come from the metal. In the other case, metal-aluminum bonds would be formed by sharing of one electron from each atom involved in the bond.

The firs t type would result in four-coordinate aluminum atoms, the preferred arrangement as seen by the penchant for aluminum to form bridged dimers and higher order aggregates. In the second case, the alu­ minum would retain its three-coordinate state and the p^ orbital would be available for p-d pi interaction.

Of course, complications can be foreseen in this scheme. The plati­ num complex is a rather "soft" base, while the aluminum compound is a

"hard" acid. In theory, strong interactions are most favorable between members of the same class and not as likely for species having different properties. An inkling of what might be expected can be found from pre­ vious work done with Lewis adducts of tris(triphenylphosphine)platinum.

Among the compounds investigated were boron tric h lo rid e and aluminum t r i ­ methyl, examples of "soft" and "hard" acids, respectively. The boron adduct proved to be more stable in this case, an adduct with the aluminum compound unable to be isolated prio r to decomposition. Whether a low valent aluminum compound, with the possibility of covalent bond forma­ tion, will be more stable is the basic question to be investigated.

Another difficulty that may arise involves the actual platinum spe­ cies that is the reactive reagent. If one triphenylphosphine ligand must be dissociated from the starting complex, affording Pt 2, then 18

a competitive reaction with the free phosphine ligand and the dialuminum

compound can be assumed. Since the degree of dissociation is minor, this

could not affect the formation of a platinum-aluminum species in itia lly .

As the amount of free ligand increased and of aluminum compound decreased during the course of the reaction, a mixture of products might result.

Resolution of this type of problem would depend largely on the relative s ta b ility of both aluminum-platinum and aluminum-phosphine species that are produced in the reaction.

One would suspect that i f covalent bonds between aluminum and platinum are formed, such a product would be more stable than any simple adduct formed between aluminum and phosphine. A possible method o f o p ti­ mizing such conditions would be to perform the reaction at an elevated temperature to remove the triphenylphosphine dissociated by sublimation and forcing any resultant aluminum-platinum species to the thermodynam­ ically most stable product.

In summary, several aspects of the chemistry of catenated aluminum compound have been examined. Attempts to prepare aluminum analogues of known diboron compounds have been undertaken by two routes in order to synthesize species with greater stability than currently known compounds.

These may prove useful starting materials in further investigations of low valent molecular aluminum compounds. Also, preparation of previously unknown aluminum-transition metal species using the dialuminum reagent as a starting material have been attempted. EXPERIMENTAL

EQUIPMENT

All manipulations of volatile, air- and moisture-sensitive materials

were carried out on a standard Stock-type vacuum line equipped with mer­

cury flo at valves and ground glass vacuum stopcocks (Figures 9 and 10).

Connections between parts of the vacuum line were made with ball and

socket type joints sealed with Kel-F fluorocarbon wax. Ground glass

stopcocks were lubricated with Apiezon-N hydrocarbon grease.

Vacuum was maintained by a Welch Duo-Seal vacuum pump in series with

a mercury diffusion pump. Under normal operating conditions, a vacuum

of 1.0 X 1 0 " 5 mm Hg could be maintained with dynamic pumping. Pressures were measured with a McCloud gauge.

Non-volatile, air- and moisture-sensitive materials were handled in an oxygen-free dry box. The original shell was manufactured by Kewanee

S c ie n tific . The mechanical and e le c tric a l systems were extensively modi­

fied by the Systems Group of the Ohio State University chemistry depart­ ment fo r improved performance.

An in e rt atmosphere was maintained by c ircu latio n through a column

containing successive layers o f activated charcoal, "Ridox" (an oxygen

scavenger), and 13X molecular sieves to rid the system of oxygen and water and solvent vapors. Concentrations as low as 3 ppm oxygen and

1 ppm water could be attained under ideal operating conditions.

19 iTfi

ro Figure 9. Vacuum Line, Front Side. o Figure 10. Vacuum Line, Back Side. ro 22

An extremely sensitive test for oxygen could be obtained by prepar­ ing a mixture of dicyclopentadienyl titanium dichloride and zinc metal in tetrahydrofuran.3? The green solution that resulted, when exposed to the dry box atmosphere, would turn orange in the presence of oxygen. The test was sensitive to less than 5 ppm of oxygen and the rapidity of the color change gave a rough indication of the concentration of oxygen.

In the absence o f oxygen, a test for the presence of water and/or solvent vapor could be conducted by burning a 25-watt lig h t bulb whose filam ent had been exposed to the dry box atmosphere. The box was deemed usable i f the filam ent remained in tac t fo r more than 24 hours. Optimum conditions allowed the bulb to burn for more than three days.

If these tests showed the atmosphere to be contaminated, the column components were regenerated. This was accomplished by heating the column to 210°C under dynamic vacuum for fiv e hours to remove solvents and water. After that period, pumping was discontinued and a mixture of 5%

H2 and 95% N2 was passed through the column while maintaining the tem­ perature at 210°C to reduce the "Ridox" c a ta ly s t. The regeneration se­ quence was performed automatically a fte r in itia tio n by the operator.

Occasionally, reagents that were less sensitive to reaction with air and moisture, such as solutions of n o n -vo latile alkyl aluminum compounds, were handled in a glove bag manufactured by I^R. An in e rt atmosphere was provided by flushing the bag three times with "hi-pure" nitrogen or argon. Handling o f reagents was performed as quickly as possible to minimize reaction with residual air or moisture in the glove bag.

Weighings outside the dry box or glove bag were made on a Sartorious analytical balance to 0.1 mg ± 0.05 mg. Under in ert atmosphere 23

conditions, weighings were made on a Torsion laboratory balance to an

accuracy of 5 mg ± 2 mg.

Reactions, for the most part, were carried out in 50 or 100 ml Pyrex or Kimax round-bottomed flasks to which a Teflon Kontes valve stopcock had been attached by means of 0-ring joints and a Viton-A rubber 0-ring

(Figure 11). This will be referred to as a standard reaction flask.

Such flasks were normally attached to the vacuum line with 0-rings and

0-ring joints. Before a reaction was attempted, the flask was attached to the line, evacuated, and checked to insure the integrity of the glass­ ware and seals. Then the flask was heated with a cool flame to purge any water or solvent that may have been adhered to the walls of the fla s k .

Reagents and solvents were added to the flask in the dry box or glove bag for non-volatile compounds, or by condensation at low tempera­ tures on the vacuum line, for volatile compounds. If necessary, the flask was frozen to the -196°C and pumped on to remove the inert atmos­ phere trapped a fte r tran s fer of substances in the dry box or glove bag.

Mixing of reagents was accomplished by adding a Teflon-coated mag­ netic spin bar to the flask and rotating it by means of a magnetic s tir­ ring motor placed beneath or to the side of the fla s k .

Separation of insoluble reaction products from those soluble was accomplished by filtr a tio n in vacuo through a scintered glass f r i t con­ tained within a filtr a tio n apparatus equipped with 0-ring joints and a

Kontes valve stopcock (Figure 12). The flask containing the m aterial to be filtered was attached to one side of the apparatus via 0-ring joints and 0-ring and another evacuated standard flask was attached to the other 24

Figure 11. Standard Reaction Bulb.

Figure 12. F iltra tio n Apparatus. 25

side to act as receiver for the filtered solution. The entire assembly was attached to the vacuum line and evacuated, the Kontes valve was

closed, and the apparatus removed from the line so to be manipulable. By

cooling the receiver side of the apparatus slightly, the solution could

be drawn through the frit.

Extraction of a mixture was accomplished by filte r in g and then con­

densing a portion of the solvent from the receiver bulb back into the

original reaction flask and washing the solid with th is solvent. This

process could be repeated as many times as necessary to insure complete extraction of soluble m aterial from the mixture.

Some reactions required that a solution of one reagent be added to a solution of a second reagent. Such reactions were carried out in a tip bulb apparatus consisting of a Y-shaped tube equipped with 0-ring jo in ts .

A flask with the solution to be added was attached to the side arm of the device and a flask containing the solution where the reaction would take place was attached to the lower 0-ring joint. The entire apparatus was attached to the vacuum line, evacuated, and then isolated from the line.

The lower flask was cooled to a temperature below that of the upper flask to aid in transfer of the solution. The rate of addition could be controlled by manipulation of the upper bulb or by adjustment of a Kontes valve stopcock if one was present. After addition of the reagent was complete, some solvent was condensed back into the upper bulb to wash the sides of the apparatus to insure quantitative transfer of the added rea­ gent.

Non-condensible gases generated in reactions were collected in an automatic Toepler system (Figure 13). The volatile materials of a 26

Figure 13. Toepler System. 27 reaction passed through two -196°C traps to isolate condensible materi­ als while allowing non-condensible materials to be pumped into a calib rated volume. This volume was divided by means of a stopcock into a small volume capable of measuring quantities as small as 1 x lOr^ mmol and a large volume capable o f measuring quantities as large as 4 mmol.

Another portion of the system consisted of a U-tube filled with a copper oxide catalyst which could be heated by means of an external re­ sistance furnace. Passing the gas through the tube resulted in oxidation of the gas to a condensible product. The temperature at which the gas was oxidized was determined by its id e n tity . Carbon monoxide was o xi­ dized to carbon dioxide at 150°C, hydrogen to water at 300°C, and methane to CO2 and H2O at 800°C.

Infrared spectra were recorded on a Perkin-Elmer Model 457 grating spectrophotometer. Gaseous samples, or liquid samples with a sufficient vapor pressure to provide a significant amount of materials in the gas phase, were contained in an air-tight gas cell. The cell was a 10 cm x

2.5 cm glass tube to which a Kontes valve and 0-rin g jo in t had been a t­ tached. The ends were capped with KBr disks sealed to the glass by means of a hard fluorocarbon wax.

Solid samples were prepared for analysis by mulling with mineral oil

(nujol) or fluorolube which had been dried and stored over 4X molecular sieves. The mull was contained between KBr p lates. Sample preparation of air- and moisture-sensitive material was carried out in the dry box and spectra run immediately upon removal from the box.

Non-volatile liquids and oils were run neat between KBr plates, tak­ ing precautions as above with air- and moisture-sensitive materials. 28

Proton nuclear magnetic resonance spectra were recorded on an EM-390

90 MHz spectrometer or on an EM-360 60 MHz spectrometer. Samples were prepared by dissolving the substance in a suitable solvent and transfer­ ring a portion of the solution to a 5-nm precision NMR tube that had been attached to an 0-ring joint. A Kontes valve was attached to the device via 0-ring joint and 0-ring, and the apparatus attached to the vacuum line. The solution was frozen to -196°C and the inert atmosphere pumped away. Closing the stopcock and warming the solution to room temperature allowed liberation of additional gas that had been dissolved in the solu­ tio n . The tube was again frozen to -196°C and the remaining gas pumped away and, while s till under dynamic vacuum, the tube was sealed off with a torch at the original point of attachment to the 0 -ring joint.

If any solid matter was present in the tube it was inverted and cen­ trifuged to insure a homogenous solution for best possible resolution of the spectrum of the soluble species.

No internal standard was employed due to the possibility of reaction with the sample or masking of absorptions by the reference peak. There­ fo re , a ll chemical s h ifts were referenced to the solvent (usually benzene) and then converted to the delta scale.

ANALYSES

Analyses'of non-condensible gases were accomplished in the Toepler system as previously described.

Aluminum was analyzed gravimetrically by precipitation with 8 -hy- droxyquinoline. The procedure was s lig h tly modified from that found in the literature.38 29

If manganese was also present, the total of these two cations was

determined gravim etrically in the same manner.

Manganese was determined qualitatively by oxidation to the perman­

ganate anion by a strongly acidic solution of potassium periodate, which

resulted in the characteristic deep purple color of that anion. Quanti­

ta tiv e analyses were performed by titr a tio n with EDTA in a solution buf­

fered to pH 10 with Eriochrome Black T as the in d ic a to r.39

As aluminum would in terfere with such a titr a tio n , i t must be re­

moved prior to determination of manganese. This is accomplished by pre­

cipitation of the aluminum as the hydroxide with ammoniun hydroxide

solution, as this leads to the formation of the water-soluble tetrahy-

droxyaluminate anion. Complete precipitation was insured by addition of

a small amount of ammonium chloride solution and boiling the mixture

until the odor o f ammonia could be detected. The mixture was then f i l ­

tered and the solution analyzed fo r manganese. The amount of aluminum was then determined by difference from the results of the gravimetric

analysis. (Note: Aluminum can be separated from a solution quantita­

tively as the hydroxide but was not analyzed for in this manner because

of difficulties in treating the precipitate once obtained.Analysis by

precipitation with 8 -hydroxyquinoline is a much more reliable method.)

Platinum was determined qualitatively by reaction of the hexachloro-

platinate anion, which is prepared by oxidation of the platinum compound with concentrated aqua regia. When potassium iodide is added to a solu­ tion of the platinum anion, an intense red color results due to forma­ tion of the hexaiodoplatinate anion by an exchange reaction.

The chloro anion solution could also be used for a quantitative 30 analysis. Addition of a solution of sodium formate and formic acid causes the platinum to be precipitated as the metal. This could then be col­ lected and weighed, as in the literature.**^

REAGENTS

Purchased Reagents

Alumina, Fisher Scientific Company, was stirred with a cyclopentane solution of triiso b u ty l aluminum, washed with cyclopentane, and pumped on to remove volatile materials.

Ammonium hydroxide, Mallinckrodt Chemical Works, was used as re­ ceived.

Benzene, reagent grade, Fisher Scientific Company, was dried over lithium aluminum hydride and distilled prior to use.

Benzophenone, J. T. Baker Chemical Company, was used as received.

Biobeads S-X4, Bio-Rad Laboratories, was pumped on under dynamic vacuum for 54 hours p rio r to use.

Boron tric h lo rid e . The Matheson Company, was fractionated through a

-78°C trap into a -126°C trap prior to use.

Calcium hydride. Bios Laboratories, was used as received.

Chloroform, reagent grade, Mallinckrodt Chemical Works, was dried over lithium aluminum hydride and distilled prior to use.

Cyclopentane, reagent grade, Aldrich Chemical Company, was dried over lithium aluminum hydride and distilled prior to use.

Diethyl ether, analytical reagent, Mallinckrodt Chemical Works, was stirred with lithium aluminum hydride and subsequently distilled onto a mixture of potassium and benzophenone from which i t was d is tille d . 31

Diisobutyl aluminum chloride, Texas Alkyls In c ., was used as received.

Diisobutylaluminum hydride, 25% by weight in heptane, Texas Alkyls

In c ., was used as received.

Dimanganese decacarbonyl, Aldrich Chemical Company, was sublimed at

50°C under s ta tic vacuum p rio r to use.

Dimethyl amine. The Matheson Company, was dried over phosphorous pen-

toxide at 0°C and subsequently passed through a -78°C trap to a -196°C

trap prior to use.

Heptane, Chemical Samples Company, was dried over lithium aluminum

hydride and distilled prior to use.

Hydrogen chloride. The Matheson Company, was passed through a -126°C

trap to a -196°C trap prior to use.

8 -Hydroxyquinoline, reagent grade, Fisher Scientific Company, was used as received.

Lithium aluminum hydride, Alfa Products, was used as received.

Methanol, reagent grade, Mallinckrodt Chemical Works, was dried over phosphorous pentoxide and distilled prior to use.

Methylene chloride, reagent grade, M allinckrodt Chemical Works, was dried over lithium aluminum hydride and distilled prior to use.

Monoethylami ne. The Matheson Company, was passed through a -78°C trap to a -196°C trap prior to use.

Nitric acid, concentrated aqueous solution, Fisher Scientific

Company, was used as received.

4-P ico lin e, reagent grade, Eastman Kodak Company, was dried over calcium hydride and distilled prior to use. 32

Potassium, Mallinckrodt Chemical Works, was cut under an in e rt a t­ mosphere p rio r to use.

Silver nitrate, reagent grade. Allied Chemical, was used as received.

Tetrahydrofuran, analytical reagent, Mallinckrodt Chemical Works, was stirred with lithium aluminum hydride then distilled onto a mixture of potassium and benzophenone from which i t was d is tille d prio r to use.

Toluene, reagent grade, Matheson, Coleman and Bell, was dried over lithium aluminum hydride and distilled prior to use.

T o lu e n e -3 ,4 -d ith io l, Aldrich Chemical Company, was sublimed under dynamic vacuum a t 65°C prior to use.

Triisobutyl aluminum, Texas Alkyls In c ., was used as received.

Trimethyl aluminum. Ethyl Corporation, was used as received.

Triphenylphosphine, Matheson, Coleman and B e ll, as received.

Prepared Reagents

Boron Compounds. Trimethoxyboron was prepared by reaction of boron trichloride and methanol in a 1:3 molar ratio, producing the desired product and HCl. The reaction was performed neat by condensing 46.8 mmol of methanol into a standard reaction bulb at -196°C and adding 15.5 mmol of boron trichloride incrementally, approximately 2 nmol at a time.

After each addition, the solution was allowed to warm to room temperature with stirring. The pressure was monitored and the reaction mixture was cooled intermittently to -126°C to permit removal of HCl before another aliquot of BCI3 was added. After the final aliquot was added, the mix­ ture was stirred for 18 hours at room temperature.

The product was fractionated through a -95°C trap to remove the last traces of HCl and then condensed into a bulb containing powdered calcium 33 hydride. Vigorous s tirrin g Over this reagent removed excess methanol.

The material was characterized by its infrared spectrum (Figure 14,

Table 1, Appendix), which matched the literature'*^ and its vapor pressure

(155 torr at 28.5°C, lit. 156 torr at 28.5°C‘*^). It was distilled from the hydride prior to use.

Tris(dimethylamino)boron was prepared by reacting boron trichloride with dimethylamine in a 1:6 molar ratio, producing the desired product and dimethyl ammonium chloride. A mixed solvent system, 60:40 CH 2CI2 :

CHCI3, was employed.44

Dimethylamine, 59.5 mmol, was condensed in to a 100 ml standard reac­ tion flask containing 20 ml of the mixed solvent at -196°C and then al­ lowed to warm to room temperature to insure solution of the reagent. The flask was again cooled to -196°C and 9.9 mmol of BCI3 was added to the flask. Warming to -78°C resulted in a solution that could be stirred.

Usually some reaction occurred in the gas phase above the solution at this point coating the walls of the flask with white (CH3) 2NH2C1. A fter one hour at this temperature, the mixture was warmed slowly while moni­ toring the pressure of the system. The reaction could be controlled by the application and removal of a -78°C bath. A fte r the pressure had stabilized, the reaction bulb was thermostatted to 0°C and the mixture s tirre d for 24 hours.

To isolate the product, the mixture was cooled to -35°C and the volatile materials at this temperature removed under static vacuum.

A fte r most of the solvent had been removed in th is manner, the flask was warmed to room temperature and all volatile materials pumped into a -78°C trap with dynamic vacuum fo r several hours. These substances were 34

fractionated through a -35°C trap to remove the remainder of the solvent

and any excess dimethylamine.

The colorless liquid in this trap was identified as the desired prod­ uct by its infrared spectrum (Figure 15, Table 2, Appendix), which matches the literature**^ and vapor pressure (4.4 torr at 25°C, lit. 4.3 torr at

25°C'*3).

Aluminum Compounds. Tetraisobutyldialuminum was prepared by reduc­ tion of diisobutyl aluminum chloride with potassium metal using a procedure similar to that described in the literature.^

Using a tip bulb apparatus, a cyclopentane solution of the aluminum compound (40.0 mmol in approximately 20 mL of solvent) was added slowly to a 100 mL standard reaction bulb containing an equimolar amount (1.56 g) of potassium metal which had been cut in to small pieces in the dry box and covered with approximately 20 ml of cyclopentane. The reaction flask was thermostatted to 0°C and the contents stirred vigorously during the addition. A brown coloration was noticed almost immediately and intensi­ fie d as more of the aluminum compound was added. The addition took about one hour, after which time the cold bath was removed and the stirring con­ tinued for about 90 hours. The product was isolated by filtration through a scintered glass frit and extracted until the solution passing through the fr it was no longer opaque.

The product was checked for purity by NMR spectroscopy in benzene solution (Figure 16, Appendix). The spectrum consisted of a multiplet centered at 2.16 ppm and two doublets centered at 1.25 ppm and 0.25 ppm.

An infrared spectrum is also presented in Figure 17, Table 3, Appendix. 35

An aliquot was taken from the solution of the product and the solvent

removed as described below. The 0.3734 g sample was hydrolyzed with di­

lute HCl and afforded 1.32 mmol of dihydrogen and 5.11 mol of isobutane.

This is a ratio of isobutane to dihydrogen of 3.8 and indicates a product

of 95% purity.

The compound was stored as a cyclopentane solution in a standard

flask. Aliquots were transferred in the dry box, after which the storage

fla s k was attached to the vacuum lin e and the in e rt atmosphere removed.

Such a solution stable for approximately one month at room temperature.

After that time the presence of a significant amount of grey precipitate

could be noticed.

Many attempts were made to obtain a crystalline product from a solu­

tion of the compound. Cooling of the solution slowly from room tempera­

ture to -78°C resulted in no precipitation of the solid product. Removal

of the solvent slowly at room temperature resulted in formation of a dark

red-brown o il.

Heating to 40-45°C eventually increased the viscosity of the o il.

After 48 hours of such treatment, the residue could be more properly termed a glass, but a powder of crystalline substance was never obtained.

Heating to a temperature higher than about 50°C resulted in decomposition, presumably via disproportionation. A grey, m eta llic solid and a color­ less liq u id were formed, aluminum metal and triis o b u ty l aluminum, respec­ tiv e ly . ^

Freeze-drying the product using benzene as a solvent was attempted with no success, the re s u lt being an o il. Mixed solvent systems such as benzene/heptane or cyclopentane/heptane gave no positive results. 36

Column chromatography was attempted to remove trace amounts of im­

purities that were thought to be the cause for formation of an o il. Use

of alumina or Bio-beads (p a r tia lly crosslinked styrene-divinyl benzene co­

polymer) columns, dried under vacuum or dried and pretreated with t r iis o ­

butyl aluminum, resulted in decomposition of the material. Hydrolysis of

the effluent from such a column gave a ratio of isobutanerdihydrogen of

less than 4.0. In several attempts, aliquots gave ratios of 2.65, 1.90,

and 2.35. These decomposition products were not investigated fu rth e r.

The best method of obtaining a pure sample seemed to be heating the

substance under vacuum fo r a rather lengthy period of tim e. In th is case,

the smaller the sample size the more efficient the purification process.

This is illustrated by a molecular weight measurement by the isopiestic method performed in th is laboratory on a small sample treated in such a manner. The results showed the compound to be monomeric in solution and

gave an experimental value that deviated from theoretical by only 4%.*+^

Diisobutyl aluminum methoxide was prepared from diisobutyl aluminum hydride and methanol in a 1:1 molar ra tio . The reaction could be performed without a solvent by virtue of the fa ct that both sta rtin g ma­ terials and the product are liquids.

The aluminum hydride (ca. 1.0 ml), obtained as a heptane solution, was transferred into a tared, evacuated standard reaction bulb by syringe in a glove bag under nitrogen atmosphere. The flask was attached to the vacuum lin e and the in e rt atmosphere and solvent were removed. Remaining was 0.1312 g (0.922 mmol) of diisobutylaluminum hydride.

The flask was cooled to -196°C and a stoichiometric amount (0.921 mmol) of methanol was condensed into the fla s k . The mixture was warmed 37

slowly to room temperature over a period of about one hour with stirring,

affording a colorless liquid.

The v o la tile m aterial was removed to the Toepler system and the non-

condensible gas measured (0.888 mmol) and oxidized at 300°C. All of the

gas proved to be dihydrogen and amounted to 96% of the theoretical y ie ld .

The non-volatile compound was identified as the desired product by

its NMR spectrum (Figure 18, Appendix). An integration of the spec­

trum showed the ratio of methyl groups on the isobutyl ligands (doublets

centered at 1.22 ppm and 1.18 ppm) to methoxy groups (absorbances at 3.31

ppm and 3.19 ppm) to be 4:1, indicating a 2:1 ratio of isobutyl to methoxy

groups bound to aluminum. An infrared spectrum of the product is pre­

sented in Figure 19, Table 4, Appendix.

Dimethyl aminodi isobutyl aluminum was prepared by reaction of d iis o -

butylaluminum hydride with dimethylamine in a 1:1 molar ra tio . The reac­

tion was analogous to that described above fo r the corresponding methoxy

compound.

A sample of the hydride weighing 0.1636 g (1.15 mmol) was treated with an equimolor amount (1.17 mmol) of dimethylamine. Dihydrogen was

evolved as the mixture warmed to room temperature and amounted to 1.06

nmol, which is 92% of the expected quantity. A small amount of unreacted

dimethylamine was tentatively identified by its infrared spectrum as the

remainder of the volatile materials recovered from the reaction.

The NMR spectrum of the product (Figure 20, Appendix) was con­ sistent with the formulation of the material as dimethyl aminodi isobutyl- aluminum. Integration of peak areas gave the expected ra tio of 2:1 fo r

isobutyl (methyl a t 1.13 ppm) to amino (methyl at 2.16 ppm) groups bound 38

to aluminum. The Infrared spectrum is presented in Figure 21, Table 5,

Appendix.

Pentacarbony 1 manganese h.ydride. This compound was prepared accord­

ing to the series of reactions:^?

Mn2 (CO)io + 2 K ^ 2 KMn(CO)s (10)

KMn(CO)s + HCl ■> HMn(C0)5 + KCl (11)

The reduction step was accomplished with potassium metal in THF solu­ tion. A standard reaction bulb was charged with 1.52 g (3.90 mmol) of dimanganese decacarbonyl, weighed in a glove bag. I t was attached to the vacuum lin e and evacuated. (Minimal time was allowed fo r th is procedure as Mn2(CO)io w ill sublime slowly at room temperature.) Approximately

15 ml of THF was condensed into the flask and the mixture was allowed to warm to room temperature and stirred until the solid had dissolved, af­ fording a bright yellow solution.

The flask was taken into the dry box where .305 g (7.80 mmol) of potassium was weighed, cut into small pieces, and added to the reaction bulb. The flask was removed from the box and reattached to the line and the in e rt atmosphere removed. The contents of the flask were stirred vigorously for 36 hours at room temperature. The bright yellow solution was filtered through a scintered glass frit to remove traces of unreacted potassium.

The THF solution of KMn(C0)5 was treated with gaseous HCl to give the final product. This was accomplished by cooling the solution to -196°C and condensing into the flask a to ta l of 7.80 mmol of HCl in three sepa­ rate aliquots. After each addition, the flask was allowed to warm to room temperature and stirred for one to two hours. After the final 39 part had been added and the solution stirred for several hours, the vola­ tiles were removed to the fractionation line. Fractionation was through

-10°C, -45°C, and -95°C traps.

A small amount o f yellow solid stopped in the -10°C trap , which ap­ peared to be unreacted Mn 2(C0 )iQ. The colorless liquid in the -95°C trap was identified as THF by its IR spectrum. The colorless liquid in the

-45°C trap exhibited an IR spectrum (Figure 22, Table 6 , Appendix) iden­ tical to that of HMn(CO)s published in the literature.**® The yield of the reaction was approximately 80%.

The acidification step was also performed on solid KMnfCOOg, obtained by removing the solvent from the filtered reaction solution. HCl was con­ densed into the flask at -196°C incrementally until it could be recovered in total after a suitable reaction time. The product was fractionated as above. The yield using this procedure was only 55%.

It was noticed that upon standing the colorless liquid gradually developed a yellow tinge to i t and a non-condensible gas was produced.

While the literature states that the compound is stable at 25°C, it ap­ pears that some decomposition to Mn2(CO)io and H2 does take place. There­ fore, only freshly prepared HMn(CO)s whose infrared spectrum matched the literature was used for reactions.

Tri(triphenylphosphine)piatinum(O). This reagent was obtained as an impure material of some antiquity and unknown origin. Recrystallization was performed in the dry box to purify the compound. The crude reagent was dissolved in a minimal amount of benzene. The solvent was removed under reduced pressure until formation of a precipitate was just notice­ able. Heptane was then added dropwise, resulting in the precipitation of 40 a bright yellow product. The solid was vacuum filtered through a scin­

tered glass fr it in the dry box and washed with several small portions of

heptane. The solid collected on the fr it was transferred to a tared

standard reaction flask, attached to the vacuum, and pumped on overnight

to remove any residual solvent.

The compound was id en tified by its IR spectrum (Figure 23, Table 7,

Appendix), which matched the lite ra tu re spectrum fo r Pt^PPhgjg.^s I t ap­

peared to be stable fo r long periods of time at room temperature in the

absence of air; no noticeable change in the bright yellow color was seen

in three months.

REACTIONS OF TETRAISOBUTYLDIALUMINUM

With Protonic Reagents

With HCl. A solution of tetraisobutyldialuminum was prepared by dis­

solving 0.1532 g (0.542 mmol) in approximately 10 ml of cyclopentane, which was condensed into a standard flas k containing the dialuminum com­

pound a t -78°C. The mixture was warmed to room temperature and stirred

to insure a homogenous solution.

The solution was frozen to -196°C and 1.18 mmol of hydrogen chloride gas was condensed into the flask. The solution was warmed slowly to room temperature over a period of two hours. Melting of the solvent allowed stirring of the mixture for most of this time.

The contents of the flask were again cooled to -196°C and the non- condensible gas present was pumped into the Toepler system. The Kontes valve of the flask was closed, the solution warmed to room temperature, and the freezing and pumping process repeated to insure quantitative trans­ fe r of the gas. Upon burning at 300°C, the gas was shown to be hydrogen 41

and amounted to 0.400 mmol.

The remaining contents of the flask were thermostatted to -78°C and the volatile materials at that temperature were removed by dynamic pump­

ing to a -196°C trap on the vacuum line and eventually transferred to the fractionation train. Fractionation of these materials was effected through traps maintained at -95°C, -126°C, and -196°C. Cyclopentane was found in the -95°C trap, while the -126°C trap contained isobutane; the la tte r amounting to 0.38 mmol. These species were id e n tifie d by th e ir respective infrared spectra. No unreacted HCl was isolated in the -196°C trap.

The residual solution had a very faint red-brown color to it and a grey solid had separated. This mixture was extracted through a scintered glass frit. The solid remaining on the frit was exposed to air with no apparent reaction taking place. It did not react with distilled water, but gas was evolved and the solid eventually dissolved when treated with dilute HCl. Upon neutralization, the solution gave a yellow precipitate with 8 -hydroxyguinoline. Quantitative analysis of this precipitate found

0.09 mmol of aluminum metal as the grey solid of the reaction.

The soluble portion of the reaction mixture was attached to the vac­ uum line and the cyclopentane remaining was removed and replaced with approximately 5 ml of benzene. A NMR sample was prepared in the dry box and the spectrum obtained is shown in Figure 24 (Results and Discus­ sion). An IR spectrum (Figure 25, Table 8 , Results and Discussion) was obtained by allowing a drop of the solution to evaporate on KBr plates in the dry box. Hydrolysis (n e u tra l) o f the remaining solution produced no hydrogen, but 1.60 mmol of isobutane. The hydrostat gave a voluminous 42 white precipitate when treated with a solution of silver nitrate.

With Dimethylamine. A standard reaction flask with 0.3184 g (1.13 mmol) of tetraisobutyldialuminum in 15 ml of cyclopentane was cooled to

-196°C and 2.20 mmol of dimethylamine condensed into the flask. The mix­ ture was warmed to 0°C and s tirre d for two hours.

After that period, the volatile materials were examined as before.

A non-condensible gas was found which amounted to 0.351 mmol and proved to be dihydrogen upon oxidation. The volatile materials that could be re­ moved at -78°C were fractionated and found to be cyclopentane and dimethyl­ amine by their respective infrared spectra. No isobutane was found.

These volatile materials were condensed back into the reaction flask and the mixture stirred at room temperature for an additional 20 hours.

Treatment of volatile materials was as before after this period. Hydrogen amounted to 0.133 mmol on this occasion and again no isobutane was found.

Volatile materials were again recondensed into the flask and an ad­ ditional 2.15 mol of dimethylamine was added to the reaction mixture.

The reaction was allowed to proceed at room temperature and intermittent checks made for non-condensible and condensible products over a period of

20 days. Table 9, found in the Results and Discussion section, gives the results o f the analyses for hydrogen. In no case was isobutane found.

After seven days a small amount of a grey solid was noticed in the bottom of the flask. A fter 20 days the amount of grey p recip itate had increased several-fold. V o la tile m aterials were checked once again and the solution, s till a dark red-brown in color, was hydrolyzed (acidic).

Dihydrogen (0.313 mmol) was found, as well as 4.60 mmol of isobutane, identified by its infrared spectrum. 43

In a separate reaction the ratio of reactants was initially 4:1,

3.75 mmol dimethylamine was added to 0.94 mmol of tetraisobutyldialuminum.

The solvent in this case was benzene and the reaction mixture was heated to 43°C.

The analytical procedure was the same as in the previous reaction.

The record of dihydrogen evolution from the reaction is given in Table 10 in the Results and Discussion section. As in the reaction at lower tem­ perature, the only gas produced was dihydrogen, ultimately totaling 0.63 nmol. Also, as before, the reaction mixture retained its dark red-brown color and a grey precipitate was noticed to form during the reaction.

With Ethyl amine. Ethyl amine was reacted with tetraisobutyldialuminum in a 1:1 mole ratio (0.775 mmol of each in benzene as before). The reac­ tion proceeded for two hours at 18°C, afte r which time the flask was cooled to -196°C and the non-condensible material pumped into the Toepler system. This gas burned as dihydrogen at 300°C and amounted to 0.269 mmol. Dihydrogen evolution was checked a fte r a to ta l of three hours, fo l­ lowing which the mixture was heated to 40°C. The reaction was allowed to proceed fo r an additional 45 hours and checked p erio d ically for produc­ tion of dihydrogen. The reaction was terminated when the solution was seen to be no longer opaque and a grey precipitate was noticed in the bot­ tom of the flask. Table 11 in the Results and Discussion section gives the results of the analysis for dihydrogen evolved during the reaction.

After ten hours total reaction time, the mixture was checked for volatile materials at -78°C, in addition to the normal procedure to col­ lect generated dihydrogen. No ethyl amine was found but 0.017 mmol of iso­ butane was found and identified by its IR spectrum. At this time, an 44

additional 0.773 mmol of ethylamine was added to the flask to bring the

original mole ratio of reactants to 2:1 (amine:Al 2). No fu rth er produc­

tion of isobutane was noticed throughout the remainder of the reaction.

With Dimethylamine Adduct of Diisobutyldimethylaminoaluminum. Di-

isobutyldimethylaminoaluminum (2.80 rrniol) was prepared from diisobutyl- aluminum hydride and dimethylamine as described above. An equimolar amount of dimethylamine was condensed into the reaction flask containing the freshly prepared aluminum compound forming a white solid. Approxi­ mately 10 ml of benzene was added to the s o lid , which dissolved, and a

NMR sample was taken from this solution in the dry box (Figure 26,

Results and Discussion). The remainder of the solution was added via a tip bulb arrangement to approximately 1.4 mmol of tetraisobutyldialuminum in 10 ml of benzene. The mixture was stirred for two hours at 27°C. Di- hydrogen was removed a t -196°C and determined by oxidation a t 300°C, amounting to 0.007 mmol.

After 24 hours stirring at room temperature, the Toepling procedure was again performed and 0.044 mmol of dihydrogen was found.

At this time, a charge of 1.40 mmol of dimethylamine was added to the reaction fla s k by condensation of the gas at -196°C and the mixture s tirred for six hours at 27°C. Non-condensible gas was removed from the flask at -196°C after that period and amounted to 0.039 mmol of dihydrogen.

The remaining materials in the flask were thermostatted to -78°C and volatile materials at that temperature pumped into a trap held at -196°C with dynamic vacuum. No separation was achieved upon fractionation through traps at -95°C , -126°C, and -196°C. An IR spectrum showed only dimethylamine to be present; no isobutane was found. The reaction was 45

terminated at this point and a NMR sample prepared in the dry box of

the solution remaining in the reaction flask and was virtually the same

as Figure 26, A very small amount of a grey precipitate that may have

been aluminum metal was noticed in the reaction flask at this time.

With Toluene-3,4-Dithiol. A tip bulb apparatus was employed to add

0.1558 g (0.997 nmol) of to lu e n e -3 ,4 -d ith io l, which had been dissolved in

approximately 10 ml of cyclopentane to a rapidly stirred solution of the

dialuminum compound, 0.293 g (1.04 mmol), in 10 ml of the same solvent.

The lower bulb was thermostatted to 0°C to facilitate the addition, which was accomplished ra p id ly . A fter condensing some solvent back into the upper bulb to wash any residual dithiol into the reaction flask, the reac­ tion was thermostatted to 0°C and allowed to proceed for eight hours.

After that period, the mixture was treated in the same manner as the HCl reaction previously described to isolate and identify non- condensible and condensible volatiles generated in the reaction. Dihy­ drogen was found and amounted to 0.627 mmol. Isobutane was also found and amounted to 0.480 mmol.

The remaining mixture consisted of a dark red-brown solution and a light brown solid. Extraction through a scintered glass frit left the insoluble material on the frit. An IR spectrum of this solid was obtained as a nujol mull (Figure 27, Table 12, Results and Discussion). Upon pul­ verization to prepare the mull, it appeared as though the solid was a mix­ ture of the brown solid originally noticed and a grey solid that was as­ sumed to be aluminum metal from its appearance. Efforts to dissolve the solid in THF and diethyl ether to affect a further separation and obtain a NMR of the brown m aterial were unsuccessful. 46

A NMR was obtained of the soluble portion of the reaction product

a fte r changing the solvent from cyclopentane to benzene (Figure 28, Re­

sults and Discussion).

Reactions with Aluminum and Boron Compounds

With Trimethyl aluminum. A standard reaction flask containing 0.1680 g

(0.594 mmol) of tetraisobutyldialuminum was cooled to -196°C and 0.821 mmol

of trimethylaluminum and approximately 10 ml of cyclopentane were condensed

into the flask. The mixture was warmed slowly to room temperature with

s tir r in g . A fter two hours a grey p rec ip itate was seen in the bottom of

the flask. The mixture was thermostatted to 0°C and stirred overnight. An

increased amount of grey precipitate was noticed after this time. The

solution was s t i l l a dark red-brown in colo r.

Separation of the two components was effected by filtratio n through

a scintered glass frit with extraction of the solid until the solution

passing through the fr it was no longer colored. The grey solid that re­ mained on the fr it gave no reaction when exposed to air and dissolved with

evolution of gas when treated with aqueous HCl. A yellow precipitate ap­

peared upon addition of 8-hydroxyquinoline to the neutralized solution.

A NMR sample was prepared from the red-brown solution but no well-

resolved spectrum was obtained. Another attempt was made by replacing most of the original solvent with deuterobenzene with similar results.

Both tubes contained some solid material.

Aqueous hydrolysis of an aliquot of the solution gave non-condensible gases th at burned as dihydrogen (0.05 mmol) and methane (1.01 mmol). Con­ densible volatile material was fractionated through a -78°C trap and shown

to be only isobutane by infrared spectroscopy, amounting to 0.973 mmol. 47

The remainder of the red-brown solution lo st its color a fte r a week.

In a separate reaction, 0.1941 g (0.687 mmol) of tetraisobutyldia­ luminum in 5 ml of cyclopentane was treated with 2.65 mmol of trimethyl­ aluminum by condensing the la tte r into a flask containing a solution of the dialuminum at -196°C. The mixture was thermostatted to 0°C and stirred for 1-1/2 hour. After that period, the solvent and other vola­ t i l e m aterials were removed under dynamic vacuum fo r nine hours a t 0°C and two hours at 25°C.

The volatile materials were fractionated through -22°C and -78°C traps to a -196°C trap. Cyclopentane and trimethylaluminum were found and identified by their respective infrared spectra. The trimethylalumi­ num was also hydrolyzed and afforded only methane as a volatile product.

The non-volatile material appeared to be a mixture of a brown oil and a grey solid. Approximately 10 ml of cyclopentane was condensed onto the sample and the mixture filte r e d and extracted once by condensing sol­ vent back into the flask to wash the solid.

The grey solid gave the same results when treated in the same manner as above. The red-brown solution gave a NMR spectrum (shown in Fig­ ure 29, Results and Discussion) and hydrolysis of the remainder solution produced less than 0.04 mmol o f dihydrogen and 1.28 mmol of methane.

Isobutane was also found but not measured quantitatively.

Attempted Preparation of Lewis Base Adducts o f Tetraisobutyldialum i­ num. Several attempts were made to prepare adducts of the dialuminum compound with donor molecules. The firs t attempt involved addition of approximately 5 ml of diethyl ether to a sample of tetraisobutyldialumi­ num. After stirring for one hour a grey precipitate was noticed. The 48 solution was filtered through a scintered glass frit to remove the solid and stirring continued for an additional 15 hours. More grey precipitate was noticed after that period and the mixture discarded.

In a second attempt, two equivalents (1.84 mmol) of diethyl ether were added to a sample of 0.920 mmol of the dialuminum compound in sev­ eral ml of cyclopentane. The mixture was stirred at room temperature for ten minutes and then the v o la tile materials were removed. Fractionation through a -78°C trap permitted identification of the materials as cyclo­ pentane and diethyl ether.

Several more attempts were made using THF as the donor compound and gave similar results. Cursory treatment of the dialuminum starting mate­ r ia l with THF seemed to give no adduct stable under dynamic vacuum; the

THF could be recovered almost entirely. Prolonged contact of the two reagents resulted in separation of a precipitate from the solution that appeared to be m etallic aluminum.

Finally, adduct formation was attempted with 4-picoline. Tetraiso­ butyldialuminum (0.156 g, 0.552 mmol) was treated with approximately 1 ml of 4-picoline in approximately 5 mL of benzene. A fter 15 minutes of stirring at room temperature, the volatile materials were removed with pumping for nine hours.

The non-volatile residue was dissolved in benzene and a ^H NMR sam­ ple was prepared in the dry box. The spectrum is shown in Figure 30 (Re­ sults and Discussion). Some solid was noticed in the sample tube after one day.

With Boron Trichloride. A benzene solution of 0.2816 g (0.997 mmol) of tetraisobutyldialuminum was treated with 4.88 mmol of boron trichloride 49 in approximately 1 nmol increments. After each increment was added, the mixture was warmed to room temperature and stirred for 24 hours. Prior to addition of the second and subsequent portions, the vapor pressure was checked for excess boron tric h lo rid e . A grey p re cip ita te was noticed a fte r the f ir s t increment was added and the amount was seen to increase as the reaction proceeded.

V o la tile materials were removed under dynamic vacuum for 30 minutes.

Benzene and boron tric h lo rid e were found and id e n tifie d by th e ir infrared spectra. The la tte r amounted to 0.812 mmol.

The non-volatile portion of the reaction mixture appeared to consist of a brown o il and a grey s o lid . The solvent was recondensed into the flask and a NMR sample prepared in the dry box by filtering the solu­ tion through glass wool into the sample tube. The spectrum is shown in

Figure 31 (Results and Discussion).

Hydrolysis of an aliquot of the residue solution with water produced dihydrogen (0.19 mmol) and isobutane (2.75 mmol). A grey solid remained in the bottom of the flask and appeared not to react with water.

The aliquot not hydrolyzed lost its intense color over a period of time (about two weeks) and eventually consisted of a faint brown solution and a grey precipitate.

With Trimethoxyboron. A cyclopentane solution of tetraisobutyldia­ luminum (0.1310 g, 0.464 mmol) was treated with 0.302 mmol of trimethoxy­ boron by condensing the la tte r into a standard fla sk containing a solu­ tion of the former at -78°C. The mixture was stirred for 25 hours at room temperature and then filte r e d through a scintered glass f r i t as in previous reactions. Passing through the f r i t was an opaque red-brown 50

solution, while remaining on the frit was a brown solid. The solid was

extracted three times with the solvent until the solution passing through

the f r i t was no longer opaque.

Attempts to mull the solid with nujol showed it to be a mixture contain­

ing a grey solid, as well as the brown material originally observed.

Upon exposure to the air, the brown solid lost its color almost immedi­

a te ly but no change in the appearance of the grey material was noticed.

The flask containing the solution was reattached to the vacuum line

and the v o la tile m aterials removed under dynamic vacuum. These m aterials

were fractionated through -45°C and -78°C traps. The material of higher

v o la t ilit y proved to be cyclopentane. The m aterial of lower v o la tility

gave an infrared spectrum that matched that of trimethoxyboron (Figure 14,

Table 1, Appendix). The NMR spectrum o f th is fraction (Figure 32, Re­

sults and Discussion) showed other material to be present.

The residue that remained in the flask was a dark red-brown o il. In

the dry box an infrared sample was prepared by placing some of the oil

between KBr disks. Figure 33, Table 13, Results and Discussion, shows

the spectrum obtained. A NMR sample was prepared by dissolving the

oil in benzene and pipetting some of the solution into a sample tube.

The spectrum is presented in Figure 34 (Results and Discussion).

The remainder of the solution was hydrolyzed with aqueous HCl and

produced 0.290 mmol o f dihydrogen and 0.568 tranol of isobutane. Methanol was also identified as a hydrolysis product by its infrared spectrum but was not q u an titatively separated from the aqueous hydrostat.

In another reaction, 0.4434 g (1.57 mmol) of tetraisobutyldialuminum was treated with 14.24 mmol of trimethoxyboron in a standard reaction 51

fla s k . No other solvent was employed; the amount of boron compound pro­

vided a mixture that could be stirred without additional solvent present.

The reaction was allowed to proceed for eight days at room temperature.

After that period, the flask was attached to a filtration apparatus

and some of the solution drawn through the f r i t . This dark red-brown

solution was washed into the receiver bulb and the red-brown o il that re­

mained on the f r i t washed back into the original reaction flask. The

apparatus was taken into the dry box and the receiver bulb replaced with

another containing to the vacuum line and the volatile materials removed

under dynamic vacuum for eight hours at room temperature. Fractionation

In the dry box an infrared sample was prepared of the oil that re­

mained in the flask by placing some of that viscous material between two

KBr plates. The spectrum was virtually identical to that obtained in the

previous reaction and shown in Figure 33, Table 13, Results and Discus­

sion. After running the spectrum, the material was exposed to the air

and the red-brown color was immediately discharged. No traces of grey

or black material were present that might indicate the presence of metal­

lic aluminum.

Another small amount o f the oil was dissolved in benzene to prepare

a NMR sample. The spectrum is presented in Figure 35 (Results and

Discussion).

The bulb containing the remainder of the oil was reattached to the

vacuum lin e and pumped on overnight at room temperature, leaving a sample

th a t appeared to be unchanged and which weighed 0.217 g. This material was hydrolyzed (pH<7) and afforded 0.479 mmol of dihydrogen and 1,90 mmol 52

of isobutane. Again, methanol was found to be present but could not be

isolated quantitatively.

A flame test of the hydrostat showed no boron to be present by the

absence of a green flame. An aluminum analysis on the hydrostat by

gravimetric determination of the 8-hydroxyquinolate found the sample to

contain 1.418 mmol o f aluminum.

With Tris(dimethylamino)boron. A standard reaction flask containing

0.2665 g (0.944 mmol) of tetraisobutyldialuminum was cooled to -196°C and

10 ml of cyclopentane added. The mixture was warmed to room temperature

and stirred for several minutes to give a homogeneous solution. This solu­

tion was cooled to -78°C and 5.98 mmol of tris(dimethylamino)boron was

added. The mixture was warmed to room temperature and stirred for two

hours. After this period, some solid was noticed on the walls of the

fla s k . I t was cooled to -196°C and approximately 8 mmol of THF was con­

densed onto the reaction mixture. The flask was again warmed to room

temperature and stirred for an additional 12 hours.

Volatile materials were removed from the reaction mixture at -22°C

under s ta tic vacuum fo r two hours. These m aterials proved to be THF and

cyclopentane from their infrared spectra.

The residue in the flask was a dark red-brown o il. It was warmed to

room temperature and pumped on with dynamic vacuum fo r six hours to re ­

move any low v o la tile m aterial present. An infrared spectrum was obtained which matched that of the starting boron compound (Figure 15, Table 2,

Appendix). A ^H NMR sample (Figure 36, Results and Discussion) was ob­

tained of the small amount of colorless liquid removed in this manner.

The appearance of the residue did not change except to become more viscous. 53

Approximately 5 mL of benzene was added to the residue and the flask

attached to a filtration device and the solution filtered. Extraction

with several portions of the solvent left a dark brown material on the

frit and a small amount of a light brown solid on the walls of the reac­

tion fla s k . The solid on the f r i t turned white upon exposure to the a ir

The red-brown solution that passed through the fr it was taken into

the dry box and a NMR sample prepared. The spectrum obtained is shown

in Figure 37 (Results and Discussion).

A portion of the solution was hydrolyzed with aqueous HCl and pro­

duced 0.798 mmol of dihydrogen and 0.985 mmol of isobutane. A small

amount of dimethylamine was also identified by its infrared spectrum as

a product of this hydrolysis. The hydrostat was tested for the presence

of boron by a flame te s t, which proved to be po sitive.

The remainder of the solution was attached to the vacuum line and

the solvent removed at room temperature. The contents of the flask were

then heated to 45°C with dynamic pumping. A fter 38 hours, the material

seemed to have undergone decomposition as evidenced by the presence of

a grey solid in the flask. A very small amount of a colorless liquid was

volatized in this procedure. It exhibited a positive test for the pres­

ence o f boron. No further work was performed on th is sample.

In a separate reaction, 0.2177 g (0.771 mmol) of tetraisobutyldi­

aluminum was treated with 3.08 mmol of tris (dimethylamino)boron and ap­

proximately 3 mL of benzene to afford a solution that could be stirred.

The reaction was allowed to proceed for 1-1/2 hour at room temperature.

After that time, the solvent and any other volatile materials were

removed at room temperature. The flask was then heated to 32°C under 54 dynamic vacuum for 18 hours, leaving a viscous brown o il.

Approximately 10 mL o f benzene was recondensed onto this residue and the resulting solution filtered through a scintered glass frit as before.

The grey solid on the fr it was washed twice with the solvent. In the dry box, samples were prepared to obtain infrared and NMR spectra. These are presented in Figure 38, Table 14, and Figure 39, respectively, in the

Results and Discussion section.

The flask containing the remainder o f the solution was reattached to the vacuum line and the solvent removed at room temperature for 3 hours.

Aqueous acidic hydrolysis produced 0.418 mmol of dihydrogen and

1.985 mmol of isobutane. Some dimethylamine was found and identified by its infrared spectrum but th is material was not isolated q u a n tita tiv e ly .

Reactions with Transition Metal Compounds

With Dimanganese Decacarbonyl. A cyclopentane solution containing

0.3276 g (1.16 mmol ) of tetraisobutyldialuminum was added via a tip bulb arrangement to a cyclopentane solution containing 0.2274 g (1.16 mmol) of dimanganese decacarbonyl. Solvent was condensed into the upper bulb to insure quantitative transfer of the dialuminum reagent.

The mixture was stirred for about three weeks at room temperature and checked occasionally for the presence of volatile materials. This was accomplished by cooling the solution to -78°C and applying a dynamic vacuum to the flask via the Toepler system. No materials volatile at th is temperature were found in these checks.

Visual examination of the solution found the appearance o f a grey p recip itate in small quantity to be present a fte r about two weeks.

Whether the amount of this material increased over the remainder of the 55 reaction period could not be determined.

After 22 days the solvent was removed from the reaction flask at

0°C, leaving a dark brown residue in the flask. Pumping at room tempera­ ture for several hours removed a yellow solid (weighing 0.2201 g) from the fla s k . An in frared spectrum of this solid was identical to th a t of the starting dimanganese compound.

With Pentacarbonylmanganese Hydride. A standard reaction flask con­ taining a sample (0.3760 g, 1.33 mmol) of tetraisobutyldialuminum was cooled to -78°C and approximately 20 mL of benzene and 0.5358 g (2.73 mmol) of pentacarbonyl manganese hydride were condensed into the bulb. The mixture was warmed to room temperature and stirred for one hour.

After that time the flask was again cooled to -78°C and the volatiles at that temperature removed to the Toepler system. A non-condensible gas was found, amounting to 0.03 mmol, which burned as dihydrogen. A con­ densible m aterial was also found which to ta le d less than 0.01 mmol.

The remaining mixture in the flask was stirred at room temperature for an additional six days. The same treatment was then applied to de­ termine volatile materials. An additional 0.16 mmol of dihydrogen was isolated. The condensible volatile material now totaled 0.19 mmol and was determined to be isobutane by infrared spectroscopy.

The opaque red-brown solution that remained was filtered through a scintered glass f r it , leaving a brown residue on the fr it which was washed six times with the solvent. An infrared spectrum of this solid was ob­ tained as a nujol mull (Figure 38, Table 15, Results and Discussion). The material proved to be insoluble in cyclopentane and THF, as well as ben­ zene. A NMR sample was prepared from the solution and is shown in 56

Figure 39 in the Results and Discussion section.

The flask containing the remainder of the solution was reattached to

the vacuum line and the volatile materials removed at room temperature

and fractionated through traps at -10°C and -22°C to a -78°C trap. The

first trap contained a small amount of a yellow solid. The second trap contained a colorless liquid (when warmed to room temperature) which

proved to be predominantly pentacarbonylmanganese hydride by comparison of its infrared spectrum with that of the starting material. The highest v o la tile material was found to be benzene.

The yellow solid was present in too small of an amount to provide a sample fo r spectroscopic examination, but was assumed to be dimanganese decacarbonyl from its color and properties.

Wi th Tri s(triphenylphosphi ne)piati num. A benzene solution (approxi­ mately 20 mL) containing 0.9929 g (1.01 mmol) of tris(triphenylphosphine)- platinum was added to a sample of tetraisobutyldialuminum weighing 0.2932 g (1.04 mmol) via a tip bulb arrangement. The mixture was warmed to room temperature and stirred for 7 days. Intermittent checks were made for the presence of v o la tile m aterials at -78°C as in previous experiments.

A fter the reaction period, the benzene was removed with dynamic vacuum and replaced with an approximately equal amount of cyclopentane.

The resulting mixture of yellow sol id "arid red-brown solution was filte re d through a scintered glass f r i t . The solid remaining on the f r i t was ex­ tracted by washing with the solvent four times until the solution passing through the f r i t was a bright yellow.

In the dry box, the receiver bulb was attached to another filtration device, the second bulb of which contained about 15 mL of heptane. This 57

new apparatus was attached to the vacuum line and the cyclopentane re­

moved, leaving a brown oil in the flask. The heptane was transferred to

the flask and afforded a red-brown solution and a yellow solid. Filtra­

tion was again effected, leaving the yellow solid on the frit.

A NMR sample was prepared from th is solution and the spectrum ob­

tained is shown in Figure 40, Results and Discussion. An infrared

spectrum of the yellow Solid from the second filtratio n was obtained as a

nujol mull and matched the spectrum of tris(triphenylphosphine)platinum

(Figure 23, Table 7, Appendix).

An infrared spectrum of the in it ia l s o lid , from the f ir s t f ilt r a t io n , was obtained as a nujol mull after that material was exposed to air

through a crack in the KBr plates. The spectrum matched that of bis-

(triphenylphosphine)platinum oxide prepared by exposure of an authentic

sample of the platinum starting material to air (Figure 41, Table 16, Ap­

pendix). A small amount of grey solid was noticed to be intermixed with

the yellow-brown solid upon mulling this sample.

An aliquot of the solution was hydrolyzed and produced 0.165 mmol of dihydrogen and 2.16 mmol of isobutane. Treatment of the hydrostat with concentrated aqua regia, followed by sodium formate and formic acid, gave no precipitate indicative of the presence of platinum. RESULTS AND DISCUSSION

INTRODUCTION

The chemistry of tetraisobutyldialuminum was examined in light of reactions designed to induce ligand exchange. The results were dominated by data from which one may in fe r that decomposition is the major reaction pathway, e ith e r by direct oxidation of the aluminum-aluminum bond or via a disproportionation involving the Al-Al bond. Both reactions result in the production of trivalent aluminum compounds (oxidation state +3). The latter reaction also results in the formation of aluminum metal.

Some s ta b iliz a tio n can be afforded the low valent aluminum compounds by substitution with ligands that can be considered pi donors. The novel product isolated in the reaction of tetraisobutyldialuminum with tr i­ methoxyboron showed substitution, yet maintained in tact the aluminum- aluminum bond. I t was not a simple replacement product but appeared to involve some association of trivalent aluminum compounds with the dialumi­ num species. It is thought that this type of interaction is important in s ta b iliz a tio n of the aluminum-aluminum bond.

Identification of the species generated from the various reactions presented a d if f ic u lt task. Indication of the presence of an aluminum- aluminum bonded species could be inferred from the intense red-brown color of the solution. However, the progress of a reaction could not be fo l­ lowed visually due to the intensity of this coloration. For example, in

58 59

preparation of the starting tetraisobutyldialuminum, the solution would

become opaque after only a small portion of the diisobutylaluminum chlo­

ride was added to the potassium. Likewise, when oxidation or dispropor­

tio nation of the aluminum-aluminum bond occurred, the solution was opaque

even after most of the dialuminum species had been determined to have

reacted.

Information from hydrolysis of the reaction solutions was thus the

only common method of determining the extent of any partial degradation

of the dialuminum species. Unfortunately, this is by d efin itio n a de­

structive technique.

Infrared spectroscopy was helpful in determination of the presence

of various ligands but not in definite characterization of new dialuminum

species. For example, compare the spectra of tetraisobutyldialuminum and

triisobutyl aluminum (Figure 17, Table 3, and Figure 43, Table 18, respec­

tively, Appendix). Virtually all of the bands seen are attributable to

the ligands and therefore the spectra are almost identical. Some subtle

shifts in band position may be seen, but for the most part this tool

could be used only in a qualitative manner. In many instances, absences

of p artic u la r bands were o f more help than the absorbances actually pres­

ent.

Proton NMR spectra were also employed in this work. Once again, the

results were less than straightforw ard, however. In theory, the chemical

shift of the methylene protons of an isobutyl ligand is in part determined

by the electronic nature of the aluminum atom to which the ligand is bound.

Comparison of the spectra of tetraisobutyldialuminum (Figure 16, Ap­ pendix) and triisobutylaluminum (Figure 42, Appendix) shows this to be 60 true. The doublet assigned to those protons in the former compound is centered at 0.25 ppm, while the corresponding absorbance fo r the triv a - lent aluminum species is centered at 0.41 ppm. The hypothesis here is that the methylene protons of tetraisobutyldialuminum are more shielded due to the higher electron density that would be associated with a low valent aluminum atom. Hence, the protons exh ib it a chemical s h ift at higher fie ld in the dialuminum compound.

Comparisons involving other aluminum species cannot be made in such an unambiguous manner. The chemical s h ift is influenced by the sum of all of the magnetic fields about the particular nucleus in the molecule.

In the case of the two isobutyl species mentioned above, the molecules are seen to be very similar, the only important difference being the pres­ ence of an aluminum-aluminum bond in the dialuminum species rather than a third aluminum-carbon bond as in triisobutylaluminum. The electronic na­ ture of the aluminum can thus be assumed as the determining factor in the re la tiv e position of the absorbance.

In other species having substituents in addition to isobutyl groups, the e ffe c t would become much more complicated and q u an titative assessment of the contributions of the various factors would be impossible to cor­ relate.so Table 17 is a compilation of the chemical shifts measured ex­ perimentally fo r the methyl and methylene protons of the isobutyl groups in some of the compounds examined in th is work.

An interesting result is found if the difference in chemical shift between the methyl and methylene protons is calculated; the values are listed in the A column of Table 17. One purpose this serves is to mini­ mize errors in calculation of the absolute values of the chemical shifts 61

Table 17

Proton NMR Data fo r Isobutyl Aluminum Compounds.

Compound Methyl‘S Methylene* at Figu

A12 (i-C^Hg)^ 1.25 0.25 1.00 16

Alfi-C^Hgjg 1.11 0.41 0.70 42

Al(i-C^Hg)(CH3)2 1.17 0.29 0.88 29

A1 ( i ”Ci).Hg)2Cl 1.16 0.57 0.59 44

Al(i-CuHg)20CH3 1.18 0.22 0.96 18

A1 (i-Ci+Hg)2N(CH3)2 1.21 0.25 0.96 20

A l2 (i”C4Hg)i|*NCgH7^ 1.38 0.58 0.80 30

^Center of doublet, in ppm.

^Difference in chemical sh ift, methyl resonance minus methylene reso­ nance.

^4-picoline adduct(?) of tetraisobutyldialuminum. 62

that might arise due to concentration effects and machine inconsistencies

from sample to sample.

A trend that is established from this investigation is that the A

value is larger for the species that have more electron density asso­

ciated with the aluminum atom. Tetraisobutyldialuminum and triisobutyl­

aluminum follow this pattern, giving values of 1.00 and 0.70, respec­

tiv e ly . Species with ligands th a t are good pi donors also seem to f i t

this trend, as seen with the methoxy and dimethyl amino-substituted com­

pounds. The quantitative aspect is lacking, as the argument above pre­

supposes, as evidenced by the fa c t that the value is the same fo r both of

these species, 0.96.

Another factor that would complicate interpretation of the A value

of the various compounds is hybridization of the aluminum atom. The

metal atoms of triisobutylaluminum and tetraisobutyldialuminum are both

three-coordinate and thus should exhibit sp2 hybridization. The other

aluminum compounds listed e x is t as dimers or higher aggregates and have

metal atoms with sp^ hybridization. This situation would be reflected in

a difference in overlap between the aluminum and the various ligands (sp2

hybridization allows for better overlap than sp3 hybridization) and

therefore differences in the transmission of electronic effects.

This second effect appears to be less important than the effect of

basicity or pi donating ability as seen by the a values of dimethyliso-

butylaluminum (0.88) and diisobutyl aluminum chloride (0.59) which are, respectively, greater and smaller than the A value of triisobutylaluminum

(0 .7 0 ). No mathematical treatment of the re s u lts , taking into account these effects, could be constructed to permit prediction of the a values. 63

The method of difference in chemical s h ift has precedence in other work. The change in in ternal chemical s h ift (derived as the a value here) was studied in connection with the strength of the donor-acceptor bond in a number of adducts o f Lewis bases and aluminum and other Lewis acid spe­ c i e s . 3 Trends followed data obtained by thermochemical analysis in some cases when a larger internal shift was found for the stronger ad­ ducts. Inconsistencies arose due to solvent effects and, in some cases, the trends were reversed in different solvents. As with the examination here, the best that could be said was that the results were seen to be self-consistent.

Therefore, for the purposes of this work, comparisons of NMR spectra must be made in a qu alitative manner. Trends drawn should in ­ volve spectra of compounds as similar as possible. Also, this tool should be used in conjunction with other data to support conclusions.

REACTIONS WITH PROTONIC REAGENTS

Tetraisobutyldialuminum was treated with a variety of protonic rea­ gents: hydrogen ch lo rid e, dimethylami ne, ethyl ami ne, and to lu en e-3,4- d ith io l. The major reaction in a ll cases was oxidation of the aluminum- aluminum bond which produced dihydrogen. Carbon-aluminum bond cleavage occurred to a varying extent, giving isobutane as a product. Some rear­ rangement reactions are believed to have occurred in solution, resulting in the fin a l lo w -v o la tile products that were id e n tifie d . One o f the rearrangement reactions was part of a disproportionation path of decompo­ s itio n of the dialuminum starting m aterial. I t proceeded at an acceler­ ated rate compared to that of the untreated tetraisobutyldialuminum. 64

With HCl

Treatment of tetraisobutyldialuminum with HCl in a 2:1 ratio resulted

in the oxidation of the aluminum-aluminum bond. Some carbon-aluminum bond

cleavage also occurred. All of the hydrogen originally added in the form

of HCl appeared as dihydrogen and isobutane. The non-volatile products

were identified as a mixture of isobutylaluminum chlorides and aluminum

metal. The reaction was rapid and appeared to be completed in minutes at

a temperature below ambient.

The highly v o la tile by-products were id e n tifie d as dihydrogen and

isobutane. A ll o f the hydrogen added as HCl was accounted fo r in the

form of these two species, 68 percent as dihydrogen and 32 percent as

isobutane. The quantity of dihydrogen found indicates that 74 percent of

the aluminum-aluminum bonds o rig in a lly present were oxidized according

to Equations 12 and 13:

Alzfi-CtHg)^ + HCl -> Alfi-C^HgjgCl + Alfi-C^HgigH (12)

Alfi-C^HgizH + HCl 4. Al(i-CitHg)2Cl + Hg (13)

The remainder of the HCl went to cleavage of aluminum-carbon bonds and

production of isobutane, as illustrated in Equation 14:

Al-(i-Ct,Hg) + HCl 4- Al-Cl + i-C^Hio (14)

The quantity of isobutane found indicates that 17.5 percent of the

aluminum-carbon bonds o rig in a lly present reacted in this manner.

No isobutene was found which would have resulted from a beta e lim i­

nation reaction, as shown in Equation 15:

Al-(i-C^Hg) ^ Al-H + i-C^Hg (15)

Hydrolysis of the colorless solution that remained after the reaction with HCl produced an amount of isobutane that was consistent with the 65

quantity of isobutyl groups expected to be intact after the initial reac­

tion (allowing for the portion of the solution used to prepare a NMR

sample). That is, the rest of the isobutyl groups were converted to iso­

butane by the hydrolysis reaction. No dihydrogen or isobutene was gen­

erated in the hydrolysis of the reaction solution.

Aluminum metal was found as a product of the in itial interaction of

HCl and the dialuminum compound. The amount recovered from the reaction

was 8.3 percent of the total aluminum originally present in the reaction

mixture. If a disproportionation reaction is responsible for the pro­

duction of aluminum metal, which w ill be examined below, i t would be

represented by Equation 16:

3A1(II)-A1(II) 4. 4A1(III) + 2A1(0) (16)

The stoichiometry above requires that 25 percent of the dialuminum com­

pound o rig in a lly present underwent disproportionation.

Thus, i t appears th at a ll of the aluminum-aluminum bonds in the in i­

tial solution react, three quarters via oxidation and one quarter via

disproportionation.

The infrared spectrum of the soluble reaction products (the solvent

was allowed to evaporate from a drop of the solution placed on a KBr disk

in the dry box) is shown in Figure 25, Table 8. It is virtually identical

to that of triisobutylaluminum (Figure 43, Table 18, Appendix), with the

addition of two bands at 473 cm'i and 498 cm“^, which are assigned to

aluminum-chloride stretching modes with regard to the spectrum of eth yl-

aluminum dichloride 486 cm-i and 502 cm -i.63

The iH NMR spectrum of the non-volatile reaction solution (Figure

24) consists of four sets of doublets which are consistent with the ir

Figure 24. Proton NMR Spectrum of Reaction Products from Treatment of Tetraisobutyldialuminum with HCl. 4000 3500 3S00 3000 1800 1400 eoo

Figure 25. Infrared Spectrum o f Products from Treatment of Tetra­ isobutyldialuminum with HCl.

cn 68

Table 8

Infrared Frequencies fo r Products from Treatment of Tetraisobutyldialuminum with HCl.

Wavenumber (cm~M ±3

2922 1073

2781 1020

2732 952

2615 820

1458 776

1401 679

1384 498

1375 473

1170 69

presence of isobutyl groups bound to aluminum. They may be grouped ac­ cording to the relative intensities. The more intense doublets centered at 1.09 and 0.49 ppm are assigned to the methyl and methylene protons of isobutyl groups which experience one environment, w hile the less intense doublets centered at 1.26 and 0.71 ppm represent sim ilar protons which experience a different environment. The more intense of the two pairs of doublets is consistent with the presence of the diisobutylaluminum chlo­ ride as a product of the reaction. While the absolute chemical shifts are not identical (those of the pure diisobutylaluminum chloride were found to be 1.16 and 0.57 ppm) the A value was found to be v irtu a lly identical, 0.60 in the reaction mixture as compared to 0.59 for the pure compound.

Diisobutyl aluminum chloride is not the only aluminum species pres­ ent. The soluble, non-volatile products are a mixture of diisobutylalu­ minum chloride and isobutylaluminum dichloride. This conclusion comes from mass balance considerations and further examination of the NMR spectrum. As above, of the aluminum o rig in a lly present, approximately

8 percent was recovered as the metal and aluminum was found in no other products except the residual solution. This leaves 82 percent in the form of A l ( I I I ) species in the solution; no lower valen t species are present based on the lack of a characteristic intense coloration and the fa c t that no dihydrogen was found upon hydrolysis o f the solution.

Present as ligands are 82.5 percent of the originally present isobutyl groups and a ll of the added chloride. The la tte r was found in no vola­ tile or insoluble material resulting from the reaction and so must be present in the solution entirely. 70

Algebraic treatment of the data obtained indicates that the solution

is a mixture that is made up of 73 percent diisobutyl aluminum chloride

and 27 percent isobutylaluminum dichloride. Recalling the NMR of the

residual solution of the reaction, the less intense pair of doublets ap­

pears to be in approximately the correct ra tio to the more intense ab­

sorptions assigned to diisobutyl aluminum chloride. The individual areas

could not be found by integration due to the proximity of the peaks, but

the la rg e r absorbances are s lig h tly more than fiv e times as ta ll as the

less intense absorbances. The algebraic treatment would require a ratio

of areas of the chloride to the dichloride of 5.4.

Unfortunately, these results are not a direct indication of the

in itial products of the oxidation and substitution reactions thus far

discussed. Exchange reactions of aluminum compounds in solution are well

known and facile,usually resulting in a distribution of available

ligands about the aluminum atoms present, as noted in the Introduction

and presented in Equations 5 and 6. Thus, the data re fle c ts the products

present after this further interaction of the initially-produced species

has occurred.

The specific reaction sequences postulated for th is system must re ly

on the known chemistry of these and related compounds. The interaction

of HCl and the aluminum-aluminum bond is straightforward and results in

the production of dihydrogen and diisobutyl aluminum chloride. This and

similar reactions, hydrolysis and methanolysis, are well known and used

as the basis for the determination of the presence of this type of metal- metal bonded species. 71

The in teractio n of HCl with the carbon-aluminum bond is responsible for the production of isobutane. This was illustrated in Equation 14.

The reaction is not that simple, however, as further reaction of the chlorodialuminum species, Algfi-C^HgigCl in this instance, is indicated.

Specifically, this species disproportionates to afford a trivalent alumi­ num species and another reactive intermediate which eventually leads to the formation of aluminum metal. Equation 17 shows this disproportiona­ tion:

Algfi-C^HgjgCl + Alfi-C^HgizCl + "Al(i-C^Hg)" (17)

The exact mechanism, whether intra- or intermolecular, cannot be deter­ mined from the experimental evidence, but a 1,2 shift is a likely possi­ bility. The chlorodialuminum species affords three different intra­ molecular rearrangements: a chloride shift to give Alfi-C^HgjgCl and

"Alfi-C^Hg)" as shown; an isobutyl shift from the unsubstituted aluminum atom of the molecule to give the same products; or a second isobutyl shift from the substituted portion of the molecule to give Alfi-C^Hg),, and

"AlCl".

Of these possibilities, the firs t would be the most favored. The chloride ligand possesses non-bonding electron pairs in its valence shell.

These may participate to create a bridged intermediate which would fa c ili­ tate the transfer of the ligand. A similar occurrence is noted in car­ bonium ion rearrangements. S h ift of a phenyl group is fa c ile in such cases due to the availability of the pi electrons of the ring, allowing formation of a relative stable intermediate.^** Such an energetically favorable transition state is unavailable to the isobutyl group. It does not possess a lone pair of electrons or a pi electron system. 72

Steric factors may also be involved. The chloride ligand is spher­

ical, more compact in comparison to the bulky isobutyl group. Attack on

the adjacent aluminum atom would be easier for the chloride ligand.

Once the s h ift is complete, or perhaps in a concerted manner, dis­ proportionation of the aluminum-aluminum bond would generate the reactive

intermediate "AUi-CjtHg)". This species would react vigorously to afford aluminum metal in a further disproportionation. The exact nature of this further reaction is ambiguous. It does not involve a beta elimination, as shown by the absence of isobutene as a volatile product. An inter­ molecular process may be involved; this would consist of several steps and other reactive intermediates. An overall equation which does not represent specific intermediate steps may be written as:

3"Al(i-C4Hg)" ^ Alfi-C^Hgig + 2A1(0) (18)

Even if this were the actual reaction involved, triisobutylaluminum would not be isolated due to the exchange reaction mentioned above. Any mecha­ nism concerning the disproportionation of the dialuminum species or the fate of an A1(I) species generated as above could not be verified directly by analysis of the products of the reaction under these experi­ mental conditions.

The above series of reactions does parallel those postulated for sim ilar diboron species. Diboron tetrach lo rid e undergoes disproportiona­ tion to give boron trichloride and a mixture of boron subchlorides, as shown in Equation 19:

nBgClt ^ nBClg + (BCl)^ n = 4 ,8 ,9 (19)

The 1,2 s h ift is the proposed mechanism for this re a c tio n .55 In i­ t ia ll y , "BCl" is generated and th is species aggregates to varying degrees 73 to form the boron subchlorides. In the current work, no evidence of alumi­ num subalkyls [oligomers of "A1 (i-Ct^Hg)"] is found. Boron is noted for its cluster-forming abilities and chloride ligands can stabilize these low valent species via backbonding through its pi o rb ita ls . Aluminum seems to prefer a four-coordinate geometry with bridging ligands due to its higher acidity.2% In addition, isobutyl ligands should not provide much stability to an aluminum cluster while creating steric problems.

A final point must be made with regard to the stoichiometry of the reaction. The amount of isobutane recovered from the initial reaction with HCl is greater than the amount of the dialuminum compound assumed to undergo disproportionation. I f one isobutyl group per dialuminum is re ­ placed before disproportionation occurs (a valid assumption if the rate of disproportionation of the monochlorodialuminum is faster than the rate of s u b stitu tio n ), then a ll of the isobutane found cannot be accounted for by this reaction alone. Of course, i t cannot be expected that aluminum- carbon bond cleavage would be specific fo r the dialuminum compound. The oxidation reaction proceeds a t a faster rate and triv a le n t aluminum spe­ cies generated in this reaction would also be susceptible to aluminum- carbon bond cleavage. This would supply the balance of the isobutane recovered from the reaction.

With Dimethyl amine

Treatment of tetraisobutyldialuminum with dimethylamine resulted in oxidation of the aluminum-aluminum bond and subsequent generation of dihy­ drogen. Carbon-aluminum bond cleavage did not occur, as evidenced by the absence of isobutane as a product. Diisobutyldimethylaminoaluminum was found as a non-volatile product. Formation of th is species is thought 74 to decrease the rate of the oxidation reaction by interaction with di­ methyl ami ne.

The absence of carbon-aluminum bond cleavage should not be taken as

unusual in this case. At lower temperatures, this reaction does not take place with trivalent aluminum alkyls. Adduct formation does occur but elevated temperatures are necessary to promote alkane elimination and nitrogen-aluminum bond form ation, 100°C in the case of t r i methyl aluminum and dimethyl ami ne.

Dihydrogen evolution from the reaction was recorded at intervals and the data is presented in Tables 9 and 10 for two different sets of reac­ tion conditions. This information is also reproduced graphically in Fig­ ure 45. The two-step mechanism responsible for the generation of H 2 is shown in Equations 20 and 21:

Alzfi-C^Hg)^ + HN(CH3)2 + Al(i-C4Hg)2N(CH3)2 + Al(i-C4Hg)2H (20)

Al(i-C4Hg)2H + HN(CH3)2 Al(i-CuHg)2N(CH3)2 + H2 (21)

The slow rate of dihydrogen evolution for the reaction studied here is at first puzzling. The reaction with HCl, as presented in the pre­ vious section, was rapid and consumed the HCl completely. The hydrolysis reaction is also very fast, although a vast excess of water is always em­ ployed in these reactions. With the amine, on the other hand, the reac­ tion proceeds at a much slower rate , which appears to decrease as the reaction continues. Even at 43°C, after seven days only two-thirds of the dialuminum compound was found to have undergone oxidation.

Hydrolysis of the reaction solution after the lengthy reaction period produced additional dihydrogen in rapid fashion. Isobutane was also found. In the reaction summarized by data presented in Table 9, the total Table 9

Dihydrogen Evolution Alzti-C^Hg)^ and HNfCHgXz at Room Temperature.

Reactant Ratio Temp. Interval Time A12 ;ami ne 0° (total time) % H2 total

1:2 0 3 hrs. 31.0

25 20 hrs. (23 hrs.) 34.4

25 19 hrs. (42 hrs.) 38.5

1:4^ 25 28 hrs. (70 hrs.) 45.3

25 23 hrs. (93 hrs.) 49.1

25 3 days (7 days) 55.2

25 13 days (20 days) 61.7

^ expected H 2 based on complete oxidation of Al-Al bonds o rig in a lly present.

^An additional two equivalents of HNMe 2 was added a fte r 42 hours. Table 10

Dihydrogen Evolution from Al 2 (i-CitHg)4 and HN(CH3)2 at 43°C.

Reactant Ratio Temp. Interval Time Al 2 :ami ne 0° (total time) % H2 total

1:4 43 2 hrs. 20.6

43 16 hrs. (18 hrs.) 52.3

43 13 hrs. (31 hrs.) 55.5

43 22 hrs. (53 hrs.) 60.1

43 37 hrs. (90 hrs.) 62.9

43 44 hrs. (134 hrs. 65.0

43 39 hrs. (173 hrs.) 67.6

\ expected H 2 based on complete oxidation of Al-Al bonds o rig in a lly present.

0\ RT 43 deg C

% Dihydrogen

60

40

50100 150200 250 300 350 400 450 500 Reaction time hours

Figure 45. Dihydrogen Evolution From Tetraisobutyldialuminum and Dimethyl amine. 78

hydrogen from both amine and hydrolysis reaction accounts for b e tte r than

90 percent of the aluminum-aluminum bonds o rig in a lly present, only 61.7

percent of which was generated in the amine reaction.

This indicates that some type of inhibitory mechanism is at work in

the system. The amount of oxidation that occurs for any particular con­

centration of the amine approaches a lim it asymptotically that is less

than total expected for the amount of amine present. One may infer that

a product of the reaction is the in h ib itin g fa c to r. Equations 20 and 21

show that the two products of the reaction are dihydrogen and diisobutyl-

dimethylami noaluminum. The former would not be a lik e ly species to

interfere in the reaction. The removal of the gas also did not cause an

increase in the rate of the reaction. The amino aluminum compound is the

logical choice as the inhibiting agent.

Data obtained from NMR studies is helpful in the determination of the reactions taking place in this system. The spectrum of the reaction mixture shortly after combination of the reagents is shown in Figure 46.

Assignments of chemical sh ifts may be made with reference to the spectra of tetraisobutyldialuminum (Figure 16, Appendix) and diisobutyl dimethyl- aminoaluminum (Figure 20). Other relevant spectra are those of the dimethylamine adduct of diisobutyldimethylaminoaluminum (Figure 26) and triisobutylaluminum (Figure 42, Appendix). The chemical shift information from these spectra is gathered in Table 21.

In the spectrum of the reaction mixture, the intense doublets cen­ tered at 1.24 and 0.04 ppm are assigned to the methyl and methylene pro­ tons of the unreacted starting material. Partially obscured by the above methyl resonance is another doublet at almost the same chemical s h ift, a Figure 46. Proton NMR Spectrum of a Mixture of Tetraisobutyldialum i­ num and Dimethylamine.

VO U l .

Figure 20. Proton NMR Spectrum of Diisobutyldimethylaminoaluminum. CO o 7 6 5 3 2 1 0

Figure 26. Proton NMR Spectrum of Diisobutyl dimethyl ami noaluminum Dimethylamine Adduct.

CO 82

Table 21

Proton NMR Data fo r Some Aminoaluminum and Related Compounds.

Compound Methyl* Methylene* Amino methyl Figure

AI2 ( i “CijHg )i^+ 1.24 0.04 - - 46

HN(CH3)2 1.12 0.20 2.21 1 7?a,b

/Ws^i-C^Hg)^ 1.25 0.25 — 16

Al(i-C,»Hg)2N(CH3)2 1.21 0.25 2.24 20

Al(i-C^Hg)2N(CH3) 2 - 1.15 0.19 2.31 26

HN(CH3)2 2.24*

Al(i-C^Hg)3 1.11 0.41 — 42

*Center of doublet.

^Free dimethylamine.

^Coordinated dimethylamine. • 83

shoulder of which is visible on the upfield side of the more intense ab­

sorption. This assumed doublet and the doublet centered at 0.20 ppm are

assigned to the methyl and methylene protons of diisobutyldimethylamino-

aluminum. The resonance due to the methyl groups of the amino ligand of

this species is visible as a singlet at 2.21 ppm. Integration of the

areas of these peaks is not possible, but the relative peak heights are

reasonable. The doublet centered a t 1.77 ppm is assigned to the remain­

ing species in solution, dimethylamine. The signal is due to the methyl

protons of the amine which is s p lit due to coupling with the amine proton.

The resonance of this la tte r proton it s e lf is not visib le because of quad-

rapole broadening.

The production of diisobutyldimethylaminoaluminum as a product of the oxidation of the aluminum-aluminum bond, as might have been expected from the results of the HCl reaction, is thus confirmed. As to the role in the inhibition of the oxidation by this product, several points must be exam­ ined. It is known that aminoaluminum species are dimeric in non-polar solvents, as evidenced by the NMR spectrom of tris(dimethyl ami no)- aluminum, which exhibits two absorbances assigned to protons of bridging and terminal dimethylamine g ro u p s.5? Molecular weight measurements in diethyl ether, a coordinating solvent, however, show the compound to be essentially m onom eric.58 a donor-acceptor complex thus can be formed. In the work here, the relevant equation would be:

[Alfi-C^HgizNfCHglglz + ZHNfCHgiz -> 2Al(i-C^H9)2N(CH3)2-HN(CH3)2 (22)

A sample of diisobutyl dimethylaminoaluminum was treated w ith dimethyl­ amine to test this hypothesis. The excess amine was pumped away and a

NMR spectrum was obtained of the solution of the product (Figure 26). The 84

spectrum shows doublets fo r methyl and methylene protons of isobutyl groups centered at 1,15 and 0.19 ppm, respectively. The amino methyl

region shows what appears to be two singlets at 2.31 and 2.24 ppm, with

the former being the more intense of the two. These are assigned to the

protons of the amino ligand and of the coordinated amine, respectively.

If the latter is actually a doublet, as seen in the spectrum of the dia­

luminum, the downfield member of the pair is obscured by the absorbance

of the amino methyl protons. Also present in the spectrum is a broad

peak at 2.73 ppm, which is assigned to the amine proton.

Integration of the total area of the methyl groups on nitrogen and

the methylene protons of the isobutyl group shows that about 80 percent

of the diisobutyldimethylaminoaluminum has formed an adduct with the d i­

methylamine. This may be due to the way that the reaction was carried

out, adding one equivalent of the gaseous dimethylamine to a sample of

the aluminum compound, which was a neat liq u id . The interaction resulted

in the formation of a white solid, and it is possible that some of the

starting aluminum species did not receive exposure to the amine before

the excess was removed. Nonetheless, the reaction shows that an adduct

is formed upon in teraction of the two reagents.

This interaction would remove dimethylamine from the reaction scheme

with respect to oxidation o f aluminum-aluminum bonds s t i l l remaining in

solution. Algebraic examination of this phenomenon shows that four moles

of dimethylamine are consumed fo r each mole of dialuminum compound that

is oxidized. Equation 20 indicates that two moles of the amine are needed

to oxidize one mole of aluminum-aluminum bonds. Equation 21 requires that each of the two moles o f the amino aluminum species produced in the o x i­ dation reaction interact with another dimethylamine. 85

From these criteria, it would be expected that in the reaction rep­

resented in Table 9, which employed an in itial ratio of amine to dialumi­

num o f 2 : 1, that the reaction would be very slow a fte r less than one-half

of the dialuminum compound had reacted. Such seems to be the case.

(Empirically, the in itial value seems a bit high and may be due to the

presence of water in the originally-added dimethylami ne; the information

listed for the second reaction in Table 10 fits a smooth curve better.)

When another aliquot of dimethylamine is added to the reaction mixture,

the rate increases for a time and is then seen to follow the same level­

ing behavior as found with the f i r s t aliq uot.

With a 4:1 ra tio of amine to dialuminum, the above scheme predicts that complete reaction should occur. Concentration effects are a likely cause for the slow rate as more of the dimethylamine is removed in the two reactions proposed. The extent of the reaction is gauged by the amount of H2 evolved, based on the amount of dialuminum originally pres­ ent, but the presence of a grey precipitate indicates that some of the dialuminum compound has undergone decomposition, making the proportion higher than that actually calculated.

Another possibility is that more than one molecule of dimethylamine can be coordinated to the amino aluminum compound. Trigonal bipyramid structures involving two Lewis base moieties per aluminum atom are known^s but steric hindrance would be severe in this case. Steric and concentra­ tion factors in the interaction between the dialuminum compound and the amine are the most probable causes for the slow ra te .

The small amount of aluminum metal that was noticed in the reaction flask is most likely due to the normal decomposition that occurs in 86

solutions of tetra is o b u ty ld ialuminum. While an enhancement of the rate

of this reaction may be due to the presence of the Lewis base (c f. the

section on attempted Lewis base adduct formation with tetraisobutyldia-

luminum), the predominant reaction occurring between the amine and the dialuminum is seen as the oxidation of the aluminum-aluminum bond.

With Dimethylamine Adduct of Diisobutyldimethylaminoaluminum

Another reaction was designed to aid in elucidating the reaction of dimethylamine and tetraisobutyldialuminum. The latter compound was treated with a solution containing the dimethylamine adduct of diisobutyl- dimethyl aminoaluminum.

After 24 hours a small amount of dihydrogen was recovered from the reaction mixture. The quantity accounted for oxidation of only approxi­ mately 3 percent of the available aluminum-aluminum bonds. When addi­ tional dimethylamine was added to the reaction m ixture, dihydrogen produc­ tion increased. An additional 2.7 percent of the total expected di hy­ drogen was evolved in one hour with the excess amine present; almost as much as was found after 24 hours with only the amine adduct as a reactant.

The fa c t th at some dihydrogen was produced with the amine adduct in ­ dicates that a dissociation of that species occurs in solution. This has been documented in other cases of donor-acceptor complexes.In these cases, a stronger adduct shows less tendency to dissociate. Dimethyl­ amine is a strong Lewis base and the complex with diisobutyldimethylamino- aluminum would be expected to be relatively stable. The results of this reaction and the original dimethylamine reaction reflect this behavior.

This series of reactions indicates that the hypothesis of reaction inhibition is valid. The product formed in the initial reaction complexes with the oxidizing reagent to reduce the rate of dihydrogen formation. 87

With Ethylamine

Tetraisobutyldialuminum was treated with ethyl amine in an effort

toward further understanding of the reaction of the dialuminum compound with amines. As before, dihydrogen was evolved in the reaction. Also found was a small amount o f isobutane (0 .5 percent of the isobutyl groups originally present). Table 11 gives a compilation of the determinations fo r dihydrogen during the course of the reaction.

The results show an accelerated rate of oxidation of aluminum- aluminum bonds when ethylamine is used as compared to the case o f dimethyl­ amine. The reaction should be expected to proceed in the same manner as that seen for the secondary amine. Equations 23 and 24 (similar to 20 and

2 1 , presented earlier) describe the production of dihydrogen in this instance:

Al2(i-C4Hg)4 + H2NC2H5 -> Al(i-Ci,H 9) 2N(H)C2H5 + A l(i-C 4Hg)2H (23)

Al(i-Ci,Hg)2H + H2NC2H5 -> Al(i-CnHg)2N(H)C2Hs (24)

The difference in the rates of dihydrogen production here and with dimethylamine can be attributed to the difference in acidity of the two amines. The primary amine is more acidic than the secondary amine.so

The rate of reaction should be higher fo r the more acidic amine. The highly acidic HCl reacted in a very short time, as seen earlier.

This would also explain the minor amount of isobutane that was iso­ lated in the reaction. With a more acidic reagent, the aluminum-carbon bonds cleavage reaction is more favorable and proceeds to a very small extent.

Another reaction may be postulated which results in the formation of dihydrogen in this case, but which is unavailable to the system involving Table 11

Di hydrogen Evolution from Al 2 (i-Ci*H9)it and H2NC2H5 .

Reactant Ratio irniol A I2 : Temp. Interval Time Interval H 2 mmol amine 0° (total time) (to ta l H2 ) % to tal

0.775:0.775 18 2 hrs. 0.269 34.7

18 1 hr. 0.028 (3 hrs.) (0.297) 38.3

40 3 hrs. 0.014 (7 hrs.) (0.311) 40.1

40 3 hrs. 0.039 (10 h rs .) (0.350) 45.2

0.775:1.548^ 40 20 hrs. 0.167 (30 hrs.) (0.517) 66.7

40 18 hrs. 0.081 (48 hrs.) (0.598) 77.2

^ expected Hg based on complete oxidation of Al-Al bonds o rig in a lly present.

^An additional 0.773 mmol H^NEt was added a fte r ten hours.

00 00 89 dimethylamine. The second proton of the ethylamino group could react with the hydride species produced in the in it ia l aluminum-aluminum bond cleaving step (Equation 23). This reaction is shown in Equation 25:

Alfi-C^HgigH + Alfi-C^HgizNfHjCzHs -v [Alfi-C^HgjgjzNCgHs + Hg (25)

The two compounds which react in th is scheme are generated a t the same time and the proximity of the species could make this reaction more favorable. A sim ilar reaction is known fo r monoalkylaminoaluminun com­ pounds, but high temperatures are necessary for the formation of species with the general formula (RNAl)^, the polyaminoalanes.®^

In this particular situation, the reaction does not appear to be favored. The infrared spectrum of the non-volatile products of the reac­ tion (Figure 59, Table 20) exhibits a band at 3275 cm'i which is assigned to the amino proton of a monoethyl ami noaluminum species which argues against further reaction of such a species with an aluminum hydride moiety.

The predominant mode of reaction is thus limited to the reactions presented in Equations 23 and 24, and the enhanced reactivity is due to the greater acidity of the ethylamine species.

With Toluene-3,4-Dithiol

Treatment of tetraisobutyldialuminum with toluene-3,4-dithiol re­ sulted in the production of dihydrogen and isobutane as volatile products.

A brown solid and a grey solid were found as insoluble products of the reaction. The solution retained its dark red-brown color after the reac­ tion appeared to be complete. The reaction is thought to be quite com­ plex. 2500 1800 I GOO 1400 BOO 600 4&0

Figure 48. Infrared Spectrum of Product of Reaction of Tetraisobutyl­ dialuminum with Ethylamine. 91

Table 20

Infrared Frequencies for Product of Reaction of Tetraisobutyldialuminum with Ethylamine.

Wavenumber (cm-i) ±3

3275 1170

2930 1152

2845 1110

2754 1059

2580 1035

1464 1011

1400 936

1373 888

1356 806

1312 666

1249 534

1201 92

As before, the production of dihydrogen and isobutane are the result

of aluminum-aluminum and carbon-aluminum bond cleavage, respectively. Di­

hydrogen recovered was consistent with oxidation of 60 percent of the

starting dialuminum compound and accounted fo r reaction of 63 percent of

the available hydrogen from the dithiol. Isobutane found totaled 11.5

percent of the isobutyl present in the solution in itially and consumed 24

percent of the available hydrogen. The discrepancy in the total amount

of hydrogen added and that recovered as volatile by-products, 13 percent,

is not readily explainable.

Spectroscopic information was also ambiguous. The infrared spectrum

of the brown solid (Figure 27, Table 12) is broad and featureless except

for bands assigned to the mulling agent. The only useful information ob­

tainable is the absence of absorptions that are typical of the aluminum

hydride moiety (ca. 1800 cm"^)®^.

The NMR spectrum o f the solution (Figure 28) exhibits absorbances

th at may be assigned to the methyl and methylene protons of isobutyl

groups, doublets centered at 1.26 ppm and 0.42 ppm, respectively. The

aromatic region is obscured by the lock signal of the solvent, benzene.

Also present are two singlets at 3.52 ppm and 3.60 ppm, consistent with the th io l protons of the starting m aterial (3.50 and 3.80 ppm). However, absent is an absorption that would be expected for the methyl protons at­ tached to the aromatic ring (2,28 ppm in the starting dithiol).

Earlier work involving this reagent with various trialkyl, trihalo-,

and mixed alkylhaloaluminum compounds showed the reaction possibilities

to be numerous and complex.®** Work employing Lewis base adducts, dimethyl-

and monoethyl ami ne and dimethyl ether, and the use of halogenated solvents. IM H I l'N X f ÜOO 4 no WAVI riilMIM

Figure 27. Infrared Spectrum of Solid Products from Treatment of Tetraisobutyldialuminum with Toluene-3,4-Dithiol.

WVO 94

Table 12

In frared Frequencies fo r Products of Treatment of Tetraisobutyldialuminum with Toluene-3,4-Dithiol.

Wavenumber (cm~^) ±3

3078 1080

3060 946

1400 852

1335 800

1252 676

1140 542

1110 501 VJ

7 6 S 4 3 2 1 0

Figure 28. Proton NMR Spectrum of Solution from Treatment of Tetra­ isobutyldialuminum with Toluene-3,4-Dithiol.

to t n 96 served to simplify the reactions. These methods were not available in this work. Protonic bases have been shown to result in oxidation of the aluminum-aluminum bond, while aprotic base seems to enhance the decomposi­ tion of the dialuminum species (to be discussed in a subsequent section).

Halogenated solvents are known to react violently with alkylaluminum com­ pound and must be avoided.

The fact that both dihydrogen and isobutane are produced is evidence that two sites of re a c tiv ity on the dialuminum are available and m u lti­ plies the possibilities of products over those obtained with a trivalent aluminum species. Various modes of intra- and intermolecular products may be envisioned.

For these reasons the investigation of the products of this reaction was not pursued to a greater extent.

REACTIONS WITH ALUMINUM AND BORON COMPOUNDS

Tetraisobutyldialuminum was treated with various aluminum and boron compounds: trimethyl aluminum, boron tric h lo rid e , trimethoxyboron, and t r i s(di methyl ami no)boron. Products consistent with occurrence of exchange reactions between the dialuminum compound and the various reagents were found. In most cases, this exchange led to the formation of unstable dialuminum species which decomposed via a disproportionation reaction.

With trimethoxyboron, a stable compound containing an aluminum-aluminum bond was isolated.

With Trimethylaluminum

The reaction of tetraisobutyldialuminum with trimethylaluminum re­ sulted in the formation of trivalent aluminum species containing both iso­ butyl and methyl ligands and aluminum metal. In no case was there isolated 97

a stable methyl-substituted dialuminum species. While such an interme­

diate is postulated, its disproportionation is thought to be rapid and

afford the identified products of the reaction. A generalized equation

for the reaction is:

Alzti-C^Hg)^ + A1(CH3)3 -> Al(i-C4Hg)3_n(CH3)n + Al° (26)

The actual triv a le n t aluminum species produced in the reaction is depend­

ent on the stoichiometry of the initial reaction mixture.

Information from the hydrolysis of the reaction solution was helpful

in determining the course of the reaction. The products of such a reac­

tion were found to be isobutane, methane, and dihydrogen. Production of

the f i r s t two m aterials is the resu lt of hydrolytic cleavage of aluminum- carbon bonds in isobutyl- and methylaluminum species, respectively, and are shown cryptically here:

Al-R + HgO ^ Al-OH + RH, R^i-C^Hg or CHg (27)

Dihydrogen is the re s u lt of cleavage of aluminum-aluminum bonds or aluminum-hydrogen bonds, as:

A l-Al + 2H2O 2A1-0H + H2 (28)

Al-H + H2O-V Al-OH + I / 2H2 (29)

In the reaction which employed a ratio of tetraisobutyldialuminum to trimethylaluminum of 3:4, the following relationships were found. Hy­ drolysis of an aliquot of the non-volatile products after two days of reaction yielded a ratio of isobutane to methane to dihydrogen recovered of 1.00:1.04:0.05, while the starting mixture was formulated to have a ra tio of isobutyl groups to methyl groups to aluminum-aluminum bonds of

1.00:1.03:0.25. There is an apparent loss of 80 percent in the amount of aluminum-aluminum bonds over the reaction period. 98

In the reaction employing a large excess of trimethylaluminum, a

q u a lita tiv e comparison must be made but the same re s u lt is obtained. The

ra tio of methyl groups to aluminum-aluminum bonds in the starting mixture

was approximately 12: 1 , w hile the hydrolysis data gave a ra tio of methane

to dihydrogen from the reaction solution of about 32:1. The actual ratio

must be significantly higher when it is realized that some of the

originally-added methyl moieties were removed from the reaction solution

as trimethylaluminum in the work-up of the reaction. Hence, it is con­

cluded that the interaction of trimethylaluminum with tetraisobutyldia­

luminum results in a decrease in the amount of the species responsible for

the production of dihydrogen upon hydrolysis.

The infrared spectrum of the non-volatile components of the reaction

solution (Figure 47, Table 19) aids in the identification of the species

responsible fo r the production of dihydrogen. The spectrum is sim ilar to

that obtained from triisobutyl aluminum (Figure 43, Table 18, Appendix).

The important information obtained from the spectrum is the absence of bands assignable to an aluminum hydride moiety. These bands are found in

the region 1680 to 1840 cm"^ for dialkyl-substituted aluminum hydrides^^ and up to 1900 cm~^ in donor-acceptor complexes.®® No such bands are ap­ parent here. Therefore, the source of dihydrogen upon hydrolysis is oxi­ dation of aluminum-aluminum bonds.

The NMR spectrum o f the reaction solution (Figure 29) shows sev­ eral intense absorbances which may be assigned as follow s. The doublets centered at 1.16 ppm and 0.29 ppm are consistent with the presence of isobutyl groups bound to aluminum, with reference to the NMR spectrum of triisobutyl aluminum (Figure 42, Appendix). The large singlet at -0.11 600 400 4000 3500 3000 3500 1800 teoo 1400 1300 800 WAVENUMBER

Figure 47. Infrared Spectrum of Products from Treatment of T e tra­ isobutyldialuminum with Trimethylaluminum. 100

Table 19

Infrared Frequencies for Products from Treatment of Tetraisobutyldialuminum with Trimethylaluminum.

Wavenumber (cm -i) ±3

2926 1047

2772 981

1478 863

1410 792

1389 700

1378 687

1273 578

1219 505

1190 438

1104 JL

. I K l I I I I i I 1 1.1 I I I. I t I I 1 I I I I I I I I I I I I I I 1 1 i 1 t I t I I ., I I 1 I I I • I I 1 . t I' I' I 1 I < 1 1 1 1 I 1 I I I 1

Figure 29. Proton NMR Spectrum of Product from Treatment o f T a tra - .« isobutyidialuminum with Trimethylaluminum. 102

ppm is indicative of the presence of methyl groups bound to aluminum

(s in g le t at -0.30 ppm in the spectrum of trim ethyl aluminums?). Caution must be taken in assigning only one environment to methyl groups bound to aluminum from this re s u lt, however.

The low temperature NMR spectrum of the reaction solution is pre­ sented in Figure 49. Apparent is the breadth of the signals obtained.

This may be due to the presence of some solid in the sample tube. Despite this complication, the signal assigned to the methyl protons shows some splitting into two separate peaks. The highest field absorbance, cen­ tered a t -0.38 ppm, is considerably broadened. Appearing at +0.63 ppm is a small sin g let. These may be assigned to methyl groups of triv a le n t aluminum dimers which occupy bridging and terminal positions. This re s u lt is consistent with the findings of studies performed on trimethylaluminum in which low temperature spectra indicated the presence of two different environments for the bridging and terminal methyl groups.&? At higher temperatures, a single absorbance was found due to the rapid exchange of the groups between the two positions on the NMR time scale. No other ab­ sorbances in this region were apparent which would indicate the existence of another environment fo r methyl groups bound to aluminum.

Another result of the reaction was the change in appearance of the solution throughout the course of the reaction. The intense coloration of the solution was maintained for several days, despite the increasing amount of grey precipitate that was noticed during as the reaction pro­ ceeded. This material was subsequently identified as aluminum metal.

The solutions that were hydrolyzed exhibited this intense red-brown color, which has been linked to the presence of dialuminum species. A ll 2 1 0 1

Figure 49. Low Temperature NMR Spectrum of Products from Treatment of Tetraisobutyldialuminum with Teimethyl O aluminum. . oo 104

of the dialuminum compounds prepared by Hoberg possessed this intense

color^»** as well as the tri aluminum species isolated by Schram.i Triva­

lent aluminum alkyls (including the hydrides), on the other hand, are

colorless in solution.

From these data, a discussion of the chemistry involved in this sys­

tem must relate to several points. First, the reduction of the amount of dihydrogen found in the reaction solution with respect to the amount of aluminum-aluminum bonds in the startin g mixture and the formation of alumi­ num metal are important. Second, the spectroscopic data, p artic u la rly the

NMR information, and the time period of the reaction should be addressed.

An overview of the reaction scheme with these points in mind would indi­ cate that a disproportionation reaction be involved and that any low- valent-aluminum-methyl species created be transitory in nature. An exchange reaction also seems apparent from the information available.

Any proposed mechanism fo r the reaction would involve several steps.

The firs t of these would be the dissociation of the trimethyl aluminum dimer into two monomeric units, or at least cleavage of one bridge bond to give a tricoordinate aluminum species.

Such atrivalent unit would associate with the dialuminum compound to give a mixed alkyl bridged system, taking advantage of the unoccupied valence site available on the low valent aluminum. This association could occur in two ways. The trimethylaluminum could interact with both aluminum atoms of the dialuminum species. The interm ediate would be formed by creating a bridging isobutyl group from one low valent atom to the triv a le n t aluminum atom and completed via a methyl bridge to the second low valent atom. This results in a five-membered ring which in­ corporates the aluminum-aluminum bond. 105

Alternately, the interaction would involve only one of the low valent

aluminum atoms. A four-membered cyclic structure with alternating alumi­

num and carbon atoms would resu lt. One methyl and one isobutyl group

would be involved in bridging the two aluminum atoms. These interactions

are shown as the f i r s t structures in the schemes presented in Figure 50.

Subsequent dissociation of either of these intermediates can occur

in two ways. The f i r s t is the breaking of the newly-formed bonds. This

is trivial and simply results in the regeneration of the original reacting

species. The second path, which involves cleavage of the bridge bonds in

the opposite manner. I.e ., cleavage of the methyl to trivalent aluminum

bond and of the isobutyl to low valent aluminum bond, results in the forma­

tion of isobutyldimethyl aluminum for both intermediates. However, the

other fragment containing the aluminum-aluminum bond would be d iffe re n t

for the two cases. The first intermediate would result in a species hav­

ing one aluminum atom with three ligands attached, linked to a second

aluminum atom with one ligand. The second scheme would create triis o b u ty l-

methyldialuminum as the product of dissociation. These paths are shown

as the second step in Figure 50.

The species generated in the firs t scheme would be very unstable and decompose to diisobutylmethyl aluminum and "Alfi-C^Hg)". The dissociation and decomposition reactions in this case could very well be concerted,

the aluminum-aluminum bond undergoing disproportionation at the same time the isobutyldimethyl aluminum dissociates from the initial intermediate, as shown in the final step in Figure 50,

The product in the second scheme, triisobutylmethyldialuminum, would also be expected to be unstable and decompose via a 1,2 shift of an alkyl 106

Al2P*z^ + AlMe^

Me Me Me Me "/ l R R o ME R R ^ / / A l —A l Al—Al / / R R R R

-AlMe2R

Me Me R R— A l* -Al—R Al-^—Al / R R R

Me

R— Al— AL— R 't ALR2ME + "A l R" R

50. Reaction Schemes for the Interaction of Tetraisobutyl­ dialuminum and Trimethylaluminum . 107 ligand to give a species such as proposed for the firs t scheme. This undergoes fu rth e r reaction to produce diisobutylmethylaluminum and

"AUi-CijHg)", again shown as the final step in Figure 50.

As discussed previously in conjunction with the reaction of the dia­ luminum compound and HCl, "AI(i-C^Hg)" undergoes fu rth e r reaction to eventually produce aluminum metal. Again consistent with the earlier case, no isobutene was found and the scheme is thought to involve rapid inter­ molecular reaction.

The two schemes o f Figure 50 are very sim ilar. The second is to be favored with regard to previous work involving the interaction of tr i­ methylaluminum with low valent boron compounds. In the reaction to pro­ duce the trialuminum species, Al 3 (CH3 ) 3 [N(CH3 ) 2 ] 2 5 by treatment of tetra- kis(dimethylamino)diboron was isolated.i This would be the result of the interaction s im ila r to the second scheme proposed while the f i r s t scheme would not permit isolation of this species.

The methyl-substituted dialuminum compound is therefore a likely intermediate in the system studies here. If present in quantity, it should give dihydrogen upon hydrolysis and exhibit a high field resonance in the NMR. Results of the experimental work show the first to be true to a small extent for the reaction solution. These results are not un­ equivocal since unreacted starting material will also generate dihydrogen upon treatment with water. Spectroscopic information tends to discount the presence o f the species. As stated e a r lie r , the low temperature spectrum indicates two environments fo r methyl groups on aluminum, which is satisfied by the presence of bridging and terminal methyl groups of triv a le n t aluminum dimers. 108

I t also seems lik e ly th a t absorption due to a methyl group bound to a

low valent aluminum atom would have a chemical s h ift a t higher fie ld than

e ith e r of the two methyl groups mentioned above. This is indicated by

comparison of the spectra o f triisobutylaluminum and tetraiso butyldia­

luminum (Figures 42 and 16, respectively. Appendix). The latter shows a

chemical s h ift for the methylene protons of the isobutyl group which is

0.16 ppm to the high field side of that seen for the trivalent species.

This phenomenon is ascribed to the d iffe re n t electro nic environments im­

posed upon the protons by the different valent states of the aluminum atoms

of the two compounds.The methylene protons are attached to the carbon

atom d ire c tly adjacent to the aluminum atom and would be expected to show

the greatest effect of a difference in electronic make-up of the aluminum

atom. Here, the methyl protons should be affected in the same way. How­

ever, no resonance a t higher fie ld is seen in the spectrum. The methyl-

substituted species are unstable and undergo rapid reaction via alkyl

shift and disproportionation.

Further indication of the relative instability of the species is given by recalling the reaction of tetraisobutyldialuminum and HCl. Sub­

s titu tio n of a chloride ligand fo r an isobutyl ligand on the dialuminum

species was proposed as one cause for the generation of isobutane found as a product of that reaction. The chlorodialuminum species was found to

be unstable to disproportionation by virtue of the fact that aluminum metal was identified as another reaction product. The disproportionation reaction was rapid in the HCl case, the reaction taking place in the time

it took the mixture to warm to room temperature. 109

A methyl group would provide less s ta b iliza tio n than a chloride; the

experience in attempts to substitute methyl ligands for chloride on diboron

tetrachloride^^ shows th is to be true. Therefore, i t can be concluded

that a methyl-substituted dialuminum species would only exist as a tran­

sient intermediate in the overall reaction scheme.

This raises the question of reaction time. The color of the solution

persists for as long as a week. Despite the fact that a very low concen­

tration of a dialuminum species is capable of imparting an intense color

to the solution, as seen in the preparation of the tetraisobutyldialumi­

num starting m aterial, one might expect the reaction to proceed at a much

greater rate. One of the other steps must be the rate-limiting step in

the reaction.

The dissociation of the trimethylaluminum dimer is known to be a

rapid reaction. Exchange reaction between aluminum and boron compounds, which depend on th is as an in it ia l step, proceeds quite readily.The

association of the trimethylaluminum monomer with the dialuminum species

is the only other step in the reaction scheme and should be considered

as the rate-determining step. Rationalization of this hypothesis is

aided by information gained from other reactions which should be presented prio r to constructing the arguments.

Attempts to Prepare Lewis Base Adducts of Tetraisobutyldialuminum

In an effort towards stabilizing the product(s) of the reaction be­ tween tetraisobutyldialuminum and trimethylaluminum, synthesis of Lewis base adducts of the dialuminum starting material was attempted.

Two general results were found in these reactions. F irs t, the ad­ ducts, i f formed a t a l l , were not stable to dissociation under dynamic n o vacuum conditions. Both THF and diethyl ether, when present for a short time, could be removed from the system by pumping. Second, prolonged con­ tact of the THF or diethyl ether resulted in the decomposition of the tetraisobutyldialuminum at a much faster rate than that seen when the material remained in contact with a non-polar, non-coordinating solvent.

This was shown by the fact that a freshly-filtered solution of the dia­ luminum compound in diethyl ether showed traces of a grey precipitate after 15 hours of stirring.

The results of the reaction tetraisobutyldialuminum and 4-picoline were slightly different than that seen with the ethers. The ^H NMR spec­ trum of the mixture (Figure 30) shows 4-picoline to be present. The singlet at 1.59 ppm is assigned to the para methyl group on the hetero­ cyclic ring. However, comparison of the integrated areas of this ab­ sorbance and the doublet assigned to the methylene protons of the isobutyl groups a ra tio of greater than 1:1 (actual 1.28). I f a 1:1 adduct is formed, then the excess picoline is present in an uncoordinated state and should exhibit a d iffe re n t chemical s h ift for the methyl protons. This is not the case. The ^H NMR spectrum o f the adduct between dimethylalumi- num chloride and 4-picoline (from which the base is not removed by pump­ ing) contains an absorbance fo r the picoline methyl group a t lower fie ld

(1.74 ppm)77 than seen in this case. If excess 4-picoline is added to the chloride system, the spectrum exhibits two signals for coordinated and uncoordinated 4-picoline.

Other reactions have shown this Lewis base to be persistent, d iffi­ cult to remove quantitatively from reaction mixtures. It may well be that the extent of dynamic pumping employed to remove the 4-p ico lin e was ...L.

Figure 30. Proton NMR Spectrum of Product from Treatment o f T etra- isobutyldialuminum with 4-Picoline. 112

inadequate. Conversely, nitrogen-containing bases are known to form stronger adducts with aluminum compounds than oxygen-containing bases.®®

A stronger interaction with 4-picoline would make it more difficult to remove than the easily dissociated ethers.

These results notwithstanding, it appears that 4-picoline acts in the same manner as the ethers to increase the rate of disproportionation of the dialuminum compound, as evidenced by the appearance o f a solid in the sample tube a fte r the spectrum was obtained.

The first observation indicates that aluminum in a low valent state is a weaker Lewis acid than a triv a le n t aluminum. Acid-base adducts in ­ volving compounds o f the la tte r type can be handled in a manner as a t­ tempted with the dialuminum without dissociation of the adduct. Indeed, the diethyl ether adduct of trimethylaiuminum is so stable that it may be sublimed in that form without s ig n ific a n t dissociation.®®

The relative Lewis acidities of these materials reflect their elec­ tronic nature. A more electro n-rich system would show lower acidity.®®

Assignment of a +2 formal charge fo r the metal atoms of the dialuminum is consistent with this.

The reasoning is also supported by the NMR resu lts. The chemical shift of the methylene protons of the isobutyl groups is found at higher field in the dialuminum starting material, as compared to triisobutyl alumi­ num as stated in the previous section. An increase in the electron density at the aluminum atom would have the effect of greater shielding for the pro­ tons of the alpha carbon.

Studies of a series of sim ilar organoaluminum compounds, triis o b u tyl- aluminum, the tetraisobutylaluminum anion, and the hexaisobutyldialuminum 113

dianion, were 0.28, -0.25, and -0.30 ppm, respectively (centers of doub­

lets in benzene).4 The chemical shift of the methylene protons moves to

higher field across the series. This parallels the increase in electron

density that would be associated with the aluminum atom in the series and

the effect that would have on the shielding of those protons.

The second observation obtained from the work with Lewis bases and

the dialuminum compound is consistent with past experimental work. Simi­

lar results were seen in the original attempts to prepare tetrasubstituted dialuminum compounds. These reactions were reductions of disubstituted aluminum chlorides performed in a polar solvent.Aluminum metal and trisu b s titu ted aluminum compounds were found as the reaction products. I t seems lik e ly , from the observations reported, that a dialuminum species was f ir s t formed and then underwent decomposition via disproportionation to afford the products obtained. The polar solvent is responsible for this, as evidenced by the successful synthesis of the desired compounds when a non-polar solvent was e m p lo y e d .

Electronic effects are implied here as well. The polar solvent would be better at stabilizing a polar transition state than a non-polar solvent.

The 1,2 alkyl shift discussed previously as the firs t step in the dispro­ portionation reaction can be envisioned as having such a transition state.

A separation of charge would occur with the aluminum atom acquiring a partial positive charge and the alkyl group a partial negative charge as the migration of the alkyl group proceeded. A Lewis base would lend sta­ b ility to this transition state by interaction with the aluminum atom and facilitate the transfer. A non-polar solvent could not provide this sort of stabilization and the 1,2 shift would be energetically more d ifficult. 114

With this argument presented, the discussion of the rate of reaction

between tetraisobutyldialuminum and trimethylaiuminum may be undertaken.

The lower Lewis acid ity of the dialuminum species would decrease the a t­

traction of that moiety for a monomeric unit of trimethylaiuminum. The

preferred association would be for recombination of two monomers. This

effect would be most noticeable in the latter stages of the reaction as

the concentration of trivalent species increases.

This reasoning, however v a lid , is not the most compelling argument

for the contention that the association of trivalent and low valent alumi­

num species is the rate-determining step in the reaction. Steric factors

are probably much more important. Surrounded by four bulky isobutyl

ligands, the aluminum-aluminum bond should be f a ir ly well insulated against

attack by another molecule. The fact that a four-atom heterocyclic struc­

ture must be attained in order for the exchange reaction to occur would make it even more difficu lt for the interaction to take place.

The s teric argument has precedence in work d etailin g the degree of association of various trialkyl aluminum species. Trimethylaiuminum was found to be virtually 100 percent associated as the dimeric form in solu­ tion by ebullioscopic measurement.As the number o f carbon atoms o f the alkyl substituent increased, the degree of association decreased. Triiso- butylaluminum was found to be virtually 100 percent unassociated in solu­ t i o n , "^2 i.e., the monomeric form of the molecule was found to be the over­ whelmingly dominant species present.

Thus, it seems that both electronic and steric effects work in con­ cert to limit the rate of reaction by hindering the association of tr i­ methyl aluminum and tetraisobutyldialuminum. 115

With Boron Trichloride

The reaction of tetraisobutyldialuminum with boron trichloride pro­ ceeded in a manner similar to that seen with trimethylaiuminum. An ex­ change reaction of chloride for isobutyl ligands occurred and subsequent disproportionation of the chloro-substituted species resulted in the formation of metallic aluminum and trivalent aluminum species. A general­ ized equation for the reaction may be represented as:

Alzfi-C^Hgit + BCI3 -> Al(i-C„H 9 ) 3_nCln + BCl 3 _n(i-Mg)^ + Al° (30)

The information obtained from hydrolysis of the reaction solution gave a ra tio of isobutane to dihydrogen of 14 .5 :1. In the starting mate­ r ia l the ratio of isobutyl groups to aluminum-aluminum moieties was 4:1

(from the molecular formula of the compound). As before, the reaction results in a decrease in the amount of the species responsible for the production of dihydrogen upon hydrolysis.

M etallic aluminum was also found as a product, as noted in Equation

30 above. This parallels work discussed earlier with the HCl and tri- methylaluminum systems and implies that disproportionation is responsible for the decrease in oxidizable species in the reaction solution.

Spectroscopic information was limited to NMR investigation of the reaction solution after filtration to remove the insoluble aluminum metal.

Figure 31 shows two doublets in the region assigned to methyl protons of an isobutyl group (compare to the doublet centered a 1.25 ppm fo r te tra ­ isobutyldialuminum, Figure 16, Appendix). Here, the doublets are cen­ tered at 1.24 ppm and 0.93 ppm, with the up field resonance being the more intense of the two. Also present is a less intense absorption centered at approximately 0.49 ppm, which may be related to the sim ilar absorbance 1 J-L

Figure 31. Proton NMR Spectrum of Product from Treatment o f T etra­ isobutyldialuminum with Boron Trichloride......

cr> 117

in the spectrum of diiscbutylaluminum chloride (doublet centered at 0.48

ppm in Figure 44, Appendix).

Analysis of the two doublets assigned to the methyl protons of the

isobutyl groups indicates that they should be assigned to moieties bounded

to two different metals, i.e., to boron and aluminum. The downfield resonance is in the region seen fo r these types of protons in isobutyl groups of alkylaluminum species. For example, sim ilar resonances appear for triisobutyl aluminum (1.11 ppm. Figure 42, Appendix), diisobutyl alumi­ num chloride (1.16 ppm. Figure 44, Appendix), and diisobutyl aluminum methoxide (1.19 and 1.22 ppm. Figure 19, to be discussed in the next sec­ tio n) .

The doublet at higher fie ld (centered at 0.93 ppm) is assigned to the methyl protons of isobutyl groups bound to b o r o n . 83 This follows first from the stoichiometry of the reaction which shows the ratio of boron to aluminum in the original mixture to be approximately 5:1. The higher fie ld resonance is the more intense of the two absorbances discussed here.

This assignment will be shown to be consistent with the results obtained in subsequent reactions as w e ll.

Therefore, the course of the reaction between tetraisobutyldialuminum is seen as analogous to th a t presented fo r the trimethylaiuminum system.

Association of boron trichloride with the dialuminum species results in the exchange of isobutyl for chloride ligands on the low valent aluminum species. The newly-formed species would then undergo disproportionation in the form of a 1,2 ligand shift and cleavage of the aluminum-aluminum bond. As presented earlier in the discussion of the reaction of tetra­ isobutyldialuminum, the shift of a chloride ligand would be the preferred 118

route due to the unshared electron pairs on the chloride, which could

facilitate the creation of a transition state by interaction with the

second aluminum atom of the dialuminum moiety.

With this in mind, the apparent slower rate of reaction of the boron

tric h lo rid e seems anomolous. I t must be recalled, however, th a t the rate -

limiting step is postulated as the association of the two reacting species.

Steric factors once again would be of influence here. Boron trichloride is approximately the same size as trimethylaiuminum and the same arguments are applicable here as the previous section.

Moreover, electronic factors may also be involved. Boron compounds are weaker Lewis acids than corresponding aluminum s p e c ie s .22 Thus, the boron trichloride would have a lower affinity for the dialuminum species than trimethylaiuminum.

Overall, the difference is not thought to be great. The important point being that the interaction does not result in the formation of a species containing a chlorodialuminum moiety that could be isolated from the reaction mixture.

With Trimethoxyboron

The reaction of tetraisobutyldialuminum with trimethoxyboron resulted in isolation of a new species, Algti-C^Hgi^tOCHg)^, %, which contains a inum- sinqle aluminum-aluminum bond. Substitution of methoxy qroups fo r iso- butyl groups on the dialuminum moiety occurred and the resulting compound is stabilized by intermolecular association with trivalent aluminum species.

The latter are generated by in itial decomposition via disproportionation of the formed substitution products. 119

Before examination of the analytical and spectroscopic information relating to U it is helpful to discuss some details of a similar com­ pound, diisobutylmethoxyaluminum. This compound was prepared from diiso- butylaluminum hydride and methanol. The NMR spectrum of th is compound, obtained shortly after its preparation is presented in Figure 18 and is more complicated than would be expected. The absorbances assigned to the methyl and methylene protons of the isobutyl groups are not doublets but pairs of doublets. The absorbance assigned to the methoxy group is a pair of singlets. The implication is that two different environments exist. Relative peak heights allow these signals to be grouped together.

One environment is represented by the group composed of the singlet at

3.31 ppm (methoxy) and the doublets centered at 1.22 and 0.205 pom (is o ­ butyl). The other environment is represented by the group composed of the singlet at 3.19 ppm and the doublets centered at 1.18 and 0.22 opm.

For the most part, methoxyaluminum compounds are trim eric but studies indicate that the firs t formed product in the sy-thesis of di­ methyl methoxyaluminum is a dimer and slowly converts to the trim er.

Sim ilar results are seen in the case of dimethylphenoxyaluminum.

A fter three months the spectrum of diisobutylmethoxyaluminum was rerun (Figure 52). The pattern and chemical shifts of the peaks are the same as before but the intensities are not. The peak in the methoxy region at 3.31 ppm is now seen to be very small compared to the absor­ bance at 3.19 ppm. With regard to the information presented above, the more intense absorption is assigned to the trimeric form of the compound.

In the spectrum of the freshly-prepared di isobutylmethnxval"rniPum

(Finur° 181 the following assignments may be made. The species assigned -L lJ :-L I i 1. I ! l_l I I : I I II . I ] I I ; . I

Figure 18. Proton NMR Spectrum of Diisobutylmethoxyaluminum . t\j o 4 3 2 1 0

Figure 52. Proton NMR Spectrum of Diisobutylmethoxyaluminum after Three Months. 122

to the singlet at 3.31 ppm (methoxy) and the doublets centered at 1.22

and 0.205 ppm (methyl and methylene of isobutyl) is the dimeric form of

diisobutylmethoxyaluminum. The other group of absorbances, singlet at

3.31 ppm and doublets centered a t 1.18 and 0.22 ppm, are assigned to the

trim eric form of the compound.

Another experiment that proved useful in the elucidation of the

results o f the methoxy system was performed. Cyclopentane/benzene-mixed

solvent solutions of tetraisobutyldialuminum and triisobutyl aluminum in

NMR sample tubes were treated separately with an excess of trimethoxy­

boron in the dry box. The tubes were sealed and the spectra of the reac­

tion recorded after a short time, after two days and after five days.

These spectra are shown in Figures 53 and 54.

Shortly after mixing, both solutions exhibited a single peak in the

region consistent with the presence of methoxy groups bound to aluminum.

Careful examination shows the peak in the spectrum o f the dialuminum reac­

tion mixture to be at slightly higher field than the counterpart in the

triisobut.vlaluminum mixture, 3.25 in the former vs 3.30 in the latter.

After two days, there was solid present in both tubes and it was

necessary to centrifuge both tubes to obtain reasonable spectra. The

spectrum o f the dialuminum sample, which contained a brown s o lid , now

consisted of a group of four peaks in the methoxyaluminum region. The

new absorbances appear at lower field (3.37 ppm) than the original peak.

The spectrum of the triis o b u ty l aluminum sample, which contained a

large amount of a white gelatinous material, showed only one peak again.

I t has the same chemical s h ift (3.30 ppm) as was recorded a fte r one hour

but the intensity of the peak has diminished. 123

1/2 hour i - JUL

2 days

-I-... : I I ■ 1______U

5 days

Figure 53. Proton NMR Spectra of a Mixture of Tetraisobutyldi alumi­ num and Trimethoxyboron as a Function of Time. 124

1 hour t y . ÜL.__

. I J I

2 days

1 ■ I

5 days

± J. I « s i I

Figure 54. Proton NMR Spectra of a Mixture of Triisobutyl aluminum and Trimethoxyboron as a Function of Time. 125

A fter fiv e days, the sample o f the dialuminum reaction showed an

increased amount of brown solid present. Some of this solid adhered to the walls of the tube and prevented a clean spectrum from being obtained.

I t can be seen, however, that the low fie ld absorptions are s t i l l present.

The triisobutyl aluminum sample was almost completely gelled after five

days. In this case, solution was separated from solid by centrifugation.

No peak assignable to methoxy on aluminum is evident in the spectrum.

Interpretation of these results must take into account the lim ita­

tions o f any comparisons th at can be made. The spectra were a ll obtained

from reactions that occurred in solution. It will be seen that the prod­

ucts of such a reaction are slightly different than those obtained by using an excess of trimethoxyboron in place of a solvent.

Nonetheless, several important observations can be made. First, the reaction of the trialkylaluminum species is not limited to a single sub­ stitution reaction. Additional exchange is possible and results in insoluble di- or trimethoxyaluminum species. Second, the amount and ap­ pearance of insoluble m aterial is very d iffe re n t in the two reactions.

The solid obtained in the dialuminum case is much less than seen with the trivalent aluminum species and it is brown in color rather than white.

Third, the methoxy compounds generated in the reaction of the tetraiso­ butyldialuminum show chemical shifts that are downfield with respect to the shifts of the tri isobutylaluminum sample, and these remain in solution for a longer period of time.

With the above information in mind a discussion of the chemistry of

the reaction of tetraisobutyldialuminum and trimethoxyboron may be under­

taken. The reaction, as described in the Experimental section, was 7G 5 4 3 2 1 0

Figure32. Proton NMR Spectrum of Volatile Materials Removed from Reaction of Tetraisobutyldialuminum and Trimethoxy­ boron. ro 127

performed using an excess of the boron reagent and s tirrin g the mixture

for almost one week at room temperature. The mixture was then partially

filtered, removing about one half of the liquid. The remaining material

was then subjected to dynamic vacuum fo r several days.

The volatile material removed from the reaction mixture proved to be

unreacted trimethoxyboron and an isobutyl-substituted boron species. The

la tte r results from an exchange reaction between the aluminum and boron reagents of the original mixture. The NMR spectrum of this volatile material (Figure 32) shows two doublets centered at 1.03 and 0.99 pom, which are assigned to the methyl and methylene protons of isobutyl groups bound to boron. Similar peaks were found in the spectra of reaction prod­ ucts of tetraisobutyldialuminum and boron trichloride (Figure 31) and the dialuminum compound and tris(dim ethylam ino)boron, to be discussed below.

Cyclopentane was then added to the non-volatile residue and the

mixture filtered, separating a dark red-brown solution from a dark

brown, viscous, oily material. The latter was present in small quantity.

It reacted immediately upon exposure to the air and gave dihydrogen

upon treatment with water. Other volatile materials from this particu­

lar hydrolysis were found to be isobutane and methanol. A trace of

boron was also found to be present in this material. The ratio of

dihydrogen to other v o la tile m aterials was found to be s lig h tly

greater than 1 . 0 .

A NMR sample could not be prepared from the viscous oil as it

proved to be insoluble in a ll common solvents. An i-fra re d spectrum of

the material was obtained and is shown in Figure 51, Table 13. The spec­

trum is very similar to that found for the soluble material isolated in Figure 51. In frared Spectrum of Insoluble M aterial from Treatment of Tetraisobutyldialuminum with Trimethoxyboron.

ÎN5 00 129

Table 13

Infrared Frequencies fo r Products o f Treatment of Tetraisobutyldialuminum with Trimethoxyboron .

Wavenumber (cm~^) ±3

2980 1324

2868 1185

2780 1163

2716 1060

2607 948

1466 821

1403 682

1380 578

1364 475 130

the reaction. Of particular interest is the band at 1363 cm~\ which

is in d icative of methoxyboron compounds with reference to the spectrum of trimethoxyboron (B-0, 1365 cm '\ Figure 14, Table 1, Appendix).

Due to the intractability of this material and the fact that it ap­ peared to be mixed with a small amount of aluminum m etal, from which i t could not be separated by common techniques, no further attempts at char­ acterization of this species was undertaken. From the results one may infer the presence of aluminum or boron in a low valent state. Similarly, the reaction of tetramethoxydiboron and trimethylaiuminum produced an in­ soluble species which analyzed as [AlBtOCHgjCHg]^.

The cyclopentane solution is of greatest interest here. The major species present is Algfi-C^Hgj^fOCHg)^, K Analytical and spec­ troscopic information obtained from this material is consistent with its formulation as a unique species containing a covalent aluminum- aluminum moiety.

Hydrolysis of a sample of shown in equation 31 with theoretical

1 + gfxsjHgO -> 3(3.0)A1(0H)3 + 4(4.2)i-C4H^Q + 3(3.1)CH40 + (l.OlH^ (31) and experimental quantities, resulted in the generation dihydrogen, iso­ butane and methanol. The firs t two materials were quantified by isola­ tio n in a known volume. Methanol could not be separated from the hydro- stat quantitatively and was determined spectroscopically. The ratio of isobutane to dihydrogen was found to be 4:1.

The hydrostat from the hydrolysis did not give a positive flame test

fo r boron nor did i t show traces o f grey or black s o lid , which would in ­

dicate the presence of metallic aluminum. The former observation elim i­

nates the boron-containing m oieties from consideration as sources o f 131

dihydrogen in the hydrolysis reaction.

Spectroscopic information, which w ill be presented and analyzed in

greater detail below, that is pertinent to this discussion focuses on

two points. Integration of the relative areas of absorbances assigned to

methoxy groups and isobutyl groups in the NMR spectrum of I_ (Figure 35)

gives a ra tio o f methoxy to isobutyl o f 3 :4 . This implies a ra tio of

methoxy to aluminum o f 1 : 1 .

The infrared spectrum of 2 (Figure 33, Table 13) exhibits no bands

in region around 1800 cm"i, which have been previously discussed as in d i­

cative of the presence of the aluminum hydride moiety.GS This fact is

relevant to the source of dihydrogen from the hydrolysis reaction of

and indicates th a t oxidation of aluminum-aluminum bonds is responsible

for the production of dihydrogen.

Analysis fo r aluminum contained in was performed by gravim etric

determination of the 8 -hydroxyquinoline compound of the metal. The ratio

of aluminum to isobutane was found to be 3:4, which indicates the ratio

of aluminum to dihydrogen is 3:1. Therefore, two-thirds of the metal

must be present in the low valent state to account fo r the production

of dihydrogen upon hydrolysis. The remaining aluminum is present in an­

other form; the most plausible state is that of an Al(III) moiety. The

ra tio of Al(II) to Al(III) is thus 2:1.

An empirical formula which incorporates all of these data may be written. is thus formulated as A13 ( 1 - 81 ^119) 1+(00113)3 and w ill be shown to be consistent with the properties of catenated aluminum compounds.

The in frared spectrum of I_ (Figure 33, Table 13) shows evidence of the presence o f a lk y l and alkoxy groups bound to aluminum. I t is very l i J_L 1 ■ > I j * I I I 7 S 4 3 2 1

Figure 35. Proton NMR Spectrum of I_, Algfi-C^Hgj^fOCHgja.

W ro 40001 1 0 0 13001400 1000 #00 #00 400 WAVCNUMOCR ICM'1

Figure 33. Infrared Spectrum of Product from Treatment of Tetra­ isobutyldialuminum with Trimethoxyboron, Neat Oil.

W CO 134

Table 13

Infrared Frequencies for Products of Treatment of Tetraisobutyldialuminum with Trimethoxyboron.

Wavenumber (cm~^) ±3

2980 1324

2868 1185

2780 1163

2716 1060

2607 948

1466 821

1403 682

1380 578

1364 475 135

s im ilar in appearance to the spectrum obtained from diisobutylmethoxy-

aluminum (Figure 19, Table 1, Appendix). The notable bands are present

at 1016 (0-C on aluminum), 469 and 441 cm ~\ These la tte r two

bands are in the region assigned to ring stretches for the cyclic struc­

tures that alkoxy aluminum compounds assume. The stronger o f these two bands appears to be at a slightly higher value than seen in dialkylalkoxy-

aluminum compounds and may indicate the presence of a term inal methoxy group.

The strong, broad absorbance a t 579 cm"^ is assigned to an AIC2 u n it, with reference to the spectra of dimethyl methoxyal umi num and di ethyl- methoxyaluminum, 686 and 655 cm"^, respectively.A s stated above, no absorptions consistent with ühe presence of hydride bonded to aluminum are present in the spectrum.

The NMR spectrum of I_ (Figure 35) shows two major areas of ab­ sorbances. Signals centered at 1.33 and 0.34 ppm are assigned to the methyl and methylene protons of isobutyl groups bound to aluminum, with reference to the spectra of tetraisobutyldialuminum (1.25 and 0.25 ppm,

Figure 16) and diisobutylmethoxyaluminum (1.18 and 0.22 ppm. Figure 18). The absorbances a t lower fie ld , singlets a t 3.40, 3.32, 3.26 and 3.19 ppm, are assigned to the methyl protons of methoxy groups bound to alu­ minum with regard to the spectrum of diisobutylmethoxyaluminum (3.31 and 3.19 ppm in Figure 18).

The singlets at 3.26 and 3.19 ppm in the spectrum of are most sim ilar to those seen in the spectrum of the triv a le n t aluminum species.

However, the absorbance at 3.26 ppm is shifted slightly upfield from the sim ilar peak in the spectrum of the triv a le n t species (found a t 3.31 136 ppm) and the in ten sities are not the same. The upfield resonance at 3.19 ppm is of much lower intensity than that of the downfield peak at 3.26 ppm in the spectrum of while the spectra of diisobutylmethoxyaluminum

(Figures 18 and 56) show the upfield peak to be of greater intensity in both cases.

Thus, the singlet a t 3.19 ppm is assigned to the trim eric form of diisobutylmethoxyaluminum. This is an impurity which is present as less than 10 percent of the material, according to the integrated area under the peak as compared to the methoxy region as a whole. The singlet at

3.26 ppm, on the other hand, represents one chemical environment of methoxy groups of K Due to the similarity of the chemical shift to absorbances found in the triv a le n t species this methoxy group most lik e ly occupies a bridging position.

A small absorbance due to the dimeric diisobutylmethoxyaluminum might be expected since the trim eric form is present. I f indeed present, the peak would be very small and obscured by the broad singlets a t lower fie ld , 3.40 and 3.32 ppm. These la tte r absorbances in the spectrum of

2 have no counterpart in the spectra of the trivalent species. They are therefore assigned to methoxy groups bound to a low valent aluminum moiety.

A structure fo r % incorporating this evidence can be proposed. I t should have three d iffe re n t environments fo r methoxy groups and a t lea s t two environments for isobutyl groups. The ratio of these ligands should be 3 :4 . I t should contain an aluminum-aluminum bond per unit and also an aluminum atom in the +3 oxidation state. Figure 55 represents a structure which is consistent with a ll these factors. R R / \ R A l A l

ME Me

A l A l / \ "Me O R R OMe

Figure 55. Proposed structure of Algfi-C^Hgj^tOCHg)^, U

CO 138

Three different environments for methoxy groups are apparent. The

bridging ligand between the two trivalent aluminum atoms is consistent

with the s in g le t found a t 3.26 ppm. I t has a chemical s h ift which is

close to that for the similar diisobutylmethoxyaluminum. The difference

is attributed to the fact that the aluminum atom has another methoxy group

associated with it rather than a second isobutyl ligand. Further

corroberation is unavailable due to the fact that insolubility of pure di-

methoxyaluminum compounds precludes NMR studies.

This s itu a tio n also makes assignment o f the two methoxy groups asso­

ciated with the low valent aluminum atoms more d if f ic u lt . The lower

field absorbances are assigned to these moieties in accordance with evi­ dence that low valent compounds generally show lower field resonances.

This is true fo r the comparison o f trimethoxyboron and tetramethoxydiboron,

as stated earlier. The former shows a singlet at 3.48 ppm (Figures 53

and 54),'while the latter exhibits a singlet at 3.68 ppm.^®

Comparison o f an aluminum system th a t can be considered analogous

and possesses bridging and term inal groups is helpful here. The NMR

spectrum of tris(dimethylamino)aluminum exhibits two resonances at 1.98

and 1.78 ppm in a ratio of 2:1, which are assigned to terminal and

bridging ligands, respectively.s?

If this trend is followed in the methoxy case, the highest field

resonance would be assigned to methoxy group bridging between the low

valent and tr iv a le n t aluminum. The th ird resonance, interm ediate in

chemical shift, is thus assigned to the terminal methoxy group associated with a low valent aluminum atom. The proximity of these absorptions, in contrast to the la rg e r separation seen in the example aluminum compound. 139

is lik e ly due to the fa c t that the bridging methoxy group in th is case

spans two different types of aluminum atoms and the resonance is shifted

to a higher fie ld by the association with the triv a le n t aluminum. Such behavior is seen in Bs[N(CH 3) 2 ] i 2Al(CH3) i 2 ,^^ in which the amino ligand is bridging between boron and aluminum and shows a chemical s h ift in te r­

mediate with respect to that seen for the ligand bonded solely to either.

Also present in this structure are three different chemical environ­

ments for isobutyl ligands. The ligands are associated with three-

coordinate low valent aluminum, four-coordinate low valent aluminum, and

four-coordinate tr iv a le n t aluminum atoms. The subtle differences in en­

vironment would account fo r the breadth o f the signals found fo r the

resonances assigned to these ligands. Indeed, the absorption consistent

with the methylene protons of the ligand, which should show the greatest

sensitivity to such different environments, is almost 1 /2 ppm wide.

Other structures, with greater or lesser degrees of aggregation are

also possible. A monomer of the structure shown in Figure 55 may be

present. More likely is the possibility that a dimer-trimer combination

exists in solution, similar to that seen for diisobutylmethoxyaluminum.

Different degrees of association would tend to broaden NMR absorbances.

The mechanism of formation of X should be consistent with the chemis­ tr y seen fo r s im ila r systems. The f i r s t step would be the same as th at

proposed in the other reactions of tetraisobutyldialuminum with other

group IIIB reagents, association of the trimethoxyboron with the dialumi­

num compound, and subsequent ligand exchange. The products of this reac­

tion would be isobutyldimethoxyboron and triisobutylmethoxydialuminum. 140

The former material was identified as a volatile by-product of the reac­ tion (Figure 32). The latter would prove to be unstable as was the case in the previous reactions in which a similar substitution was determined to have occurred. The second step is therefore the disproportionation of this species to diisobutylmethoxyaluminum and "Al (i-Ci^Hg)".

This reactive species would most lik e ly in te ra c t with the excess trimethoxyboron available. It is most probably the source of the insolu­ b le viscous o il th a t remained on the f r i t a fte r removal of the soluble . species during the work-up of the reaction mixture.

Returning to the mechanism of formation of X» the firs t formed diiso­ butylmethoxyaluminum exists as a dimer. It w ill react with the excess trimethoxyboron in another exchange reaction to yield isobutyldimethoxy- aluminum. Normally, this material would be expected to be insoluble.

However, association of the material with a dialuminum species would cause it to remain in solution in this aggregated form. •

Just what form the dialuminum species which is complexed has a t th a t time is ambiguous. The disproportionation reaction in the previous cases

is thought to be rapid. While methoxy ligands are known to lend con­

siderable stability to diboron species, the disproportionation of a methoxy-substituted dialuminum moiety should s till be considered to be a

rapid reaction.

Most likely then, the species coordinated would be unreacted tetra­

isobutyldialuminum. In te ra c tio n w ith a low valent aluminum atom through

a methoxy bridge is the^expected method. Further reaction of one portion

of this molecule is possible. The low valent aluminum atom that is s till

three-coordinate can associate with addition of trimethoxyboron and undergo 141

an exchange reaction to acquire a methoxy ligand. The difference in the

system a t th is point in the reaction scheme is that the pathway to decom­

position, the 1,2 shift of a ligand, and subsequant disproportionation,

is no longer available. The adjacent aluminum atom is four-coordinate and

does not possess a vacant site to accept transfer of the ligand.

Further substitution of methoxy for isobutyl ligands at the only

site of aluminum unsaturation would be much slower, if occurring at a ll.

The low valent aluminum atom, shown to be a re la tiv e ly weak Lewis acid due to its greater electron density, would become even less so w ith the

addition of a strong pi donating ligand such as the methoxy group.

The structure and mechanism for formation of I have been elucidated which are consistent with the data obtained and the known chemistry of

similar systems. Probably the best way of considering this species is as

a stable intermediate in a more complicated scheme which involves fur­

ther decomposition.

This brings up a final point in the discussion of this system, reac­

tion conditions. In the specific case resulting in formation of a large excess of trimethoxyboron was added to a sample of tetraisobutyldialumi­

num. The liq u id methoxy compound served as a good medium to perm it s t i r ­

ring and obviated the need for a solvent to perform the reaction. After

approximately one week at room temperature, the volatile products were

removed by dynamic vacuum fo r several more days. The resu lting o il was

dissolved in a non-polar solvent and what may be termed "selective" or

fractional filtration performed to isolate a relatively pure sample of L

The solvent was removed as quickly as possible from the sample after f i l ­

tration. Slightly different NMR spectra and analysis were obtained if 142 a solvent such as cyclopentane was employed to perform the reaction or i f the selective f il t r a t io n technique was not used. Figures 58 and 34 show the spectra of the reaction products in these two cases. In Figure

58 note the low intensity of the absorbances at lowest field. Decomposi­ tion o f ^ is promoted by the presence o f a solvent. In Figure 34 note the great in te n s ity o f the absorbance a t highest fie ld of the methoxy region (3.19 ppm) indicating the presence of much diisobutylmethoxy­ aluminum. Selective f ilt r a t io n was used to decrease the concentration of this impurity by more than a facto r o f fiv e to obtain the sample of

L The point here is that the reaction conditions are an important factor in determining the final composition of the reaction mixture.

By using the stated procedure the material was isolated in better than 90 percent purity although less than 50 percent yield. It was stable for several days in benzene solution and several weeks as a semi- solid oil at room temperature.

The mode of decomposition is most likely dissociation of the triva­ lent aluninum moiety and disproportionation of the remaining dialuminum species in the same manner as set forth in the reaction scheme for pro­ duction of diisobutylmethoxyaluminum.

With Tris(dimethylamino) boron

The reaction of tetraisobutyldialuminum and tris (dimethylamino)boron in

the presence of an excess of the boron compound did not result in the

isolation of a dimethyl amino-substituted dialuminum species although i ! I I I I I I I 1 I I I I I I I I I I I I r ...... I I I I , l.l : i I I j_ L

Figure 58. Proton NMR Spectrum of Product from Treatment of Tetraisobutyldialuminum with Trimethoxyboron in Cyclopentane. 4^ CO 1 I ! I I 1 I I 1 I I 1 I I I I I I I I I I I I I I 1 1 ■ I I I I ■ ' ' ' I I ■ ■ I ■ I I ! ■ I I I I I I I ■ . I .

Figure 3 4 . Proton NMR Spectrum of Product from Treatment of Tetraisobutyldialuminum with Trimethoxyboron, Fully Extracted. 145

ligand exchange was found to take place due to the presence of isobutyl

boron species in the reaction mixture. A boron-free residue could be ob­

tained from the reaction mixture but only by dynamic pumping at tempera­

tures higher than ambient.

The reaction conditions and work-up were essentially the same as

those employed with success in the methoxy system discussed above, with

the exception of the higher temperature need to remove the boron species.

Confirmation of the latter was obtained from the NMR spectrum of the

species pumped from the reaction mixture (Figure 36). The singlet at 2.61

ppm is assigned to the methyl protons of a dimethylamino group bound to

boron. A negative result from a flame test for boron on a portion of the

non-volatile residue further substantiates this claim.

Figure 36 also supports the idea that an exchange of dimethylamino

ligands for isobutyl ligands occurs when the starting materials are com­

bined. The less intense pair of doublets centered at 1.04 and 0.92 ppm

are consistent with absorbances expected fo r methyl and methylene protons

of isobutyl groups bound to boron. Similar results were noticed in the

spectra of the reaction products of tetraisobutyldialuminum and boron t r i­ chloride (methyl proton resonance centered a t 0.93 ppm. Figure 31) and

trimethoxyboron (methyl at 1.03 ppm. Figure 32).

Hydrolysis of the soluble, non-volatile product of the reaction pro­ duced isobutane and dihydrogen. A typical ratio of isobutane to dihydro­ gen was 8:1 but varied as much as 30 percent in several trials from sepa­ rate attempts to perform the reaction. Upon treatment with water, the dark red-brown color was in stantly discharged but a small amount of grey or black solid was noticed in the bottom of the flask. Stirring of the 7 6 5 4 3

Figure 35. Proton NMR Spectrum of V o la tile M aterials Removed from Reaction of Tetraisobutyldialuminum and Tris(dimethyl- amino)boron. 147 solution for one or two hours resulted in the disappearance of this solid, but successful isolation of the material could not be effected.

The infrared spectrum of the soluble oil (Figure 38, Table 14) is consistent with isobutyl and dimethylamino groups attached to alumi­ num by reference to the spectrum of diisobutyl dimethyl ami noaluminum (Fig­ ure 21, Table 5, Appendix). The important bands are found at 532 cm'i, which is assigned to the Al-N stretch;?^ 1006 cm -i, which is assigned to the NC 2 s tre tc h ;15 and the broad absorbance 674 cm"i, which is assigned to the AIC 2 stretch. No bands typical of amino boron compounds are pres­ ent (B-N stretch at 1521 cm"i).^5 Also absent are any bands in the region of 1800 cm”i , which would indicate the presence o f an aluminum hydride moiety. This last observation supports the conclusion that the produc­ tion of dihydrogen is a re s u lt of oxidation of aluminum-aluminum bonds.

The iH NMR spectrum of the soluble o il is shown in Figure 39 and con­ sists of three major absorbances. A singlet at 2.19 ppm is consistent with the presence of an amino group attached to aluminum (sin g let a t 2.24 ppm in the spectrum of di i sobutyldimethyl ami noaluminum. Figure 20, Appen­ dix). The doublet centered at 1.17 ppm and the doublet centered at 0.25 are assigned to the methyl and methylene protons of an isobutyl group bound to aluminum, with respect to the same fig u re (1.21 and 0.25 ppm).

Minor peaks present are not well resolved. An absorbance appears to be present at 2.01 ppm and a broad shoulder is seen on the downfield side of the methyl of isobutyl resonance. This la tte r corresponds to the sim i­ la r absorbance found in the spectrum of the product from a sim ilar reac­ tion to be discussed subsequently. The former is most likely due to the presence of terminal dimethylamino groups on a trivalent aluminum species. 4000 3500 2500 3000 1800 1600 120 0 100 0 800 WAVENUMBER (CM1

Figure 38. Infrared Spectrum of Products from Treatment of T etra­ isobutyldialuminum with Tris(dimethylamino)boron, Neat O il.

4^ 00 149

Table 14

Infrared Frequencies for Products of Treatment of Tetraisobutyldialuminum with Tris(dimethylamino)boron.

Wavenumber (cm~^) ±3

2932 1036

2849 1006

1462 940

1400 923

1376 896

1359 807

1315 762

1300 740

1227 678

1175 636

1150 616

1113 535

1058 490

432 Figure 39. Proton NMR Spectrum of Products from Treatment of T etra­ isobutyldialuminum with Excess Tris(dimethylamino)boron, U1 o 151

Similarly, tri s(dimethylamino)aluminum shows singlets at 1.98 and 1.78

for bridging and terminal amino group protons.is

Integration of the relative areas proved difficult due to the breadth

of the absorbances and the fact that several overlapped. Comparison of the areas of the amino methyl groups and the methylene groups gives a ratio of approximately 1:2. The shoulder of the isobutyl methyl peak is less than one-half as intense as the methyl peak itself.

When this reaction was attempted in the presence of THF, in hopes of stabilizing the product, a very complex reaction mixture resulted. The non-volatile product(s) exhibited properties consistent with the presence of both boron and aluminum. An exchange reactio n , as seen in a ll of the previous reactions with group IIIB compounds, is again proposed as evi­ denced by data which shows isobutyl and dimethylamino groups bound to both boron and aluminum. Some THF is associated with the reaction product, but in an indeterminate stoichiometry. Heating of the material induced de­ composition, as evidenced by the change in appearance of the sample.

The ^H NMR spectrum of the benzene soluble portion of the non-volatile reaction residue (Figure 37) shows a complex array of absorbances. Id e n ti­ fiable are the following: the broad triplets centered at 3.64 and 1.76 ppm (partially obscured) are due to the presence of THF. The singlet at

2,68 ppm, which appears to have a shoulder at slightly higher field , is assigned to a dimethylamino group associated with boron, with reference to the starting material (recall Figure 36). By the same token, the group of signals centered about 1.08 ppm and consisting of four peaks, is con­ s is te n t with the presence of isobutyl groups on boron. Figure 37. Proton NMR Spectrum of Products from Treatment of Tetra­ isobutyldialuminum with Tris(dimethylamino)boron in cn Cyclopentane in the Presence of THF. ro 153

Other absorbances may be assigned to organoaluminum moieties. Two

doublets, one centered at 1.32 ppm and the other centered at 0.50 ppm,

re fle c t the presence of isobutyl groups bound to aluminum by analogy to

the spectrum of d i i sobutyldi methy1ami noalumi num (1.21 and 0.25 ppm for methyl and methylene, respectively. Figure 20, Appendix). With reference

to the same spectrum, the s in g le t at 2.38 ppm is assigned to the methyl

protons of the dimethylamino group attached to aluminum.

The remaining absorbances appear between 1.40 and 1.58 ppm. The

singlet at 1.58 ppm may be due to cyclopentane that was not completely removed from the reaction residue. The broad absorbance adjacent to the singlet to the high field side is not typical of the spectra of the com­ pounds thus far discussed. It appears to be in the region of methyl pro­ tons of isobutyl groups.

When the solvent is removed from this residue, a dark brown oil re­ mains. I t loses its color rap id ly upon exposure to a ir . A flame test of the oxidized material exhibits a green color which substantiates the presence of boron-containing species in the product.

No common chemical technique could be employed to separate the reac­ tion residue into components. Heating of the mixture resulted in the liberation of additional low-volatile boron compounds. This was also accompanied by decomposition, presumably via a disproportionation reac­ tion, since the presence of grey solid which proved to be metallic alumi­ num was noticed to form during the heating process.

The results of these two reactions indicate that, as in previous cases, ligand exchange does take place. The amino-substituted dialuminum compound could not be isolated from the other products of the reaction 154 without inducing decomposition o f the newly-formed species. This is in ­ dicated by the fact that while the NMR spectrum of the second reaction product exhibited a low field absorbance that is consistent with the pres­ ence of an amino group attached to a low valent aluminum atom, boron- containing compounds were also present. Techniques th at were successful in removing the boron species also resulted in the loss of the low field signal and growth of an absorbance consistent with the presence of an amino-substituted trivalent aluminum species.

The basic reaction scheme in th is system is analogous to that pro­ posed for the three previous cases involving tetraisobutyldialuminum and an aluminum or boron species. The f ir s t step involves an exchange reac­ tion of dimethylamino and isobutyl ligands:

Al2(1-C4Hg)4 + B[N(CH3)z ]3 -v Al2 (1 -Ci^Hg)3N(CH3 )2 + B[N(CH3) 2] 2 (i-M g) (31)

The dialuminum compound may dissociate in a second step, a reaction that is aided by the elevated temperatures necessary to remove the boron compound as a volatile material from the reaction mixture.

Algfi-C^HgigNfCHgiz Al(i-C 4Hg)2N(CH3)2 + "Al(i-C^Hg)" (32)

The intermediate "Al(i-Ci^Hg)" undergoes further reaction, most likely with excess tris(dimethylamino)boron, to give a low valent species that is insoluble and probably polymeric, as proposed in the methoxy case. Heat­ ing of the reaction mixture causes a further disproportionation of this species resulting in the formation of aluminum metal.

When THF was present in the reaction mixture, the reactions illu s­ trated by Equations 31 and 32 s till occurred. An added complication was that with the Lewis base included in the system the boron-containing spe­ cies could not be completely removed under the same conditions which allowed them to be pumped away in the absence of THF. 155

REACTIONS WITH TRANSITION METAL COMPOUNDS

Tetraisobutyldialuminum was treated with several transition metal

compounds: dimanganese decacarbonyl, pentacarbonylmanganese hydride, and

tris(triphenylphosphine) platinum, with the intention of forming and char­

acterizing metal-aluminum bonded species previously unknown. In general,

these attempts were marked by a lack of interaction between the dialumi­

num compound and the various metal species. Steric hindrance seems to be a major factor in the inhibition of these reactions. In the platinum

system, a reaction between the basic phosphine ligand and the dialuminum compound was noted.

With Dimanganese Decacarbonyl

Treatment of tetraisobutyldialuminum with dimanganese decacarbonyl in a non-polar solvent at room temperature resulted in essentially no reaction between the two species, even after contact of the reagents for three weeks. The manganese compound would be removed from the reaction mixture intact and virtually entirely (greater than 96 percent) by subli­ mation at ambient temperature. Some decomposition of the dialuminum spe­ cies was noticed, as evidenced by the formation of a grey precipitate in the reaction flask. This behavior is typical of solutions of tetra­ isobutyldialuminum over extended periods of time and is not thought to be due to interaction with the manganese compound.

Two possible sites of attack are available on the dimanganese spe­ cies, the metal-ligand bond and the metal-metal bond. The metal-ligand bond is usually inert to substitution unless forcing conditions are used, irradiation or heating.®® The metal-metal bond is subject to reduction as seen in the scheme to prepare the pentacarbonyl hydride species.^7 An 156 a lk a li metal, sodium or potassium, is employed in this reaction to pro­ duce the manganate s a lt. Evidently, the reducing power of the aluminum- aluminum bond is not sufficient to accomplish this.

Such a reaction would not be favored on steric grounds in any event.

Both metal-metal bonds are well insulated by the attached ligands.

With Pentacarbonylmanganese Hydride

Treatment of tetraisobutyldialuminum with two equivalents of penta­ carbonylmanganese hydride at room temperature in benzene resulted in the production of small amounts of dihydrogen and isobutane. A small amount of a brown solid was found as an insoluble product of the reaction. Di- manganese decacarbonyl was also isolated from the reaction residue.

The production of dihydrogen may have its source in two d iffe re n t reactions. Oxidation of aluminum-aluminum bonds is one possible reaction.

Decomposition of the manganese hydride is a second path, resu lting in the formation of a manganese-manganese bond with liberation of dihydrogen.

The total dihydrogen found accounted fo r 14 percent of the hydrogen available from the hydride. A portion of this is thought to be due to the decomposition o f the pentacarbonylmanganese hydride because of the small amount of the decacarbonyl compound that was isolated after the reaction.

The production of isobutane is due to cleavage o f the carbon-aluminum bond of the dialuminum species. It accounted for approximately 7 percent of the available hydrogen in the starting mixture. No isobutene was found as a volatile product of the reaction.

The infrared spectrum of the insoluble brown m aterial from the reac­ tion (Figure 40, Table 15) is characterized by broad, poorly-resolved absorbances. The presence of a manganese compound in some form is 7000 1000 1400 10" X) 800 600 WAVi N U M IU II (CM 1

Figure 40. Infrared Spectrum of Solid from Treatment of Tetraiso­ butyldialuminum with Pentacarbonylmanganese Hydride. 158

Table 15

Infrared Frequencies for Solid from Treatment of Tetraisobutyldialuminum with Pentacarbonylmanganese Hydride.

Wavenumber (cm~^) ±3

2850 1148

2015 1057

1910 1000

1852 800(br)

1500(br)' 636

1172 159 confirmed by the band in the carbonyl region a t 1950 cm"^. The absorbance due to a manganese hydride moiety (at 1785 cm'i) is absent in this com­ pound. An isobutyl aluminum moiety seems to be present due to strong ab­ sorbance a t 634 cm 'i, consistent with the spectrum o f triisobutylaluminum

(AlCz at 675 cm"^ in Figure 43, Table 18, Appendix). Other bands not due to the mulling agent are too weak to be discernible.

Attempts to determine the relative proportions of aluminum and man­ ganese by the method outlined in the Experimental section were apparently thwarted by the presence of m etallic aluminum in s o lid . The values ob­ tained, 3.6 and 3.21, fo r the ra tio of aluminum to manganese seem unrea­ sonable.

The NMR spectrum of the soluble portion of the reaction mixture

(Figure 41) shows three resonances. The doublets centered at 1.29 and

0.32 ppm are consistent with methyl and methylene protons of isobutyl groups bound to aluminum, found at 1.25 and 0.25 ppm in tetraiso b u tyld ia­ luminum (Figure 16, Appendix). The third resonance appears at -7.54 ppm and is assigned to the hydride of the unreacted starting manganese com­ pound.

The reactions in this case do not lead to a stable species containing a manganese-aluminum bond. The complexity of the system can be a t t r ib ­ uted to the fact that two reactions of the dialuminum compound occur, at least one of which involves disproportionation, to give metallic aluminum and the unidentified insoluble materials.

The slow reaction rates are due to steric hindrance, other acidic species such as HCl and toluene-3,4-dithiol give the same volatile by­ products, dihydrogen and isobutane, but react more readily. Further I I I I I I I L l I I I I I I I I I I I , I ! I I I I I -6

Figure 4 1 . Proton NMR of Solution from Treatment of Tetraisobutyl dialuminum with Pentacarbonylmanganese Hydride. § 161

complications are created by this phenomenon due to the low stability of

both of the starting materials.

The indeterminate results achieved here parallel the previous work

attempted in this area.33 Successful isolation of a manganese-aluminum moiety appears to require additional s ta b iliz in g forces than afforded with alkyl-su b stitu ted aluminum compounds.

With Tris(triphenylphosphine)platinum (0)

Treatment of tetraisobutyl dialuminum with tris(triphenylphosphine)-

platinum in benzene at room temperature for one week resulted in the asso­ ciatio n of a ligand from the platinum species with the dialuminum com­ pound and subsequent disproportionation o f the la tte r species. The platinum compound is thought to undergo reaction in a known manner.

Work-up o f the reaction mixture was accomplished by changing to a non-aromatic solvent and filtering the solution to separate the insoluble material. The solution obtained, when hydrolyzed, gave a ratio of iso­ butane to dihydrogen of 13.5:1. No platinum was found in the hydrostat from the solution.

The NMR spectrum of the solution is shown in Figure 56. Present are doublets centered a t 1.19 and 0.64 ppm, which are assigned to the methyl and methylene protons of isobutyl groups bound to aluminum, with reference to the spectrum of triisobutylaluminum (1.11 and 0.41 ppm. Fig­ ure 42, Appendix). Much less intense is a doublet centered at 0.27 ppm which is assigned to the methylene protons of unreacted tetraisobutyl- dialuminum (0.25 ppm in Figure 16, Appendix).

These results indicate th at the predominant aluminum species present is in the trivalent state. The A value of 0.55 is much closer to that of y-oK. U a .

Figure 56. Proton NMR Spectrum o f Product from Treatment o f Tetra- isobutyldialuminum with Tris(triphenylphosphine)plati- num. ro 163

triisobutylaluminum (0 ,7 0 , Table 17) than that o f the dialuminum species

(1.00). The dark red-brown color of the solution is due to a small amount

of tetraisobutyldialuminum which is present (and which accounts fo r the

small amount of dihydrogen evolved upon hydrolysis).

The infrared spectrum of the solid (Figure 57, Table 16) shows no

bands typical of alkyl aluminum compounds (A1C2 a t 675 cm"^ in t r iis o ­

butylaluminum). I t resembles the spectrum of the startin g platinum com­

pound (Figure 23, Table 7, Appendix).

The reaction scheme proposed fo r this system involves dissociation

of a triphenylphosphine ligand from the three-coordinate platinum,35 ac­

cording to Equation 33:

[P(C6Hs)3]3Pt t [P(C6Hs)3]2Pt + P(C6Hs)3 (33)

This ligand interacts in much the same way as that seen with other Lewis

bases in a previous section, i.e ., it accelerated the disproportionation

reaction of the dialuminum species, creating a trivalent aluminum species and aluminum metal.

A stronger interaction is possible between the trivalent aluminum spe­ cies and the phosphine ligand and i t is e ffe c tiv e ly removed from the platinum equilibrium situation. Bis(triphenylphosphine)platinum is not stable in the absence of excess phosphine and undergoes an associative reaction to produce the tetrameric compound [Pt P(C 6H5 ) 3]it.®^ The excess phosphine is free to reestablish the equilibrium of Equation 33. No evidence of oxidative addition to the platinum was found nor is a Lewis acid-base adduct formed.

Work done in th is la tte r area found trim ethyl aluminum to form the weakest adduct to tris(triphenylphosphine)platinum in a series, including 3000 1200 1000 WAVENUMBER(CM') WAVENUMBER (CM ')

Figure 5 7 . Infrared Spectrum o f Solid from Treatment of T etra­ isobutyldialuminum with Tris(triphenylphosphine)plati , num a fte r Exposure to A ir. 165

Table 16

Infrared Frequencies for Solid from Treatment of Tetrai sobuty 1di a1umi num wi th Tri s( t r i phenylphosphine)piatinum.

Wavenumber (cm~^) ±3

3044 1170

2935 1154

2844 1091

2753 1060

2580 1028

2271 1014

1960 995

1811 940

1663 913

1585 859

1572 840

1481 810

1460 786

1438 687

1400 532

1375 509

1357 435

1312 166

Silicon tetrafluoride and boron trichloride.36 it was thought that tr i­

methyl aluminum was too "hard" an acid to interact strongly with the soft

base. Tetraisobutyldialuminum should be a softer acid due to its lower

oxidation state, but s till may exhibit the same tendancy as trimethyl- aluminum toward the platinum.

Also, a negative factor is steric hindrance of the bulky ligands of both metals. This appears to be a factor in other systems discussed

previously and may well lim it the interaction of the dialuminum to one with the free phosphine only. SUMMARY

The chemistry of tetraisobutyldialuminum is dominated by the tend­ ency toward disproportionation of the aluminum-aluminum bond. Unlike diboron analogues, the metal-metal bond is the most reactive moiety of the dialuminum molecule. Therefore, facile routes to a variety of substituted low valent aluminum compounds are not a v aila b le , such as reaction with protonic reagents.

Reactions designed to induce substitution of ligands lead, in many cases, to disproportionation of the dialuminum species. Some stabiliza­ tion can be afforded by ligands with pi donor ability, such as methoxy groups, but the substituted species actually isolated in this case was also stabilized by interaction with a trivalent aluminum species.

Also in contrast to the diboron compounds were the fin a l products of disproportionation. Instead of cluster formation, as seen with boron, the tendency was to reduction to the zero valent metal.

The major facto r in the s ta b iliza tio n of tetraisobutyldialuminum ap­ pears to be s te ric in nature. This is also the reason fo r the slow rate of reaction in many instances. Electronic effects are not as significant but do play a role in directing the course of the reactions.

167 APPENDIX

168 MW BOO 600 400

Figure 14. Infrared Spectrum of Trimethoxyboron, Gas Phase.

o> to 170

Table 1

Infrared Frequencies for Trimethoxyboron

Wavenumber (cm~^) ±3

2955 1365(br)

2872 1190

2206 1032

2140 693

2070 660

1490 529 40Q0 3500 3000 IMO 1300 4 0 0

Figure 15. Infrared Spectrum of Tris(dimethylamino)boron, Gas Phase 172

Table 2

Infrared Frequencies for Tris(dimethylamino)boron

Wavenumber (cm~M ±3

2978 1220

2844 1193

2776 1114

1510 1062

1451 905

1398 _jvJ

Figure 16. Proton NMR Spectrum o f Tetraisobutyldialuminum

CO MOO 1 000 #00 • 0 0 ICM

Figure 17. Infrared Spectrum of Tetraisobutyldialuminum, O il. 175

Table 3

Infrared Frequencies for Tetraisobutyldialuminum

Wavenumber (cm~^) ±3

2980 1169

2875 1074

2780 1019

1464 950

1401 821

1382 775

1367 690

1326 438

1185 3900 9500 IMO 14009000 1900 1000 •0 0 •0 0 400 WAVENUMBER (CM 'I

Figure 19. Infrared Spectrum o f Diisobutylmethoxyaluminum, Neat Liquid .

' s i 177

Table 4

Infrared Frequencies fo r Diisobutylmethoxyaluminum

Wavenumber (cm~^) ±3

2934 1151

2845 1059

2757 1030

1461 1010

1400 977

1375 810

1358 664

1316 622

1174 4000 9000 3900 3000 WOO 1400 1900 MO MO 400

Figure 21. In frared Spectrum of Oiisobutyldimethylaminoaluminum, , Neat Liquid.

00 79

Table 5

Infrared Frequencies for Diisobutyldimethylaminoalurainum

Wavenumber (cm~^) ±3

2933 1154

2878 1113

2847 1060

2785 1038

2589 1008

1461 937

1399 896

1376 808

1359 673

1318 600

1228 434

1200

1177 9 00 0 7D00 laoo teoo 1400 #00 600 400

Figure 22. Infrared Spectrum of Pentacarbonylmanganese Hydride.

00 o 181

Table 6

Infrared Frequencies fo r Pentacarbonylmanganese Hydride

Wavenumber (cm -i) ±3

2950 1176

2492 1014

2428 981

2380 913

2217 855

2115 725

2090 655

2005 600

1939 571

1783 469

1267 4000 2000 laoo isoo uoo 1200 1000 BOO 600 400 WAVFNUMBER (CM 1

Figure 23. In frared Spectrum of Tris(tripheny1phosphine)platinum .

CO ro 183

Table 7

Infrared Frequencies for Tris(triphenylphosphine)platinum

Wavenumber (cm~^) ±3

3032 1060

2908 1027

2840 990

1580 964

1475 900

1460 834

1430 735

1373 711

1298 687

1260 609

1172 520

1148 510

1083 425 Figure 42. Proton NMR Spectrum of Triisobutylaluminum

4^0 3 3000 30001600 1600 (400 1000 600 600 400 WAVE NUMBER (CM*) WAVENUMBER (CM *)

Figure 43. Infrared Spectrum of T riiso butyl aluminum, Neat Liquid.

00 un 186

Table 18

Infrared Frequencies for Triisobutylaluminum

Wavenumber (cm~^) ±3

3187 1071

2920 1019

2780 950

2731 917

2610 821

1455 775

1400 675(br)

1382 551

1373 492

1324 430

1188 390

1169 i I » M » X. L ,L1 I I a a L i_ L jJL

Figure 4 4 . Proton NMR Spectrum of Diisobutyl aluminum Chloride .

CO 188

LIST OF REFERENCES

1. Schram, E. P .; H a ll, R. E .; Glore, J. D ., J. Am. Chem. Soc. (1969), 91, 6643.

2. Skell, P. S.; McGlinchey, M. J ., Angew. Chem. Internat. Edit., Engl. (1975), 14, 195.

3. Hoberg, H .; Krause, S ., Angew. Chem. (1976), 88, 760.

4. Hoberg, H.; Krause, S., Angew. Chem. Internat. Edit., Engl. (1978), 12, 949.

5. Wartik, T.; Moore, R.; Schlesinger, H. I., J. Am. Chem. Soc. (1949), 71, 3265.

6. Urry, G.; Wartik, T.; Schlesinger, H. I., J. Am. Chem. Soc. (1952), 74, 5809.

7. Lanthier, G. F .; Kane, J .; Massey, A. G ., J. Inorg. Nucl. Chem. (1971), 33, 1569.

8. Hursthouse, M. B.; Kane, J . ; Massey, A. G., Nature (1970), 228, 659.

9. Schram, E. P.; Urry, G., Inorg. Chem. (1963), 2, 405.

10. Peterson, L. L.; Brotherton, R. J., Inorg. Chem. (1963), 2, 423,

11. Trefonas, L.; Lipscomb, W. N ., J. Chem. Phys. (1968), 28, 54.

12. B iffa r , W.; Noth, H .; Pommerening, H ., Angew. Chem. In tern at. E d it., Engl. (1980), 1, 56.

13. Holliday, A. K.; Massey, A. G.. Chem. Rev. (1962), 62, 303.

14. B iffar, E.; Noth, H.; Pommerening, H., Angew. Chem. Internat. Edic.. Engl. (1980), 1, 56. ------

15. Amero, B. A .; Schram, E. P ., Inorg. Chem. (1976). 15, 2842.

16. Crompton, I . R ., "Analysis of Organoaluminum and Organozinc Com­ pounds," Pergamon Press, Oxford, 1968, p. 199.

17. Cohen, M.; G ilb e rt, J. K .; Smith, J. D ., J. Chem. Soc. (1965), 1092,

18. Koster, R .; Bruno, G ., Annal en (1960), 629, 89. 189

LIST OF REFERENCES (c o n t.)

19. Hoffmann, E. G ., Trans. Faraday Soc. (1962), 58, 642.

20. Jeffery, E. A.; Mole, T., Aust. J. Chem. (1973), 26, 739.

21. Ruff, J. K., J. Org. Chem. (1962), 27, 1020.

22. Purcell, K. F.; Kotz, J. C., "Inorganic Chemistry," W. B. Saunders Co., Philadelphia, 1977, p. 229.

23. K ris to ff, J. S .; Nelson, N. J . ; S hriver, D. F ., J. Organomet. Chem. (1973), 49, C82.

24. Kotz, J. C.; Turnipseed, C. D., Chem. Comm. (1970), 41.

25. Petz, W. J ., Organomet. Chem. (1973), 55, C42.

26. Brunner, H .; Wail es, P. C .; Kaesz, H. 0 ., Inorg. Nucl. Chem. Letters (1965), 1, 125.

27. Kroll, W. R.; McNicker, G. B., Chem. Conro. (1971), 591.

28. Rettig, S. J .; Storr, A.; Thomas, B. S.; Trotter, J., Acta. Cryst. (1974), 830, 666.

29. Forder, R. A.; Prout, K., Acta. Cryst. (1974), 830, 2312.

30. Schwarzhans, K. E .; S teiger, H ., Angew. Chem. (1972), 84, 587.

31. Petz, W.; Schmid, G ., J. Organomet. Chem. (1972), 35, 321.

32. LaPlaca, S. J., Inorg. Chem. (1969), 8, 1928.

33. Kinstle, G., Ph.D. Dissertation, The Ohio State University, Columbus, Ohio, 1976.

34. Ib id .

35. Birk, J. P.; Halpern, J.; Pickard, A. L ., Inorg. Chem. (1968), 7, 2672. ------

36. Durkin, T. S., Ph.D. Dissertation, The Ohio State University, Columbus, Ohio, 1971.

37. U.S. Patent 3,545,930 (1967). 190

LIST OF REFERENCES (co n t.)

38. Kolthoff, I. M.; Sandell, E. B.; Meehan, E. J.; Bruckenstein, S., "Quantitative Chemical Analysis," 4th Ed., The MacMillan Company, New York, 1971, p. 600.

39. Basset, J .; Denney, R. C.; Jeffery, G. H.; Mendhan, J., "Vogel's Textbook of Quantitative Inorganic Analysis," Longman, New York, 1978, p. 330.

40. Ibid, p. 434.

41. Srehla, G ., "Vogel's Textbook o f Macro and Semimicro Q u alitative Inorganic Analysis," 5th Ed., Longman, New York, 1979, p. 516.

42. Lehmann, W. J.; Onak, T. P.; Shapiro, I., J. Chem, Phys. (1959), 30, 1215.

43. Stull, D. R., Indus, and Eng. Chem. (1947), 39, 269.

44. Dorogy, E. E ., Personal Communications.

45. Becher, H. J .; Baechle, H. T., Z. Phys. Chem. (1966), 48, 359.

46. Schram, E. P ., Unpublished Studies

47. Edgell, W. F.; Risen, W. M., Jr., J. Am. Chem. Soc. (1966), 83, 5451

48. Adams, D. M. ; Squire, A., J. Chem. Soc. (A) (1968), 2817.

49. Plummer, J. F .; Schram, E. P ., Inorg. Chem. (1975), 14, 1505.

50. Huheey, J. E., "Inorganic Chemistry," 2nd Ed., Harper and Row, New York, 1978, p. 181.

51. Ibid, p. 142.

52. Browstein, S .; Smith, B. C .; E hrlich , G. ; Laubengayer, A. W., J. Am. Chem. Soc. (1959), 81, 3826.

53. Mole, T., Aust. J. Chem. (1963), 16, 801.

54. H arris, J. M.; Wamser, C. C ., "Fundamentals of Organic Reaction Mechanisms," John Wiley and Sons, In c ., New York, 1976, p. 162.

55. Purcell, K. F.; Kotz, J. C., "Inorganic Chemistry," W. B. Saunders Co., Philadelphia, 1977, p. 456. 191

LIST OF REFERENCES (c o n t.)

56. Davidson, N .; Brown, H. C ., J. Am. Chem. Soc. (1942), 64, 316.

57. Amero, B. A., Ph.D. D issertation, The Ohio State U n iversity, Columbus, Ohio, 1975.

58. Wiburg, E .; May, A ., Z. Naturforsch. (1955), 10b, 234.

59. Schmidt, D. L.; Flagg, E. E., Inorg. Chem. (1967), 6, 1262.

60. Day, A. R .; J o u llie , M. M ., "Organic Chemistry," D. Van Nostrand, Princeton, New Jersey, 1960, p. 565.

61. Carey, A. A., Ph.D. Thesis, The Ohio State University, Columbus, Ohio, 1980.

62. Satchel 1, D. P. N.; Satchel 1, R. S., Chem. Rev. (1969), 69, 251.

63. Maslowski, E., J r., "Vibrational Spectra of Organometal 1ic Compounds," John Wiley and Sons, Inc., New York, 1977, p. 40.

64. Carey, A. A ., Ph.D. Thesis, The Ohio State U niversity, Columbus, Ohio, 1980.

65. Carey, A. A ., Personal Communications.

66. Schomburg, Von G.; Hoffman, E. G., Zeit. fur Electrochemie (1957), 61, 1110.

67. Ramey, K. C.; O'Brien, J. F.; Hasegawa, I . ; Bouchert, A. E ., J. Phys. Chem. (1965), 69, 3418.

68. Smith, C. A.; Wallbridge, M. G. H., J. Chem. Soc. A (1970), 2675.

69. Popov, A. P.; Forost, M. P.; Kurbatov, R. N.; Korneev, N. N.; Badaev, V. K ., J. Appl. Chem. U.S.S.R. (1973), 46, 142.

70. Mole, T .; J e ffe ry , E. A ., "Organoaluminum Compounds," E lsevier, Amsterdam, 1972, p. 37.

71. P itz e r, K. S .; Gu.towski, H. S ., J. Am. Chem. Soc. (1946), 68, 2204.

72. Yamamoto, 0.; Hayamizu, K., J. Phys. Chem. (1968), 72, 822.

73. Mfinn, Von G. ; Haaland, A.; Weidlein, J., Z. Anorg. Allg. Chem. (1973), 398, 231. / ‘ ■' ■'92 I

LIST OF REFERENCES (c o n t.)

74. Mole, T., Aust. J. Chem. (1966), 19, 381.

75. Owens, R. M., J. Organomet. Chem. (1973), 55, 237.

76. Jefferey, E. A.; Mole, T.; Aust. J. Chem. (1968), 21, 2683.

77. Dorogy, W. E., Ph. D. D issertation, The Ohio State University, Columbus, Ohio, 1982.

78. Ib id .

79. Maslowski, E ., J r., "Vibrational Spectra of Organometallie Compounds," John Wiley and Sons, In c ., New York, 1977, p. 158.

80. Cotton, F. A.; Wilkinson, G., "Inorganic Chemistry," 3rd Ed., Interscience, New York, 1972, p. 701.

81. Gillard, R. D.; Ugo, R.; Caraiti, F.; Cenini, S.; Bonati, F., Chem. Comm. (1966), 869.