A Mechanistic Investigation of Gelation. The Sol-Gel Polymerization of Bridged Silsesquioxane Monomers. - ~ -.

Transcription

1 = -, ! :.?s. -,-.-,7 m--=.<: k.-/-: , -1 Shea/Loy [.9 5/9 N4xwxb=-/l3f3LT A Mechanistic Investigation of Gelation. The Sol-Gel Polymerization of Bridged Silsesquioxane Monomers. - ~ ~~@o Kenneth J. Shea * and Douglas A. Loy+* ~@ j > t 4!JJ #Department ofchemistry, Universiv ofm&nia Irvina Irvines CA92697-@!!2 ~ ~ fcatalysts and Chemical Technologies Department, Sandia National L.uboratories, Albuquerque, NM Abstract: The study of a homologous series of silsesquioxane monomers has uncovered striking discontinuities in gelation behavior. An investigation of the chemistry during the early stages of the polymerization has provided a molecular basis for these observations. Monomers containing from one to four carbon atoms exhibit a pronounced tendency to undergo inter or intramolecular cyclization. The cyclic intermediates have been characterized by 29Si NMR, chemical ionization mass spectrometry and isolation from the reaction solution. These carbosiloxanes are local thermodynamic sinks that produce kinetic bottlenecks in the production of high molecular weight silsesquioxanes. The formation of cyclics results in slowing down or in some cases completely shutting down gelation. An additional finding is that the cyclic structures are incorporated intact into the final xerogel. Since cyclization alters the s~cture of the building block that eventually makes up the xerogel network, it is expected that this will contribute importantly to the bulk properties of the xerogel as well.

2 This report was prepared DISCLAIMER as an account of work sponsored by an agency of the United States Government. Neither the United States Government nor any agencythereof, nor any of their employees, make any warranty, express or implied, or assumes any legal Inability or responsibility for the accuracy, completeness, or usefulness of aw information, apparatus, product, or process disclosed, or represents that its use would not infringe privately owned rights. Reference herein to any specific commercial product, process, or service by trade name, trademark, manufacturer, or otherwise does not necessarily constitute or imply its endorsement, recommendation, or favoring by the United States Government or any agency thereof. The views and opinions of authors expressed herein do not necessarily state or reflect those of the United States Government or any agency thereof.

3 DISCLAIMER Portions of this document may be illegible in electronic image products. Images are produced from the best available original document.

4 --, !k.-..~.:f,?.- ---,7- --?-~-- --= T?,T ShealLoy Introduction Hybrid materials lie at the interface of thg organic and 4 Their synthesis offers exceptional op-portunities to not only combine the important properties fi-om both worlds, but to create entirely new compositions with truly unique propepies. The size of the organic and inorganic domains can range from traditional composites prepared from physical mixtures of datively large, micron size, particles to nanophase materials that include substances with organic and inorganic domains that are dispersed homogeneously at the molecular level. A representative of this latter group are the bn dged polysilsesquioxanes?-22 a family of hybrid organic-inorganic materials prepared by the hydrolysis and condensation of monomers containing a variable organic bridging group with two or more trifunctional silyl groups. Trialkoxysilyl groups are the most common silyl functionality although other groups, such as silyl chlorides, have served as precursors. (Scheme 1).3 4 Inorganic Oxide Precursor I n (EtO)@~Si(OEt)3 i II Variable Organic Component Scheme 1. Dried monolithic bridged polysilsesquioxane xerogels and aerogels and a schematic of the sol-gel polymerization of monomers which afford the organically bridged polysilsesquioxane network polymers. The organic group, covalently attached to the trialkoxysilyl groups through Si-C bonds, can be varied in length, rigidity, geometry of substitution, and functionality. This variability provides an opportunity to explore how the organic structural unit contributes to bulk properties such as porosity, thermal stability, optical clarity, 2

5 t * ShealLoy chemical resistance, and dielectric constant. Representative monomers are shown in Figure 2. 0 % --. -~ // :RO)3S \ / i(or)a R0)3s n & \ / (or) (RW \ / i(or)~ n=l,ls,g 54 n = 2, 25 6 \/ n = 3, ,6 (RO)3Si~~i(OR)3 n ~Ro)N.(oR)3 A n= l,& ~14,15 n=2,? (RO)3 +-l- Si(OR)3 m (Ro)3~o--si(oR)3 m = 5, 9, & m = 6, 10MO G 3 m = 8, ll, G 3 (RO)#F~Si(OR)s ~ = 10, 12~.l~ls R m = 14, 13 G 3,59 (RO)$ii Jo-+) R R ;=:;:;:l,,= m (RO) Si(OR)3 (RO) Si(OR)3 1 Y 2214,15 21 Q (EtO)&f Si(OEt)3 rl 29 (RO)@f12Si(OR)3 (RO)3flSi(OR)3 c 2~7,8, IC-13 2~14,15 (RO)3 d--f ~Si(OR)3 (RO) Si(OR)3 +--l 267,8,1 C-13,20 277:6,1 CF13,X (RO)~si(0R)3 (RO)~Si(OR)3 D 28 3,Z 29 3VZ Figure 1. Monomeric precursors to bridged polysilsesquioxanes with rigid (A), long flexible (B), 1- and 2-carbon alkylene (C), and 3- and 4-carbon (D) bridging groups. 3

6 I, SheaJLoy Sol-gel polymerization of these hybrid monomers involves a series of hydrolysis reactions forming ~ilanol~functionalized intermediati%%%iil condensation reactions that produce oligomenc and cyclic intermediates. These species react to produce higher molecular weight polymers that eventually form a three-dimensional infinite netiork which span the dimensions of the reaction vessel. At this point the solution loses fluidity and reaches the gel point (Scheme 2). The wet gel is then dried to afford a xerogel (AY gez). During the drying process the gel undergoes significant shrinkage resulting in partial collapse of pore structure. At fiist glance, the bridged polysilsesquixoxme system appears to fall into the same reaction pathways observed with organotrialkoxysilanes, RSi(OR )32324or tetralkoxysilanes.25 z6 However, the bridging configuration will be shown to provide additional, un-anticipated sol-gel reaction pathways that are directly dependent on the nature of the organic bridging group. Scheme 2. Schematic representation of the early stages of oligomerization of bridged silsesquioxane monomers. The condensation reactions continue leading eventually to the formation of an infinite network (gel point). One objective of contemporary materials chemistry is to develop an understanding of how the molecular building block determines the bulk properties of the resultant material. Since bulk properties of materials, such as porosity and modulus, have both chemical and physical origins, the problem is formidable. Bridged polysilsesquioxanes provide an exceptional opportunity to study relationships between the starting monomer and final xerogel morphology. There are several reasons for this. First, they readily react under a standard, uniform, set of sol-gel conditions to afford gels. Second, their modular construction permits access to a wide range of monomers which allow systematic variation of the organic bridging group. Third, they have a range of interesting and highly reproducible bulk properties (such as high surface area and porosity) that are strongly coupled to the organic bridging group. Fkally, they have a rather simple morphology, that is, they are true molecular composites that are 4

7 Shea/Loy t, homogeneous to the molecular level with no discemable phase separation between organic and inorganic domains.27 In this Account we discuss the in.uence ofsmull, stitematic-~terbations of the organic fragment on the formation of bridged silsesquioxane gels. We have focused on the early stages of the process to document the chemicai events that lead up to and include gelation. These studies have produced insight as to how these small perturbations affect the polymerization chemistry and ultimately, the final xerogel structure. Gelation of Bridged!Ylsesquioxane Monomers Typical bridged polysilsesquioxane monomers are hexafunctional. Experience has shown- that man~ of-these monomers (Figure 1), even as very dilute solutions (ea. 0.1 M), rapidly polymerize and form gels. Our interest in the chemistry of gelation was peeked by observations of anomalous gelation times during a sol-gel polymerization study of cx,m alkylene-bridged silsesquioxane monomers (9-13, 21, 24, 26, 27; Equation 1).&8It was noted that changing the length of the alkylene-bndging group by a single methylene group could result in a change in gelation time from hours to over six months!68although gelation is not a primary phase transition, it was nevertheless quite reproducible under standard conditions. Eq.1 +tis~m r 1 m (EtO)3S I-# Si(OEt)3 The response in gel time to such small changes in the alkylene bridge is likely to have a significant influence on the network that was being produced and ultimately on the xerogel morphology. It was decided therefore to study this series in more detail.iw 13:o Hexafunctional silsesquioxane monomers can be grouped into three broad classes based on their solution chemistry and gelation behavior. Scheme 3 below summarizes these findings. The fh-st group (Scheme 3, left branch) incorporates monomers that gel rapidly even as very dilute solutions. The second group of monomers are decidedly slower to gel than the first group (Scheme 3, counterclockwise path). The remaining group (Scheme 3, right branch) show little tendency to gel under most conditions. 5

8 , SheafLoy Dimerization,* 1 E~ ~Et- - - EtO-~~S~Et =C30r HO OEt o --- C4 k\ Intramolecular o=cl Intramolecular or C2 Cyclization Scheme 3. Representation of the oligomerization reactions of bridged polysilsesquioxane monomers. The monomers are grouped into three broad categories with regard to their gel times, those that gel rapidly (left branch), those that are much slower to undergo gelation (counterclockwise path), and those that show little tendency to gel (right branch). The gelation behavior corresponds to a standard set of polymerization conditions (0.4 M monomer, 6 equivalents water, 10.8 mol% HC1 catalyst. Monomers That Gel Rapidly The first class consists of monomers that have bridging groups that are either short and stiff (Figure 1; A) or long and flexible (Figure 1; B).&13$181,4- Bis(triethoxysilyl)benzene (1) and 1,10-bis(triethoxysilyl) decane (12) are representative examples. Their gelation times, together with gelation times of bridged monomers from the two other classes of monomers and TEOS (>2x the concentration)z~ are summarized in Table 1. Bridged monomers in this first group rapidly form gels even at very low concentrations (O.lM) regardless of solvent or catalyst.c s.

9 SheaJLoy Table 1. Gelation times for the sol-gel polymerization for various bridged polysfisesquioxane monomers under standard sol-gel conditions (0.4 M monomer, 6 equivalents water, 10.8 mol% EIC1 catalyst). For~parison, the gelation ties for tetraethoxysilane at 1.0 M concentration are also given. - Monomer EtOH/ HC1 Catalyst THF/ HC1 Catalyst Si(OEt)d (1.0 M) 300 days RapidGellingMonomers:Short/StiffandLongFkxible t e ( )3s \ / 10 minutes 1 hour 1 \ (EtO)3S %i ~Si(OEt)3 24 hours 1 minute 10 \ Cl andc2monomers:slowergeltimes f (EtO)3S H,Si(OEt) hours 50 hours 21 (EtO)3S ~Si(OEt)3 720 hours 96 hours +3 A (EtO)3Sl 25 Si(OEt)3 120 hours 240 hours C3andC4Monomers:SlowestGelTimes (EtO)3S Si(OEt)3 H 4000 hours 4300 hours 263 (EtO)3S +-t ~Si(OEt) hours 27 (EtO)3S~Si(OEt)s 29 >2000 hours >9000 hours

10 ShealLoy We wished to establish the pattern of oligomerization for these fast gelling monomers for comparison with related monomers that exhibit deviations from this behavior. However, since gelation occurs rapidly, it is diffidult to%?~w the solution chemistry in real -time. TOovercome this difficulty we have developed a procedure that employs sub-stoichiometic quantities of water for the sol-gel polymerization. These conditions permit monitoring of the reaction by solution phase 29SiNMR and chemical ionization mass spectrometry in real time. Under these modified conditions the low molecular weight intermediates that are formed during the early stages of polymerization can be detected. The results of these investigations are summarized below. The sequence of events in the initial stages of sol-gel polymerization of 1,10- bistriethoxysilyldecme (12), is straightforward. Both mass spectrometry and 9Si NMR studies revealed the rapid accumulation of hydrolysis products ( F) and linear dimers ( l ).Their assignment was based on a systematic downjield shift of 2-3 ppm in the 9Si NMR upon each ethoxysilane hydrolysis to a silanol and an approximate 5-8-ppm upfiekl shift for each condensation to form a siloxane bond. The subsequent accumulation of linear and branched timers and higher condensation oligomers proceeds rapidly even under these substoichiometric reaction conditions. In a short time these higher molecular weight species exceed the detection limits of mass spectrometry and the rapid onset of gel formation precludes study at longer reaction times. A similar pattern emerges with 1,4-bistriethoxysilyl benzene (l). Solution gsi NMR and chemical ionization mass spectrometry reveals a rapid burst of hydrolyzed monomer followed by dimer and timer formation. Although the analytical methods and onset of gelation preclude following the reaction further, this process proceeds unabated with the rapid build up of higher molecular weight branched oligomers. Conspicuously absent is evidence for the accumulation of significant quantities of cyclic monomers and dimers during the early stages ofpolymerization. Apparently cyclization is entropically and/or sterically unfavorable in these systems. The xerogels produced from these polymerizations exhibit typical patterns of T, Tz, and T3 absorption in the 2 SiNMR. These spectra were compared with spectra of xerogels prepared from closely related monomers (vide inji-a). The rapid gelation of hexafunctional monomers such as 1 and 12 serve as a benchmark for the gelation behavior of related monomers. With this in mind the gelation results of an extended series of alkylene bridged monomers produced a number of surprises. Figure 3 summarizes the gelation times for the series of alkylene bridged monomers under acidic conditions (9-13, 21, 24, 26, 27).Z9 There are clearly three different distinct regions on the graph delineated by gel times. The gelation times for the methylene- and ethylene exhibit a sharp increase as the length of the bridging group increases from two to three methylenes. The extension of one additional methylene group to the five-atom bridge returns the gelation time to normal, that is, similar to the longer-chain fast gelling monomers. What s the reason for the unusually long gelation times? By identifying the oligomers produced during the early stages of-sol-gel polymerization we have obtained insight to the origin of the anomalous gelation behavior.

11 Shea/Loy. P _ (EtO)$~ Si(OEt)3 To -EtOH I I i -4o ppm ~H To TO T1 I ~. +! 1 I I 1 1 I ~pti 2_4sk f I 1 I 1 i I ppm Figu~ 2. 29SiNMR spectra for 1,10-bis(triethoxysilyl)decane 12 (top) and its sol-gel solutions with 1 equivalent HZO (middle) and 2 equivalents HZO(bottom) under acidic conditions. The two peaks downfield of the absorption from the monomer To peak (6= ppm) are from hydrolysis products. Peaks from more condensed silsesquioxanes ( 1 1and Tz) appear upfield with multiple peaks due to the unhydrolyzed and hydrolyzed species.

12 Shea/Loy ,,,,,,,,, -- _ n & # #,,,,...,!!A;;~:i!..,...,,,,,!,~,, =.!..,.... +;5,Oo, t.,, o,,,,,, o...,,!i,!i,,..,!...o!o %$.0..3! !....?.!.0 ~ (EtO)$im~&~~E~3V ~ j ;s@ m....,,,...,,...! ,.,..., > !! o O...! o Alkylene Length, n Figure 3. Graph of gelation times for alkylene-bridged polysilsesquioxanes under standard sol-gel conditions with HC1 (squares) and NaOH (circles) catalysts. Under basic conditions gelation proceeds rapidly. Under acidic conditions, small changes in the bridging alkylene chain length produce profound changes in gel time. Cl and C2 Bridged Monomers Short chain bistriethoxysilyl monomers such as the methylene- and ethylenebridged monomers 21 and and the vinylidene- and (Z)-ethenylene-bridged monomers 22 and deviate significantly in gel time from their longer chain relatives (Figure 3). The modified reaction conditions (sub-stoichiometric water, acid catalyst) were employed to monitor the reaction during the early stages of polymerization. It was observed that in addition to both branched and linear oligomers, the solution chemistry of both Cl monomers 21 and 22 and C2 monomers 24 and 25 were found to include rapid Cyclodimerization. The formation of substantial quantities of cyclic dimers marks a significant departure from what was observed with long chain 10

13 Shealhy Accounts of Chem calresearch and stiff bridging groups which rapidly gelled (1-19). Indeed, monocyclic eight member rings and bridged bicyclics dominate the composition of the methylene-bridged monomer 21. (Figure 4). Monomer 22, a 1,1-vinyliden&bridged%&@hdkoxysilane), also formed an eight-membered dimer 30 with two exocyclic methylene groups. (Scheme 4) This dimer was isolated from preparative scale reactions. and its 2 Si NMR and mass spectrometric characteristics were consistent with the predominant species observed in the sol-gel of 22. Gelation of 22 under these conditions has not been observed after more than two years (Table 1). M 100$ 2? a Cyclic Dimer (CD) so 45 M ls M / LA a 1 L ( 4 LD 533 LD 5 FT 69 CT 7s3 Cyclic Trimer (CT) Al., ~~~~ 400 7b 860 9Ao lioo nfz H2 #+#si S<o= f~ i Sii%i ~? H2{ ) Monomer (M) Linear Dimer (LD) SKO, S K #i H2 Fused Cyclic Trimer (H) Figure 4. Chemical ionization mass spectrum of the initial stages of the acid catalyzed sol-gel polymerization of bis(triethoxysilyl)methane (21). The ring size of cyclic dimers and trimers can not be unambiguously assigned by mass spectrometry.

14 , -m- 7.. ;, ,-.,,.#-4r77? <E%7F...,.-,Z7U-:. -- ShealLoy Accounts of Chem cal Research n w&t * 6n H20 =? ~g;t -2EtoH EtO. 22 Scheme4. Sol-gel polymerization ofvkylidene monomer 22 fomedcyclic dfier~in sufficient quantities to be isolated. The ethylene- and ethenylene-bfidged monomers 24 and 25, on the other hand, were found to be siphoned off to bicyclic dimer intermediates (Scheme 5). A high chemical yield of the bicyclic dimer 31 can be isolated when 24 is reacted with a two equivalents of water under acidic conditions. Even greater yields of the unsaturated bicyclic analog are recoverable from the hydrolysis of 25 with two equivalents of water. In all cases involving Cl and C2 carbon atom bridging groups, the cyclic derivatives of 21,22,24,25 have a reduced level of functionality. The observation of significant populations of cyclics is an important factor in the gelation time since cyclization reduces the level of ji.mctionality and does not contribute to network formation. EtO, ~ OEt EtO-~ ~<c)et ~ EtO EtO H@ 23 E~n. Et Et% ~ % ~ OEt EtO+ F OH ~ EtO-f ~Et&l;EtEto~OEt EtO EtO EtO OEt EtO / / \\?E&PEt \ ~P- \ OEt Et&A Q. EtO 0- J GEL OEt 31 Scheme 5. Sol-gel polymerization of monomer 24 was shown by %i NMR and mass spectrometry to dimerize and cyclize to afford bicyclic dimer 31 as an intermediate involved in the slowing of network growth. In addition, some cyclics, such as the bridged bicyclic dimer 31 found in the ethylene bridged system (24), is inherently less reactive in subsequent hydrolysis and condensation steps due to the presence of tertia~ alkoxysilane linkages. This reduced reactivity further slows down the approach to gelation. Despite the reduced functionality and reactivity, gelation eventually does occur in these systems. Gelation times for the unsaturated ethenylene-bridged monomer 25 are comprable to those for 24, but substantially longer than that observed a long alkylene-bridged monomer, such 12

15 Shea/Loy as 10, under the same conditions (Table 1). The vinylidene bridged monomer 22 does not gel under acidic conditions and takes nine months to gel under basic conditions. The observation of substantial amounts-cyclic mid blcy~h~ners during the early stages Ofpolymerization is noteworthy. However it is also important to establish the relevance of the oligomers to the final structure of the network. Substantial literature exists docm enting the formation of cyclic oligomers in silica polymerization.x2gunder conditions of kinetic network growth, a working hypothesis is that the oligomers, including the cyclic and bicyclic structures, present during the early stages of polymerization, are incorporated into the final xerogel network. In most cases this is extremely difficult to verify since the spectroscopic signatures of the individual species are not unique and therefore cannot be identified in the xerogel. This is particularly true since the tools for characterizing highly condensed xerogels at the chemical level (for example solid state NMR) are relatively low resolution. In contrast, we have found that a number of the cyclic and bicyclic bridged silsesquioxane oligomers do have distinctive signatures in the 29SiNMR that allow one to track the structure all the way through to the xerogel. A case in point is the bridged bicyclic silsesquioxane (31). The bridged bicyclic silsesquioxane 31 exhibits two 29Siresonances at and ppm, several ppm downfield from typical 9Si resonances found in aliphatic silsesquioxane monomers. The xerogel from the ethylene-bridged monomer 24 has a very atypical 29SiNMR spectra consisting of a broad absorption at ppm. This peak is several ppm downfield from what would be expected for a typical aliphatic bridged polysilsesquioxane xerogel. Compare it for example with the 9Si NMR spectra of xerogel from the 1,10-decylene-bridged monomer 12. The differences between the two spectra can be attributed to incorporation of bridged bicyclic subunits into the xerogel network. Indeed, the 29SiNMR of a xerogel prepared from the bicyclic precursor 31 under basic conditons is superimposable with that of the spectra obtained from the ethylene-btidged monomer 24.

16 Shea/Loy Accounts oj Chemical Research /l- /J\ A (E,O,S & 43 tp i ~ i(oet)3 ~Et EtO ~ Gel Et 31 I i I I I 1 I I I [ I I I o oo -120 PPm Figure 5. %i solid state NMR of the dried xerogel formed from the acid-catalyzed polymerization of monomer 12 (top spectra) and a spectrum of the dried xerogel from monomer 24 (bottom spectra). See text for an explanation of the differences in the two spectra. The spectroscopic results reveal that the early stages of the step-growth polymerization of the Cl bridged monomers 21 and 22 and the Q bridged monomers 24 and 25 are dominated by bimolecular dimerization reactions. These cyclic dimers have reduced reactivity slowing, but not stopping, progress towards gelation. In the case of the ethylene-bridged and the ethenylene-bridged monomers, the signature of bridged bicyclics present in the dried xerogels, provides direct evidence that the oligomers formed in the early stages of polymerization are being incorporated in the network. C3 and C4 Monomers A substantially more dramatic effect is observed when the bridging group consists of three or four bridging methylenes. Under the standard acidic sol-gel conditions, these monomers fail to gel after 6 months!1g13this is truly remarkable when compared to the gelation time of 1 min for the 1,10-decylene bridged monomer (12) (HC1/THF). How can such a small change in the monomer result in such a profound change in gelation time? The answer is in the chemistry of the sol-gel polymerization. In the presence of 1 equivalent of water both the C3 and C4 monomers are rapidly siphoned off to the intramolecular cyclization product. Indeed both cyclization products can be prepared and isolated in high chemical yield by the addition of one equivalent of

17 Shea/Loy water.. The mass spectrum and %i NMR of these reactions are remarkably simple (Figure 6). Upon addition of a single. equivalent of wat~r t~ ~eslene:bridged monomer (26), the monocyclic Silsesquioxane 32 is formed quantitati~ely. Furthermore, no higher molecular weight condensation products are observed and the polymerization reaction runs into a wall. The failure of these monomers to gel must reflect the reduction in number of functional groups (6-> 4) of the cyclic product and its intrinsic lower reactivity towards subsequent condensation reactions. Indeed, when a second equivalent of water is added the sole product is the bicyclic dimer (33) and its hydrolysis products (Figure 6). Crystalline bicyclic dirner (33) can readily be isolated and its structure has been confhmed by X-ray crystallography. It is important to note that these conditions (2 eq water) are sufllcient to gel longer alkylene chain monomers A very simikm pattern of reactivity is observed for the four carbon bridged monomer (27) (Scheme 6). Rapid and almost quantitative intramolecular cyclization to produce the seven-membered cyclic carbosiloxane (34) is observed. This is followed by slower formation of a bridged bicyclic dimer (35) when the second equivalent of water is added (Equation 4). Similarly, the (Z)-2-butenylene-bridged monomer (29) also readily cyclizes to afford the analogous unsaturated cyclic monomer and tricyclic dimer.13 Gelation of both butylene13 and 2-butenylene systems,17 when it eventually occurs (ea. six months or more!), produces xerogels that bear the signature of a bridged bicyclic. This is illustrated in figure 7. The typical %i pattern of a silsesquioxane xerogel is shown at the top of the stacked plot. The decylene-bridged xerogel is shown at the top (for reference), the butylene-bridged xerogel is shown immediately below and the %i NMR of a xerogel prepared from base-catalyzed condensation of the cyclized bridged monomer 33 is shown at the bottom. (Base catalysis was used so that experiments could be completed in reasonable amounts of time). The absorption at approximately -60 ppm in the spectra of the xerogels derived from 27 and 34 are very similar to what one would predict for a network comprised of bridged bicyclic units. These resuits provide additional support for the concept that cyclic oligomers are incorporated in the growing network.

18 ShealLoy To ppm (EQS 0 lcvclic i(oet]247 Ppm nlacyclic ppm] T1cycliG (2Q + 1 HZO 1 I I t I i I i ppm mt#. d3et P EtO- OEt 8 33 T2cYCliC 51 ppm [T2acyclic 60 ppml 2qyclic ppm Figure 6. Solution %i NMR of monomer 26 (top) in the presence of 1 (middle) and 2 equivalents (bottom) of water and acid catalyst. The spectra consist of almost exclusively monocyclic and bicyclic products. Et0)3s ~s@et) Scheme 6. Cyclization of 27 to afford seven-membered cyclic 34 and tricyclic ,-w,.., ~- :- -.Z??.?.?~.~,~z;y-.. j;fl> --?-

19 slleaby - _ - -.-Y A l\ L--l- A (EtO)&- H Si(OEth 4 L oo -120 ppm Figure 7. %i solid state NMRof the dried xerogel formed by the acid-catalyzed polymerization of monomer 12 (top spectra), a spectra of the dried xerogel prepared by the base-catalyzed polymerization of 27 (middle spectrum) and the xerogel prepared from the base catalyzed polymerization of the cyclic monomer 34 (bottom spectra). See text for an explanation of the differences. With the identification of these cyclic and bicyclic structures it is now possible to explain the discontinuities in gelation times (Figure 8). The formation of stable mono 17 ~!.,,,.>+.<~ L ;.L,..,?/.J.-,; :...:---,-:,>-yq.<.-.,\~-.:t.. ~:~.. ---,~:-

20 ShealLoy and bicyclic carbosiloxane.,,,,,,.,,,,,,,, intermediates not only results in reduced functionality of the monomers but also lowers reactivity in subsequent condensations, particularly under acidic conditions. Shortly after the reaction begrns, the stiuc-~re-;f=monomer has been funneled to a.more stable and less reactive species. :ti [ iii1 :.-. -,,,,.,.t,,..,..., ~Et~:gE,,,,.,,,*[ o Alkylene Length, n Figure 8. Plot of gel time of the alkylene-bridged silsesquioxane series as a function of chain length. Structures refer to the dominant species present in the polymerization reaction shortly after initiation. Conclusion The study of a homologous series of silsesquioxane monomers has uncovered striking discontinuities in gelation behavior. An investigation of the chemistry during the early stages of the polymerization has provided a molecular basis for these observations. Monomers containing from one to four carbon atoms exhibit a pronounced tendency to undergo inter or intramolecular cyclization. The cyclic intermediates have been characterized by %i NMR, chemical ionization mass spectrometry and isolation from the reaction solution. These carbosiloxanes are local thermodynamic sinks that produce kinetic bottlenecks in the production of high molecular weight silsesquioxanes. The formation of cyclics results in slowing down or 18

21 Shea/Loy in some cases completely shutting down gelation. An additional finding is that the cyclic structures are incorporated intact into the final xerogel: Sinc~lization alters the structure of the building black that eventually makes ~p the xerogel network, it is expected that this- will contribute importantly to the bulk properties of the xerogel as well. Acknowledgments We are grateful to support from the Department of Energy and in part National Science Foundation. Sandia is a multiprogram laboratory operated by Sandia Corporation, a Lockheed Martin Company, for the United States Department of Energy under Contract DE-AC04-94AL References 1) Ellsworth, M. W.; Gin, D. L. Polym. News 1999,24, ) Judeinstein, P.; Sanchez, C. J. Mater. Chem. 19%, 6, ) Loy, D. A.; Shea, K. J. Chem. Rev. 1995,95, ) Corriu, R J. P.; Leclercq, D. Angew. Chem., ht. Ed. Engl. 19%, 35, ) Shea, K. J.; Loy, D. A.; Webster, O. J. Am. Chem. Sot. 1992,114, ) Small, J. H.; Shea, K. J.; Loy, D. A. J. Non-Cryst. Solids 1993,16, ) Oviatt, H. W., Jr.; Shea, K. J.; Small, J. H. Chem. Mater. 1993,5, ) Loy, D. A.; Jamison, G. M.; Baugher, B. M.; Russick, E. M.; Assink, R. A.; Prabakar, S.; Shea, K. J. J. Non-Cryst. Solids 1995,186, ) Oviatt, H. W. Jr.; Shea, K. J.; KalIuri, S.; Shi, Y.; Steier, W. H.; Dalton, L.R. Chem. Mater. 1995,7, ) Loy, D. A.; Carpenter, J. P.; Myers, S. A.; Assink, R. A.; Small, J. H.; Greaves, J.; Shea, K.J. J. Am. Chem. Sot. 1996,118, ) Loy, D. A.; Carpenter, J. P.; Myers, S. A.; Assink, R. A.; Small, J. H.; Greaves, J.; Shea, K. J. Polym. Prepr. (Am. Chem. Sot., Div. Polym. Chem.) 1996,37, ) Loy, D. A.; Carpenter, J. P.; Myers, S. A.; Assink, R. A.; Small, J. H.; Greaves, J.; Shea, K.J. Mater. Res. Sot. Symp. Proc. 1996, 435(Better Ceramics through Chemistry VII: Organicllnorganic Hybrid Materials),

22 . Sheaby 13) Loy, D. A.; Carpenter, J. P.; Alarn, T. M.; ShaltouQ R; DorhouL P. K.; Greaves, J.; SmaI1,J. H.; Shea, K J. Y.Am. Chin. SOC.1999,12, 5413-&2S ) Carpente~, J. I?; Yamanaka, S. A.; Mcclain, M. D.; Imy, D. A.; Greaves, J.; Shea, K J. Polym. Prepr. (Am. Chem. Sot., Div. Polyrn. Chem.) 1998,39, ) Loy, D. A.; Carpenter, J. P.; Yamanaka, S. A.; McC1ain, M. D.; Greaves, J.; Hobson, S.; She~ K. J. Chem. Mater. 1998,10, ) Hobson, S. T.; Shea, K. J. Chem. Mizter. 1997,9, ) Shaltout, R. M.; Loy, D. A.; Carpenter, J. P.; Dorhout, K.; Shea, K. J. Polym. Prepr. (Am. Chem. Sot., Div. Polym. Chem.) 1999,40, ) Loy, D. A.; Beach, J. V.; Baugher, B. M.; Assink, R. A.; Shea, K. J.; Tran, J.; Small, J. H. Chem. Mater. 1999,11, ) Loy, D. A.; Beach, J. V.; Baugher, B. M.; Assink, R. A.; Shea, K. J.; Tran, J.; Small, J. H. Mater. Res. Sot. Symp. Proc. 1999, 576(Organic/Irwrganic Hybrid Materials II), ) Myers, S.A.; Assink, R. A.; Loy, D. A.; Shea, K. J. Perkin 2,2000, ) McClain, M. D.; Loy, D. A.; Prabakar, S. Mater. Res. Sot. Symp. Proc. 1996, 435(Better Ceramics through Chemistry VII: OrganiclInorganic Hybrid Materials), ) Shaltout.j R. M.; Loy, D. A.; Carpenter, J. P.; Dorhout, K.; Shea, K. J. Polym. Prepr. (Am. Chem. Sot., Div. Polym. Chem.) 1999,40, ) The formation of 6- and 8-membered cyclic siloxanes in the sol-gel polymerization of silicate precursors such as TEOSZX and trifunctional organosilanes (T-resins24) is well documented. These cyclic siloxane structures are responsible for the high concentrations required for gelation and for the high degrees of condensation observed in their xerogels.3z Cyclic 6- and 8- membered ring siloxane formation may also be taking place during sol-gel polymerization of bridged silsesquioxane monomers. However, the higher level of functionality of these monomers does not allow these cyciic siloxanes to inhibit the gelation of bridged polysilsesquioxanes. This distinguishes the behavior of hexafunctional bridged silsesquioxane monomers from tetrafunctional silicate precursors and trifunctional organosilanes. 24) a) Baney, R. H.; Itoh, M.; Sakakibara, A.; Suzuki, T. Chem. Rev. (Washington, D. C.) 1995,95, b) Voronkov, M. G.; Lavrent yev, V. I. Top. Curr. Chem 1982, 102, c) Loy, D. A.; Baugher, B. M.; Schneider, D. A. Polym. Prepr. (Am. Chem. Sot., Div. Polym. Chem.) 1998,39,

23 . ShealLoy 25) Brinker, C. J.; Scherer, G. W. Sol-Gel Science: the physics and chemis~ of sol-gel processing; 1st Ed.; Academic Press, Inc.: San Diego, PP99-.l.i3f&_. - 26) Ng, L.V.jThompson, P.; Sanchez, J.; Macosko, C. W.; McCormick, A. V. Macromolecules 1995,28, ) Schaefer, D. W.; Beaucage, G. B.; Loy, D. A.; Ulibarri, T. A.; Black, E.; Shea, K. J.; Buss, IL J. Mater. Res. Sot. Symp. Proc. 1996, 435(Better Ceramics through Chemistry VII: OrganiclInorganic Hybrid Materials), ) Baugher, B. M.; Loy, D. A.; Prabakar, S.; Assink, R. A.; Shea, K. J.; Oviatt, H. Mater. Res. Sot. Symp. Proc. (1995), 371(Advances in Porous Materials), ) During acid-catalyzed sol-gel polymerizations, the rate of condensation becomes slower with higher degrees of condensation. See refs. 25 & 26. Thus, cyclics formed early in the sol-gel polymerization can slow gelation. Under base-catalyzed conditions, the rate remains the same or increases with condensation. Even when cyclics form, they will be incorporated into the network.