@prefix vivo: . @prefix edm: . @prefix ns0: . @prefix dcterms: . @prefix skos: . vivo:departmentOrSchool "Science, Faculty of"@en, "Chemistry, Department of"@en ; edm:dataProvider "DSpace"@en ; ns0:degreeCampus "UBCV"@en ; dcterms:creator "Oakley, Richard Thomas"@en ; dcterms:issued "2011-05-11T22:06:11Z"@en, "1970"@en ; vivo:relatedDegree "Master of Science - MSc"@en ; ns0:degreeGrantor "University of British Columbia"@en ; dcterms:description """The isolation of nitrilohexaphosphonitrilic chloride is reported in this work, and its structure shown to be a tricyclic condensed ring. The physical and chemical properties of this molecule are unlike the monocyclic phosphonitrilic chlorides. The reasons for this are not fully understood, but are thought to involve the weakness of the central P-N bond; the crystal structure and mass spectrum of the molecule show that this central bond is long and weak. Substitution reactions, common in the monocyclic chlorides, bring about either a partial or total break-down of the tricyclic P₆N₇ skeleton. In the case of electron-withdrawing substituents (e.g. -F,OMe), complete loss of the phosphonitrilic skeleton is achieved, and, in the case of electron-donating groups (-NMe₂), substitution onto the ring brings about cleavage of one of the internal P-N bonds. The molecule that is formed during the dimethylamination of P₆N₇Cℓ₉ is unique in phosphonitrilic chemistry and its suggested structure visually emphasizes the weakness of the central P-N bond in P₆N₇Cℓ₉. A series of molecular orbital calculations on the P₆N₇Cℓ₉ molecule were carried out in this study, and it is suggested that the observed structural trends in the molecule can be attributed to the properties of the phosphorus 3d[subscript]yz orbital on the bridgehead atoms."""@en ; edm:aggregatedCHO "https://circle.library.ubc.ca/rest/handle/2429/34476?expand=metadata"@en ; skos:note "THE CHEMISTRY OF NITRILOHEXAPHOSPHONITRILIC CHLORIDE by RICHARD THOMAS OAKLEY Sc.(Hons.), University of B r i t i s h Columbia, 1 A THESIS SUBMITTED IN PARTIAL FULFILMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE i n the Department of Chemistry We accept t h i s t hesis as conforming to the required standard THE UNIVERSITY OF BRITISH COLUMBIA August, 1970 In p r e s e n t i n g t h i s t h e s i s i n p a r t i a l f u l f i l m e n t o f the r e q u i r e m e n t s f o r an advanced degree at the U n i v e r s i t y o f B r i t i s h C o lumbia, I a g r e e t h a t the L i b r a r y s h a l l make i t f r e e l y a v a i l a b l e f o r r e f e r e n c e and s t u d y . I f u r t h e r agree t h a t p e r m i s s i o n f o r e x t e n s i v e c o p y i n g o f t h i s t h e s i s f o r s c h o l a r l y p u rposes may be g r a n t e d by the Head o f my Department o r by h i s r e p r e s e n t a t i v e s . I t i s u n d e r s t o o d t h a t c o p y i n g o r p u b l i c a t i o n o f t h i s t h e s i s f o r f i n a n c i a l g a i n s h a l l not be a l l o w e d w i t h o u t my w r i t t e n p e r m i s s i o n . Department The U n i v e r s i t y o f B r i t i s h Columbia Vancouver 8, Canada ABSTRACT • ^ The i s o l a t i o n of n i t r i lohexaphosphoni t r i i i c chloride i s reported i n th i s work, and i t s structure shown to be a t r i c y c l i c condensed r i n g . The physical and chemical properties of th i s molecule are unl ike the monocyclic phosphoni t r i l i c ch lor ides . The reasons for th i s are not f u l l y understood, but are thought to involve the weakness of the centra l P - N bond; the c rys t a l structure and mass spectrum of the molecule show that th i s central bond i s long and weak. Subst i tu t ion react ions, common i n the monocyclic ch lor ides , br ing about ei ther a p a r t i a l or t o t a l break-down of the t r i c y c l i c P^Ny skeleton. In the case of e lec t ron-withdrawing substituents C e-g- -F,OMe), complete loss of the phospho-n i t r i l i c skeleton i s achieved, and, i n the case of electron-donating groups ( - N I V i ^ ) > subs t i tu t ion onto the r i ng brings about cleavage of one of the in te rna l P - N bonds. The molecule that i s formed during the dimethylamination of P^N^CA^ i s unique i n phosphoni t r i l i c chemistry and i t s suggested structure v i s u a l l y emphasizes the weakness of the central P - N bond i n PJsLCJL.. o / y A series of molecular o r b i t a l ca lcula t ions on the P - N - C i l - molecule o / y were car r ied out i n th i s study, and i t i s suggested that the observed s t ruc tura l trends i n the molecule can be a t t r ibuted to the properties of the phosphorus 3d^ z o r b i t a l on the bridgehead atoms. - i i -TABLE OF CONTENTS Page ABSTRACT i TABLE OF CONTENTS . . i i LIST OF TABLES . v LIST OF FIGURES •••••• vi> ACKNOWLEDGEMENTS _ v i i i CHAPTER 1. INTRODUCTION 1 CHAPTER 2. PREPARATION OF NITRILOHEXAPHOSPHONITRILIC CHLORIDE 10 I. Condensation Reactions 10 II. High Temperature Rearrangements 11 EXPERIMENTAL •. 15 (a)-(d) Thermal Decomposition of Phospho-n i t r i l i c High Polymer 15 (e) Reaction of Lithium N i t r i d e with P ^ 6 ^ 1 2 ^ (f) Reaction of (Me 3Sn) 3N with ^ 6 ^ 6 ^ ^ 1 2 ••• 2 1 (g) Reaction of (Me 3Si) 3N with P^^Cl^ 21 CHAPTER 3. PROPERTIES OF NITRILOHEXAPHOSPHONITRILIC CHLORIDE 23 31 I. P n.m.r. Spectrum 23 II. Mass Spectrum 25 III . U l t r a - v i o l e t Spectrum 31 IV. V i b r a t i o n a l Spectra 32 V. Base Strength Measurements ... 38 VI. Lewis Acid Adduct Formation 42 - i i i -Page CHAPTER 4. SUBSTITUTION REACTIONS OF NITRILOHEXAPHOSPHO-NITRILIC CHLORIDE 43 I. (a) F l u o r i n a t i o n of P 6 N ? C £ g 43 ( i j\" Reaction of P ^ C ^ with KS0 2F i n Cyclohexane . ... 46 ( i i ) Reaction of P,N,C£.„ with Activated KF ....tt.1.2.. 47 C i i i ) Reaction of PJvLCfl,. with Activated KF 48 Civ ) Reaction of P 6 N y C £ g with AgF 48 I. (b) Methoxylation of P 6 N y C £ 9 49 II . Amination of VMnCln 52 o / y I I I . The Molecular Structure of P ^ (NH) (NMe 2) 8C£ 2 56 (a) Infra-red Spectrum • 57 (b) Mass Spectrum ••••• 57 31 (c) P n.m.r. Spectrum 58 (d) '''H n.m;r. Spectrum 62 (e) Conformational Analysis 69 IV. Summary 72 CHAPTER 5. THE MOLECULAR AND ELECTRONIC STRUCTURE OF NITRILO-HEXAPHOSPHONITRILIC CHLORIDE 74 I. O r b i t a l Interactions i n V,N.Cln 74 II. Symmetry-based Calculations on P^N^Citg .... 78 I I I . The Molecular Structure of P ^ C ^ 83 IV. Detailed Hiickel M.O. Calculations on P^N^C£„ 89 - iv -Page APPENDIX I. EVALUATION OF OVERLAP INTEGRALS IN P.N_C£ n ... 97 6 7 9 APPENDIX II. DETAILS OF INSTRUMENTAL TECHNIQUES 101 REFERENCES • 103 - V -LIST OF TABLES Table 3:1. 3 1 P chemical sh i f t s for (PNCXJ- , and P.N.,CJL,. 3:2. Fragmentation pattern of P^N^Cilg and P ^ 6 ^ 1 2 3:3. Relat ive abundances of a l l ions i n c y c l i c series i n fragmentation pattern of PgN^C&g 3:4. Vib ra t iona l spectra of P^N^Cl^ 3:5. Vib ra t iona l spectra of P,.N.CJL „ . . . . • o o 1 z 3:6. Assignments for the v (PCJ^) in-phase mode i n phospho-n i t r i l i c chlorides . . f ^ ? > 3:7. Average LPNP i n (PNC£ 2 ) 3 _ 5 . . . . . . . . . 4 :1 . Data from 3 1 P n.m.r. .spectrum of P ^ (NH) (NMe 2 ) g C£ 2 . . 4:2. Data from 1 H n.m.r. spectra of P ^ (NH) (NMe 2) gC£ 2 4:3. Relationship between 6 ^ and K & for some secondary amines . . 5:1. Charge density and bond order differences between PgN^ and P f i N ? systems as derived from symmetry based ca lcula t ions . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5:2. Bond angles and bond lengths i n P^N^CJlg 5:3. Observed and g-adjusted Tr-contractions i n P^N^C£g . . . . 5:4. Calculated charge densi t ies i n P^N^CX^ using a l l -3d-o rb i t a l s on phosphorus 5:5. Calculated bond orders i n P^N^C£g using a l l 3d-orbi tals on phosphorus . - v i -LIST OF FIGURES Page Figure 1.1. Structures of some condensed r i ng inorganic molecules . 2 1.2. T r i c y c l i c structure proposed for P^NjClg 6 1.3. Possible structures of parent molecules of ions i d e n t i f i e d i n mass spectra 6 2.1. Furnaces used i n preparation of P^NyC£g . . 18 3.1. Possible structure of parent molecule of ions based on P^N^ nucleus - 27 3.2. Var ia t ion i n p a r t i t i o n r a t i o of phosphoni t r i l i c chlorides between hexane and cone. H^SO^ 40 4 .1 . Apparatus used i n attempted f luor ina t ion of P^N yC£ g . . . . 50 4.2. Suggested structure of P ^ ( N H ) (NMe 2) 8CJi 2 56 4.3. Gross structure of P,N £(NH)(NMe„) 0C£. indicated by i t s 31 z a z P n.m.r. spectrum 59 31 4.4. P chemical sh i f t s of some t r imer ic dimethylamino-chlorophosphonitri les 61 4.5 . : H n.m.r. spectra of P f i N 6 (NH) (NMe 2) gC& 2 •. . . 63 4.6. Two possible conformations for the P,N,(NH)(NMe 0) 0C£„ i i b o z o z _ A molecule . . . 7 0 4.7. Suggested stereochemistry of nuc leophi l ic attack at a bridgehead phosphorus atom i n P^N^CJig • . . . . 71 5.1. Overlap schemes for Tr-bonding at a non-bridgehead phosphorus atom i n P^N yC£ g • 76 5.2. Overlap schemes for out-of-plane TT-bonding at a bridge-head phosphorus atom i n P^NyC£g . . . . 77 5.3. Energy levels for the PgN^ and ? 6 N y systems 80 5.4. Orb i t a l nomenclature used i n symmetry based ca lcula t ions on P ^ C J L , 81 - v i i -Page Figure 5.5. Nomenclature for bonds and atoms i n P^N_C£rt 82 o / y 5.6. Molecular structure of P,N_C£_ . . . . . . . . . 84 o / y A . l . Rotated coordinate scheme used i n the evaluation of overlap in tegra ls at the P R atoms i n the IT system . . . 98 - v i i i -ACKNOWLEDGEMENTS Professor N . L . Paddock has been my supervisor during the course of th i s work. The debt of gratitude that I owe him I w i l l always remember. For h i s help, advice, and patience, I can only offer my thanks. I would also l i k e to thank Dr. R.D. Spratley for many helpful discussions and for h i s useful advice. I am very grateful to the fol lowing people for the i r constructive comments and assistance i n the course of th i s Thesis: Mr. E. B ich le r , Mr. R.W. Harrison, Mr. T.N. Raganathan, Dr. J . Serreqi , and Mr. C . J . Stewart. My thanks also go to Miss V. Ormerod for typing th i s thes i s , and to the National Research Council of Canada, for f inanc ia l assistance. F i n a l l y , but not l eas t , I would l i k e to thank my friends and my family for t he i r help and encouragement during the las t year. To Mum and Dad - 1 -CHAPTER 1 INTRODUCTION Inorganic r ing systems, of the general type (AB) n , have been prepared between many elements of the f i r s t two rows of the Periodic Table. However, p o l y c y c l i c r i ng systems are much rarer . These poly-c y c l i c compounds are usual ly unstable and are formed i n small y ie lds as decomposition products of reactions involv ing monocyclic systems. In most instances where the i s o l a t i o n of a po lycyc l i c condensed r ing species has been reported, no discussion of the chemistry of these compounds has been given. Condensed r i ng species i n inorganic chemistry are of in teres t not only because of the i r novelty, but because the i r chemistry may be very different from that of ordinary c y c l i c systems, the properties of bridgehead atoms often being quite unusual i n comparison to that of non-bridgehead atoms. In siloxane chemistry, l i n e a r , c y c l i c and crossl inked species are a l l known. In borazine chemistry, the pyro lys i s of cyclot r iborazine at 380°C gives hydrogen and a mixture of condensed species,^ one of which i s believed to have the structure shown i n F i g . 1.1a. B- t r i ch lorocyc lo t r iboraz ine reacts with methylmagnesium bromide to give the expected B-trimethyl de r iva t ive , CMeB-NH)^. • The other product, 2 which i s not expected, i s again a naphthalene analogue (see F i g . 1.1b). - 2 -Condensed r ing silazanes have been prepared, and the compound shown i n F i g . 1.1c has been i so la ted as a minor product during the react ion of cyclohexamethyltrisilazane with potassium hydroxide at ~180°C. The ammonolysis of methyldichlorosilane y i e l d s , as a minor product, 4 the compound i l l u s t r a t e d i n F i g . l . l d . In carbon-nitrogen r i ng systems, the pyro lys i s of melamine (NH2-C=N)g at 350°C y ie lds a com-pound whose structure i s suggested to be a condensed r ing^ (Fig . L i e ) . Thus, wi th in any heterocycl ic r ing system (AB) n , t he re ,ex i s t , .at most, only a few condensed r ing species,- and very l i t t l e i s known of the physical and chemical properties of these species i n comparison to the i r monocyclic congeners. The family of compounds known as the c y c l i c phosphonitri les contain the same basic repeating unit -£ P^2=N The c l a s s i c a l react ion used i n the preparation of c y c l i c phosphonitri les i s that between phosphorus pentachloride and ammonium ch lor ide , and was f i r s t 6 carr ied out by Stokes i n 1897:-PCZC + NH.C£ —*• -(NPCA„) + 4HC£ L l 5 4 n 2'n Unlike a l l other inorganic r i n g systems, and a l l organic annulenes, r ing s izes corresponding to n = 3,4,5 . . . are formed i n th i s reac t ion . This fundamental property of these compounds, namely the ease of preparation of different r i ng s i ze s , i l l u s t r a t e s the unique nature 7 8 of the bonding i n these compounds. The o r i g i n a l theories ' on the - 3 -H H H H H H Me Me a). b). H Me Me x 1e-Si Si- Me H ^ , , / S i ^ , / . ~ N .Me / ^ H X 0 .^ H M ^ M r f ? * H - \\ . ^ / k M e / -Me Me Me Me Me H Me C>- H D>-r N N fi i H i ! ' e ) . F i g . 1.1. Suggested structures of some inorganic molecules that are believed to be condensed r ing compounds. 6 - 4 -bonding i n phosphonitri les invoked the use of d-orbi ta ls to explain th i s unique type of bonding, the difference i n symmetry properties of d- and p -o rb i t a l s being used to account for the general s t a b i l i t y of a l l r ing s i zes . The chemistry of the phosphonitri les has been studied extensively and l igand replacement reactions are numerous. However, up to the present time, a l l experimental and theore t ica l studies have been r e s t r i c t ed to monocyclic der iva t ives . In the course of the work carr ied out by Stokes, a compound was i so la ted with the molecular formula P^NjCHg. This compound was made as a minor product i n the react ion between phosphorus pentachloride and ammonium chloride (equation 1.1). This reac t ion , as carr ied out by Stokes, was done i n sealed glass tubes at a temperature of 200°C. More recent workers have done th i s experiment i n re f lux ing solvents at temperatures less than 140°C, and the compound P^N^CZ^ has never been i so la ted during these react ions. The chemistry of th i s compound has, up u n t i l now, never been studied. I ts s t ructure, both molecular and e lec t ron ic , has not been known. In 1955, Krause apparently obtained the compound P^N^C&g by the 9 thermal depolymerization of a phosphoni t r i l i c chloride high polymer. However, few de ta i l s were given by him on the character izat ion of th i s compound and no studies were made of i t s chemistry. At th i s time, the structure that was suggested for th i s molecule was a t r i c y c l i c con-densed r i ng (F ig . 1.2) but no evidence was given i n support of th i s - 5 -s tructure. Studies on the mass spectra of the phosphoni t r i l i c c h l o r i d e s ^ have shown evidence of the existence of ions , the parent molecules of which are thought to have condensed r ing structures. One of these structures corresponds to that proposed for P^N^CAg (Fig . 1.2). The proposed structures of the other condensed r i ng molecules are given i n F i g . 1.3. Thus, at the beginning of th i s work, there was only a small amount of scattered information on condensed r i ng phosphoni t r i l i c chemistry. Because of the lack of information i n th i s f i e l d , the purpose of th i s present work has been to study the chemistry of condensed r i ng phosphoni t r i l i c chlor ides . During th i s work, the i s o l a t i o n of P^N^CJig has been carr ied out and the compound i s now 31 f u l l y characterized. V i b r a t i o n a l , mass and P n.m.r. spectra con-firmed the condensed r ing structure of the molecule, the proper name of which i s 3,7,11 ,nitr ' i lo-nonachlorocyclohexaphosphonitri le. The name assigned to the compound by Stokes was n i t r i lohexaphosphoni t r i l i c ch lor ide , and th i s nomenclature w i l l be used throughout th i s work. Even though the compound was prepared i n only very small quant i t ies , the chemical properties of the molecule were studied and several l igand subs t i tu t ion reactions were attempted, i n order that the chemistry of P^ijCZ^ might be related to that of the ordinary c y c l i c phosphoni t r i l i c chlorides (PNC£„) . Ligand subs t i tu t ion reactions with the ordinary phosphoni t r i l i c chlorides are common. Fluor ina t ion of the chlorides with potassium Cl Cl \\ / Cl Cl N Cl P N' / Cl Cl Jk—ci Cl F i g . . 1.2. T r i c y c l i c condensed r ing structure proposed for n i t r i l o -hexaphosphonitri l ic ch lor ide . F i g . 1.3. Possible structures of parent molecules'corresponding to ions i den t i f i ed i n mass spectra. - 7 -f luorosulphi te produces the phosphoni t r i l i c f luorides (PNF2) 12 amination of both the chlorides and the f luorides y ie lds p a r t i a l l y and ful ly^ '\"*\"^ aminated phosphonitr i les . Aryl''\"\"' and a l k y l \" ^ derivat ives can be prepared by react ing a phosphoni t r i l i c f luor ide with the appropriate Grignard reagent. The preparation of phosphoni t r i l i c derivat ives with different subs t i tu t ion patterns has also received at tent ion. For example, subs t i tu t ion of a chlorine l igand i n (PNCi!^^ by a f luor ine atom increases the e l e c t r o p h i l i c i t y of the phosphorus atom at which the subs t i tu t ion has occurred. Further subs t i tu t ion i s therefore favoured, and i n the case of f l uo r ina t i on , th i s subs t i tu t ion i s always geminal. On the other hand, amination of the phosphoni t r i l i c r i ng deactivates the r i ng to further subs t i tu t ion , espec ia l ly at the substi tuted phosphorus atom. Further subs t i tu t ion i s therefore found to take 12 place i n a non-geminal fashion. Unlike that of the ordinary c y c l i c ch lor ides , the chemistry of n i t r i lohexaphosphon i t r i l i c chloride i s l i m i t e d . The introduction of the central nitrogen into the P^N^ r ing system has a dras t ic effect on the properties of the molecule. The fragmentation pattern and c rys ta l structure of the compound indicate that the in te rna l PN bond i s weak. Various subs t i tu t ion reactions were attempted on the condensed r ing ch lor ide . When r i ng ac t iva t ing groups are used (-F, -OMe), the t r i c y c l i c r i ng structure of the PN skeleton i s destroyed. Reaction with deact ivat ing groups (-NMe2) i s less dest ruct ive , and - 8 -only p a r t i a l r i ng cleavage occurs, leading to the i s o l a t i o n of a com-pound i n which the t r i c y c l i c PN skeleton has been l o s t , but i n which some c ross l ink ing of the r i ng i s s t i l l apparent. The differences between the condensed r i ng chloride and the ordinary c y c l i c phosphoni t r i l i c chlorides are therefore quite marked. The l a t t e r are stable to subs t i tu t ion , whereas the former i s not. The physical and chemical properties of P^N^Ci^ suggest that the reason for t h i s l i e s i n the weakness of the in te rna l PN bond, which appears to be weaker than an exocycl ic PCJl bond. Simple Hiickel molecular o r b i t a l calcula t ions have been carr ied out on an idea l i zed condensed r ing framework, using the out-of-plane d-orbi ta ls on phosphorus and the out-of-plane o r b i t a l on nitrogen to form Tr molecular o r b i t a l s . In so far as lone pa i r d e r e a l i z a t i o n of the central nitrogen takes place only through the out-of-plane TT-system, any effect which th i s d e r e a l i z a t i o n has on the in-plane Tr-system w i l l only be secondary. Since i t i s thought that i t i s the in-plane TT-system which controls the chemical r e a c t i v i t y of phospho-n i t r i l e s , the effect of the central nitrogen on the properties of the molecule should only be minimal. The fact that th i s i s not the case, and that the resul ts of base strength measurements and trends i n bond-lengths are the reverse of those predicted by the M.O. ca lcu la -t ions on the out-of-plane TT-system, indicates that more deta i led ca lcu la t ions , as wel l as more experimental information are necessary, before a complete understanding of th i s system i s accomplished. - 9 -In the fol lowing chapters, the points out l ined above w i l l be f u l l y discussed. In chapter (2^ a f u l l account i s given of the various methods, both successful and unsuccessful, that were used i n the preparation of the condensed r ing ch lor ide . - 10 -CHAPTER 2 PREPARATION OF NITRILOHEXAPHOSPHONITRILIC CHLORIDE The o r i g i n a l i s o l a t i o n of P^N yC£ g by Stokes was achieved during a high temperature react ion . In the present work, several directed syntheses from N^P^CA.^ have been attempted without success, and at present the only route to th i s compound seems to be v i a high temperature rearrangements. For the purposes of th i s thes i s , the discussion of the types of preparation can be divided into two parts , the f i r s t involv ing attempted d i rec t condensation react ions, and the second involv ing high temperature rearrangements. I . Condensation Reactions Throughout the course of th i s work, many comparisons were made between n i t r i lohexaphosphon i t r i l i c chloride and the open r i ng hexa-phosphoni t r i l i c ch lor ide . Because of th i s analogy, a series of experiments was carr ied out on the (PNCJ^)^ molecule i n attempt to add an extra nitrogen atom to the PN skeleton. Lithium n i t r i d e has 17 been used to make t e r t i a r y amines, and was used here i n an attempt to condense the P ,N, r ing with the e l iminat ion of l i th ium chlor ide , o o The use of amino- derivat ives of the group IV metals as aminating agents has been reported i n the l i t e r a t u r e , ^ ' - ^ and dimethylamino-- 11 -t r imethyls i lane and -stannane have been used e f fec t ive ly to prepare 20 monofunctional dimethylamino-derivatives of phosphoni t r i l i c f luor ides . In view of the strength of the M-CZ bond (M = S n , S i ) , when compared to the M-N bond, i t was hoped that the react ion of (Me^M)^ with hexa-phosphoni t r i l i c chloride might produce the condensed r i n g , fol lowing the react ion path suggested i n equation 2 .1 . (R 3M) 3N + CPNC£2)6 • P 6 N 7 C£ Q + 3R3MC£ (2,1) None of these reactions produced any P^N^C£g, nor was there any ind ica t ion that th i s type of d i rec t preparation would do so. There i s an obvious s t e r i c s t r a in involved i n forming the molecule by th is type of reac t ion , and i t i s now thought that preparations of th i s type are unfeasible. I I . High Temperature Rearrangements The two previous reports of the i s o l a t i o n of P^N^Cl^ have involved the use of high temperatures. As has already been stated, the prepara-t i on of (PNCJ^) i n re f luxing solvents does not produce any condensed r ings . The mechanism of the react ion between phosphorus pentachloride and ammonium chloride has been studied by Becke-Goehring and 21 22 coworkers. ' E s sen t i a l l y , the mechanism involves the formation of l inea r cations of the type [C£3P=N-PCJi3] + . These cations ei ther lengthen by further reaction or cyc l i ze into the neutral phosphoni t r i l i c - 12 -chlor ides . I f condensed r ing species were formed i n th i s react ion, / P CV + at some stage an ion of the form [=C£„P-N ,] would have to be X P C £ 2 = present. That th i s i s not the case implies t h a t t h e [=P=N-P=] linkage i s stronger than the [EP-N^ J moiety. Although r i ng rearrangements of the phosphoni t r i l i c chlorides have not been observed at low temperatures, transannular processes are noticeable i n the reactions of, e spec ia l ly , the larger r i ng s ize phosphonitr i les . The f luor ina t ion of hexameric chloride with potassium f luorosulphi te always y ie lds a small amount (-5% of a l l f luorinated products) of tetrameric f luor ide . A l s o , the mass spectra of phosphoni t r i l i c ch lo r ides ' ^ show that transannular processes are very important i n the fragmentation pattern of hexameric, heptameric, and octameric chlor ides . For example, 40.6% of a l l fragments i n the mass spectrum of CPNCJ^)^ contain the (PN)^ nucleus. A rearrangement by some s o r t o f t r a n s a n n u l a r p r o c e s s i s t h e r e f o r e i n d i c a t e d . N / p % i _ P ^ % N/% II i t | - * 2 || | 12.2) Pyrolys is of any phosphoni t r i l i c chloride at a temperature >300°C for several hours produces a polymer, the structure of which 23 can be r a t i ona l i zed i n terms of a crossl inked matrix with a molecular weight of the order of 60,000. The formation of the polymer proceeds i n two steps. The f i r s t step i s the formation of an apparently - 13 -uncrosslinked long chain polymer which then cross l inks to form the polymeric inorganic rubber. As a resul t of the high temperature experiments carr ied out on the monocyclic chlorides and described below, the fol lowing conclusions were drawn. Depolymerization of phosphoni t r i l i c chloride high polymer i n vacuo invar iab ly produces (PNCJ^^ a n j 4 a n a \" sometimes, but not always, a trace of P^N^Cilg. No other products are obtained, no matter which phosphoni t r i l i c chloride i s used to produce the high polymer. For instance, P^N^Cilg i s even recovered from the depolymerization of polymer which has been prepared from t r imer ic and tetrameric chlor ides . When the crossl inked high polymer i s extracted into benzene, no condensed r i ng chloride i s found i n the benzene. These observations suggest that the P^N^Ciig i s a product of the depolymerization process, rather than being formed as a side product of the polymerization step. An X-ray d i f f r a c t i o n study of the high polymer has shown that i t 24 consists of a long chain h e l i x . Thus the formation of n i t r i l o h e x a -phosphoni t r i l i c chloride from the high polymer may be viewed i n the fol lowing way. I II I ll I 11 I II J> J> / P ^P ^ P .P .P / P N ^ N N ^ N ^ N ^ N ^ N ^ N I 11 1 i l 1 i ! 11 St. S, / N ^ N S, I I] I II I II I II .... 12.3) ^ / p / p \\ A p N N N W ^ N ^ * N II N N \" ^ N ^ N ^ ^ \" N < I II I II I P P P P P P - 14 -The resu l t s of a recent study by Todd et. a l . ' support th i s conclusion. They found that the slow, control led polymerization of (PNCiy^ at 250°C produces a pale yellow polymer which, upon extract ion with l i g h t petroleum, y ie lds some (PNCS^^ 5 a n c j ^ a s wel l as the crossl inked high polymer. The mechanism of the formation of these different r i ng sizes i s thought to involve a polymerization-depolymerization equi l ibr ium process of the l inear polymer. In the vapour phase, phosphoni t r i l i c chlorides are remarkably) s table . After passing the vapour of (PNCJ?^^ through a s i l i c a furnace at 600°C, the condensate remained unchanged and no rearrangements occurred. A pyro lys i s i n the l i q u i d phase was carr ied out by passing a l i q u i d mixture of homologues (PNCJl^J > mean n ~5, through a v e r t i c a l s i l i c a furnace at d u l l red heat (~600°C). Inside the furnace the l i q u i d (PNCi!^^ rap id ly polymerized. Depolymerization quickly followed, and the products of depolymerization were forced out of the furnace under a s l i g h t pos i t ive pressure of ni trogen. Although th is sort of experiment did produce a trace of P^N y C£ g , further inves t iga-tions along th i s l i ne were not pursued, since considerable breakdown of the polymer occurred i n the furnace, with the formation of phosphorus pentachloride and elemental phosphorus. The most r e l i a b l e way of producing the condensed r ing was found to be through the polymerization of the c y c l i c chlorides at 300°C to give the crossl inked high polymer. The subsequent d i s t i l l a t i o n of polymer y ie lded , as a f i n a l f r ac t ion , a very small quantity of - 15 -P^N^C&g. The rate of polymerization was never reproducible, as trace amounts of impuri t ies catalysed the react ion, e.g. water increased the rate of polymerization and the presence of a i r decreased i t . The r ing s ize of the s ta r t ing chloride also affected the rate of polymerization. (PNCJ^Jj takes up to 8 hours to polymerize at 300°C whereas (PNCJ^)^ takes only h a l f an hour. I t i s not known i f adaptations of these processes w i l l improve the y i e l d of the condensed r ing compound. In general, from a poly-merization involv ing 200 grams of phosphoni t r i l i c ch lor ide , only 0.2-0.5 gm of P^NyC£g could be recovered. The high temperature chemistry of phosphoni t r i l i c compounds has not received d i rec t a t tent ion, and consequently, new compounds, s imi l a r to P^N y C£ g , such as the P^N^C£y and P y NgC£^ found by mass spectrometry may remain to be discovered. EXPERIMENTAL (a) Thermal Decomposition of Phosphoni t r i l ic High Polymer Several hundred grams of phosphoni t r i l i c chloride o i l (PNCJ^^ were added to a large round-bottomed f lask which was f i t t e d with a s t i l l - h e a d , a short a i r condenser, and a receiver f l ask . Keeping the react ion vessel at room pressure, the o i l was heated i n a sand bath to a bath temperature of about 300°C. Trimeric phosphoni t r i l i c chloride b o i l s at 257°, but at a bath temperature of 300°C and even higher, d i s t i l l a t i o n of th i s compound was almost n e g l i g i b l e . After - 16 -heating the o i l at th i s temperature for several hours, the o i l slowly-darkened and f i n a l l y polymerized into a dark, rubbery polymer. The length of time required for polymerization was a function of the pur i ty and composition of the o i l . F i l t r a t i o n of a so lu t ion of the o i l i n pe t ro l ether (30-60°) through s i l i c a gel or Clearsorb columns removed a l l polar impuri t ies and produced a colourless l i q u i d on evaporation of the solvent. This o i l was more d i f f i c u l t to polymerize since the presence of polar impuri t ies reduced the polymerization time. The actual process of polymerizat ion, once s tar ted, i s quite rap id . The change from a dark, but mobile, l i q u i d to a rubbery polymer, takes less than a minute. The presence of some polymer appears, therefore, to catalyse further polymerizat ion. Depolymeriza-t i on of th i s rubbery material was achieved by heating the polymer i n vacuo. Depolymerization was best at a pressure of <0.1 to r r and at a bath temperature of 300°C. The use of higher temperatures and higher pressures reduced the y i e l d of the depolymerization products, which were t r imer ic and tetrameric phosphoni t r i l i c chlorides and a trace of P^N^CJlg. Even under high vacuum, complete depolymerization took about 48 hours. The f i n a l f r ac t ion , which was the condensed r i ng compound, usual ly s o l i d i f i e d i n the condenser arm, and was i so la ted by washing the condenser with pe t ro l ether (30°-60°C) to remove the monocyclic chlorides) and then with benzene, i n which P,N_C£„ i s soluble . The receiver f lask was also washed i n the same way, i . e . f i r s t l y with pe t ro l ether, and then wi th benzene. - 17 -The residue l e f t i n the d i s t i l l a t i o n f lask was a l i g h t f laky black ash, the composition of which i s unknown. In a t y p i c a l depoly-merization react ion 200 grams of ( P N C i ^ ^ were polymerized and the polymer cracked to y i e l d 170 grams of a mixture of tetrameric and t r imer ic chlorides ( in a r a t i o of -2:1) and 0.30 grams of the condensed r i ng ch lor ide . The y i e l d of P^N^CJlg was va r i ab le , and i n some reactions no condensed r i n g was recovered. The compound was i d e n t i f i e d by i t s melting point , 235°C ( l i t e ra tu re va lue 6 237.2°C), and by analysis (calculated for P 6 N ? C £ g ; P, 30.84; N, 16.29; CH, 52.87; found N, 16.19; Cl, 52.81). The mass spectrum (which w i l l be discussed i n chapter I I I ) of the compound completed the character izat ion of the compound. The next three experiments were, l i k e the f i r s t , a l l pyro lys i s reactions carr ied out on phosphoni t r i l i c chlor ides , and although they did produce a l i t t l e P^N^Cl^, were not as r e l i a b l e nor as e f f i c i en t as the f i r s t process. (b) This experiment involved the use of a v e r t i c a l s i l i c a glass column (see F i g . 2.1a) with a cons t r i c t ion at the lower end to support a packing of hengar granules. The upper end of the column was so designed that phosphoni t r i l i c o i l could be introduced to the column under a s l i g h t pos i t ive pressure of ni t rogen. The lower end of the column was heated to d u l l red heat (~600°C) with an e l e c t r i c furnace and phosphoni t r i l i c chloride oil-was added dropwise at the top of the column. Immediate polymerization of the o i l occurred, and a depoly-(PNCi> ) i n 2 n } a), N, in. - 1! b). (PNC£0) i n 2 n Er a y ,x x x * ^ x l To Vacuum L i q u i d Nitroger Cold Trap C) % *U v k, U A i r leak P5= F i g . 2 . 1 . Diagrams of furnaces used i n the preparation of N i t r i l o -hexaphosphonitr i l ic Chlor ide . - 19 -merization-polymerization process ensued. Any v o l a t i l e products passed out of the bottom of the heated zone, where they were co l lec ted as l i q u i d s . At the furnace temperature used, breakdown of the (PNCJ?^^ unit was apparent, as i n th i s type of experiment, phosphorus pentachloride was emitted from the column and red phosphorus s o l i d i f i e d i n the receiver f l ask . I f lower temperatures were used to stop th is breakdown, depolymerization of the high polymer could not be achieved and the column would quickly become blocked with polymer. In a t y p i c a l experiment of t h i s type, 100 grams of (PNCJ^)^ were pyrolysed to give 55 grams of ( P N C ^ ^ (n = 3,4 and 5) and 0.10 grams of P^N^C£g. Separation and i s o l a t i o n of the condensed r ing compound was done using the techniques described i n experiment (a). (c) This experiment was designed as a modification of experiment (b) and involved the use of the copper furnace i l l u s t r a t e d i n F i g . 2.1b. The bottom of the furnace was heated to d u l l red heat with a bunsen burner. Phosphoni t r i l i c chloride o i l was dropped onto the hot hengar granules on the bottom of the furnace. Polymerization, depolymeriza-t i on and vaporizat ion of the o i l then occurred and any v o l a t i l e products were co l lec ted through a short a i r condenser. No P^N^CZg was recovered from th i s type of p y r o l y s i s , and indeed much of the phosphoni t r i l i c o i l was decomposed, with less than 50% of the s ta r t -ing material being recovered. - 20 -(d) A pyro lys i s react ion was attempted i n the gas phase, to see i f r ing rearrangement occurs i n the vapour. Using the apparatus shown i n F i g . 2.1c, tetrameric chloride was d i s t i l l e d through a hor izonta l s i l i c a glass furnace packed with hengar granules, and heated to ~600°C. The products of the react ion were col lec ted i n a dry i c e / acetone trap. I t was found that the tetrameric chloride had d i s t i l l e d unchanged through the apparatus, and no c y c l i c rearrange-ments had taken place i n the gas phase. (e) Reaction of Lithium Ni t r i de with Hexaphosphpnitrilic Chloride A so lu t ion of (PNC& 2) 6 (8.8 g . , 12.6 mmol.) i n 50 ml. of t e t r a -hydrofuran was added, over 10-15 minutes, to a s t i r r e d s lu r ry of l i t h ium n i t r i d e (0.44 g . , 12.6 mmol.) i n 50 ml. of tetrahydrofuran. The mixture was then gently heated under re f lux for 2-1/2 hours, after which time the so lu t ion was f i l t e r e d under ni trogen, and most of the tetrahydrofuran d i s t i l l e d of f . 100 ml- of benzene were added' and immediate p r ec ip i t a t i on (of LiC&) followed. The so lu t ion was allowed to se t t l e and some of the c lear so lu t ion drawn of f and allowed to evaporate. The product l e f t on evaporation was i den t i f i ed by i t s infra-red spectrum as being (PNC^)^ . T n e gummy white prec ip i ta te could not be characterized, but a strong O-H stretching band i n i t s inf ra- red spectrum: suggests that hydrolysis of the phospho-n i t r i l i c r i ng had occurred. No P,N_C£Q was recovered. - 21 -(f ) Reaction of T r i s C t r i m e t h y l s t a n n y l ) amine with Hexaphosphpnitrilic Chloride A so lu t ion of (Me^Sn^N (3.0 g . , 5.9 mmol.) i n 30 ml. of benzene was added over a period of 20 minutes to a s t i r r e d so lu t ion of the phosphoni t r i l i c chloride (4.1 g . , 5.9 mmol.) i n 50 ml .o f benzene. The react ion mixture was kept under an atmosphere of nitrogen and allowed to b o i l gently under re f lux overnight. On d i s t i l l a t i o n of the solvent, there remained a gum, h a l f of which was then heated to 250°C at a pressure of 0.5 t o r r . This heating of the gum produced a white lumpy s o l i d , which had the charac te r i s t i c smell of t r imethyl t i n ch lor ide , but which was insoluble i n pe t ro l ether, chloroform, benzene, d i l u t e mineral acid and d i lu t e a l k a l i . The mass spectrum of the pyrolysed material showed a pattern charac te r i s t i c of (PNCJ^)^. To that h a l f of the gum which was not pyrolysed, a l i t t l e pe t ro l ether was added. This produced an amorphous white p rec ip i ta te which was insoluble i n hot a l k a l i , n i t r i c ac id , chloroform and benzene. The infra- red spectrum was charac te r i s t i c of a phosphoni t r i l i c com-pound, with a broad peak at 1340 cm ^, as wel l as others at 955 cm ^ (med, broad), 790 cm * (w), 755 cm ^ (med), 601 cm ^ ( s t ) , 525 cm * (med) and 470 cm * (w). However, i t was not the condensed r i ng compound, and no evidence of the formation of P^N^CJig was detected at any stage i n the reac t ion . (g) Reaction of Tr i s ( t r imethy l s i ly l )amine with Hexaphosphpnitril ic Chloride A so lu t ion of (Me,Si)_N (1.4 g . , 6.0 mmol.) i n 25 ml, of benzene - 22 -was added over a period of 10 minutes to a so lu t ion of hexaphospho-n i t r i l i c chloride (4.1 g . , 5.9 mmol.) i n 50 ml-of benzene and the mixture s t i r r e d and heated under ref lux i n an atmosphere of dry ni trogen. After 24 hours, the benzene was d i s t i l l e d and the neat l i q u i d heated to 200°C at a pressure of 0 .5 ; to r r . fo r one hour. On cool ing , the res idual l i q u i d was i d e n t i f i e d as unreacted (PNOp^. With the i s o l a t i o n of n i t r i lohexaphosphoni t r i l i c chloride achieved, the compound was subjected to a thorough spectroscopic analys is . The resul ts of the various physical measurements made on the molecule are discussed i n the fol lowing chapter. Although no PgNyC£g was i so la ted from the above-mentioned d i rec t syntheses, enough of i t was i so la ted from the pyrOly t ic reactions to allow a thorough spectroscopic analysis to be made of the molecule. The resul ts of the various physical measurements made on the molecule are discussed i n the fol lowing chapter. - 23 -CHAPTER 3 PROPERTIES OF NITRILOHEXAPHOSPHONITRILIC CHLORIDE Although, up to now, no s t ruc tura l information has been avai lable on the P^N yC£ g molecule, a t r i c y c l i c structure has been proposed, mainly from simple valence considerations. That th i s was indeed the 31 correct structure was confirmed by P n.m.r. spectrometry. In th i s present chapter, de ta i l s are given of the resu l t s obtained from the i n .m. r . , mass, u l t r a v i o l e t , and v ib ra t iona l spectroscopic invest igat ions carr ied out on the molecule. 31 I . P n.m.r. Spectrum Because of the low s o l u b i l i t y of P^N^CJlg i n a l l common solvents, 31 the P n.m.r. spectrum proved very d i f f i c u l t to obtain. Saturated solutions i n hot benzene and xylene f a i l e d to produce a s i gna l . F i n a l l y , a 40% solu t ion of the compound i n molten naphthalene (at 150°C) produced a weak but de f in i t e spectrum, consis t ing of two peaks of equal i n t ens i ty . The peaks d id exhibi t f ine structure but the signals were too weak to allow any reso lu t ion of th i s f ine structure 31 to be made. The P chemical sh i f t s of these two peaks are given i n Table 3:1 together with those of the monocyclic phosphoni t r i l i c 31 chlor ides . The P n.m.r. pattern of two signals of equal in tens i ty i s quite d e f i n i t i v e and can only be interpreted i n terms of a molecule - 24 -Compound (PNC£ 2 ) 3 CPNCil 2) 4 CPNC£2)5 ' (PNC£ 2 ) 6 P 6 N 7 C£ g 6 p p.p.m. r e l a t i ve to 85% H 3 P0 4 -20.0 +7.0 +17.0 +16.0 -20.2,+3.5 TABLE 3.T. P n.m.r. chemical sh i f t s for the phosphoni t r i l i c chlorides (PNC£ 2 ) 3 _ 6 and for P 6 N y C£ g . Values for (PNC£ 2 ) 3 _ 6 are from reference 26. having two sets of three equivalent phosphorus atoms. The t r i c y c l i c condensed r ing structure i s unique i n th i s regard, with one s ignal corresponding to the bridgehead (=P-C£) phosphorus atoms (see F i g . 1:2), and the other corresponding to the non-bridgehead (=PC£2 type) phosphorus atoms. The in terpre ta t ion of phosphorus chemical sh i f t s i s d i f f i c u l t , as angular as wel l as e lec t ronic factors can contribute to the s h i f t . However, the environment of the =PC&2 phosphorus atoms i n the t r i c y c l i c condensed r i ng structure i s very s imi l a r to that of the phosphorus atoms i n (PNC£ 2 ) 3 > and the phosphorus chemical sh i f t s are expected to be s i m i l a r . Thus the resonance at 6 =-20.2 p.p.m. i s assigned to the =PC£2 type phosphorus atoms of the condensed r i n g . The s ignal at 6 = +3.5 p.p.m. i s therefore assigned to the bridge-head phosphorus atoms. On a simple f i r s t order bas i s , lone pa i r - 25 -de loca l i za t ion from the central nitrogen atom would be expected to increase the chemical sh i f t of the bridgehead phosphorus atoms with respect to the =PC&2 t y P e > a n d th i s idea i s i n complete agreement with the above assignment. The t r i c y c l i c condensed r i ng structure shown i n F i g . 1.2 was thus established for P^NyC£g, and the subsequent X-ray c rys t a l structure analysis (discussed i n chapter 5) confirmed th i s assignment. I I . Mass Spectrum The fragmentation pattern of th i s compound i s of in teres t for several reasons. Not only does the mass spectrum serve as an excellent ana ly t i c a l too l for the character izat ion of the compound, but in te res t ing comparisons may be made between the fragmentation pattern of th i s compound and those of the ordinary c y c l i c chlor ides , comparisons which may be related to the chemistry of the two types of compound. As i n the mass spectra of the ordinary c y c l i c ch lor ides , (PNCJ^)^ fragments with an even number of electrons are more s table , and therefore more abundant, than fragments with an odd number of electrons. The fragments can be divided into three classes . Those with the formula CP N ,C£ j n + have stable s ing ly charged ions for even numbers x x+1 y • J • of chlorine atoms. For doubly charged ions i n th i s ser ies , the abundances of fragments are greatest for those with an odd number of chlorine atoms. The second type of fragment that i s seen i s of the - 26 -form (P N C£ ) n + , which corresponds to a fragment with a c y c l i c s t ructure. For these fragments abundances alternate with va r i a t i on i n y so that for singly-charged species, ions with an odd number of chlorine atoms predominate whi ls t for doubly charged species, those with an even number of chlorines predominate. The t h i r d type of fragment has the general formula (P N ^CH ) n + and these fragments are thought to be l i nea r . This type of ion i s known i n preparative chemistry, the (P^^Cilg)* ion being an intermediate i n the preparation of phosphoni t r i l i c ch lor ides . These l inear fragments, as would be expected, have abundance charac te r i s t i cs which are the reverse of those observed for the c y c l i c ions. Namely, for s ing ly charged species, ions with an even number of chlorine atoms predominate, whi ls t for doubly charged species, those with an odd number of chlor ine atoms are more abundant. Thus, the important point of the above argument i s that there i s general correspondence between the chemical s t a b i l i t y of an ion and i t s abundance i n a mass spectrometer. Fragments of the type ( P x N x + i ^ ^ y ) n + a r e thought to have con-densed r i ng structures, with the central nitrogen atom, bound to three d i s t i n c t phosphorus atoms, being maintained i n the structure of the ion . The abundance r a t i o of ions with odd and even numbers of electrons i s consistent with these fragments being completely c y c l i c . Within th i s ser ies of ions , fragments with formulae corresponding to three values of x (= 6,5 and 4) are found. Frag-- 27 --ments based on the PgNy and the P^N^ nucle i have already been iden t i f i ed i n mass spectra, but the ions based on the P^N^ nucleus have not been seen before. These ions are believed to be based on a tetrameric r i ng system (see F i g . 3 .1) , with an extra nitrogen atom bound to three separate phosphorus atoms. In the mass spectra of the C I F i g . 3 .1. ' Possible structure of parent molecule of ions based on the P . N r nucleus. 27 + phosphoni t r i l i c f luor ides , the series C T ^ ^ ^ ^ n 6^ ^ n = ^-12) has been seen. The smallest ion i n th i s se r ies , P ^ N ^ ^ * , may be related to the P^N^C£^ n + ions , except that i n the former, there would be two nitrogen atoms t r i p l y bound to the tetrameric r i ng system, one being above the plane of the r i n g , and the other being below i t . The abundance of ions based on the P.N,, and P r N , nucle i i s much 4 5 o o less than that of ions based on the parent P^N y nucleus, which cleaves p re fe ren t i a l ly into c y c l i c fragments based on the PgN^ and P 4 N 4 r i n g . Table 3.2 shows the fragmentation pattern of P^NyC£g, and for comparison, the fragmentation pattern of (PNC£„) . In th i s Table, P 6 N ? C i l 9 FRAGMENT P PN P 2 P 2N P 3 N 2 P 3 N 3 P N 4 3 P .N. 4 4 P N 4 5 P N 5 5 P 5 N 6 P 6 N 6 P 6 N 7 TOTAL Condensed Ring 1.3 4.2 17.8 23.3 C y c l i c 2.4 19.4 9.6 1.1 0 32.5 Linear 14.8 1.2 18.2 7,4 2.3 43.9 P 6 N 6 C £ 1 2 FRAGMENT P PN P 2 N P 3 N 2 P 3 N 3 P N 4 3 4 4 P N 5 4 P N 5 5 P N 6 6 TOTAL C y c l i c 0.4 .40.6 10.3 15.1 12.0 78.4 Linear 6.2 9.0 3.8 2.3 0.1 21.4 TABLE 3.2. The fragmentation pattern of P^NyC£g and P^N^C£^2 into condensed, c y c l i c and l inea r species. The number given i s the t o t a l percentage of ions of the type shown, i r respect ive of chlorine content or charge. For convenience, chlorides of phosphorus are included with the l inear se r ies , and the PN series with the c y c l i c compounds. - 29 -a l l ions , s ingly or doubly charged, belonging to a pa r t i cu l a r parent nucleus P X N^ are grouped together. Several conclusions can be made by comparing the two patterns. The (P N nC£ ) n + series only accounts for 23.3% of the t o t a l frag-v x x+1 y ments, a high proportion of the fragments being small l inear ions. This indicates a general i n s t a b i l i t y of the condensed r i n g skeleton with respect to the c y c l i c and, more e spec ia l ly , l inea r species. (PNCi^)^ breaks down into mainly P^N^ fragments, which shows the importance of a transannular effect i n cleaving the P^N^ nucleus into two P^Ng un i t s . In the c y c l i c fragments formed from the breakdown of P^NjClg, P^N^ uni ts again predominate. The high proportion of ? 4 N 4 uni ts does not f ind an obvious explanation. The actual mechanisms involved i n the breakdown of the phospho-n i t r i l i c chlorides are not understood. Nonetheless, some deductions can be made from the fragmentation pattern of the condensed r i n g . Within the series of c y c l i c fragments, the most abundant fragments are ( P 4 N 4 C £ 7 ) + and ( P 3 N 3 C £ 5 ) + (see Table 3 .3) . Any mechanism which i s postulated for the formation of these ions must involve the cleavage of the in te rna l PN bond and the loss of the. bridgehead (HPC£) phosphorus atoms. I f =PC&2 type phosphorus atoms were l o s t , the resultant ions would have too few chlorine atoms. In the formation of the ( P 4 N 4 C £ 7 ) + i on , two PC£ type phosphorus atoms must be l o s t . Of the three nitrogen atoms which must be los t from P^N-CJt,, i n order to form the (P.N.C£_) + fragment, one must be o / y 4 4 / PARENT FRAGMENT FORMULA CHARGE RING ON ION P N 5 5 P 5 N 5 a 7 PCN_CJL D O O P 5 N 5 C£ 5 2.8 0 1.5 + 0 0.7 0 ++ P N P 4 N 4 C£ y P 4 N 4 C *6 P 4 N 4 C £ 5 P 4 N 4 C £ 4 P 4 N 4 C A3 23.5 1.0 9.7 0 3.7 + 0 2.2 0 2.0 0 ++ P N 3 3 P„N_C£, J J D P 3 N 3 C£ 5 P 3 N 3 C£ 4 P 3 N 3 C£ 3 P 3 N 3 C 1 2 P 3 N 3 Cil 0 39.0 2.4 37.5 0 4.2 + 0 0 1.7 0 0 0 ++ TABLE 3 3. Relative abundances of a l l ions in the various c y c l i c series i n the fragmentation pattern of P,N-CJl n . The numbers given are not percentages and represent only the in tens i t i e s of peaks r e l a t i ve to each other. - 31 -the central nitrogen atom. It i s believed that the loss of th i s nitrogen atom i s achieved i n a process which involves the cleavage of the c y c l i c structure of the condensed r i ng nucleus to give a l inear species which can then ei ther r e cyc l i z e into the smaller P^N^ and PjN^. un i t s , or form a series of l inear fragments. The high proportion of l inea r fragments from P^N yC£ g (43.9%), i n comparison to (PNC£ 2 ) 6 (21.4%) supports th i s hypothesis. In so far as the fragmentation pattern of P^NyC£g i s a guide to i t s chemical proper t ies , several in te res t ing comparisons can be made. Loss of the condensed r ing nucleus i s favoured i n the mass spectrum, with the formation of a large proportion of l inear fragments (much more so than i n the monocyclic chlorides) and small \"cycl ic fragments. The central nitrogen atom, and the bridgehead phosphorus atoms, seem to be los t i n these processes. These resu l t s indicate a greater chemical i n s t a b i l i t y for P^N yC£ g i n comparison to the monocyclic ch lor ides , and, as the resu l t s of chapter 4 w i l l show, the chemistry of the condensed r i ng compound i s l imi ted by the i n s t a b i l i t y of the P^Ny nucleus. I I I . U l t r a - v i o l e t Spectra The monocyclic phosphoni t r i l i c chlorides do not show any U .V .absorp-tions above 2100 X, where TT-»IT* t r a n s i t i o n s are normally expected. Below o 2100 A, n-*rr* t rans i t ions on the chlor ine ligands do occur, but these t rans i t ions have not received any specia l a t tent ion . The U.V. spectrum of PJ^7CiQ was recorded i n cyclohexane and a c e t o n i t r i l e so lu t ions , and - 32 -no spectrum was observed above 2100 A. The appearance of an absorption o below 2100 A was not iced, ind ica t ing that the type of U.V. t rans i t ions i n PgN^CJlg are s imi l a r to those observed i n the monocyclic chlor ides , and that the addi t ion of the central nitrogen does not produce any eas i ly observable TT-HT* t r ans i t i ons . IV. V ib ra t iona l Spectra Several de ta i led studies have been made on the v ib ra t iona l assign-or «.!. + • • 28-32 , . . . 28,29,33 , , . „ . . . ments of the t r imer ic and tetrameric phosphoni t r i l i c chlorides but only tentat ive assignments have been made for the penta-27 meric and hexameric chlor ides . The infra-red (mull and solut ion) and Raman (powder) spectra of P^N^CJlg were recorded. These spectra are given i n Table 3.4. The v i b r a t i o n a l spectra of (PNCl^)^, de ta i led i n Table 3.5, act as an in te res t ing comparison. Whilst complete v ib r a t i ona l assignments are nearly se t t led for N^P^CJl^, for the larger r i ng s i zes , assignments by different authors vary.considerably. For any molecule, v ib ra t iona l assignments can be given which f i t the observed frequencies, but the reasons behind the assignments are d i f f i c u l t to j u s t i f y completely. J u s t i f i a b l e assign-ments can r e a l l y only be given when the spectra of a large number of der ivat ives of a pa r t i cu l a r compound have been studied, and p o l a r i z a -t i o n measurements are ava i l ab le . In the case of P,N_C£A, no substi tuted der ivat ives have been made, o / y Only a powder Raman spectrum has been obtained, so that no po la r i za t ion - 33 -RAMAN SOLID -1 cm INFRA-RED NUJOL MULL cm INFRA-RED SOLUTION cm ASSIGNMENT 149 (3.5) 169 (7.0) 288 (0.4) 280 w 332 (7.0) 330 w 6(PCil 2) in-phase 411 (0.6) 402 v.w 442 (4.0) 435 m 430 w v (PC£„) in-phase sym 2 r 507 (0.8) 503 525 548 v.w w sh 557 (1.0) 564 St 562 m v(P-C£) in-phase 661 (1.5) 592 660 v. St St 603 m 662 VSym^ C V °\" t _° f -phase 807 (1.5) 804 870 903 1188 m St v.w sh 805 w 858 sh 880 m A^ r i n g vib'n (see text) -1280 v.St 1279 st degen r i n g vib'ns -1320 sh TABLE 3.4. Vib ra t iona l spectra of n i t r i lohexaphosphoni t r i l i c chlor ide . For the Raman spectrum, the figures i n parenthesis indicate the r e l a t i ve i n t ens i t i e s of peaks with respect to each other. t - 34 -RAMAN -1 INFRA--RED INFRA-RED ASSIGNMENT SOLID cm NUJOL MULL cm SOLUTION cm 128 CL5) 154 Ci.o) 167 CO.8) 198 (0.4) 275 (1.5) SCPCl^) in-phase 343 w 430 m 451 (10.0) v (PC£„) in-phase sym 2 r 463 st 465 w v (PC£_) 1st degen. sym 2 6 518 m 518 w 578 (1.0) 584 sh -580 sh 602 st 602 st V (PC£_) out-of-748 st 748 m sym 2 , J phase 830 m 853 m 1432 v. st 1427 st degen. r ing v i b ' n TABLE 3.5. Vibra t iona l spectra of hexaphosphonitri l ic ch lor ide . For the Raman spectrum, the figures i n parenthesis indicate the r e l a t i v e in t ens i t i e s of peaks with respect to each other. - 35 -measurements are ava i l ab le . The low s o l u b i l i t y of the compound i n a l l solvents has made the detection of a Raman spectrum i n solut ion unfeasible. Attempts to take Raman spectra of the melt also f a i l e d . Because of t h i s dearth of information, rather than quote v i b r a -t i ona l modes and then match frequencies to them, i t i s the purpose of th i s study to take the most prominent frequencies from the Raman and infra-red spectra, and to correlate them with the generally accepted assignments i n the smaller c y c l i c phosphoni t r i l i c ch lor ides . The appearance of the infra-red spectrum i s s imi l a r to that observed i n the monocyclic ch lor ides , the strong broad band at 1280 cm being charac te r i s t i c of a l l phosphoni t r i l i c compounds and assigned to a degenerate r ing v i b r a t i o n . The large number of coincidences i n the Raman and infra-red spectra suggest that the molecule has low symmetry. 31 I f one assumes the presence of the 3-fold axis indicated by P n.m.r. spectrometry, the symmetries D^, D^^ and D^^ can be excluded since these do not lead to Raman and infra- red coincidences of t o t a l l y symmetric v ib ra t ions . This leaves or C^v as being the only symmetries the molecule can possess. That the bridgehead chlorine atoms are a l l c i s to one another i s also indicated s ince, i f not, the resultant C s symmetry would produce a much more complex spectrum. The two most prominent features of the Raman spectrum are the two intense peaks at 332 cm * and 442 cm ^. In the monocyclic ch lor ides , the most intense peak i n the Raman spectrum i s the polar ized band corresponding to the PC£„ in-phase symmetric s t retching v i b r a t i o n . - 36 -Table 3.6 shows the frequencies of these peaks i n the Raman spectrum of the monocyclic ch lor ides . On the basis of these assignments, the Compound (PNC£ 2 ) 3 (PNC£ 2) 4 ' (PNC£ 2 ) 5 (PNC£ 2 ) 6 P 6 N ? C£ 9 S 4 C 2h symv 2J in-phase cm 365 387 391 413 439 451 442 References 29,30 31,32 29,33 29; 33 29 This work This work TABLE 3:6. Assignments for the symmetrical in-phase PC&2 s t retching v ib ra t ion i n phosphoni t r i l i c ch lor ides . band at 442 cm i n the Raman spectrum of P^N^CJlg i s assigned to the V (PC£„) in-rphase mode. - sym 2 r 32 In (PNC£ 2 ) 3 , the v S y m ( P G £ 2 ) out-of-phase mode occurs as a very strong band i n the infra- red spectrum at 530 cm Although i t i s not forbidden i n the Raman spectrum, i t does not appear there. On these grounds the analogous mode i n P6N7CJ£g i s assigned to the absorption at 592 c m - 1 i n the infra-red spectrum. S i m i l a r l y , the band at 602 cm\" i n the infra-red spectrum of (PNCZ^)^ i s assigned to the v S y m C P 0 ^ 2 ^ out-of-phase mode.in that molecule. The strong peak at 332 cm * i n the Raman spectrum i s assigned to the in-phase PC&2 scissors v i b r a t i o n . The assignment of t h i s mode i n the monocyclic series i s uncertain. It i s expected to - 37 -occur at a lower frequency than the symmetric s t retching motion. In 32 the t r imer ic ch lor ide , Emsley did not assign th i s mode to any 30 -1 frequency. Califano has assigned i t to a peak, at 100 cm and 31 -1 Hisatsune to a peak at ~304 cm . In (PNF,,)^ 4 , the in-phase PF 2 scissors v ib r a t i on l i e s about 100 cm * below the symmetric in-phase stretching mode. The in tens i ty of the peak at 332 cm i n the Raman spectrum of P^NyC£g suggests that i t i s an mode and that the above assignment i s v a l i d . The next assignment that can be made concerns the strong infra- red band at 564 cm ^. This i s thought to be the symmetric P-CZ s t re tching motion of the bridgehead chlorine atoms. In P^N^P^CH, the P-CZ v i b r a --1 32 t i o n i s assigned to a peak at 637 cm . The corresponding v ib ra t ion -1 34 i n PgN^Br^CJi occurs at 563 cm . In mononuclear phosphorus (V-J com--1 -1 35 pounds, v(P-C£) varies between 513 cm and 567 cm . The assignment of v (P-CJlj at 564 cm * seems l o g i c a l , sym 6 A recent review of the infra- red absorption frequencies of PN bonds indicates a broad range of frequencies between 700 cm ^ and 950 cm ^ as being the s t re tching frequency of the PN single bond. In P_N_C£ rNMe„, 3 7 v(P-N) occurs at 711 c m - 1 . In P,N^C£ r t the infra-red o 3 b I 6 / 9 peak at 804 cm * weakens i n so lu t ion and i s a strong peak i n the Raman spectrum at 807 cm ^. The condensed r i ng PN nucleus has one more A^ r i n g s t re tching mode than does the open r i ng P^N^ nucleus. This e s sen t i a l ly involves the symmetric s t re tching motion of the in ternal PN bonds. On the basis of the above information, the infra-red - 38 -absorption at 804 cm i s assigned to th i s v i b r a t i o n . Although further assignments could be made here, t he i r val id i ty-would be questionable. The prominent features of the spectra have been given prel iminary assignments, and the resul ts are consistent with the P 6 N 7 C£ g molecule having C 3 v symmetry. The shape of the molecule i s therefore completely defined, and the bridgehead P-C£ bonds must a l l be c i s to one another.. V. Base Strength Measurements Although base strength i s a chemical property, the in terpre ta t ion of th i s property i s i n terms of a physical phenomenon, and therefore th i s discussion of the base strength of P^N^CJlg has been grouped with the other physical properties of the molecule. The phosphoni t r i l i c chlorides are weak bases, the r ing nitrogen atoms act ing as the basic centres. To the extent that lone pa i r d e r e a l -i z a t i o n from the r i ng nitrogen atoms affects the base strength of these molecules, then the measurement of the r e l a t i v e base strength of the condensed r i ng compound and the open r i ng (PNCi^Dg molecule gives some ins ight into the difference i n the bonding of the two compounds. The r e l a t i ve values of the base strengths of the phosphoni t r i l i c chlorides (PNCJ^) (n = 3-6) have been obtained by the measurement' of the p a r t i t i o n ra t ios of the chlorides between hexane and concentrated 38 sulphuric ac id . A s imi l a r study was carr ied out on the condensed r i n g ch lor ide , using (PNC£ 2), to standardize the sulphuric ac id . By - 39 -varying the concentration of the sulphuric ac id , the change i n the p a r t i t i o n r a t i o was measured as a function of the concentration of the ac id . The resu l t s of these measurements, together with those found for the monocyclic ch lor ides , are shown i n F i g . 3.2, where the p a r t i t i o n r a t io i s shown Con the v e r t i c a l a x i s ) , p lo t ted against the Hammett a c i d i t y function HQ. The HQ function represents the a b i l i t y of a solvent to donate a proton to a neutral base and i s defined by equation 3.1. The physical chemistry behind th i s separation i s not H = pK + - log fflJ- -(3.1) U M [ B J f u l l y understood, and at present there i s no explanation for the TBH +1 va r i a t i on of the slopes i n F i g . 3.2. When — = 1, i . e . along the I B ] x - a x i s , HQ = pKgj_|+ and thus the intercept of the various l ines with the x-axis gives an approximate measure of the d i s soc ia t ion constant for a pa r t i cu l a r compound. Although the quotation of absolute values of Kg^+ have l i t t l e meaning, the resu l t s show that the base strength of the condensed r i ng chlor ide i s less than that of the open r ing hexameric chlor ide by a value of ApKg^+ - 0.1. The increasing r i n g angle at nitrogen with increasing r ing s ize i n phosphoni t r i l i c chlorides (see Table 3.7) i s symptomatic of increased d e r e a l i z a t i o n of the nitrogen lone pa i r electrons, with a consequent weakening of the i r basic character. In planar molecules, 2 bonding of the sp lone pa i r hybrid on nitrogen i s ch i e f l y with the F i g . 3.2. Va r i a t i on i n the p a r t i t i o n r a t i o of phosphoni t r i l i c chlorides •; between hexane and concentrated sulphuric acid with change i n concentration of the sulphuric ac id . - 41 -Compound (PNC£ 2 ) 3 (PNC£ 2 ) 4 (PNC£ 2) 5 Ring angle at nitrogen 120° 132° 149° Reference 39 40 41 TABLE 3:7. Average r ing angle at nitrogen i n the monocyclic phospho-n i t r i l i c chlorides (PNC£ 2 ) 3 _^. in-plane d o 2 o r b i t a l on phosphorus.. In non-planar systems, however, x -y overlap can also occur into the out-of-plane d-orbi ta ls on phosphorus. Ring nitrogen lone-pair d e r e a l i z a t i o n i s therefore cont ro l led by several factors , and the extent to which i t occurs i s a balance of these factors . A f u l l e r in terpre ta t ion of these resu l t s i s given i n Chapter 5. _3 Method:- 10 molar solutions of P 6 N 7 C£ g and (PNC£ 2 ) 6 were made up i n cyclohexane solvent . The p a r t i t i o n ra t ios were measured by mixing 30 ml of the cyclohexane solu t ion with 30 ml of concentrated sulphuric ac id , shaking the mixture for 2 minutes, and then f inding the resultant concentration of chloride i n the cyclohexane layer by evaporating a known volume of so lu t ion to dryness and weighing the residue. The con-centrat ion i n the acid layer was determined by difference. A l l measure-ments were made i n t r i p l i c a t e . Each concentration of acid was standard-ized by f inding the p a r t i t i o n r a t i o of hexameric chloride i n the acid - 42 -and f i t t i n g the point so obtained onto the o r i g i n a l p lot for hexaraeric chloride. 3** V I . Lewis Acid Adduct Formation The a b i l i t y of the phosphoni t r i l i c chlorides to form Lewis a c i d : adducts i s r e a l l y a measure of the i r base strengths, and to th i s extent they are only weak bases, since only a few such adducts have ever been prepared. (PNCK^g forms a 1:2 complex with aluminium t r i c h l o r i d e 43 and a 1:1 complex with aluminium tr ibromide. The larger r i ng s ize chlorides do not form complexes and an attempt to react P^N^C&g with aluminium tribromide d id not produce any complex. The d e r e a l i z a t i o n of the lone pa i r on the central nitrogen i s therefore considerable, and the molecule as a whole must be viewed as being less basic than CPNC£ 2) 3. Method:- A so lu t ion of resublimed aluminium tribromide (0.039 g. 0.145 mmol) i n 9.0 ml of carbon disulphide was added, under an atmosphere of dry ni trogen, to a s t i r r e d so lu t ion of P^N^Cilg (0.087 g . , 0.145 mmol) i n 10 ml of carbon disulphide . After h a l f an hour, there was no v i s i b l e sign of a react ion . The carbon disulphide was d i s t i l l e d o f f and 0.083 g. of the P, N_C£_ recovered from the residue • 6 7 9 by extract ion into benzene. - 43 -CHAPTER 4 SUBSTITUTION REACTIONS OF NITRILOHEXAPHOSPHONITRILIC CHLORIDE A series of subs t i tu t ion reactions was attempted i n th i s study. The resul ts of these reactions show that, unl ike the monocyclic phospho-n i t r i l i c ch lor ides , the condensed r i ng compound i s not stable to sub-s t i t u t i o n . The weakness of the in te rna l PN bonds, indicated i n the mass spectrum, i s apparent i n the chemistry of the compound. Attempted introduct ion of electron-withdrawing groups (-F, -OMe) into P^N^Ci^ brings about a collapse of the P^N y skeleton, and only when electron-releasing groups are introduced does the condensed r ing compound form stable products. The fol lowing discussion i s divided into several par ts . The f i r s t part describes the resu l t s of the f luor ina t ion and methoxyla-t i on react ions, and the second part describes the resu l t s of dimethyl-amination of the condensed r ing compound. In the t h i r d sect ion, a discussion i s given of the spectra and structure of PgN^(NH) QWk^igCi^-I (a ) . F luor ina t ion of Ni t r i lohexaphosphoni t r i l i c Chlor ide . Monocyclic phosphoni t r i l i c f luor ides are v o l a t i l e l i q u i d s , ^ and i t was consequently expected that P^N^F^ would be s i m i l a r , and hence easy to separate from other react ion products. I ts properties were 19 also expected to be informative, since F has a nuclear sp in , and more s t ruc tura l information would be obtainable spectroscopical ly from the - 44 -f luor ide than the chloride of the condensed r ing phosphonitr i le . Because of these fac ts , many attempts were made to f luor inate the condensed r ing ch lor ide . A var ie ty of reaction, conditions and f l u o r i n a -t i n g agents were used. Potassium f luorosulphi te has been used success-11 43 f u l l y , as has argentous f luor ide , to f luor inate phosphoni t r i l i c 44 chlor ides . Act ivated potassium f luor ide has been used to f luor inate phosphorus chlor ides . A l l these reagents were used, but none produced any compound which could be properly characterized. Some of the f luor ina t ion reactions were done using open systems. In these cases, although no f luorides were i so l a t ed , the condensed r i ng chloride was los t i n the react ion, with the formation of ion ic ch lor ine . At f i r s t no explanation could be given for these f a i l u r e s . F i n a l l y , an exper i -ment was carr ied out using a closed system with a l i q u i d nitrogen trap to condense out any gaseous products. The products of th i s react ion, which was carr ied out using s i l v e r f luor ide as the f luor ina t ing agent, were a l l ei ther gases or v o l a t i l e l i qu ids at room temperature. I t i s thought that there are several products of th i s reac t ion , a l l of which are f u l l y f luor ina ted , and a l l of which appear to have los t the condensed r ing PgN^ nucleus. The number of products made e f f i c i en t separation impossible. Trap to trap f rac t ionat ion was attempted using dry ice/acetone i-97°C), dry i c e / C C ^ (-23°C), dry ice/chlorobenzene (-45°C) cold t raps, but no f ract ionat ion was achieved. The appearance of PQR patterns i n the vapour-phase i n f r a -red spectra of the v o l a t i l e products indicated that some of the products - 45 -had low molecular weights, i . e . low enough to allow ro ta t iona l fine structure to be observed. The magnitude of the s p l i t t i n g of the PQR branches was such that the distance between the maxima of the P and R branches was, i n a l l cases, about 20 cm \\ For l inear molecules, the separation, Av, between the maxima of the two branches i s given by the formula'**': Av = /8kTB = 2.358 yflT (4.1) he where T i s the absolute temperature and B i s the ro ta t iona l constant of the molecule (B = ) . This formula only holds r igorously for 8TT 2IC l inea r molecules, but may be used as an approximate guide for non-l inear molecules. Using equation 4 . 1 , the value of B corresponding to a Av of 20 cm\" 1 i s B = 0.24 c m - 1 . The B value for PF_ i s B = o 3 o 46 0.26084. Although a quant i ta t ive comparison i s f u t i l e , the value of B = 0.24 cm ^ does indicate that the s ize of the molecule i s not vas t ly different from PF^-Mass spectra resul ts again indicated a mixture of products, mainly of low mass number, the most intense peaks corresponding to the fragments ( P 2 F ) + and ( N 2 F ) + . The charac te r i s t i c i so topic pattern of chlorine did not appear i n any fragment, showing that complete f l u o r i n a -t i o n had taken place, and although the average molecular weight of the product (by vapour density) was 114, fragments of mass number >400 were observed i n the mass spectrum i n low abundance. It i s therefore l i k e l y - 46 -that a l i t t l e P 6 N y F g (M = 455) was formed, but i t could not be characterized because of the sca rc i ty of peaks i n the range m/e = 200-350, which made accurate mass counting impossible. Very l i t t l e pos i t ive information was obtained on the types of molecules present i n the product mixture. The important conclusion to be drawn i s that the f luor ina t ion react ions, as carr ied out i n th i s work, break the condensed r i ng PN skeleton to produce v o l a t i l e products of unknown composition. An attempt was made to f luor inate the condensed r ing compound p a r t i a l l y , using a method s imi l a r to that used i n pre-47 paring the tetrameric chloride f luorides P.N.F C£ n . This react ion r 6 4 4 x 8-x produced only unreacted P^N^CJlg and a trace of v o l a t i l e decomposition products of the type already discussed. Experimental i ) Reaction of Ni t r i lohexaphosphoni t r i l i c Chloride and Potassium Fluorosulphite i n Cyclohexane. A s lu r ry of P^N^Cilg (3.5 g . , 3.85 mmol) and potassium f luoro-sulphite (6.5 g . , 11.0 mmol) i n 125 ml. of cyclohexane were heated under ref lux i n an anhydrous atmosphere. The re f lux condenser was connected i n series to a cold trap immersed i n a dry ice/acetone slush bath. After 16 hours, the react ion mixture was cooled and f i l t e r e d to remove insoluble s a l t s . The cyclohexane solvent was d i s t i l l e d o f f to y i e l d a c r y s t a l l i n e deposit which was i den t i f i ed as P^NyCJlg (1.4 g . , 2.3 mmol) by i t s infra-red spectrum. No other fract ions were i so la ted i n th i s - 47 -d i s t i l l a t i o n , ind ica t ing that no P^NyFg was present. The insoluble prec ip i ta te was dissolved i n water and a few pe l l e t s of caustic soda added. The solu t ion was evaporated to dryness and the residue dissolved i n 250 ml. of water. A small al iquot of th i s so lu t ion gave a pos i t ive resul t when tested for chloride ion , ind ica t ing that sub-s t i t u t i o n of the chlorine ligands had occurred. Another al iquot of t h i s so lu t ion was oxidized with a few drops of cone, n i t r i c ac id , and treated with a few drops of a so lu t ion prepared by d i s so lv ing ammonium and magnesium chlorides i n d i l u t e aqueous ammonia. The appearance of a white p rec ip i ta te confirmed the presence of the phosphate ion i n the o r i g i n a l so lu t ion , ind ica t ing that decomposition of the phosphoni t r i l i c r i ng had occurred. i i ) Reaction of Hexaphosphonitril ic Chloride with Activated 44 Potassium Fluoride Freshly prepared \"act ive\" potassium f luor ide (2.5 g . , 43 mmol) was added to a small 3-necked f lask containing f i n e l y ground (PNCJ^Jg (2.0 g . , 2.8 mmol). A ref lux condenser was attached to the apparatus. The condenser was then connected i n series to a cold trap (immersed i n a dry ice/acetone slush bath) which was i t s e l f connected to a s i l i c a gel drying tube. The react ion mixture was s t i r r e d mechanically and heated i n an o i l bath to a bath temperature of 145°C. The reaction mixture soon turned into a paste which, after h a l f an hour at 145°C, began to bubble. After about a minute, the bubbling lessened and the - 48 -react ion was allowed to continue for 4 hours, after which time the o i l bath was removed from the react ion vessel and the whole apparatus allowed to coo l . No v o l a t i l e products were present i n the cold trap. The react ion vessel was washed with 3 x 10 ml.of d ie thy l ether. The ether extract was f r a c t i o n a l l y d i s t i l l e d to y i e l d (PNF^)^ CO.70 g . , 48% of theory), b o i l i n g point 147°C.U i i i ) Reaction of Ni t r i lohexaphosphoni t r i l i c Chloride with Activated Potassium Fluoride In an analogous experiment to that described above, P^N^Cilg (1.07 g . , 1.78 mmol) and act ivated potassium f luor ide (3.6 g . , 62 mmol) were s t i r r e d together and heated to 150°C for 48 hours. There was no v i s i b l e evidence of any react ion occurring at any time during th is per iod. At the end of th i s time, no v o l a t i l e products had been col lec ted i n the cold t rap. The react ion mixture was extracted with d ie thy l ether, benzene and a c e t o n i t r i l e . On d i s t i l l a t i o n of these solvents no other products were found. The react ion residue was completely water soluble and gave a pos i t ive test for chloride ion when treated with a c i d i f i e d s i l v e r n i t r a t e , ind ica t ing that subs t i tu t ion of the condensed r i ng compound had taken place. iv ) Reaction of Ni t r i lohexaphosphoni t r i l i c Chloride with Argentous Fluoride A mixture of f i n e l y ground P^N^C£g CO.800 g . , 1.34 mmol) and f i n e l y ground s i l v e r f luor ide (3.5 g . , 27.5 mmol) were added to a 50 ml . - 49 -one-necked f lask and the f lask connected i n series to a cold trap (immersed i n l i q u i d nitrogen) and a vacuum system (see F i g . 4 .1) . At a pressure of atmosphere, the reactants were heated to ~180°C i n a sand bath for one hour. No d i s t i l l a t e was noticed on the walls of the cold t rap. Over a period of 15 minutes, the temperature of the sand bath was raised to 240°C. A condensate slowly col lec ted on the walls of the cold t rap. The heating was continued for 3 hours, at the end of which time, the react ion mixture residue was a dark grey s o l i d (pure AgF i s l i g h t ye l low) . The s i l v e r f luor ide plug (see F i g . 4.1) was heated to about 100°C with an infra-red lamp and the v o l a t i l e products of the react ion passed through t h i s plug ten times by a series of trap to trap vacuum d i s t i l l a t i o n s . The product (0.290 g.) was col lec ted i n a sealed tube. This product was gaseous at room temperature, and i t s average molecular weight was found to be 114 (the molecular weight of P 6 N y F 9 i s 455). I (b) . Methoxylation of Ni t r i lohexaphosphoni t r i l i c Chloride The f u l l y methoxylated c y c l i c phosphonitr i les (PNfOMe^) 48 49 (n = 3-8) have been prepared. ' These compounds are colourless c r y s t a l l i n e s o l i d s , except for (PNCOMe^)^ a m j j, which are both l i q u i d s . They are soluble i n most organic solvents, but not i n water. However, i t i s possible to extract them from an organic so lu t ion with IN hydrochloric ac id . Heated under vacuum, the methoxyphosphonitriles undergo a thermal rearrangement, to give a hygroscopic glass , the F i g . 4 .1 . Apparatus used i n the attempted f luor ina t ion of P ^ C J l g . - 51 -infra- red spectrum of which indicates the presence of a P=0 bond. A Kirsanov type rearrangement i s believed to take place during t h i s reac t ion . \\ / MeO J) N y N I (4.2) J? OMe .P OMe OMe OMe These compounds are made by a simple metathetical react ion with sodium methoxide. Completely anhydrous reagents and conditions must be used, since the presence of moisture d r a s t i c a l l y reduces the y i e l d of product. The preparation of a condensed r ing methoxyphosphonitrile was attempted 31 1 i n t h i s work. P and H n.m.r. spectrometry would have been p a r t i c u l a r l y informative i n deducing the structure of any compound i so l a t ed , and the thermal properties of a condensed r ing methoxide would be in te res t ing i n comparison to those of the monocyclic methoxides. Despite elaborate precautions for the exclusion of moisture i n the reactants, no phospho-n i t r i l i c compound was obtained from the reac t ion . The y i e l d of i on ic chlorine from the react ion corresponded to complete subs t i tu t ion of the r ing chlorine atoms. Concurrent or subsequent r i ng cleavage i s believed to have occurred, with the loss of the P,N_ skeleton. o / Method:- Anhydrous reagents were prepared as fo l lows. \"Super-dry\" methanol was produced by t r i p l y d i s t i l l i n g 100% methanol from sodium. Benzene (solvent) was dried by doubly d i s t i l l i n g spectroscopic grade benzene from sodium. Sodium methoxide was prepared*^ by d i sso lv ing - 52 -clean, dry sodium (0-150 g . , 6.53 mmol) i n 10 ml of \"super dry\" methanol. The excess of methanol was d i s t i l l e d o f f to leave a white powder of sodium methoxide. Under an atmosphere of dry ni trogen, 15 ml of dry benzene were added to the powdered sodium methoxide and the mixture s t i r r e d into a s lu r ry with a magnetic s t i r r e r . A so lu t ion of n i t r i lohexaphosphoni t r i l i c chloride (0.300 g, 0.50 mmol) i n 10 ml of dry benzene was / s l o w l y added to the sodium methoxide s l u r r y . The react ion mixture was gently heated under re f lux for 4 hours, after which time the s l u r ry was cooled and 20 ml of d ie thy l ether were added to the mixture. The re su l t ing suspension was washed with cold d i s t i l l e d water (4 x 25 ml) and th i s water co l lec ted and analysed for Cl ion by p r ec ip i t a t i on as AgC£. The y i e l d of Cl ion was 96% of that amount equivalent to complete subs t i tu t ion of the r i n g . The res idual ethereal so lu t ion was then extracted with IN hydro-ch lo r i c acid (2 x 25 ml) , the extract neutral ized with sodium bicarbonate, and re-extracted with chloroform. This so lu t ion was dried over anhydrous calcium chloride for 24 hours, decanted and allowed to evaporate. No phosphoni t r i l i c compound was obtained from th i s extract . The o r i g i n a l ethereal so lu t ion was also dried and evaporated, and no phosphoni t r i l i c compound was i so la ted from th i s so lu t ion . I I . Amination of Ni t r i lohexaphosphoni t r i l i c Chloride The react ion between phosphoni t r i l i c chlorides and many different types of amines has been studied. In most cases the react ion i s a - 53 -simple metathetical subs t i tu t ion involv ing the e l iminat ion of the amine hydrochloride. P a r t i a l l y aminated phosphoni t r i l i c chlorides can also be made by react ion of the two reagents i n the appropriate s to i ch io -metric r a t i o . The successive subs t i tu t ion of chlorine ligands by amine groups i s i n a non-geminal fashion, and subs t i tu t ion of the r ing tends to deactivate the r i ng to further subs t i tu t ion . The f u l l y dimethyl-aminated phosphonitri les [PNCNN^j^^ $ have been prepared 5 1 and a l l are colourless c r y s t a l l i n e s o l i d s . / Because the react ion proceeds eas i ly at room temperature, i s not sensi t ive to moisture, and because the substituent deactivates the r i ng to further reac t ion , dimethylamination of the condensed r ing compound was attempted. Even with an excess of dimethylamine, complete sub-s t i t u t i o n of the chlorine ligands could not be achieved. The compound that was i so la ted from th i s react ion was f i n a l l y characterized as being 3,11,imino-3,11,dichloro-octadimethylaminocyclohexaphosphonitrile, the molecular formula of which i s P,N,(NH)(NMe 0) 0C&„. The course of the o o z o I dimethylamination react ion may be wri t ten simply as: - 54 -A discussion of the structure of P^N^ CNH) Q M ^ D g C J ^ i s given i n the fol lowing sect ion. A s imi l a r react ion was attempted with diethylamine. No c r y s t a l l i n e product could be i so la t ed , and since the react ion was done on a small sca le , p u r i f i c a t i o n of the res idual o i l could not be made. In the monocyclic phosphoni t r i l i c dialkylamides, increasing s ize of the a l k y l group lowers the melting point of the compound and i t i s thought that th i s i s the reason that no c r y s t a l l i n e product could be i so la ted from the react ion of diethylamine with the condensed r i ng ch lo r ide . The course of the react ion with diethylamine i s thought to be the same as that with dimethylamine. The o i l gives a pos i t ive test for ch lor ine , showing that subs t i tu t ion of the r i ng i s incomplete. The infra-red spectrum of the o i l was t y p i c a l of a phosphoni t r i l i c dialkylamide, with a \"P=N\" absorption at - 1 2 6 0 cm * and a mul t ip le t of C-H stretching peaks i n the 2 9 0 0 - 3 0 0 0 cm * region. The \"^H n.m.r. spectrum ( 6 0 MHz) of the o i l i n carbon te t rachlor ide solut ion indicated at least p a r t i a l subs t i tu t ion . Signals from methyl and methylene protons were observed. The former appeared as a t r i p l e t C>J^_LJ ~ 7 c . p . s . ) at 5 = 1 . 0 6 p.p.m. and the l a t t e r , as\" a broad unresolved mul t ip le t (caused by 1 H - 1 H and 3 1 1 P- H coupling)and making i t impossible to d i s t ingu i sh inequivalent methylene groups appeared at 6 = 3 . 1 p.p.m. As section I I I of th i s chapter w i l l show, the resu l t s of the react ion with dimethylamine are very important i n the understanding of the chemistry of P,N_C£Q. In the course of th i s react ion the - 55 -condensed r i ng PN skeleton i s broken. Attack by a nucleophile at the bridgehead phosphorus atoms leads ei ther to the loss of a chlorine l igand or to the cleavage of the in te rna l P-N bond. That the in te rna l P-N bond i s broken i n preference to the P-C£ bond indicates the chemical weakness of the P^N_ nucleus to chemical subs t i tu t ion . b / Method:- Reaction of Dimethylamine with Ni t r i lohexaphosphoni t r i l i c Chlor ide . 12 ml of a cold (0°C) 10% solu t ion of dimethylamine i n benzene (equiv. to 0.804 g . , 17.9 mmol of HNMe2) were slowly added to a cold (0°C), s t i r r e d so lu t ion of P ^ C J l g CO.509 g . , 0.845 mmol) i n 15 ml of benzene. Immediate t u rb id i t y was noticed. After 3 hours, the react ion mixture was f i l t e r e d , and the react ion mixture and prec ip i ta te washed with 3 x 5 ml of benzene. The weight of the prec ip i ta te (0.550 g.) was equivalent to 0.68 mmol of dimethylamine hydrochloride. The benzene so lu t ion was d i s t i l l e d to leave a pale yellow o i l , which did not c r y s t a l l i z e on cool ing . The o i l was dissolved i n pe t ro l ether (30°-60°C), f i l t e red ,and the so lu t ion evaporated to leave a s o l i d product (y ie ld 0.460 g.) which could be r e c r y s t a l l i z e d from a c e t o n i t r i l e to produce co lour less , f laky c rys ta l s (m-pt. 161-163°C) of 3,11,imino-3,l l ,dichloro-octadimethylaminocyclohexaphosphonitri le (calculated for P ^ (NH) (NMe^gCJ^:- P, 26.32; N, 29.75; C, 27.18; H, 6.94; Cl, 10.01. Found:- N, 29.71; C, 27.17; H, 6.91; Cl, 9.97.) Unlike the parent condensed r i ng ch lo r ide , th i s compound i s soluble almost to the point of m i s c i b i l i t y i n benzene, chloroform, - 56 -carbon te t rachlor ide and pe t ro l ether (30°-60°) . E f f i c i en t r e c r y s t a l l -i z a t i o n can only be made from a c e t o n i t r i l e . Twinning of the crys ta ls i s evident, and the crys ta ls are also t r i b o - e l e c t r i c . The inf ra- red spectrum of the compound was recorded on samples i n a nujol mull and a carbon te t rachlor ide so lu t ion . The prominent peaks i n the mull spectrum are: 504 (s t ) , 540 (m), 658 (w), 682 (w), 730 (m), 771 (m), 781 (sh), 838 (w), 863 (w), 891 (w), 929 (m), 984 ( v . s t ) , 1062 (m), 1188 ( v . s t ) , 1231 (s t ) , 1280 (v . s t ) , 1456 (m), 2790 (w), 2880 (w), 2920 (m). I l l . The Molecular Structure of 3,11, imino-3,11,dichloro-octadimethyl- aminocyclohexaphosphonitrile The structure of th i s compound, which i s discussed i n the fo l low-ing sect ion, i s unique i n phosphoni t r i l i c chemistry. Although the actual conformation of the molecule has not been completely established, the structure of the molecule i s believed to be that shown i n F i g . 4.2. F i g . 4 .2 . Suggested structure of P 6 N (NH) (NMe2)gC£ - 57 -In the fol lowing discussion of the molecule, rather than use the f u l l name of the compound, the molecular formula P.N.(NH)(NMe.,) 0C£. w i l l be D O Z o Z used as a notation for the compound.. a) Infra-red Spectrum The infra- red spectrum of th i s compound shows a complex v ib ra t iona l pattern cha rac te r i s t i c of phosphoni t r i l i c dimethylamides. The \"P=N\" stretching frequency (at 1280 cm *) i s considerably lower than i n the open r i ng hexameric dimethylamide, i n which v(P=N) occurs at 1340 cm The large difference i n frequency indicates that P^N^(NH) (NMe2)gC£2 has very different symmetry and bonding from the open r ing dimethylamide and should not be regarded as a simple der ivat ive of the open r ing system. Neither the mull nor the so lu t ion spectra show an observable (N-H) v i b r a t i o n . The reason for the absence of th i s band i s thought to be intramolecular hydrogen bonding, which would sh i f t v(N-H) to lower frequency and also broaden the absorption peak. b) Mass Spectrum The mass spectrum of th i s compound i s d i f f i c u l t to discuss i n d e t a i l , f i r s t l y because of the large number of peaks (caused by proton abst ract ions) , which makes a de ta i led assignment of fragments impossible, and secondly because there are no analogous compounds with which to compare i t . The parent peak consists of a mul t ip le t based at m/e 706. The mass number 706 corresponds to the formula P^N^(NMe2)gC£2 (taking CI = 35). The mass number of P N f i(NH)(NMe 2) gC£ 2 i s 707 (taking CH = 35), - 58 -but the parent ion i s an odd electron.species . The parent peak i t s e l f , at m/e 707, has an intensi ty, of 24% (± 1%) of the P - l peak. The natural 15 13 abundance of N i s 0.38% and that of C i s 1.1%. On th i s bas is , the 13 15 calculated contr ibut ion to the parent peak from C and N i n the [P 6 N y CNMe 2 ) g C£ 2 ] + fragment i s 23% that of the P - l peak. The contr ibut ion of the parent ion to the m/e 707 peak must therefore be very smal l , and the N-H bond must be weak and ea s i l y broken. The absence of a parent ion i s not uncommon. The fragmentation 52 pattern of diphenylamine does not exhibi t any peak corresponding to the ion [C^H^-NH-C^H^] +, and the peak corresponding to the ion ( [C^Hj.-N-CgHj.]+ i s the most intense peak i n the whole mass spectrum. Presumably loss of a proton i s favoured since the r e su l t ing ion frag-ment w i l l : be s t a b i l i z e d by resonance with the two phenyl groups. The 53 mass spectrum of heptasulphur imide, S yNH, exhibi ts a s imi l a r phenomenon, the in tens i ty of the (S yNH) + ion being only 5% of that of the (S 7 N) + i on . The higher mass range part of the spectrum consists mainly of fragments of the type P^N^C^ (NMe 2) x (x = 8-2). In th i s part of the spectrum, loss of dimethylamine groups i s more favoured than loss of ch lor ine , thereby ind ica t ing a stronger P-C£ bond than P-NMe2 bond. 31 c) P n.m.r. Spectrum 31 The P n.m.r. spectrum i s deta i led i n Table 4 :1 . Resolution of 31 1 31 31 the 3 peaks i s poor because of P- H and P- P coupling. In the case of the low- f i e ld s ignal at 6 = -13.0 p .p .m. , the peak i s resolved into - 59 -Type of Phosphorus Atom P l P P P P P 4 ' 5' 6 6 p p.p.m. r e l a t i ve to 85% H 3 P0 4 -13.7 -3.7 +1.9 Intensity r a t i o 1 2 3 Structure T r ip l e t JP _ P - = 4 0 C - p ' S * broad unresolved s ingle t broad unresolved s ingle t TABLE 4 :1 . Data from P n.m.r. spectrum of P N (NH)(NMe ) C£ F i g . 4 .3 . Gross structure of P,N 6 (NH)(NMe 2 ) g C£ 2 indicated by 31n P n.m.r. spectrum. - 60 -a t r i p l e t caused by P- P coupling (Jp,p = 40 ± 2 c .p . s . ) P- H coupling i s unresolved and not measurable. The two h i g h - f i e l d s ignals are broad (half-height width - 100 c .p . s . ) unresolved peaks with no dis t inguishable f ine s tructure. The spectrum indicates the gross structure given i n F i g . 4 .3 . The lack of reso lu t ion i n the h i g h - f i e l d s ignals i s most l i k e l y caused by the fact that phosphorus atoms to which these signals are assigned can couple with more than one type of phos-phorus atom, thereby causing a complex coupling pattern which i s unresolvable. The P^ phosphorus atom, however, only couples with one type of phosphorus atom, and the simpler coupling scheme i s , therefore, resolvable . The chemical sh i f t s of the P^, P,- and P^ atoms are not required to be equivalent by symmetry. This type of occurrence i s not unusual. In monosubstituted hexameric f luorophosphonitr i les , a l l the =PF2 phosphorus atoms, except those adjacent to the substi tuted phos-phorus atom, have i den t i ca l chemical s h i f t s , even though they are not symmetrically equivalent. 31 The in terpre ta t ion of P chemical sh i f t s i n terms of the e lec t ro-nega t iv i t i e s of ligands i s unrewarding i n many phosphoni t r i l i c der iva-t i v e s . A l igand can affect the electron density at phosphorus i n two ways. A conjugative effect between phosphorus and the l igand can ei ther donate or withdraw electron density from phosphorus, whi ls t a TT-inductive effect can e i ther suppress or increase the d e r e a l i z a t i o n of the lone pa i r electrons on the r i ng nitrogen atoms, thereby changing the TT-electron density at phosphorus. In f u l l y substituted phospho-31 n i t r i l e s , the P chemical sh i f t moves upf ie ld with increasing e lec t ro -- 61 -54 negat iv i ty of the l igand. S i m i l a r l y , i n the series P^N^C^X (X = F , OEt, Ome, CI, NMe 2), the chemical sh i f t of the substituted phosphorus atom moves upf ie ld with the increasing e lec t ronegat iv i ty of X . 5 5 In these compounds, the influences of a TT-inductive effect seem to be the most important. 31 The cor re la t ion of the P chemical sh i f t s of V.UnCln and o / y 31 P^N^(NH)(NMe2)gC£2 i s not understood. At best, the P n.m.r. spectrum of P^N^(NH)(NMe2)gC&2 serves as a way of determining the gross symmetry of the molecule, but gives only a l i t t l e ins ight into the e lec t ronic structure of the molecule. The posi t ions of the three peaks i n the spectrum r e l a t i v e to each other are d i f f i c u l t to r a t i ona l i z e simply i n terms of a Tr-inductive effect . S imi lar d i f f i c u l t i e s ar ise i n the 31 P n.m.r. spectra of p a r t i a l l y dimethylaminated t r imer ic phosphoni t r i l i c chlor ides . F i g . 4.4 shows three such der ivat ives and the de ta i l s of 31 56 the i r P n.m.r. spectra. A Tr- inductive effect which would account B Me2N NMe2 Me2N NMe, Me2N C£ c i s CI NMe, 6 = -22.0 p.p.m. 6^ = -20.4 p.p.m. F i g . 4 .4. 6 = -23.9 p.p.m. 6 R = -30.5 p.p.m. 6 = -27.6 p.p.m. 31 P chemical sh i f t s of some t r imer ic dimethylamino-chlorophosphonitri les ( a l l values r e l a t i v e to 85% H 3 P 0 4 ) . - 62 -for the observed chemical sh i f t s i n compound A, cannot account for the sh i f t s i n compound B, where the introduct ion of an extra -NMe2 group has increased the chemical sh i f t of the atom from i t s o r i g i n a l value i n compound G. 31 Anomalies, therefore do occur i n the P n.m.r. parameters of dimethylamino-chlorophosphonitriles, and the in terpre ta t ion of chemical sh i f t s i n terms of Tr-electron density i s not always rewarding. d) *H n.m.r. Spectrum The *H n.m.r. spectra (at 60 and 100 MHz) of P N (NH) (NMe2)gCJl are shown i n F i g . 4 .5 . The phosphorus decoupled proton spectrum i s also shown, together with assignments of the various peaks. At 100 MHz, the resolu t ion of the peaks i s much better than at 60 MHz. The signals corresponding to the H a and H^ protons have collapsed into a s ingle peak i n the 60 MHz spectrum. The r e l a t i ve 31 i n t ens i t i e s of the signals i n both the coupled and P decoupled spectrum are i n the r a t i o of H : H, : H : H, = 1 : 2 : 3 : 2. In the 3 1 P a D c d decoupled spectrum, taken using a noise band with a band-width of 31 1 2000 Hz at a frequency of 40.481000 MHz, P- H decoupling i s completely achieved and the ordinary high resolu t ion spectrum, consis t ing of 4 doublets, collapses into 4 s ingle peaks, corresponding to 4 different types of proton i n the molecule. The *H spectra have been interpreted i n terms of the molecular structure shown i n F i g . 4 .5 . Although there are two conformations of the molecule which are consistent with *H n.m.r. spectra, a considera-t i on of the mechanism (which w i l l be discussed la te r ) of the formation - 63 -Ha Hb Hc Hd 5 ' . 1 0 1 5 ' 20 ' 25 5 10 15 20 25 F i g . 4 .5. • The H n.m.r. spectra of. PgNgCNH) (NMe'^ .gCJl' . Spectra B and C are taken at f i e l d strengths of 10.0 MHz. and 60 MHz. respect ive ly . 31 1 Spectrum A i s the P- H decoupled spectrum (taken using a 2000 Hz. noise band at 40.48100 MHz.). The figure on the top r ight of the diagram shows the assignments of the various peaks. - 64 -of the molecule has shown that the structure shown i n F i g . 4.5 i s the most probable conformation. The chemical sh i f t s and coupling constants of the various types of proton are given i n Table 4 .2 . I t i s important to note that the largest coupling constant i s J p „ = 12.7 c .p . s . Both i n mononuclear phosphorus(V) Type of proton H A H B H c HD Intensi ty r a t i o 1 2 3 2 6 p.p.m. 2 626 2.606 2.576 2.551 J p _ H c .p . s . 10 7 10.7 10.7 12.7 TABLE 4:2. Data from n m.r. spectra of PgN^ (NH)(NMe2)i compounds, and i n the dimethylamino-derivatives of t r imer ic phospho-n i t r i l i c c h l o r i d e , 5 ^ Jp^ values decrease with increasing aminolysis and the value of Jp^ acts as an excellent c r i t e r i o n for d is t inguish ing between =PC£NMe2 and ^ ( N l ^ ^ groupings. For the former, J p H i s never below 16.3 c . p . s . , and for the l a t t e r , i s never above 13.9 c .p . s . In - 65 -the series (PN(NMe 2 ) 2 ) n (n = 3-6), J p ^ decreases from 11.2 c .p . s . (for 58 n = 3) to 10.0 c .p . s . (for n = 6) . The effect of r ing s ize on Jp^ i s smal l , and the fact that i n P f i N 6 (NH) (NMe 2) gC£ 2 J <.12.7 c .p . s . proves that there are no =PC£NMe2 groups i n the molecule, and i s a further confirmation of the structure shown i n F i g . 4 .5 . The in terpre ta t ion of the \"*\"H n.m.r. spectra of phosphoni t r i l i c derivat ives i s i n most cases hampered by complex unresolvable coupling. 5 8 For instance, the simple c y c l i c dimethylamides show a deviat ion from the simple doublet pattern which i s expected from a f i r s t order t rea t -ment. The spectra consist of a sharp doublet between which l i e s a broad peak which, i n the case of (PN(NMe 2 ) 2 ) 3 accounts for 25% of the t o t a l in tens i ty of the s ignal and, i n the case of (PN(NMe 2 ) 2 ) g , accounts for a l l of the s i g n a l , with the outside doublet being completely unresolved. The ca lcu la t ion of t r ans i t i on energies and the r e l a t i ve in t ens i t i e s of each of the hydrogen nuc le i i n phosphoni t r i l i c molecules i s very complex and has not yet been achieved. Analysis of simpler spin 59 systems has shown that the \"odd\" appearance of the spectra i n high symmetrical spin systems such as c y c l i c phosphoni t r i l i c dimethylamides i s due en t i r e ly to t he i r symmetry.6^ The resolut ion of the *H n.m.r. spectrum of P^N^(NH)(NMe2)gC&2 into separate peaks i s ind ica t ive of the lack of s t ruc tura l symmetry i n the molecule. Long range coupling i s n e g l i g i b l e , and the lack of chemical, and therefore magnetic, equivalence of the phosphorus atoms i n the molecule eliminates the complexities which occur i n more symmetric systems. A simple f i r s t - 66 -order treatment can therefore be used to interprete the H n.m.r. spectrum i n terms of the structure of the molecule. Another in te res t ing feature that th i s molecule shows i s that some pairs of dimethylamino-groups attached to the same phosphorus atom have different chemical s h i f t s . This d i s t i n c t i o n of dimethylamine groups d i f f e r i ng only i n t he i r conformation at a phosphorus atom i s unique i n phosphoni t r i l i c chemistry. In (PNCNfr^J^J^ there are a x i a l and equatorial -NMe2 groups (which have different exocycl ic P-N bond lengths), but a l l protons i n the molecule resonate at the same frequency. In the monocyclic dimethylamide chlor ides , when two -NMe2 groups on a pa r t i cu l a r phosphorus atom are i n different environments, the i r ~~H chemical sh i f t s are s t i l l i d e n t i c a l . The fact that there i s a d i s t i n c -t ion between -NMe2 groups i n P^N^(NH) (NNfc^gCJ^ must be a resu l t of the molecule being held i n a r i g i d conformation. The reason for t h i s i s thought to be intramolecular hydrogen bonding, which holds the molecule r i g i d . Hydrogen bonding i s thought to occur between the (N-H) proton and the a x i a l -NMe2 groups on the r i n g . The (N-H) proton s i g n a l , which i s not shown i n F i g . 4 .5 , occurs at 6 = 6.15 p.p.m. at 60 MHz. I t i s a weak broad peak, and although in tens i ty comparisons are d i f f i c u l t , the expected r a t i o of 1:48 i s not inconsistent with the observed experi-mental in tegra l curve. Hydrogen bonding of the (N-H) proton would tend to withdraw electron density from the dimethylamino-group and thereby lower the chemical sh i f t of the protons attached to that group. The observed chemical sh i f t s are consistent with th i s theory, the extent of the hydrogen bonding depending on whether the -NMe ? protons are i n the - 67 -L & or pos i t ion (see F i g . 4 .5 ) . The equatorial L -NMe2 groups do not in teract with the (N-H) proton and a l l the protons resonate at a s ingle frequency. The H^ protons are i n different environments i n the molecule, but as i n the case of P^H^CZ^QMe^)^ (Fig . 4.4B), they have i den t i ca l chemical s h i f t s . Some understanding of the two in te rna l P-N bonds i n PgN^CNH) ( N I ^ ^ C J ^ can be gained from an examination of the pos i t ion of the chemical sh i f t of the CN-H) proton i n secondary amines. Table 4:3 shows the va r i a t i on of 6 ^ i n a number of organic secondary amines. For some of the amines, the K i s also given. The resul ts indicate that the c l chemical sh i f t of the (N-H) proton i s a sensi t ive measure of the basic character of the nitrogen atom. Conjugation of the CN-H) group to a r y l groups decreases the basic character of the nitrogen atom and at the same time lowers the chemical sh i f t of the CN-H) proton. A l k y l groups, however, suppress lone pa i r d e r e a l i z a t i o n , producing a stronger base and a high chemical sh i f t for the CN-H) proton. That the chemical sh i f t of the (N-H) proton i n P f i N 6 (NH) (NMe2)gCJo2 occurs at 6 = 6.15 p.p.m. indicates considerable lone pa i r d e r e a l i z a t i o n from the nitrogen atom into the phosphoni t r i l i c ir-systems. The (N-H) proton i s ac id i c i n nature, and th i s information, coupled with the fact that the mass spectrum of the molecule indicates that t h i s proton i s eas i ly l o s t , suggests a stronger in te rna l P-N bond i n P,NA(NH) (NMe9)„CJl9 than i n PJ\\LC£ Q . - 68 -6NH K a Compound p.p.m. o-NO_C,rLNHC,H,_ z o o o o 9.39 p-NO„C,HcNHC,Hc r 2 6 5 6 5 8.20 CH3CO-NH-CH2CH3 8.08 C,H r-NH-C,H,_ 6 5 6 5 5.42 1.62 x 10\"1 C^Hr-NH-CH„C,Hr 6 5 2 6 5 3.78 C 6 H 5 -NH-CH 3 3.34 1.41 x 10\" 5 C 6 H 5 - N H 2 3.32 2.34 x 10\" 5 C.HC-NH-CH„CHT 0 b Z o 3.03 7.59 x 10\"6 (C 6 H 5 -CH 2 ) 2 NH 1.78 C 6 H 5 -CH 2 -NH-CH 3 1.00 CH 3-NH-CH 3 1.85 x 1 0 \" 1 1 [(CH 3) 2CHCH 2] 2NH 0.79 1.23 x 1 0 \" 1 1 Reference 61 62 TABLE 4:3 . Relationship between (N-H) proton chemical sh i f t and base strength of some secondary amines. - 69 -e) Conformational Analysis As was mentioned above, the n.m.r. spectra can be interpreted i n terms of two molecular conformations, both of which are shown i n F i g . 4.6. The two conformations d i f f e r i n the r e l a t i ve posi t ions of the phosphorus and nitrogen atoms i n the ten-membered r i n g . In structure I , the 3 adjacent r ing phosphorus atoms are above the mean molecular plane and, i n structure I I , they are below i t . In both structures, i t i s the a x i a l dimethylamino-groups that are involved i n hydrogen bonding, and i t i s th i s intramolecular in te rac t ion which holds the two conformations r i g i d . Although structures I and II both account for the observed ''\"H n.m.r. spectrum, the former i s bel ieved, on the basis of general mechanistic considerations, to be the correct s t ruc-ture . Although the stereochemical course of replacement reactions i n phosphonitri les i s not f u l l y understood, the resul ts of chlorine exchange between chloride ion and chlorophosphonitri les suggest an 63 mechanism. Kine t ic studies on the dimethylamination of CPliCH^)^ 64 suggest a penta-coordinated phosphorus atom as an intermediate, with the attacking nucleophile and the leaving group being i n the a x i a l pos i t ions . In general, most replacement reactions at phosphorus i n mononuclear phosphorus compounds are accompanied by an inversion of conf igura t ion ,^ 5 and the same inversion i s believed to occur i n phospho-n i t r i l e s , i n so far as subs t i tu t ion i n these compounds i s effected by a bimolecular reac t ion . - 70 0 = CL O = NMe2 O =.N o =H I I F i g . 4..6. Idealized structures for the P . N . (NH) (NMe„)0C&„ molecule. ' 31 i Both structures are consistent with P and H n.m.r. data, but structure I i s believed to be the correct structure (see t e x t ) . - 71 -The mechanism of the complete dimethylamination of P^N^CJlg i s not known. The weakness of the in te rna l PN bonds has already been discussed, and i t i s thought that attack at a bridgehead phosphorus by a dimethylamine group leads to preferent ia l cleavage of the in te rna l P-N bond rather than the P-Cl bond. I f also the mechanism of sub-s t i t u t i o n involves nucleophi l ic attack at these phosphorus atoms, leading to a penta-coordinated intermediate and inversion of configurat ion, the stereochemistry of the f i n a l product can read i ly be established. As F i g . 4.7 shows, nuc leophi l ic attack at a bridgehead phosphorus atom, HNMer NMer F i g . 4 .7 . Suggested stereochemistry of nuc leophi l ic subst i tu t ion at a bridgehead phosphorus atom i n P,N_C£_. o / y with subsequent cleavage of the in te rna l P-N bond, leads to a product where the r ing phosphorus atoms are above the mean molecular plane. - 72 -Thus, the in terpre ta t ion of t h i s mechanism suggests that the conforma-t ion shown i n F i g . 4.61 i s the true conformation P-N, (NH)(NMe„)0C£„. O b 2 o 2. IV. Summary The extent of t h i s study of the chemistry of P^N yC£ g has been l imi t ed by the small amount of the compound that was avai lable and by the very nature of the compound i t s e l f . Unlike the monocyclic phosphoni t r i l i c ch lor ides , P^NyC£g i s unstable to subs t i tu t ion . Whether the attack of an approaching nucleophile occurs at the bridgehead or r ing phosphorus atoms i s not known. To the extent that lone pa i r d e r e a l i z a t i o n from the central nitrogen atom deactivates the bridge-head phosphorus atoms, then to th i s extent the r i ng phosphorus atoms w i l l be r e l a t i v e l y more e l e c t r o p h i l i c i n nature. To the present time, only one der ivat ive of the condensed r i ng compound has been i so la ted . The fact that i n th i s compound the condensed r i n g skeleton has been broken, v i s u a l l y emphasizes the chem-i c a l weakness of the P^N y skeleton. In the react ion with dimethylamine, one of the in te rna l PN bonds i s broken. That the r i ng cleavage process stops at th i s stage re f lec t s the strengthening of the two remaining in te rna l P-N bonds which occurs as a resu l t of the loss of the f i r s t . The fact that subs t i tu t ion at the two remaining bridgehead phosphorus atoms i n P^N^CNH) (NMepgCJ^ does not occur, even with an excess of dimethylamine, must be a resul t of the complete deact ivat ion of these phosphorus atoms. - 73 -In the case of r ing ac t iva t ing nucleophiles C e-g. f luor ine) further subs t i tu t ion seems to occur, with subsequent collapse of the phosphoni t r i l i c skeleton. In organic chemistry, subs t i tu t ion at a bridgehead carbon atom v i a an SN 2 mechanism i s s t e r i c a l l y impossible. In P^NjClg, the same s i tua t ion may also occur, and the subs t i tu t ion at bridgehead phosphorus atoms may be k i n e t i c a l l y disal lowed. Attack by f luor ide at the bridgehead atoms seems to break the P-N bonds rather than the P-C£ bonds. Further work may yet produce a f luorinated condensed r i ng molecule. The appearance of high molecular weight fragments i n the products of one of the f luo r ina t ion reactions attempted i n th i s work indicate that th i s may be poss ib le . The observed charac te r i s t ics of the condensed r i ng molecule are en t i r e ly unexpected, and further work i s necessary before a complete understanding of th i s compound i s achieved. At th i s po'int I would l i k e to thank the fol lowing people, who provided me with technical assistance i n the course of th i s work: Mr. P. Borda, for carrying out microanalyses; Mr. G. Gunn, for taking mass spectra; Mr. R. Burton and Miss P. Watson, for running various n.m.r. spectra; and Dr. L .D. H a l l and Mr. P. Steiner, for running 31 P decoupling n.m.r. experiments. - 74 -CHAPTER 5 THE MOLECULAR AND ELECTRONIC STRUCTURE OF NITRILOHEXAPHOSPHONITRILIC CHLORIDE The t r i c y c l i c condensed r i ng shape of P^NjCig was established by 31 P n.m.r. and v ib ra t iona l spectroscopy and C^ symmetry was suggested for the molecule. The recent X-ray c rys t a l structure analysis*^ of the molecule confirmed th i s assignment. In the course of t h i s study, Huckel molecular o r b i t a l ca lcula t ions were carr ied out on an idea l ized planar condensed r i ng P^N y framework. The resu l t s of these ca lcu la t ions , and the X-ray c rys t a l structure analys is , are discussed i n th i s chapter. I . Orb i t a l Interactions i n P-N-CJl^ 6—7—9 Throughout the course of t h i s work, comparisons have been made between the condensed r i ng P^N yC£ g molecule and the open r ing P^N^C£^ molecule, and th i s comparison again becomes useful i n considering the bonding i n P^N y C£ g . In phosphoni t r i l i c compounds i n general, i f the valencies of phosphorus and nitrogen are taken as f ive and three respec t ive ly , the a-framework of the c y c l i c -PX2=N- uni t can be 3 2 described using sp hybr id iza t ion at phosphorus and sp hybr id iza t ion at ni trogen. The formation of a Tr-bond, necessary to sa t i s fy the valencies of phosphorus and nitrogen, then involves the remaining 2p o r b i t a l on nitrogen and a 3d-orb i ta l on phosphorus. - 75 -In P^N^CJlg, there are two types of phosphorus atom. At the non-bridgehead (SPCJ^) type atoms, the orientat ions of the d-orb i ta l s are analogous to those that occur i n the monocyclic phosphoni t r i les . F i g . 5.1 shows the o r b i t a l overlap schemes for the d-orb i ta l s of a EPC&2 phosphorus atom with valence o rb i t a l s of ni trogen. As can be seen i n t h i s diagram, two types of TT-bonding can occur. Convention has i t that the out-of-plane system i s termed the IT system. The d . and d^ z o rb i t a l s on phosphorus are involved i n th i s system, overlap of these o rb i t a l s being ( in planar systems) with the nitrogen p z o r b i t a l . The in-plane TT-system i s termed the TT S system, and involves overlap of ( p r inc ipa l ly ) the d^2 y.2 a n < l d o r b i t a l s on phosphorus with the 2 sp lone pa i r hybrid on ni trogen. At the bridgehead atoms of P^^Cl^, TT- interactions of the phosphorus d-orb i ta l s with the r i ng nitrogen atoms are the same as the non-bridgehead atoms. However, the addi t ion of the central nitrogen atom allows a further TT - in teract ion between the p z o r b i t a l on that nitrogen and a phosphorus d - o r b i t a l . As F i g . 5.2 shows, the overlap of t h i s central nitrogen p z o r b i t a l with a phosphorus d -o rb i t a l i s highly se lec t ive . The in-plane d-orb i ta l s (d x 2 2 a n c l &Xy) do not interact with the out-of-plane p o r b i t a l . In the TT system, the d o r b i t a l i s orthogonal to the p o r b i t a l , and only the d o r b i t a l gives a non-z yz zero overlap with the central nitrogen (see F i g . 5.2) . Therefore, any TT-effect that t h i s atom has on the rest of the r i ng can only be made through in te rac t ion of th i s atom with the d^ z o r b i t a l on phosphorus. - 78 -Since the only formal difference between the „ system and the b o 12 P^NjCig system i s the inse r t ion of t h i s extra ni t rogen, the difference i n the TT-bonding should not be great. The TT s systems i n both molecules are formally the same and any differences i n molecular bonding can only occur i n the TT^ system. The extent of the differences depend on the r e l a t i ve importance of the two components of the TT system. Both i n the TT and TT systems, the two separate components are not required to be equally involved, and experimentally th i s i s found to be so. Overlap 67 68 ca lcu la t ions , ' have shown that i n the TT system, d -p overlap i s a xz z 2 the more important component, and i n the TT s system, d^,2 y 2 ~ s P overlap i s the more important. I I . Symmetry-based M.O. Calculat ions on P^N^Cflg Because the only difference between the open r i ng P^N^ system and the condensed r i ng P^N^ system i s the addit ion of the central nitrogen i n the l a t t e r , and because TT-bonding from th i s nitrogen can only occur into one phosphorus d - o r b i t a l , a simple Hiickel M.O. ca l cu la t ion was carr ied out on the P^N^ and P^N y phosphoni t r i l i c skeletons, ca lcula t ions which e f fec t ive ly measured the difference i n bond orders and charge densi t ies between the two systems. For both molecules, a planar structure was assumed, where a l l r i ng angles were taken as being 120°. Overlap was only considered between the phosphorus d^ z o r b i t a l and the nitrogen p o r b i t a l , and a l l overlaps were taken as being equivalent. For the system, the energy levels that are formed are s imi l a r to - 79 -those found for the phenalenyl system.^^'^^ For any phosphoni t r i l i c r ing s i z e , the secular equations are most conveniently set up u t i l i z i n g the ro ta t iona l symmetry of the molecule. For in teract ions involv ing the d^ z o r b i t a l at phosphorus and the p z o r b i t a l at ni t rogen, the secular determinant for an open r i ng system becomes: a D -E 26cos— P n 26cos— aX I-E n N = 0 (5.1) where £, the r ing quantum number, takes on the values 0, ±1 , ±2 . . . ± ( n - l ) / 2 (n odd) and 0, ±1 , ±2 . . . n/2 (n even), and n i s the number of repeating PN uni t i n the phosphoni t r i l i c r i n g . In th i s ca l cu l a t i on , the two Coulomb parameters, dp and a^, are expressed i n terms of an e lec t ronegat iv i ty difference p, such that = + p 3 ; p i s here a r b i t r a r i l y set at p = 2. For the c y c l i c P^N^ system, the energy levels deduced from the secular determinant (equation 5.1) are: E = a ± \\/cos2^- + %L (5.2) where a i s the average Coulomb parameter (a = (otp + a^)/2) . These energy levels are shown i n F i g . 5.3a. In the PgN y condensed r ing system, the PN skeleton has 3-fold ro ta t iona l symmetry. In the C_ point group, the central nitrogen p - 80 -- 81 -o r b i t a l has symmetry, and therefore can only combine with the three A^ levels of the P^N^ system. Using the o r b i t a l nomenclature indicated i n F i g . 5.4, the three A^ intermediate o rb i t a l s with which the central • i *12N-^^|S|(')2 cb I , I cb y N r * N b *8 7 6^ F i g . 5.4. Orb i t a l nomenclature used i n symmetry based calcula t ions on P.N_C£_. o / y nitrogen p z o r b i t a l Of^) can combine are: (a) = ^ + (j>5 + n) (C) ^3 = /6 C(j>2 + ^6 + * 1 0 + *4 + *8 + ( J )12 ) (5.3) Mixing of these four A^ levels produces a new 4 x 4 A^ secular determ-inant. The so lu t ion of th i s determinant produces the energy levels shown i n F i g . 5.3b, the degenerate E levels and the non-bonding - 82 -Bond A A P N B A P N B B Difference between P .B .O . ' s of levels i n P ,N, and P.N_ systems 6 6 6 7 +0.095 +0.030 +0.444 Atom P A N A P B N B Difference between charge densi t ies i n A^ levels of P..N, and P.N_ systems 6 6 6 7 +0.027 +0.010 +0.093 -0.426 TABLE 5:1. Charge density and p a r t i a l bond order (P.B.O.) differences between P^N^ and PgN y systems as derived from symmetry based ca lcu la t ions . A pos i t ive difference indicates P^N y > P N . 6 6 F i g . 5.5. Diagram showing the nomenclature used i n describing the various types of bonds and atoms i n the P,N 7C£ Q molecule. - 83 -l eve l being unchanged from those found i n the P^N^ system. Within th i s TT-system, the changes i n charge densi t ies on the r i ng atoms, and the changes i n p a r t i a l bond orders on moving from the open r i ng P^N^ system to the condensed r i ng P^N^ system were ca lcula ted , and the resul ts are shown i n Table 5:1. The nomenclature used for d i f f e ren t i a t ing between the various types of atoms and bonds i n PgN^ structure i s i l l u s t r a t e d i n F i g . 5.5. The important conclusion which these resul ts suggest i s a strengthening of the P.N bond with respect to the P n N. bond (see A A D A F i g . 5 .5) . A considerable ir-bond order i s also indicated for the PgNg bond, suggesting that th i s bond must be quite short. As w i l l be discussed i n the next sect ion, the observed bond lengths i n P^N^CJig are i n complete disagreement with the simple argument described above. More deta i led calcula t ions were carr ied out, as described i n the f i n a l section of t h i s chapter, and they lead to a better cor re la t ion of theory and experiment. I I I . The Molecular Structure of P.N_C£f, — o—/—-y 66 The molecular structure of t h i s compound i s shown i n F i g . 5.6, with the relevant s t ruc tura l data given i n Table 5:2. The molecule i s not planar, the mean planes of the three six-membered rings (which are themselves not planar) s loping down from the central nitrogen atom. The symmetry of the molecule i s s l i g h t l y d is tor ted i n the c r y s t a l , as one of the P-C£ bonds i s almost coaxial with a two-fold axis wi th in F i g . 5.6. The molecular structure of n i t r i lohexaphosphoni t r i l i c ch lor ide . Structural data for the molecule i s given i n Table 5.2. - 85 -BOND ANGLES degrees BOND ANGLES degrees BOND LENGTHS (A) C£1-P1-C£2 101.5 P5-N4-P6 123.4 P l - N l 1.569(6) N1-P1-N3 116.5 P6-N3-P1 124.8 P1-N3 1.579(6) C&3-P2-N2 104.8 P2-N2-P4 119.0 Pl-C£l 1.988(3) N1-P2-N2 110.3 P2-N2-P6 121.0 Pl-C£2 1.970(3) N1-P2-N7 111.8 P4-N2-P6 119.8 P2-N1 1.549(6) N2-P2-N7 110.2 P2-N2 1.720(6) C£4-P3-CJt5 101.3 P2-N7 1.568(6) N7-P3-N6 118.1 P2-C£3 2.002(3) C£6-P4-N2 104.5 P3-N7 1.557(6) N6-P4-N2 110.1 P3-N6 1.569(6) N6-P4-N5 113.8 P3-C&4 1.977(3) N2-P4-N5 111.4 P3-C£5 1.990(3) C£7-P5-C£8 101.9 N5-P5-N4 116.1 P4-N6 1.578(6) C£9-P6-N2 102.8 P4-N2 1.733(6) N4-P6-N2 111.4 P4-N5 1.539(6) N4-P6-N3 111.8 N2-P6-N3 110.8 P4-C&6 2.002(3) P1-N1-P2 125.8 P5-N5 1.574(6) P2-N7-P3 127.3 P5-N4 1.580(6) P3-N6-P4 123.2 P5-C£7 1.970(6) P4-N5-P5 178.6 P5-C&8 1.981(3) P6-N4 1.560(6) P6-N2 1.715(5) P6-N3 1.553(6) P6-C£9 2.008 (3) TABLE 5.2 . Bond angles and bond lengths i n n i t ro lohexaphosphoni t r i l ic chloride (see F i g . 5 .6) . Standard deviations for bond lengths are given i n parentheses. - 86 -the unit c e l l , thus causing packing forces to twis t the molecule out of ideal symmetry. Possibly the most important fact that i s seen i n the c rys t a l structure i s the length of the in te rna l PN bond, which i s considerably longer than the r i n g PN bonds. As i l l u s t r a t e d i n F i g . 5.5, the r i ng PN bonds f a l l into two classes , the P^N^ type and the PgNg type. The P-N bond lengths i n monocyclic phosphonitr i les range between 1.50-1.60-A. In the phosphoramidate ion , (PO^NHg) (where there i s no p o s s i b i l i t y of lone pa i r d e r e a l i z a t i o n re inforc ing the a-bond), the o 71 P-N s ingle bond length i s 1.77 A. Although the shortness of the phosphoni t r i l i c bond i s ind ica t ive of some Tr-bond character, the bond shortening i s also contributed to by a a-hybr id iza t ion effect . The PN 3 s ingle bond i n the (PO^NH^) ion i s based on sp a-hybr id iza t ion at both centres, the bond angles being close to the tetrahedral value. In phosphoni t r i l i c compounds, the r ing angle at phosphorus i s usual ly close to 120°, and at nitrogen varies between 120° and 150°. Changing the bond angle at e i ther atom has the effect of changing the a-hybr id iza t ion at that atom. The important point to consider i s that i n phosphoni t r i les , up to 40% of the shortening of the phosphoni t r i l i c P-N bond with respect to a P-N s ingle bond can be a t t r ibuted to such a-effects , and any discussion of the TT-bonding i n a pa r t i cu l a r molecule can only be made after a-effects have been considered. A l l the r i n g bond lengths i n P^NjCHg are given i n Table 5:3. The actual TT -cont rac t ions (with reference to the s ingle bond length of - 87 -1.77 A) are also given. Based on a study made of the C-C bond length 72 73 i n various organic compounds, Craig and Paddock have calculated O adjusted values of the atomic r a d i i of phosphorus (1.069 A) and O nitrogen (0.701 A) as the a-hybr id iza t ion of the two atoms i s changed. The adjustments depended on the r i ng angle at ei ther atom, the s ize of the adjustment being that f rac t ion of the bond length by which a C-C 72 bond would change given a s imi l a r angle change. Using the resul ts of these ca lcu la t ions , the a-corrected TT-contractiohs for the r i ng PN bonds of P^NyC£g have been estimated, and these values are also given i n Table 5:3. Both the actual and a-corrected TV-contractions show the same trend, namely that the P^N^ type bonds have less u-character than the PgNA bonds. A smaller ir-bond order i s therefore indicated i n the former. This experimental observation i s the reverse of that predicted by the simple theore t ica l model described i n the previous sect ion. As has been mentioned already, the PgNg bond i s long (1.723 X), and when corrected for a -hybr id iza t ion , i s even longer (1.748). The length of th i s bond suggests only a l i t t l e Tr-character (ir-contraction =• 0.02 A) for the bond, despite the fact that the sum of the bond angles at the central nitrogen i s 359 .8° . In tetrameric^ 4 and hexameric^ phosphoni t r i l i c dimethylamides, [PN(NMe 2) 2]4 t n e average exocyclic PN bond lengths are 1.68 A and 1.67 X respec t ive ly . Thus the exocycl ic PN bond i n these molecules i s stronger than the PgNg bond i n P^N^CJlg, despite the fact that the sum of the bond angles at the exocyclic nitrogen i n the P A N A Bond P-N. Bond D A Bond length 0 TT-contraction A Bond length o Tr-contraction A 0 A Observed a-adjusted A Observed a-adjusted 1.569 0.20 0.14 1.578 0.19 0.15 1.574 0.20 0.13 1.539 0.23 0.19 1.580 0.19 0.13 1.560 0.21 0.18 1.579 0.19 0.13 1.553 0.22 0.18 1.569 0.19 0.14 1.5.49 0.22 0.19 1.557 0.21 0.14 1.568 0.20 0.15 1.571 0.195 0.135 1.558 0.21 0.17 TABLE 5:3. Ring bond lengths and Tr-contractions i n P ^ N ^ C ^ . Both observed and a-adjusted TT-contractions are tabulated (Tr-contractions are estimated with reference to P-N single bond length of 1.77 A). The las t f igure i n each column i s an average value. - 89 -dimethylamid.es i s less than 359 .8° , ind ica t ing less than complete lone pa i r d e r e a l i z a t i o n . The exocycl ic bonds i n P^NyC£g also show an in teres t ing anomaly. The P-C£ bonds (2.00 A) are s l i g h t l y longer than the =PC£2 bonds (1.98 A). Considering the a-hybr id iza t ion at both the P. and P D phosphorus atoms, the HPC£2 bonds have more p-character than the P-C£ bonds. Neglecting, for a moment, the Tr-character i n these bonds, the increased s-character i n the P-C£ bonds should make them shorter than the =PC&2 bonds. That t h i s i s not the case indicates that , i n as much as exocycl ic T T -bonding does occur, i t i s much stronger at the P^ atoms than the P_ atoms. The d i spa r i t y between the simple theore t ica l model of the bonding and the observed s t ruc tura l features i s therefore s ign i f i can t and con-s i s t en t . The reversal of bond order trends between theory and exper i -ment shows that the simple theory of the bonding i s inadequate i n describing the s t ruc tura l features of P^ijCig. Because of t h i s anomaly, further, more deta i led ca lcu la t ions , were carr ied out, and are now described. IV. Detai led Hiickel M.O. Calculat ions on P^N-CJL. ——o—/—-y These molecular o r b i t a l ca lcula t ions were carr ied out using a l l the phosphorus 3d-orbi tals (except the 3 d z 2 ) and the nitrogen p z and Py o rb i t a l s (see F igs . 5.1 and 5.2) , the molecule being assumed to be planar. No e x p l i c i t consideration was made of the chlorine - 90 -ligands and only the P^N y nucleus was examined. The usual Hiickel approximations were appl ied, i . e . non-nearest neighbour overlap was ignored and the motion of a l l the electrons was completely uncorrelated. Suitable Coulomb parameters were chosen by reference to recent 76 calcula t ions on the monocyclic phosphoni t r i l i c f luor ides . For a l l the nitrogen atoms, values of a(Np z) = a(Np^) = -11 e.v. were used, and for the phosphorus atoms, values of a(Pd^ z ) = a(Pd^2 yi) = -6 e - v -were taken. Suitable Coulomb parameters for the phosphorus d^ z and d o rb i t a l s were believed to be important i n the ca lcu la t ions , since the i r value was thought to depend on whether the pa r t i cu la r phosphorus atom was a bridgehead or non-bridgehead atom. The reason for th i s dependence of a on the type of phosphorus atom i s thought to be a resu l t of an exchange in te rac t ion between the chlor ine ligands and the various phosphorus 3d-orb i ta l s . At the bridgehead (P^) atoms, the chlorine l igand l i e s close to the z-axis and therefore does not des tab i l i ze any of the d-orb i ta l s under consideration. At the non-bridgehead CpA) atoms, the two (P-Cjl) bonds l i e close to the lobes of, and therefore d e s t a b i l i z e , the d o r b i t a l . Because the lobes of the ' yz d o r b i t a l are also close to the (P-CJt) bond axes at the P. atom, xy • A the effect of exchange interact ions on t h i s o r b i t a l were also con-sidered. The effect of exchange des t ab i l i z a t i on on the above-mentioned o rb i t a l s was introduced into the ca lcula t ions by varying the value of the Coulomb parameter. - 91 -A l l resonance in tegra ls were essen t i a l ly for pTr-d.Tr in te rac t ions , and the formula used for the ca lcu la t ion of these integrals was: / V N 1 J C-6) x C - l l ) This allowed the g's to depend on the a ' s , as they must, and i s broadly consistent with the conclusions obtained from a recent study of the 76 i n e o v e naD i n t e g r a l s , t i n terms o f a p r i m i t i v e pTr-diT o v e r l a p i n t e g r a l o f S(pTr-dTr) = 0.3636, monocyclic CPNF2)n se r ies . The overlap in tegra l s , . , were expressed and a p r imi t ive pa-da overlap in tegra l of SCpa-da) = 0.1277. An out l ine of the ca lcu la t ion of the various S ^ ' s i s given i n Appendix I . The charge densi t ies and bond orders derived from the so lu t ion of the 37 x 37 secular determinant are given i n Tables 5:4 and 5:5. The effect of exchange interact ions on the s t a b i l i t y of various phosphorus o rb i t a l s i s to produce a s ign i f i can t change i n the trends i n bond orders. Des tab i l i z ing the d^z Cand the d ) o r b i t a l on the atoms (by se t t ing aCd ) = a(d ) = -2 e .v . ) has the effect of decreasing the Tr-bond order y z ' ^ xy a of the P^N^ bond with respect to the PgN^ bond, which i s consistent with the trend observed i n the actual molecule. As the values of the p a r t i a l bond orders i n Table 5:5 show, the concept of o r b i t a l s e l e c t i v i t y i s v a l i d , and some o rb i t a l s on phosphorus (d' x z and d^2 yi) a r e more important i n TT-bonding than others (.^yZ a n ( l d ) . However, i n so far as the d o r b i t a l i s used i n TT-bonding, i t xy J ' yz & ' - 92 -Atom Orb i t a l Charge Density A a B b V d xz d yz d o 2 x -y^ d xy 0.07480 0.02696 0.04818 0.03455 0.07480 0.002945 0.04818 0.003796 E 0.18449 0.12973 P B . d xz d yz d o 2 x -y d xy 0.7479 0.07155 0.06977 0.01018 0.07607 0.07193 0.07126 0.01043 Z 0.22629 0.22969 N A Pz p y 1.899 1.919 1.910 1.933 N B Pz 1.860 1.860 TABLE 5:4. Calculated charge densi t ies on various atoms i n P^NyC£ using a l l phosphorus 3d-orbitals (except d 2) • (a) Coulomb parameters for P.(d ) = P.(d ) = -6.0 e.v. J r A y z ' A xy (b) Coulomb parameters for p A ( d y Z ) = P A ^ X y ^ = ~ 2 *^ e , V -(For other Coulomb parameters, see Text.) - 93 -Bond Orb i t a l Bond Order Overlap A a B b A A d -p xz r z 0.2002 0.2002 d -p yz *z d ? 2 - P x -y r y d -p xy *y 0.07021 0.1262 0.08986 0.01348 0.1262 0.01733 Z 0.4865 0.3572 P N B A d -p xz rz 0.2002 0.2024 d -p yz *z d 2 2 _ P x -y r y d -p xy *y 0.06682 0.1851 0.02617 0.06757 0.1877 0.02659 0.4783 0.4843 P N 1 1 d -p yz *z 0.2553 0.2551 TABLE 5:5. Calculated bond orders for the various types of bonds i n P 6 N 7 C £ g . 6.0 e.v. 2.0 e.v. (c) The d e f i n i t i o n of p a r t i a l bond order differed from that used i n calcula t ions on monocyclic phosphoni t r i les . For the purposes of t h i s study, the p a r t i a l bond order, (P.B.O.) , between two atoms r and s, was defined as: ' (P.B.O.) = 2 £ c. c. 6 ^ r , s . i r i s rs ' l = a l l occupied o rb i t a l s (a) Coulomb parameters for P. (d ) = P. (d ) = -^ J * A ^ y z ' A ^ xy ' (b) Coulomb parameters for P^Cd^) = ^A^xy-* = ~ (For other Coulomb parameters, see Text.) - 94 -i s thought that the contr ibut ion of t h i s o r b i t a l i s greater at the P D a than at the atom, f i r s t l y because of exchange des t ab i l i za t i on at the P^ atom, and secondly because the d^ z o r b i t a l i s s t a b i l i z e d on the Pg atom by the influence of the central nitrogen atom. Another in te res t ing point concerns exocyclic TT-bonding at the Pg atoms. At these atoms, exocycl ic TT-bonding can (by symmetry) only occur between the chlorine l igand and the d and d o rb i t a l s on • xz yz phosphorus. The d^ z o r b i t a l i s strongly involved i n forming a r ing TT-system, and therefore i s less involved i n the exocyclic TT-bonds. I f the d ^ o r b i t a l i s also strongly involved i n r ing TT-bonding at the Pg atom, then i t also w i l l only form.weak TT-bonds to ch lor ine . That\" the observed values of the bond lengths of the P-C£ bonds are greater than those of the PC&2 bonds suggests that there i s l i t t l e or no TT-bonding i n the exocyclic P-C£ bonds and adds credence to the above argument involv ing a strong in te rac t ion of the P g d^ z o r b i t a l i n the r ing Tr-systems. The P^N,, bond order i s smaller than those of the P.N. and P„N. B B A A B A bonds, but i s s t i l l quite substant ial and varies l i t t l e with changes i n the a parameters of the d-orb i ta l s on the P^ atom. The unusually o long PgNg bond length (1.723 A) cannot be properly explained by these ca lcu la t ions , and the reason for i t may wel l be s t e r i c rather than e lec t ron ic . The base strength measurements, described i n Chapter 3 are also d i f f i c u l t to explain i n terms of e lec t ronic fac tors . The base strength - 95 -of any pa r t i cu la r phosphoni t r i l i c compound i s contributed to by two e lec t ronic effects . The a-hybr id iza t ion of the lone pa i r hybrid on nitrogen i s a function of the r i n g angle at ni trogen, so that the smaller the r ing angle, the greater i s the s-character of the nitrogen lone pa i r hybrid . An increase i n the s-character of t h i s lone pa i r o r b i t a l decreases the basic character of the lone p a i r . Thus, considering a-effects alone, a decrease i n the r i n g angle at a nitrogen atom i s accompanied by a decrease i n the base strength of the lone pa i r on that nitrogen atom. TT-effects also contribute to the basic character of the lone pa i r electrons. Because of t h i s , an increasing r i ng angle at ni trogen, which i s symptomatic of greater lone pa i r d e r e a l i z a t i o n into the r i ng TT^ system, i s expected to be accompanied by a decrease i n the base strength of the lone p a i r . Thus, a- and Tr-effects produce trends i n the base strengths of phosphonitri les which are the opposite of each other. Whilst a-effects increase base strength with increase i n ^PNP, rr-effects decrease them, and the resultant base strength of any pa r t i cu la r molecule w i l l be a balance of these two factors . As had already been mentioned, any formal difference i n the ir-bonding of the PgNg and P^N_, system can only ar ise i n the system. Since lone-pair d e r e a l i z a t i o n occurs through the TT s system, the effect of changes i n the TT system on the base strength can only be secondary EL ( i . e . through a TT /TT i n t e rac t ion) . The differences i n basic properties - 96 -of P^N_C£_ and P.I\\LC,SL„ are therefore most l i k e l y to be caused by a 6 7 9 6 6 12 a-effect rather than a if.-effect. Since the r i ng angle at nitrogen i n P^NyCilg i s small (125 .5° ) , the s-character of the lone pa i r hybrid i s l i k e l y to be greater than that found i n P^N^CJi^j and thus the observed base strength of P^NyC£g (which i s less than that of ^^^^12^ i s explainable. However, anything more than th i s empirical conclusion cannot be made at present, since there exists, no quanti tat ive theory for separating the a- and u-factors which control the base strength of phosphoni t r i les . I would l i k e to thank Dr. K.A.R. M i t c h e l l and Mr. R. Bruce for t he i r invaluable advice and assistance i n performing the Hiickel M.O. ca lcu la t ions . - 97 -APPENDIX I Evaluation of Overlap Integrals for P^N^C&g In the Huckel M.O. ca lcu la t ions , a l l overlap integrals between phosphorus and nitrogen were expressed i n terms of a l inear combination of a p r imi t ive piT-dTr overlap, where S(pTT-dTr) = 0.3636 and a p r imi t ive pa-da overlap, where S(pa-da) = 0.1277. These values were based on o r b i t a l exponents of 1.95 for the 2 p i T o r b i t a l at ni trogen, 1.40 for the 3diT o r b i t a l at phosphorus and a P-N bond length of 1.52 K ( this value of the P-N bond length has been used for some studies on the phosphoni t r i l i c f luor ides , but the s l i g h t l y longer phosphoni t r i l i c bonds i n P^NyC£g were not considered to be important from the point of view of t h i s study). In the out-of-plane TT-system, overlap in tegrals were obtained quite simply using the fol lowing formulae: At the P. atom, S(d -p ) v xz r z ' S(d -p ) k yz r z ' ± S(pTT-dTr) cos30° - S (pTT-dlT) cos60° At the P D atom, D for overlap with N . , S(d -p ) v xz r z S(d -p ) ^ yz *z S(d -p ) yz *Z ± S(pTT-dTT) cos30° + S(pTT-dTT) cos60° for overlap with N + S(pTT-dTr) In the in-plane ir-system, the evaluation of overlap in tegrals required a ro ta t ion of coordinates using the scheme shown below. In - 98 -th i s diagram (Fig . A . l ) , the standard coordinate scheme at the Pg and F i g . A . l . Rotated coordinate scheme used i n the evaluation of overlap in tegrals at the Pg atoms i n the TT system. atoms (Fig. 5.7) i s rotated into a new system such that: /3 At N A , At Pg, 2 Z N + 2 X N X = /3 1 2 Zp ~ 2 X P 1 /I 2 Z P + 2XP Therefore, the atomic o rb i t a l s i n the new coordinate system have the form: - 99 -2 P x = R 2 p x X 3 d z 2 = f R 3 d C Z ^ ) 2 P y = R 2 p x Y 3 d x y = R 3 d XZ 2 P Z = R 2 p x Z 3 d Y Z = R 3 d Y Z 3 d x 2 - Y 2 = R 3 d ^ 2 - Y 2 ) 3 d X Y ' = R 3d C X Y ^ Using th i s transformation and a s imi l a r one at the atom, the overlap in tegra ls between phosphorus and nitrogen were calculated as being: 3 /3 At the P A atom, SCd x 2 _ Y 2 - P y ) = j SCpTT-d-rr) - - SCpa-da) (sign S (d X Y -p y ) = * j S(p7r-dTT) + | S(pa-da) (sign ±) 3 J% At the Pg' atom, S ( d x 2 _ Y 2 - p Y ) = - j S(pTr-dTr) - g- S (pa-da) (sign /3 3 •' S (d X Y -p y ) = j SCp^-du) - I S (pa-da) (sign ±) The ca lcu la t ion of these overlaps was a f a i r l y simple algebraic process, e .g. - 100 -At N A , 2pY = - I 2p z + f 2p x A t PB> 3 d x 2 _ Y 2 = \\ 3 d z 2 _ x 2 - f 3 d x z /3 1 i t i s easy to show that 3d 2 2 = o\" 3d 2 - y 3d 2 2 and therefore z —x <6 z ^ x ~y overlap of 2p at N. and 3d 2 2 at P R becomes: 7 2P Y 3 dx2-Y? d T = / 2p + \" f 2p ) ( j 3d 2 - I 3d 2 2 - ^ 3d ) dx ' J 2 r z 2 v 4 z z 4 x - y z 2 x = - ^ (pa-da) - j Cp-iT-dTT) . Other S . . ' s were obtained s i m i l a r l y . - 101 -APPENDIX II Detai ls of the instruments and techniques used i n recording the various types of spectra were not discussed i n the main text of th i s thes is . A summary of pertinent de ta i l s w i l l now be given. i ) Infra-red spectra : - These were a l l recorded on a Perk in-Elmer 457 grat ing spectrophotometer, and ca l ibra ted against a standard polystyrene spectrum. Nujol and hexachlorobutadiene mull spectra were taken using cesium iodide p la tes . Solut ion and gas phase spectra were recorded using potassium bromide windows. i i ) Raman spectra : - These were a l l recorded on a Cary 81 Laser Raman spectrophotometer. No depolar izat ion measurements were taken. i i i ) U l t r a - v i o l e t spectra:- These were a l l recorded on a Cary 14 spectrophotometer, using quartz c e l l s and a s ingle beam technique. iv ) Mass spectra:- These were a l l recorded at 70 e.v. on a A . E . I , type M.S.9 mass spectrometer, samples being admitted through conventional i n l e t systems. v) N.m.r. spectra : - *H n.m.r. spectra were run at 60 MHz on a Varian T-60 n.m.r. spectrometer and at 100 MHz on a Varian HA-100 31 n.m.r. spectrometer. P n.m.r. spectra were recorded at 40.1 MHz - 102 -31 1 on a Varian HA-100 n.m.r. spectrometer. P- H decoupling experiments were carr ied out using a recent ly described double resonance technique. a See L .D. H a l l and R. Burton; Can. J . Chem., 48, 59 (1970). - 103 -REFERENCES (1) A.W. Laubengayer, P .C. Moews, and R .F . Porter , J . Amer. Chem. S o c , 83, 1337 (1961). (2) D.C. Carpenter and L.O. Brockway, J . Amer. Chem. S o c , 58_, 1270 (1936). (3a) K.A. Andrianov and G.Y. Rumba, Izv. Akad. Nauk. SSSR Otd. Khim. Nauk., 1313 (1962). (3b) K.A. Andrianov, B.A. Ismailov, A.M. Kononov and G.V. Kotrelev, J . Organometal. Chem., 3, 129 (1965). (4) E.A. Semenova, D.Y. Zhinkin , and K.A. Andrianov, Izv. Akad. Nauk. SSSR Otd. Khim. Nauk., 2036 (1962). (5) M. Takimoto and T. Funakawa, Kogyo Kagaku Zasshi , 66_, 797 and 804 (1963). (6) H.N. Stokes, Amer. Chem. J . , 1_9, 782 Q.897) . (7) D.P. Craig and N . L . Paddock, Nature, 181, 1052 (1958). (8) D.P. Cra ig , J . Chem. Soc. , 997 (1959). (9) H . J . Krause, Z. Elektrochemie, 59_, 1004 (1955). (10) C .E . Brion and N . L . Paddock, J . Chem. Soc. (A), 388 (1968). (11) A . C . Chapman, D.H. Paine, H.T. Searle, D..R. Smith and R.F .M. White, J . Chem. S o c , 1768 (1961). (12) K. John, T. Moeller and L . F . Audrieth, J . Amer. Chem. S o c , 82^ 5616 (1960). (13) K. John, T. Moeller and L . F . Audrieth, J . Amer. Chem. S o c , 83_, 2608 (1961). (14) A.M. de Ficquelmont, Ann. Chim., 1939 ( i i ) , 1_2, 214. (15) T. Moeller and F. Tsang, Chem. and Ind . , 361 (1962). (16) G. Tesi and P . J . S lo ta , Proc. Chem. S o c , 404 (1960). (17) W.L. Lehn, J . Amer. Chem. S o c , 86, 305 (1964). - 104 -(18) E .A .V . Ebsworth and H . J . Emeleus, J . Chem. S o c , 2150 (1958). (19) T.A. George and M.F. Lappert, J . Chem. Soc. (A), 992 (1969). (20) T. Chivers , R.T. Oakley and N . L . Paddock, J . Chem. S o c , i n press. (21) M. Becke-Goehring and E. Fluck, Angew Chem., 74_, 382 (1962). (22) M. Becke-Goehring and W. Lehr, Z. Anorg. Allgem. Chem., 327, 128 (1964). (23) H.R. A l l c o c k , R . L . Kugel and K . J . Valen, Inorg. Chem., 5_, 1709 (1966). (24) E. G i g l i o , F. Pompa and A. Ripamonti, J . Poly. S c i . , 59, 293 (1962). (25) G. A l l e n , C . J . Lewis and S.M. Todd, Polymer, 11, 31 (1970). (26) L . G . Lund, N . L . Paddock, J . E . Proctor and H.T. Searle, J . Chem. S o c , 2542 (1960). (27) C .E . Brion and N . L . Paddock, J . Chem. Soc. (A), 392 (1968). (28) L.W. Daasch, J . Amer. Chem. S o c , 7(3, 3403 (1954). (29) A .C . Chapman and N . L . Paddock, J . Chem. S o c , 635 (1962). (30) S. Cal i fano, J . Inorg. and Nuclear Chem., 24_, 483 (1962). (31) I .C . Hisatsune, .Spectrochim. Acta , 21_, 1899 (1965). (32) J . Emsley, J . Chem. Soc. (A), 109 (1970). (33) I .C . Hisatsune, Spectrochim. Acta , 25A, 301 (1969). (34) R. Stahlberg and E. Steger, Spectrochim. Acta , 23A, 2057 (1967). (35) R.A. Chittenden and L . C . Thomas, Spectrochim. Acta , 21_, 861 (1965). (36) R.A. Chittenden and L . C . Thomas, Spectrochim. Acta , 22_, 1449 (1966). (37) R. Stahlberg and E. Steger, Spectrochim. Acta , 23A, 2005 (1967). (38) N . L . Paddock and D.R. Smith, unpublished r e su l t s . (39) F. Pompa and A. Ripamonti, Ricerca S c i . , 29_, 1516 (1959). (40) R. Hazekamp, T. Migchelsen and A. Vos, Acta C r y s t . , L5, 539 (1962). - 105 -(41) A.W. Schleuter and R.A. Jacobsen, J . Amer. Chem. S o c , 88, 2051 (1965). (42) G.E. Coxon and D.B. Sowerby, J . Chem. Soc. (A), 3012 (1969). (43) R. Ratz and Ch. Grundman, J . Inorg. Nucl . Chem., 16^ , 60 (I960). (44) Von F. See l , K. Rudolph and R. Budenz, Z. Anorg. Allgem. Chemie, 341, 196 (1965). (45) G. Herzberg, \"Infra-red and Raman Spectra\", Van Nostrand C o . , Inc . , 1962, p. 391. (46) G. Herzberg, \"Elec t ronic Spectra of Polyatomic Molecules\", Van Nostrand Co . , Inc . , 1966, Table 68, p. 618. (47) J . Emsley, Ph.D. Thesis, London, 1964. (48) B.W. Fitzsimmons and R.A. Shaw, J . Chem. S o c , 1735 (1964). (49) F. R a l l o , Ricerca S c i . , 35, 1134 (1965). (50) A . I . Vogel , \"A Textbook of P r a c t i c a l Organic Chemistry\", 2nd e d i t . , 1948, Longmans, London, p . .798. (51) S.K. Ray and R.A. Shaw, Chem. and Ind. , 53 (1959); N . L . Paddock, unpublished r e su l t s . (52) A. Cornu and R. Massot, \" F i r s t Supplement to Compilation of Mass Spectral Data\", Heyden and Son L t d . , London, 1967, p. 16C. (53) R.W. Harr ison, unpublished r e su l t s . (54) N . L . Paddock, Quart. Rev., 113, 168 (1964). (55) F. Heatley and S.M. Todd, J . Chem. Soc. (A), 1152 (1966). (56) R. Keat, S.K. Ray and R.A. Shaw, J . Chem. S o c , 7193 (1965). (57) R. Keat and R.A. Shaw, J . Chem. S o c , 4802 (1965). (58) G. A l l e n , D . J . O l d f i e l d , N . L . Paddock, F. R a l l o , j ; Serreqi and S.M. Todd, Chem. and Ind. , 1032 (1965). (59) R.K. Har r i s , Can. J . Chem., 42, 2275 (1964). (60) R.K. Har r i s , Inorg. Chem., .5, 701 (1966). - 106 -(61) \"Sadtler Standard Spectra\", Sadtler Research Laboratories, i . r . spectrum #68. (62) \"Handbook of Physics and Chemistry\", Chemical Rubber Co . , 49th e d i t . , 1969, p. D87. (63) D.B. Sowerby, J. Chem. S o c , 1396 (1965). (64) B. Capon, K. H i l l s and R.A. Shaw, 4059 (1965). (65) R.F . Hudson, Adv. Inorg. Chem. Radiochem., 5_, 347 (1964). (66) W. Harrison, unpublished re su l t s . (67) D.P. Craig and N . L . Paddock, J. Chem. S o c , 4118 (1962). (68) K.A.R. M i t c h e l l , J . Chem. Soc. (A), 2683 (1968). (69) D.H. Reid, Quart. Rev., 19, 274 (1965). (70) A. Strei tweiser and C.A. Coulson, \"A Dict ionary of TT-electron Calcu la t ions\" , W.H. Freeman, San Francisco, 1965, p. 206. (71) E. Hobbs, D.E.C. Corbridge and B. R a i s t r i c k , Acta C r y s t . , 6, 621 (1953); D.W.J. Cruickshank, Acta C r y s t . , 17_, 671 11964). (72) M.J. Dewar and H.N. Schmeising, Tetrahedron, 11_, 96 (1960). (73) D.P. Craig and N . L . Paddock, unpublished r e su l t s . (74) G.J. Bul len , J. Chem. S o c , 3193 (1962). (75) A.J. Wagner and A. Vos, Acta C r y s t . , B24, 1423 (1968). (76) G.R. Branton, C .E . Br ion, D.C. Frost , K.A.R. M i t c h e l l and N . L . Paddock, J. Chem. S o c (A), 151 (1970). - 107 -"@en ; edm:hasType "Thesis/Dissertation"@en ; edm:isShownAt "10.14288/1.0059299"@en ; dcterms:language "eng"@en ; ns0:degreeDiscipline "Chemistry"@en ; edm:provider "Vancouver : University of British Columbia Library"@en ; dcterms:publisher "University of British Columbia"@en ; dcterms:rights "For non-commercial purposes only, such as research, private study and education. Additional conditions apply, see Terms of Use https://open.library.ubc.ca/terms_of_use."@en ; ns0:scholarLevel "Graduate"@en ; dcterms:title "Chemistry of nitrilohexaphosphonitrilic chloride"@en ; dcterms:type "Text"@en ; ns0:identifierURI "http://hdl.handle.net/2429/34476"@en .