764
Macromolecules 1993,26, 764-771
Synthesis and Properties of High Polymeric Phosphazenes with (Trimethylsily1)methylSide Groups1 Harry R. Allcock' and William D.Coggio Department of Chemistry, The Pennsylvania State University, University Park, Pennsylvania 16802 Received March 30, 1992; Revised Manuscript Received September 30, 1992
ABSTRACT Poly(organophosphazenes) that bear Me3SiCH2 side groups were prepared via the reaction of [(trimethylsilyl)methylllithium with polytphenylfluorophosphazene) (INP(Ph)F(NPF2)21,). Fluorine replacement occurred under mild conditions to yield polymers that contained up to 70% of the side groups as alkylsilyl units. Subsequent reaction with CF3CH20Na gave air- and moisture-stablepolymers with the basic stoichiometry of [NP(Ph)~,~~OCH~CF~)1.6s-,(CH2SiMe~),1.. The polymers were analyzed by 31P, 'H, and IgF NMR spectroscopy,gel permeation chromatography (GPC), and differential scanning calorimetry (DSC). Replacement of the fluorine atoms by MeaSiCHz side groups was accompanied by some skeletal cleavage. However, high molecular weight polymers were obtained by limiting the reaction time and temperature. Evidence was obtained that the organosilicon side groups are the sites of cross-linking reactions when the polymers we heated at temperatures above 200 O C .
Agrowinginterest exists in the synthesis and properties cleavageof organosiliconunits fromthe chain unlessspecial of new macromoleculesthat contain main-group inorganic precautionsare taken. Thus, limits exist to the percentage elements such as phosphorus, nitrogen, and silicon. loading of organmilicon unita in polymers prepared by Polymerssuch as poly(organosiloxanes),polysilanes, polythis route. silazanes,and poly(organophosphazenes)2-'2are the focus (2) Polymers Prepared by Metalation of Alkylphosof considerablecurrent researcheffort. Much of this effort phazene Polymers Followed by Reaction with Organois directed toward the development of new elastomers, silicon Halides. Neilson, Wisian-Neilson, and co-workfibers, biomedical materials, lithography coatings, and e r have ~ synthesized ~ ~ organosilicon-phosphne ~ ~ ~ ceramics.12-16 polymers via the deprotonation of polycmethylphenylphw p h n e ) by n-BuLi, followed by reaction with a chloroIn a series of recent papers2-" we have describedseveral new classes of polymers that possess a phosphazene organmilane. The maximumof organosilicongroup along backbone and organmilicon side groups. These polymers the polyphosphazene chain was reported to be about 60%. may be viewed as molecular hybrids of polyphosphazenes (3) Polymers Prepared by the Reactions of Organoand polysiloxanes or polyeilanes, with molecular and silicon Chlorides with (Lithioary1oxy)phosphazenePolymaterials properties that are some combination of the mers. This method is similar to the one reported in (2) phosphorus and silicon based polymer systems. For except that the para position of an aryloxy side group is example, a valuable feature of polyphosphazenes is the the locueof metal-halogen exchange,with the lithioaryloxy ease with which properties can be fine-tuned by facile derivative then being the site for coupling to an organovariations in the side groups linked to the phosphorus siliconhalide or ring openingof a cyclosiloxane.8 Polymers atoms. Thus,properties such as solubility, crystallinity, with up to 76% of the side groups bearing organmilicon and materials flexibility can be controlled easily. The units were prepared. presence of organosilicon side groups offers the prospect ( 4 ) Reactions of Chlorophoaphuzene High Polymers of hydrophobic surface character, biomedical compatiwith Aminosiloxane Reagents. Dependingon thereaction bility, and poeeibly enhanced thermooxidative stability. conditions, phosphazene polymers in which 100%of the side groups are aminoalkyl linear-siloxane units can be The possibility also exists that the presence of silicon in the side group structure may provide a means for the prepared by this method.9 The only known drawback to preparation of cross-linked matrices that contain phosthese polymers is the relatively high cost of the (amiphorus, nitrogen,and siliconin ceramic-typestrU~tures.~~J~noalky1)siloxanereagents. Nevertheless, this is a highly attractive route to new elastomers and materials with The previously known examples of hybrid organophosphazene-organosilicon polymers fall into the following tailored surface properties. (5)Use of Hydrosilyhtion Reactions for the Coupling categories. of Organic or Organometallic Side Groups to (Aminosil(1) Polymers Prepared by the Ring-Opening Polymerizationof Cyclic Phphuzenes Thut Bear both Halogen ory)phosphazenePolymers. This approach builds on the and Orgonosilicon Side Groups. This route requires that chemistry used in method 4 but adds the option that the organosilicon side units themselves become the sites for the halogen side units in the intermediate polymer must be replaced by reaction with an alkoxide or aryloxide linkage to interesting and potentially useful tertiary nucleophile. For example, Allcock, Brennan, and Graaskastructures such as vinylsiloxanes,epoxycross-linkingunits, etc." mp4c synthesized cosubstituent phosphezene polymers with both akylsilyl and trifluoroethoxy side unita via the In thispaper we describea sixth approach which involves ring-openingpolymerization of gem-N&(CHzSiMe&C4 the reactions of (organ0eilyl)lithium reagents with polyfollowed by further reaction with sodiumtrifluoroethoxide. (fluorophoephazenes). The logic behind this approach is A drawback of this approach is that a maximum of only as follows. An intuitively simplemethod for the synthesis two alkylsilylside groups can be introduced per three P-N of organmilicon-substituted polyphosphazenb is via the reactions of metallo-organoeiliconderivatives with polyrepeating units (only 33% of all side groups). Moreover, the macromolecular substitution step (for example, the (dihalogenophosphazenes) such as poly(dich1orophosreaction with sodium trifluoroethoxide) can lead to phazene). However, detailed earlier work in our program 0024-929719312226-0764$O4.O0/0
0 1993 American Chemical Society
Macromoleculee, Vol. 26,No.4, 1993
Polymeric Phoephazenee with Me3SiCH2 Side Groups 761 Scheme I*
Q
N4
‘N
I
F-P / \NO F
Q
,CHzSiMq
F P’
N//p\N
!-F \
+ 2.0
MqSiCH2Li
m/400c
____)
F
N4
I
p@F
F/ \N/
\F
F-p
F
II
F-p
t
F
P’ ‘N
I
1,
P
’%’
CH2SiMq ‘CHzSiMe3
1
0
SiMq
p,:
MqSMzC MqSM2C’
NH4CI or MeOH
N ‘
N@
- I %’ p
11 ,CH2SiMc,
Mc~SM~C,
4
Mc3SiCHzLi
P
\CHzSiMe3
4
!,
I
CHzSiMe3
Mc3SiHzC’
\CHzSiMc, 5
The regiochemistry shown for the metal-hydrogen exchange reaction for 4 is an example only. The deprotonation is not specific to any particular MeaSiCHp side group. (I
revealed that most organometallicreagents react with poly(dichlorophosphazene)not only to replace the halogen atoms but also to cleave the phosphorus-nitrogen backbone The skeletal cleavage process can, however, be retarded by the use of poly(difluorophosphazene) as a macromolecularintermediate.= Moreover, as we reported recently, the cyclic trimer (WF2)3reacts with MeaSiCHzLi to give complete fluorine replacement without ring c1eavage.lo Thus, it seemed possible that fully or highly silylated polyphosphazenes might be accessible via the reactions between MeaSiCH2Li and fluorophosphazenehigh polymers. The only problem to extending the reaction of (WF2)3 to the high-polymer level was the insolubility of high polymeric (NPF2)n in all the normal solvents appropriate for organometallicchemistry.= However, poly(phenylfluorophosphazene)is soluble in common organic solvents,and so thispolymer was employed in place of poly(difluorophosphazene). Several questions were kept in mind throughout this investigation. (1) Does MeaSiCaLi react with [NP(Ph)F(WF&]n without causingsignificantphosghorusnitrogen bond cleavage? (2) If backbone cleavage does take place, at what stage in the substitution process does it occur? (3) Can MesSiCH2 side groups be incorporated into the polymer together with cosubstituenta such as trifluoroethoxy (CFsCH20) or phenoxy (PhO)? (4) What is the behavior of these polymers at elevated tempemtures under conditions that might lead to cross-linking and ultrastructure formation?
Results and Discussion Mods1 Reactions. (a) Replacement of Fluorine by MeaSiCHz Groupr. Although the reactions of (WF2)3 with Me&CH&i have been studied, no information existed about the reactions of N&F@h (1) with the same reagent. In particular, it was necesaary to determine if the phenyl group in 1 might sterically retard the replacement of the geminal fluorine &tomby MesSiCH2. Thus the model reaction between MeaSiCHzLi and N3P3F$h was studied.
The replacement of the fluorine atoms in 1 by MeaSiCH2 groups was monitored by NMR spectroscopy and gas cbmatography/mass spectrometry (GC/MS). The reaction was found to proceed rapidly in THF at 40 OC. As shown in Scheme I, the addition of 2.0 equiv of Me3SiCH2Li to a THF solution of 1 yielded a mixture of products. The compounds were identified by GC/MS as gem-NsPflPh(CHaSiMe3) (2) and gem-NsPsFgh(CH2SiMe& (3). Integration of the peak areas suggested that the reaction mixture consisted of 2 (16%) and 3 (86%). Further reaction of this mixture with excess Me&iCHzLi gave the fully substituted product NsPsPh(CH2SiMes)a (5) in high yield. However, the formation of a complez intermediate, shown as 4 in Scheme I, was detected in the reaction mixture by NMR spectroscopy. This compbx was identified by a series of broad peaks in the NMR spectrum between 34 and 26 ppm. Intermediate 4 is formed via a competing metal-hydrogen exchange reaction. The yield of 5 was m a x i m i 4 by the addition of a proton source, such as methanol or NKC1, to the reaction mixture (see Scheme I). The details of this metalhydrogen exchange reaction have been discused in previous p~blications.~OJ7-1Q The high-yield synthais of 5 provided support for the expectation that the phenyl group in 1 would not limit substitution of the fluorine atoms in the high polymer. (b)C-Si Bond Cleavage Studies. Compounds 2 and 3 are mod& for the replacement of fluorine by trifluoroethoxy groups. In earlier work Allcock, Brennan, Graaskamp, and Dm394*‘3described the baee-catalyzed cleavage of the C-Si bonds in wmpounds such aa [NP(CH@iMe&(NsC12)dnby CF&H&Na to generate phcmphazenee which contained P-CHa unite in place of P - C H a i M e g r o w , “hie cleavage proceee is retarded by steric effectS.lo Thus, we were concerned that the macromolecukm formed from 2 or 3 may undergo C-Si bond cleavage in the preeence of CFaCHzONato produce polymers with methyl grmps in place of alkylailyl side groups.
Macromolecules, Vol. 26, No.4, 1993
766 Allcock and Coggio
6
2
3
+ NaOCH2CF3
"
3
0
NO REACTION
(3)
48 h
It was found that 2 reactedwith exceaa CFsCH20Na in THF to undergo"fluorine replacement by trifluoroethoxy side groups and that this reaction wm accompanied by C-Si bond cleavage to produce gem-NaPs(OCH2CFs)PhICHd (6) (seeScheme11,reaction 1). However, compound 3 reacted with excess CFsCHzONa under the same conditione to give NsPs(OCHaCFs)sPh(CH2Si)2 (7) in high yield. No C-Si bond cleavage producta were detected by either GC/MSor SIP NMR analyaia (Scheme11,reaction 2). Similarly, no reaction took place when 6 wlls expoeed to excess CFsCHzONa in THF at 66 OC for 48 h (Scheme 11, reaction 3). Characterization data are listed in the Experimental Section.
Theseresulteindicatethat,althoughC-Sibondcleavage is possible with relatively unhindered speciea (Scheme 11, reaction l), compounds that have intermediate steric crowding (Le., 3) undergo fluorine atom replacement by CFsCHzO in preference to C-Si bond cleavage. Thie provided encouragement for the view that, at the highpolymer level, replacement of fluorine by trifluoroethoxy may be poesible without seriousside reactionsthat involve the organosilicon groups.
~ r o m o b ~ u l a€&tuttion# r between [NP(Ph)F(NFFl)r],, and lUe&iCH,Li. The reactions between 8 and Me&iCH&i (v8shown in Scheme III. As in the emallmolecule model reactoae, fluorine atom replacement OcCuTrBd rapidly at 40 O C in THF. However, in contrast to the emall-moleculereactiom, complete replacement of thefluorineatomebyMe&iCH2&ro~pedidnotoccurunder theee conditions. Two facton appear to limit the extent of the organometallic substitution reaction. First, when fluorine replacement reaches a certainw,the polymer precipitabsfromthereactionmedium,andthieesesnWy ternrinatee the reaction. This behavior t ascribed to metal-hydrogen e " g e reactions, similar to those identified in the cyclic trimer model eyetems. Thia side reaction would lead to a buildup of anionic charge on the polymer, with a resultant decrease in solubility. Second, the steric hindrance associated with the polymer chain
may slow the rate of fluorine replacement, eapecially after each phosphorus atom beara one organosilicon side group. The presence of reaidual P-F bonds could sensitize the polymeratohydrolysbandcroes-linking. H m t h e y w e r e not isolated but were used as reactive intermediates for alt~~tti~halQgen-replecementreactions.'2.13 Th~thpolymers were treated with CFsCHzONa in the presence of a 100% exma of CFsCH20H. The free alcohol protonated any methylene groups that had undergone metal-hydrogen exchange. After this step,the polymers redissolved, and this suggmted that the initial insolubility wata not due to cr~ee-linking.~ Attempta were also made to replace the remaining fluorine atoms by phenoxy side groups. However, complete replacement of the fluorine atoms did not occur in THF at 66 OC, as determined by 'QFNMR spectroscopy. Thia reeult is attributed to the steric restrictions experienoed by the bulky phenoxy anions. PolymwChorPctarirrtion. Aaerieaofpdymerswith different ratios ofalkyleilyl to trifluoroethoxysidegroups wata prepared (amScheme III) to study the effect of the MdiCHz side group on the polymer propertim. T h e p o l y " were purified by sequentialprecipitationsfrom concentrated THF eolutions into water and hexane followed by Soxhlet extraction with hexane for 24 h. The polymers (Sa-) were ieolated ata light brown, air- and moisture-stable materiale. Characterization data for 9a-e were obtained by 'H, leF, and 31PNMR spectroscopy. Proton and phsephorue NMR signah were broadened significantly because of solution viecoeity effects. Polymers 9a-e were f o d to have the basic stoichiometry of [NP(Ph)o.s,(CHzSih4e~),(OCHzCFs)te&l,,, with the value of I: being deduced by 'HN M R epectmecopy. Theside group ratios were estimated by the area of the SiMea resonance at 0.1 ppm to the CFsCH20 resonance at 4.2 ppm.= Characterizationdata are listed in Table I. The 31PNMR epectra of 9 h showed threemqjor typea of broad featureless resonances, assigned to P(Ph)(CHr
Macromolecules, Vol. 26, No.4, 1993
Polymeric Phoephazenee with MeaSiCHz Side Groups 767 Scheme I11
L
1
8
n - l5oO
9d x =0.63 9e x = 1.13
compd
T~(00
[ N P ( P ~ ) o . ~ (8) ~FI.~~
Table I Analytical Data for Polymers 31P NMR" M, (xio-4) signal shift (ppm) PFP -23.0 7.0 -8.1 3.8 -8.0 12.5 3.1-2.1 -8.0 12.5 3.1-2.1 -8.0 12.5 3.1-2.1 -8.0 12.5 3.1-2.1
a
signal
lH NMR shift (ppm)
Ph OCHzCF3 Ph OCHzCF3 Ph SiMe3 OCHzCF3 Ph SiMea OCHzCF3 Ph SiMe3 OCHzCF3 Ph SiMe3
1.3-6.8 4.1 1.3-6.8 4.1 7.3-6.8 0.3 4.1 7.3-6.8 0.3 4.1 1.3-6.8 0.3 4.1 1.3-6.8 0.3
ratio 2.0 1.0 2.1 1.2 1.0 1.5 1.0 1.8 1.2 1.0 3.4 1.0 1.3
1.0
R = CHzSiMea, R' = CHzSiMe3 or OCH2CF3.
SiMes) (12.3 ppm), P(CHzSiMe3)2, P(OCH2CF3)(CH2SiMes)(3.1-1.2ppm), andP(OCHzCF3)z(-8.2ppm) units. The *OFNMR spectra contained a broad singlet at 38.4 ppm from the CF3 groups, with no evidence for residual P-F bonds (54 ppm, JPF= 800 Hz) in 9a-e. From these analyses it was concluded that the polymers were at least 95% substituted by either the MesSiCH2 or the CF3CH20 side groups. Thermal analyses of the polymers were obtained by differential scanning calorimetry. Polymer 9a has a T,of -33 OC. Polymers 9b-e were amorphous, rigid materials with Tis between +51 (9c) and +73 "C (%I, depending on the loading of CHzSiMe3 units.2s Infrared analysis showed bands consistent with the assigned structuresz7 (see Table I for further details). Attempts To Bring about Complete Substitution. Several experiments were carried out to determine the upper limit for the replacement of fluorine atoms by the alkylsilyl side groups. Analyses of reaction mixtures in which THF solutions of MesSiCHzLi alone were added to a THF solution of 8 suggested that a maximum of
approximately 70% of the fluorine atoms could be replaced. Surprisingly,despite the presence of unreacted P-F bonds, these polymers were sufficiently stable to atmosphericmoistureto allow purification and subsequent analysis by NMR spectroscopy. However, extended exposure to the atmosphere caused these materials to become insoluble, presumably because of P-F bond hydrolysis and subsequent cross-linking. The lSFNMR spectra of these species showed two broad resonances, which were consistent with the presence of residual P-F bonds. When 8 was added to a large excess (-150%) of MeaSiCHzLi under refluxing conditions for 12 h, full replacement of the fluorine atoms apparently occurred.ls However,GPC analysisshowedthat the polymer molecular weight had declined to M, < 16 OOO. Molecular Weight Decline during Substitution. It was important to establish the degree to which the (organosily1)lithiumreagent attacked the polymer backbone during the initialfluorinereplacement process. Thus, different samples of polymer 8 were treated with different amounts of MeaSiCHzLi. The remaining fluorine atoms
Macromolecules, Vol. 26, No.4, 1993
768 Allcock and Coggio
incorporation of alkylsilyl side groups (9b), the weight retention of the polymer increased to 27% . This trend of higher weight retention with higher incorporationof Mea150 600 SiCHzunits continued for the other polymers in the series (see Table 11). .-lCn 0 0 3 0 The residues from the pyrolyses were glassy, graphite.c 100 400 x like materials. Infrared spectra of the residues in KBr a c e pellets showed only broad absorptions at 1030,1100,and 0 II: 1280 cm-l. A solid-state l3C NMR spectrum consisted of 50 200 a single broad resonance centered at 120 ppm, which was consistent with a graphite-likestructure. No evidence for the presence of S i 4 groups was found by solid-state l3C 0 0 NMR analysis. A solid-state 31PNMR spectrum of the 0 50 100 150 200 residue showed that at least two different types of Mol o/o RLi phoephorus environmentswere present. Both signalswere Figure 1. Relationship between GPC-averagemolecular weight in the range of 12.0-4.0 ppm and were consistent with (similar to M,)and the mole percent of MeaSiCHeLi used for alkylphosphazene structures. No resonances associated preparationof polymers9a-e. The related curve for the variation with the presence of CF3CH20 groups were detected. It in repeating units provides a correction to take into account the was apparent from these results that the Me3SiCH2 units change in side group structure as subetitution occurs. GPC samples were prepared at a concentration of 2% bulk weight in in the polymers had the effect of increasing the gross THF. thermal stability29and weight retention. Although the organosilicon-containing polymers 9b-e underwent an Table I1 initial weight loss at temperatures that were lower than Weight Loss for Polymers Heated First at 200 OC and Then for 9a, [NP(OCH&F3)2],, or [NPMez], (see Table II), to 900 O c a these control samples volatilized completely at temper% % w t loss tot % w t loss onset atures above 450-500 "C. An explanation for these compd MexSiCH2 at 200 O C at 900 O C ("C) differences is as follows. 466 N"OCHzCFa)zln 0 4.0 97 First, the control polymers, [NP(OCH2CF3)21n and 532 [NP(CH3)21n 0 14.2 99 [NPM&, are known to undergo thermally-induced 9a 0 90%). See Table I for 31PNMR analysis. an AB2 pattern with no evidence of P F coupling. Excess methanol Polymer 8 was dissolved in dry THF (1.0 g/80 mL) and was was added slowly to the reaction mixture (Caution: a vigorous warmed to -40 "C. To this solution was added 6,5 equiv of exotherm will result) and the mixture was stirred for 5 min. The MesSiCHzLi (1 M solution in pentane). After approximately pentane was removed under vacuum, and the remaining solution 20% of the lithium reagent had been added, the reaction became was poured into water and extracted twice with CHzC12. The light yellow, and precipitated polymer adhered to the sides of organic layers were combined and dried over MgS04. Solvent the reaction flask. Vigorous stirring was employed until all the removal produced a light yellow oil, which was purified by column lithium reagent had been added. After approximately 30 min, chromatographyon silica gel using hexane:CHzClz (5:l). Isolated the mixture was treated with 6.5 equiv of CFsCH20Na and 6.5 7.3 g (72% yield). equiv of CF3CH20H in THF and was refluxed for 5 h. The Analytical data. 31PNMR (CDC4, ppm) A B p , 6~ = 25.2 (P(CHr polymer redissolved,and the solution was then concentrated and p unresolved. lH SiMed~),bg = 16.8 (PPh(CHzSiMe3));2 J ~ = precipitated into water. The polymer was redissolved in THF ~ NMR (CDCla, ppm) 7.2-6.8 (m, phenyl), 1.23 (d, CH2, 2 J p = and purified further by multiple reprecipitations from THF into 17.7Hz),O.l(s,SiMe3). WNMR(CDC13,ppm) 128-118(phenyl), water and hexanes. Full characterization data are listed in Table ~ 1.4 Hz), 0.6 (8, SiMe3). Mass 27.5 (d, CH2, Jpc = 81 Hz, 3 J p = I. spectral anal.: calcd, 647; found 647. Anal. Calcd for Analysis of Molecular Weight vs Percent MerSiCHzLi C26HmNaP3SiS: C, 48.17; H. 9.35; N, 6.48. Found: C, 47.70; H, (9a-e). Polymer obtained from a single polymerization of 1was 9.20; N, 6.30, used in this experiment to ensure that the initial molecular weight Synthesis of gem-N;P;Ph(CH,)(OCH2CFS)4(6). Comof polymer 8 was the same for all reactions. Thus 5.0 g (16.3 pound 6 was formed in approximately 15% yield from a reaction mmol) of 8 was dissolved in exactly 100 mL of dry THF to give in which 2.0 equiv of MesSiCHnLi was added to N3P35Ph (1) in THF. The intermediate product, N ~ P ~ F ~ P ~ ( C H (2), Z Swas ~ M ~ ~ a) standardsolution containing0.05g/mL of 8. From this solution four 15-mL portions (0.75 g, 12.2 mmol in terms of P-F bonds) identified by GUMS analysis of the reaction mixture. Subsewere transferred to four separate reaction flasks equipped with quent treatment of the product mixture containing 2 and 3 with a graduated addition funnel, a condenser, and a magnetic stirrer CF3CHzONaresulted in complete fluorine replacement together bar. The polymer solutionswere then diluted with an appropriate with C-Si bond cleavage of the Me3SiCHz group in 2. Compound amount of THF such that the final volume of each reaction was 6 was isolated from the reaction mixture by column chroma100 mL. An appropriate amount of MeaSiCHnLi (1M solution tography (further details are provided below). Compound 6 has in pentane), calculated to replace either 0,15,25,50, or 75% of been synthesized previously by an alternative method.5 the fluorine atoms in 8, waa transferred to the addition funnels. Analytical data: NMR (CDC13, ppm) AB2, 6~ = 31.2 (t, Because of the competing metal-hydrogen exchange reaction, P(Ph)CH3), 6~ = 15.2 (d, P(OCHzCF3)z); 'Jpp = 32.9 Hz. 'H the actual incorporation of the MesSiCHz group was lesa than NMR (CDC13, ppm) 7.8-7.5 (m, phenyl), 1.74 (dt, CH3, 2 J p = ~ calculated. The polymer solutions were then warmed to 40 "C, 14.6 Hz, 'JpH = 2.OHz). NMR (CDCl3,ppm) 135-132 (phenyl), and the Me3SiCHzLi was added. After 30 min the reaction 22.0 (dt, CH3, Jpc 126 Hz, Vpc = 3.0 Hz). Mass spectral anal.: mixtures were treated with CF3CHzONa (15 mmol) containing calcd, 623; found, 623. CF3CHzOH(15 mmol) in THF and were stirred for 12 h at 60 Synthesis of gem-NsPs(OCHtCFs)sPh(CH~SiMes)2 (7). To "C. The polymers were purified by precipitations into water a reaction flask equipped with an addition funnel, condenser, and hexanes, followed by Soxhlet extraction with methanol for and a stirrer bar were added 2.5 g (9.0 mmol) of 1 and dry THF 24 h. Characterization data are listed in Table I. GPC eamples (200 mL). To this solution was added Me3SiCHzLi (18.1mL, 2.0 were prepared by dissolving 50 mg of polymer in 5 mL of THF. mmol) in pentane over a 30-min period during which time the reaction mixture turned light yellow. A 31PNMR spectrum of Acknowledgment. We thank the Air Force Office of the reaction mixture showed a complex series of peaks consistent (3). The reaction with the structuregem-N3P3F3Ph(CH&3iMe3)2 Scientific Research for financial support of thie work. We mixture was stirred at room temperature for 8 h to ensure also thank A. J. Benesi and R. Dudenhoefer for assistance complete formation of the product. A portion of this reaction with the solid-stateNMR and TGA experiments, J. Blank mixture was saved for GC/MS analysis. It was concluded that and L. Collins for conducting the TGNMS and GC/MS 85% of the product was 3, and the remaining compound was experimente, and C. J. Nelson for assisting with some of identified as gem-N3P3F4Ph(CHzSiMe3)(2). To the reaction the experimente. mixture containing 2 and 3 was added 6 equiv of CF3CH20Na. The solution was stirred at 25 "C for 12 h and heated further for References and Notes 6 R at 66 "C to ensure full substitution. The reaction was terminated when no further change was detected by 31PNMR (1) This paper is the eleventh from our laboratory on organaeilicon spectroscopy. The THF was removed under reduced pressure, derivativesof phosphazenes. For previous papere in this aeries, and the resultant light yellow oil was chromatographed on a see refs 2-11.
Macromolecules, Vol. 26, No.4,1993 (2) Allcock, H. R.; Brennan, D. J.; Allen, R. W. Macromolecules 19811,18,139. (3) Allcock, H. R.;Brennan, D. J.; Graaakamp, J. M. Organometallics 1986,6,2434. (4) Allcock,H.R.;Brennan,D. J.;Graaskamp, J.M. Macromolecules 1988,21,1. (5) Allcock, H. R.;Brennan, D. J.; Dunn, B. S.; Parvez, M. Znorg. Chem. 1988,27,3226. (6) Allcock, H. R.; Brennan, D. J. J. Organomet. Chem. 1988,341, 231. (7) Allcock, H. R.;Brennan, D. J.; Dunn, B. S. Macromolecules 1989,22,1534. (8) Allcock, H. R.; Coggio, W.D.; Archibald, R. S.; Brennan, D. J. Macromolecules 1989,22,3571. (9) Allcock, H. R.; Coggio, W. D. Macromolecules 1990,23,1626. (10)Allcock, H. R.; Coggio, W. D.; Parvez, M.; Turner, M. L. Oraganometallics 1991,10,677. (11) Allcock, H. R.; Nelson, C. J.; Coggio, W. D. Organometallics 1991,10,3819. (12) (a) Allcock, H. R. Chem. Eng. News 1985 (March 18),63,22. (b)Zeldin, M.; Wynne, K. J.; Allcock, H. R., Eds. Inorganic and Organometallic Polymers; ACS Symp. Ser. 360; American Chemical Society: Washington, DC, 1987. (c) Mark, J. E.; Allcock, H. R.; West, R. C. Inorganic Polymers; Prentice-Hall: Englewood Cliffs, NJ, 1991. (13) (a) Phosphorus-Nitrogen Compounds; Academic Press: New York, 1972. (b) Also see: Allcock, H. R. In Comprehensive Polymer Science; Pergamon Press: New York, 1989; Vol. 4, Chapter 26. (14) Allcock, H. R.; Kwon, S.; Riding, G. H.; Fitzpatrick, R. J.; Bennett, J. L.Biomateriak 1988.9, 509. (15) See ref 13b, Chapters 3 and 11, Ad references therein. (16) Wynne, K. J.;Rice, R. W. Annu. Rev. Mater. Sci. 1984,14,297. (17) (a) Neilson, R. H.; Hani, R.;Wisian-Neilson, P.; Meister, J. J.; Roy, A. K.; Hagnauer, G. L. Macromolecules 1987,20,910.(b) Allcock, H. R.; Allen, R.W.; OBrien, J. P. J. Am. Chem. SOC. 1977,99,3984. (18) Wisian-Neilson, P.; Ford, R. R.; Neilson, R. H.; Roy, A. K. Macromolecules 1986, 19, 2089. Also see: Neilson, R. H.; Wisian-Neilson, P. Chem. Reu. 1988,88,541. (19) Allcock, H. R.; Desorcie, J. L.; Riding, G. H. Polyhedron 1987, 6,119. (20) Evans, T. L.; Patterson, D. B.; Suszko, P. R.; Allcock, H. R. Macromolecules 1981,14, 218. (21) Allcock, H. R.;Evans, T. L.; Patterson, D. B. Macromolecules 1980,13,201. (22) Paddock, N. L.; Ranganathan,T. N.; Todd, S,M. Can.J.Chem. 1971,49,164.Ala0 see: Paddock, N. L.; Ranganathan, T. N.; Todd, S. M. Znorg. Chem. 1973,12,316.
Polymeric Phosphazenes with MeaSiCHp Side Groups 771 The preparation of poly(difluorophosphane)requires the use of a high-pressure autoclave. When formed, the polymer ie soluble only in fluorinated solvents. See: Evans, T. L.; Allcock, H. R. J. Macromol. Sci. 1981,A16 (l),409. This solubility behavior haa been detected previously in other polyphosphazenesystemsknown to contain anionic sidegroups. See ref 8 and Wisian-Neilson,P.; Islam, M. S.M13CrOmOleCUk8 1989,22,2026. Although an integration of the peaks in the 'H NMR spectrum provided evidencefor the stoichiometry of these polymers,their ceramic nature prevented the use of elemental analysis to corroborate the 1HNMRdata. Carbon, hydrogen,and nitrogen analyses were consistently low because of incomplete combustion. Such results for ceramic materials have been reported previously. See: Wu, H.; Interrante, L. V. Chem. Mater. 1989, 1, 564 and references therein. No Tgwas detected for 9b by DSC analysis. However, 9b was more flexible than 9c-eand may perhaps possess a Tgbetween those of 9a and 9c. Infrared spectral analysis gave absorption bands that were consistentwiththe presenceof aryl, alkyl,Si
