Application of solid-state NMR (13C and 29Si CP/MAS NMR) spectroscopy to the characterization of alkenyltrialkoxysilane and trialkoxysilyl-terminated polyisoprene grafting onto silica

Application of Solid-State NMR (13C and 29Si CP/MASNMR) Spectroscopy to the Characterization ofAlkenyltrialkoxysilane and Trialkoxysilyl-TerminatedPolyisoprene Grafting onto Silica Microparticles

DANIEL DEROUET,1 SYLVIE FORGEARD,1 JEAN-CLAUDE BROSSE,1 JOEL EMERY,2 JEAN-YVES BUZARE2

1 Chimie et Physique des Materiaux Polymeres (Unite Mixte de Recherche no. 6515), Universite du Maine,avenue Olivier Messiaen, BP 535, 72017 Le Mans, France

2 Laboratoire de Physique de l’Etat Condense (UPRES A CNRS no. 6087), Universite du Maine, avenue Olivier Messiaen,72017 Le Mans, France

Received 4 February 1997; accepted 12 August 1997

ABSTRACT: The solid-state Nuclear Magnetic Resonance (NMR) was used to character-ize surfaces of silica gels chemically modified by alkenyltrialkoxysilanes and trialkoxy-silyl terminated 1,4-polyisoprenes. The formation of covalent bonds created betweenalkoxy functional groups from alkenyltrialkoxysilane or trialkoxysilyl-terminated 1,4-polyisoprene and silanol groups on silica was clearly demonstrated by means of 13Cand 29Si CP/MAS NMR spectroscopy. Quantitative data, including calculation of thegrafting yields in relation with the initial silanol concentrations, were also obtainedby using solid-state 29Si-NMR leading to a final well-defined characterization of thesilica surfaces. A relatively good agreement was noticed between the grafting yieldscalculated from 29Si-NMR spectra and those determined from other analytical tech-niques such as Wijs titration or elementary analysis. The reactivity of the various silicasilanols towards each coupling agent was clearly characterized and estimated, as werethe proportions of the various grafted structures formed at the silica surface. q 1998John Wiley & Sons, Inc. J Polym Sci A: Polym Chem 36: 437–453, 1998Keywords: functionalized silica; functional polyisoprene; grafting onto silica; polyiso-prene-grafted silica; alkenyl-bonded silica; 13C and 29Si CP/MAS NMR; grafting charac-terization; surface characterization

INTRODUCTION tomer interactions and in mechanical propertiesof the vulcanizates can be noticed.

To develop new materials for these applica-Organic molecules or polymer-grafted silica gels tions, we have prepared various polyisoprene-were used for many years in a large number of grafted silica fillers. Many articles were publishedapplications: for example, in chromatography as about the silica silylation using functionalized si-stationary phases,1–3 for heterogeneous sup- lane reagents9 and polymeric coupling agents10

ported catalysis,4–6 or for rubber reinforcement.7,8but, to our knowledge, grafting of polydienes onto

In this last case, by grafting some coupling agents silica particles was not yet described. Only poly-onto the filler surface, changes in the filler–elas- diene-coated silica gels have been nowadays

tested as stationary phases for chromatography11

or as reinforcing fillers for rubber technology.12Correspondence to: D. Derouet

Synthesis of linear 1,4-polyisoprene-grafted sil-Journal of Polymer Science: Part A: Polymer Chemistry, Vol. 36, 437–453 (1998)q 1998 John Wiley & Sons, Inc. CCC 0887-624X/98/030437-17 ica particles was previously described.13 Chemical

437

8G66 022T/ 8G66$$022T 12-11-97 12:07:37 polca W: Poly Chem

438 DEROUET ET AL.

bonding of polymer (polyisoprene) layers onto sil- silanol groups concentration: 5.6 mmol/m2) is acommercial silica gel usually used as stationaryica surfaces was obtained by reaction of silanol

groups from silica surface with alkoxysilane func- phase for adsorption liquid chromatography.Prior to use, it was heated for 12 h at 393 K undertions of a trialkoxysilyl-terminated polyisoprene

(alkoxy Å methoxy or ethoxy) [eq. (1)] . vacuum (1002 mbar).Allyltriethoxysilane, allyltrimethoxysilane, vi-

nyltriethoxysilane, vinyltrimethoxysilane, hexa-methyldisilazane (HMDS) (all from Roth–So-chiel) and n -butyllithium (1.6M in hexane, Ald-rich) are commercially available. They were usedas received.

Isoprene (Aldrich) was dried over lithium al-uminium hydride under inert atmosphere be-

Bu

n

OHOHOH

Silica R 5 Me or Et

(RO)‹Si1

OOR

SiOOH

Bu (1)n fore use.

Toluene, dichloromethane, tetrachlorometh-The purpose of the present article is to reportane, and tetrahydrofuran were purified accordingthe procedure used to access to a qualitative andto classical methods and carefully redistilled justquantitative characterization of the silica sur-before use.faces chemically modified by the trialkoxysilyl-

terminated polyisoprenes. This was realized bymeans of the Cross Polarization/Magic Angle Apparatus and MethodsSample Spinning Nuclear Magnetic Resonance

All reactions were carried out under dry argon(CP/MAS NMR) technique. The study was real-atmosphere in dry glassware according to the pre-ized with silica gels grafted (1) with variousviously described procedure.13

alkenyltrialkoxysilanes: allyltriethoxysilaneTitration of carbon–carbon double bonds was(ATES), allyltrimethoxysilane (ATMS), vinyl-

performed by iodine addition according to the Wijstriethoxysilane (VTES), and vinyltrimethoxysi-method.14

lane (VTMS); and (2) with triethoxysilyl (PI/The specific surface area of the Kieselgel S sil-SiOEt) and trimethoxysilyl (PI/SiOMe)-termi-

ica was determined at the ‘‘Laboratoire de Chimienated polyisoprenes. The study with the alkenyl-des Surfaces’’ (CNRS–URA no. 1428, Paris,trialkoxysilanes was done with a view to collectFrance) by using the BET (‘‘Brunauer, Emmettsignificant information useful to analyze theand Teller’’ ) method consisting in measuring N2grafting of the trialkoxysilyl-terminated polyiso-adsorption at 1207C under high vacuum (1002

prenes onto silica. The solid-state NMR tech-mbar) for 12 h.15

nique, and particularly 29Si CP/MAS NMR, wasThe silanol group surface concentration of theused at first to characterize the creation of chemi-

Kieselgel S silica was determined at the ‘‘Labora-cal bonds between silica and the coupling agent,toire des Materiaux Organiques a Proprietes Speci-and so to confirm grafting onto silica particles,fiques’’ (CNRS–UPR no. 9031, Vernaison, France)and afterwards, to establish the surface struc-by measuring the amount of ethane evolved by reac-tural composition of the grafted silica gels to com-tion with triethylaluminium.16

pare the reactivities of the various silica silanolsThe elementary analyses (Si, C, O) were madetowards each coupling agent, and then to give an

by the ‘‘Service Central d’Analyse du CNRS’’ (Ver-approach of the mechanisms involved during thenaison, France).grafting process. To verify the validity of the mea-

Solid-state 29Si and 13C CP/MAS NMR mea-surements made from the 29Si CP/MAS NMRsurements were carried out on a Bruker MSL300spectra, the grafting yields were compared tospectrometer at 7.05 Teslas with samples in dou-those obtained from: (1) titration of carbon–car-ble air bearing 4 mm rotors of ZrO2. Magic Anglebon double bonds according to the Wijs titration14

Spinning was performed at 5 kHz spinning rate.and (2) elementary analysis.For 29Si NMR measurements (at 59.63 MHz) ,the proton 907 pulse was 3.5–4.4 ms and the rep-EXPERIMENTALetition time 2 s. For CP curves 1024 FIDs were

Materials accumulated. The contact time was fixed at 5 msin all experiments. For 13C-NMR measurementsKieselgel S (Riedel-de-Haen; particle size: 63 to

200 mm; specific surface area: 480 m2/g; surface (at 75.47 MHz), the proton 907 pulse was 3.5 ms


POLYISOPRENE GRAFTING ON SILICA PARTICLES 439

and the contact time 3.5 ms. Solution 13C-NMR spectively) is introduced onto 1 g of Kieselgel Ssilica (K0) dispersed in 10 mL of dry toluene. Thespectra were performed in CDCl3 on a Bruker

AC 400 spectrometer (100.62 MHz). All NMR mixture is refluxed for 15 h under argon. Afterreaction, the ungrafted silane is removed by ex-spectra were refered to tetramethylsilane

(TMS) . Chemical shifts (d ) were given in parts traction with dichloromethane in a Soxhlet for 48h. Finally, the grafted silica gel is dried for 15 hper million (ppm).under vacuum at 407C and then characterized bymeans of 13C and 29Si solid-state NMR. Grafting

Synthesis of Trialkoxysilyl-Terminated yields were calculated from the 29Si solid-state1,4-Polyisoprenes NMR spectra and by using the Wijs titration.‘‘Living’’ poly(isoprenyllithium) with Mn Å 800(calculated by 1H-NMR, after deactivation with

Grafting of End-Capping Agentmethanol) was prepared in toluene at 407C for 2h by anionic polymerization of isoprene with n - The VTMS46%- and PI/SiOMe14%-grafted silicabutyllithium as initiator.13,17

gels were submitted to an end-capping treatmentTrialkoxysilyl-terminated 1,4-polyisoprenes were with hexamethyldisilazane (HMDS). The reac-prepared using optimized experimental conditions tions were carried out in CCl4 for 24 h at roompreviously established.13 The poly(isoprenyllith- temperature under argon. At the end, the silicaium) was reacted with tetramethoxysilane or tetrae- particles were extracted in a Soxhlet with CH2Cl2thoxysilane (the used molar ratios r Å [‘‘living’’ for 48 h and then dried for 15 h under vacuumchains RCH02 Li/]/[tetraalkoxysilane] were respec- (1002 mbar) at 407C.tively 0.4 and 0.6) in THF, at 0207C for 4 h and

then at room temperature for 18 h. Solvents (tolu-ene, THF, and hexane) and residual Si(OR)4 were 29Si NMR Calculation of G Concentrations of theextracted under vacuum. The product was then di- Grafted Structures Formed at the Silica Surface18,19

luted in dichloromethane and purified by filtrationafter precipitation with methanol or ethanol. After First of all, it is important to consider that each

grafting-derived structure is the result of the reac-a new dissolution in dichloromethane and extrac-tion of residual solvents under vacuum, the dried tion of the coupling agent with the silanol groups

at the silica surface. In the initial step, the num-final polymer was analyzed by NMR DPn Å 8 from1H-NMR). ber of silanol groups per area unit that can theo-

retically react with the coupling agent is then thevalue found for the K0 phase (Kieselgel S un-

Grafting of Alkenyltrialkoxysilanes or grafted silica) G0 Å 5.6 mmol/m2 (obtained by theTrialkoxysilyl-Terminated Polyisoprenes onto Silica BET method). The possible structures (T1 , T2 ,

T3 , Q2 , and Q3) , formed after alkenyltrialkoxysi-The grafting of the different coupling agents(alkenyltrialkoxysilanes or trialkoxysilyl-termi- lane or trialkoxysilyl-terminated polyisoprene

grafting onto silica, are given in Figure 1.nated polyisoprenes) onto silica was carried out asfollows. The coupling agent (silane and polymer In the 29Si NMR spectra, each structure is char-

acterized by a peak whose area (symbolized by A )concentrations are given in Tables III and V, re-

Figure 1. Grafted structures formed at the silica surface with their correspondingconcentrations.


440 DEROUET ET AL.

is assumed to be proportional to the concentration(G ) of the corresponding silicon. FA (Q2)

A (T3) G Å 32r

GOH(Q2)

GSilane(T3)(6)

GOH(Qx) represent the concentrations of the var-ious residual Si{OH: the (Qx) subscripts corre-spond, respectively, to geminal (x Å 2) and free

GOH(Q2) ÅG0

F1 / 12rFA (Q3)

A (Q2) G/ 12rFA (T1)

A (Q2) G(7)(x Å 3) silanols.

GOH(Silane) represents the Si{OH concentrationof the structures formed after condensation ofonly one silanol group of the geminal silanol

/ FA (T2)A (Q2) G / 3

2rFA (T3)

A (Q2) GGgroups.Gsilane(Tx) represent the Si{O{Si{R * siloxy

concentrations of the Tx structures, respectively,formed onto the silica surface (see Table II) after G0

OH(Q2) ÅG0

F1 / 12rFA0(Q3)

A0(Q2) GG(8)

condensation of one (x Å 1), two (x Å 2), or three(xÅ 3) alkoxy groups from alkenyltrialkoxysilaneor v-trialkoxysilyl polyisoprene with, respectivelyone, two, or three silica silanols. GOH(Silane) Å 1

2r(G0OH(Q2) 0 GOH(Q2) ) (9)

In all cases, the whole concentration of silicasilanols before grafting (G0 Å G0

OH(Q2) / G0OH(Q3) ,

By recombining all these equations, it is thenthe last symbols being the respective concentra-possible to determine the concentrations of each oftions of geminal and free silanols in the K0 initialthe six species characterized at the silica surfaceungrafted silica) can be related, after grafting,(residual geminal and free silanols; mono-, bi-, orto the sum of the concentrations of the differenttridentate silylated groups) (see Tables III and V).silanols and siloxy functions at the final grafted

silica surface [eq. (2)] .The silicon from Q2 silanols bearing two hy- Definition and Calculation of the Graft

droxyl groups, the ratio [A (Q2) /A (Q3)] can be Concentration and the Grafting Yieldlinked to the ratio of the corresponding concentra-

The graft concentration ngraft ( in mmol/m2) wastions, as shown in eq. (3).defined as the number of mol of grafts (alkenylOn the other hand, the ratios [A (Q2) /A (T1)] ,or polyisoprene) per m2 of silica surface. It was[A (Q2) /A (T2)] , and [A (Q2) /A (T3)] can be alsocalculated from the following relations dependingwritten as a function of the concentrations of sila-on the analytical technique:nols and grafted functional groups [eqs. (4) – (6)] .

Finally, combining eqs. (2) – (6), eq. (7) is devel-(1) in the case of 29Si CP/MAS NMR:oped.

As the initial concentration G 0OH(Q2 ) of gem-

ngraft Å GSilane(T1) / 12GSilane(T2) / 1

3GSilane(T3)inal silanols can be calculated from the K029Si

NMR spectrum with the relation described in eq.(2) In the case of the Wijs method: (a) if alkenyl-(8) , the concentration GOH(Silane) is deduced fromtrialkoxysilane coupling agents are used:eq. (9) .

GO Å GSilane(T1) / GSilane(T2) / GSilane(T3) ngraft ÅN

(mSiO02 NrM )r

1S/ GOH(Q2) / GOH(Q3) / GOH(Silane) (2)

(b) If trialkoxysilyl-terminated polyisoprenes areFA (Q2)A (Q3) G Å 1

2r

GOH(Q2)

GOH(Q3) / GOH(Silane)(3)

used:

ngraft ÅN

(mSiO02

NDPn

rM )r

1Sr

1DPn

FA (Q2)A (T1) G Å 1

2r

GOH(Q2)

GSilane(T1)(4)

FA (Q2)A (T2) G Å GOH(Q2)

GSilane(T2)(5) where N is the number of mol of carbon–carbon

double bonds in the titrated sample (Wijs titra-



tion), mSiO2 (g) is the weight of grafted silica in oxysilane reagents), which indicates the presenceof residual alkoxy groups.the titrated sample, M (grmol01) is the molecular

weight of the alkenyltrialkoxysilane or the trial-koxysilyl-terminated polyisoprene, S (m2

rg01) is 29Si CP/MAS NMRthe specific surface area of the silica gel (480

This method was proven to be very useful to studym2rg01) , and DPn is the polymerization degree of

the grafting onto silica because spectra are easythe polyisoprene graft (DPn Å 8).to analyze. Resolution of signals is very good andso the relation between Si chemical shifts and cor-(3) In the case of elementary analysis:responding assigned structures onto silica surfaceis easy to establish.

ngraft Å%Cr100/acrMc

F1 0 %Cr100acrMc

rMGr1S In Figure 3, 29Si CP/MAS NMR spectra of the

different organosilanes grafted onto silica gels (thegrafting yield—calculated by the Wijs method—isgiven in index in the formula abbreviations) arecompared to that of the K0 ungrafted one.where ac is the number of carbons and Mc (12

The Si chemical shifts and the assigned corre-grmol01) is the carbon molecular weight.sponding chemical structures are summarized inThe grafting yield can also be related to theTable II.initial silanols and calculated by using the follow-

The 29Si CP/MAS NMR spectrum of the un-ing relation:treated silica (K0) shows three signals at 090.2,0100.2, and 0108.9 ppm, which are respectively

T Å ngraftr100G0

Å ngraftr1005.6 assigned to Q2 geminal silanols, Q3 free silanols,

and Q4 siloxane groups.On the 29Si CP/MAS NMR spectra of the modi-

fied silica gels, two groups of peaks are alwaysRESULTS AND DISCUSSION found. As expected, the first one, from 090.0 to

0110.0 ppm, is representative of the silica surfaceAlkenyl-Bonded Silica Gels silicon atoms unmodified during the grafting pro-

cess. Comparing to the K0 spectrum, an increase13C CP/MAS NMRof the signal of the siloxane groups is observedas well as a reduction of those corresponding toThe silica gels obtained after grafting of various

alkenylsilane coupling agents were characterized geminal and free silanol groups. In the same time,it is interesting to see that Q2 peak totally disap-by means of 13C CP/MAS NMR. The resonance

assignments of the corresponding chemical struc- pears when grafting yields are higher. The secondgroup of peaks, in the range 050 to 080 ppm,tures are shown on the spectra given in Figure 2.

In all cases, well-resolved spectra were obtained. proves that the silica surface is effectively chemi-cally modified by the alkenyltrialkoxysilane re-It was noticed that a good similarity exists be-

tween the chemical shifts of the carbons from the agent. The reaction of silica surface silanols witha trialkoxysilyl functional group can lead to threegrafted structures and those of the corresponding

carbons of the starting alkenyltrialkoxysilane types of new structures onto the silica gel. TheT1 , T2 , and T3 structures are the result of the(Table I) . Whatever the conditions of NMR analy-

sis (in CDCl3 standard solutions for the starting reaction of respectively one, two, and three alkoxygroups of a same silane molecule with the silanolsalkenyltrialkoxysilanes or at solid-state for the al-

kenyl-bonded silica gels) , chemical shifts for the at the silica surface (see Fig. 1 and Table II) .The dSi chemical shift of the silicon of thesame corresponding carbon are always very near.

On other hand, 13C CP/MAS NMR analysis was grafted alkenylsilyl group depends on the numberof siloxane bonds created between this silicon andused to check the completion of the grafting onto

silica and so to demonstrate that the reaction be- the silica: the signals are shifted about 10 ppmbetween the mono- and bi-dentate silicon (respec-tween alkoxy functions from silane reagent and

silica silanols is never complete. This was proven tively, T1 and T2) and between the bi- and tri-dentate silicon (T2 and T3) .18,19 For a graftedon the 13C CP/MAS NMR spectra by a signal near

58.8 ppm (grafting of alkenyltriethoxysilane re- structure of same type (T1 , T2 , or T3) , the signalcharacteristic of a silicon bearing an allyl groupagents) or 49.0 ppm (grafting of alkenyltrimeth-


442 DEROUET ET AL.

Figure 2. 13C CP/MAS NMR spectra of alkenyl-bonded silica gels; (a) VTES, (b)ATES, (c) VTMS, and (d) ATMS.

is always displaced to the lower field in compari- matization of the limit forms in the case of thevinyldialkoxysilyl grafted group is given in eq.son with that of a silicon with a vinyl group. This

deshielding of the vinylsilyl silicon can be attrib- (10) (R Å Me or Et). It shows that the siliconatom is bonded to the neighboring carbon by auted to a charge delocalization involving the sili-

con d orbitals: unlike the allylsilyl structure, the double bond responsible for the observed de-shielding.20,21vinylsilyl one is stabilized by mesomery. A sche-



Table I. 13C-NMR Chemical Shifts (d in ppm) of Alkoxysilane Structures before (in CDCl3 Solution) and after(CP/MAS NMR) Grafting onto Silica

Chemical Shifts

Coupling Agent NMR Method dCaa dCb

a dCca dCd

a dCea

VTES in CDCl3 58.5 18.2 137.1 129.3 —CP/MAS NMR 58.8 16.0 135.2 128.5 —

ATES in CDCl3 58.6 18.25 114.7 132.7 18.3CP/MAS NMR 58.9 15.8 113.5 131.3 16.3

VTMS in CDCl3 50.6 — 138.0 126.6 —CP/MAS NMR 48.9 — 134.3 128.8 —

ATMS in CDCl3 50.8 — 115.1 132.1 17.0CP/MAS NMR 49.1 — 113.2 131.0 17.95

a References of carbons are those used in Figure 2.

yields of the coupling agents grafted onto silica,was realized by using 29Si NMR peak area simula-tions. The main notions about these calculationsare developed in the experimental part. The re-sults of the surface concentration calculations in

RO©Si©CH®CH¤ RO2 ¤Si®CH©CH1

OR

©O©Si©O©

O

(10)RO

©O©Si©O©

O

relation to the respective alkenyl-bonded silicagels are summarized in Table III.

Because a large proportion of silanol groups Generally, whatever the coupling agent or thestill remains unreacted with the alkoxysilane re- grafting density at the silica surface, the spectraagent, an end-capping reaction with hexamethyl- of the grafted silica gels, compared to the un-disilazane (HMDS) was tested on the VTMS46% grafted one (K0), show an area decrease of thevinyl-bonded silica. The purpose was to deactivate peaks corresponding to Q2 geminal and Q3 freethe residual free silanol groups from the grafted silanols and an area enhancement for that of thesilica, to use it as a stationary phase for liquid siloxane groups. On the other hand, it was noticedchromatography. The overall reaction of HMDS that grafting results depend on the initial cou-with residual silica surface silanols occurs as de- pling agent concentration used in comparisonscribed in eq. (11). with the silanol concentration at the surface of

the K0 silica.

Gi ° G0

OHOH 1 (CH‹)‹SiNHSi(CH‹)‹OH

OSi(CH‹)‹OSi(CH‹)‹ 1 NH‹OH

(11)

If the Gi initial coupling agent concentration(e.g., Gi Å 1.63 and 1.38 mmol/m2 of silica, respec-The analysis of the final product by means oftively, for VTES29% and ATMS21%) is smaller thansolid-state NMR revealed that end-capping is notG0 , the reactivity of the silanol groups towardsvery efficient because ungrafted silanol groups arethe coupling agent depends on the alkoxy groupalways present. Nevertheless, the action of thenature. With the alkenyltrimethoxysilane deriva-end-capping with HMDS can be evaluated fromtives (e.g., ATMS21%), there is no significant dif-the 29Si CP/MAS NMR spectrum. Indeed, the end-ference of reactivity between geminal and free si-capped silica shows a new signal at /14.0 ppmlanols (tQ2 graft É tQ3 graft É 37%). On the othercorresponding to the trimethylsilyl anchored

groups (so called M). hand, with the alkenyltriethoxysilane reagents(e.g., VTES29%), reactivity of Q2 silanols towardsTo obtain more information about the evolution

of the reactions involved during the silica silyla- ethoxysilyl functions is 1.35 times higher thanthat of Q3 silanols. It is also interesting to notetion, a quantitative determination of the concen-

trations of the different structures at the silica that when Gi ° G0 , only T1 and T2 structures areformed. The proportions of T1 and T2 structuressurface (especially, T1 , T2 , and T3) , as well as the


444 DEROUET ET AL.

Figure 3. 29Si CP/MAS NMR spectra of alkenyl-bonded silica gels; (K0) ungraftedsilica, (a) VTES, (b) ATES, (c) VTMS, and (d) ATMS.



OH

Coupling agent

263.3 252.0 261.2 252.0

270.5 259.7 270.2 260.3

— —

279.0 269.6

— —

14.0 —

T⁄ T¤ T‹ M

VTES ATES VTMS ATMS

Table II.≥™ªSi CP/MAS NMR Analysis of the Various Alkenyl-Bonded Silica Gels: Silicon Chemical Shifts (d in ppm) in Relation with the Grafted Structures

OR

R9

O©©Si©OROH

OH

CH‹

CH‹

O©©Si©CH‹OH

OH

ORSi

R9OO

OO©Si©R9

O

at the silica surface are approximately equivalent These last results can surprise because the in-crease of trialkoxysilane reagent concentration(tT1Å tT2É 50%) when the coupling agent used is

an alkenyltriethoxysilane (VTES29% or ATES28%). should normally favour the formation of the T1

monodentate structures and not that of the T2On the contrary, disubstitution, which leads tothe formation of T2 structures, is favored when bidentate structures, or even less the T3 tri-

dentate ones. In fact, the explanation is given bythe silane reagent is an alkenyltrimethoxysilane(ATMS21%): two methoxysilyl functions from considering the results obtained at low concentra-

tion where it is demonstrated that, in these condi-same silane molecule react with two silanols fromthe silica surface (tT1 É 30%; tT2 É 70%). tions, a maximum of two alkoxysilyl functions of

the same reagent molecule can react with the sil-ica silanols [eqs. (12) and (13)] .Gi ¢ G0

If Gi (e.g., Gi Å 8,17 mmmol/m2 of silica, forATES43%) is higher than G0, the reactivity of thesilanols towards the coupling agent depends on thenature of the alkenyl functional group. In the caseof vinyl derivatives (VTMS22% and 46% and VTES33%),the reactivity of the Q2 and Q3 silanols towards thealkoxysilyl functions is the same, whatever the{OR alkoxy groups. For example, for VTMS22%

({OR Å {OMe): tQ2 graft Å tQ3 graft É 35% and forVTES33% ({OR Å{OEt): tQ2 graft Å tQ3 graft É 44%.

(12)

(13)

©O©Si©O©Si©O©

Si

O O OH

R9RO

©O©Si©O©Si©O©

Si

O O O

R9

Si

OR OR

22 ROH

23 ROH

RO

R9

©O©Si©O©Si©O©

HO OH OH

On the other hand, with allyl derivatives(ATMS43%), the silanols Q2 are 0.8 time less reactive Then, the reaction of the third alkoxysilyl function

with a third silica silanol appears impossible. Thisthan the Q3 ones. It is also noted that when Gi

¢ G0, the disubstitution reaction is always favoured is justified by the heterogeneous distribution of silicasurface silanols and the very low probability of exis-whatever the coupling agent. However, the propor-

tions of T1, T2, and T3 structures seem to depend tence of three silanols sufficiently close to each otherto be able to react with a same alkenyltrialkoxysilaneon the nature of the alkoxy groups. Indeed, even

though the addition of vinyltriethoxysilane onto K0 molecule. Consequently, the presence of T3 struc-tures, characterized in the case of reactions of alken-(VTES33%) only leads to T1 and T2 structures, the

formation of the three T1, T2, and T3 structures is yltrimethoxysilanes with silica at high initial concen-trations in silane, can be explained only by the forma-clearly demonstrated with the trimethoxysilyl re-

agents (VTMS22% and ATES43%), but the T3 yield tion of three Si{ O{Si siloxane bonds involvingtwo different reactions: (1) condensation betweenis always very low (e.g., ATES43%: tT1 Å 38%, tT2

Å 53%, and tT3 Å 9%). methoxysilyl functions of the silane and silanols from


446D

ER

OU

ET

ET

AL

.

Table III. 29Si CP/MAS NMR Characterization of the Alkenyl-Bonded Silica Gel Surfaces: Calculation of the Respective Silanol and Alkenyl GroupConcentrationsa

Concentrations (G in mmol/m2 of Silica)

GOH(Q2) GOH(Q3) GSilane(T1) GSilane(T2) GSilane(T3)Reference Gi Coupling AgentProduct (mmol/m2 of silica) (tQ2 graft

) (tQ3 graft) GOH(Silane) (tT1 graft

) (tT2 graft) (tT3 graft

) ngraft GSilane(M)

K0d — 1.14 4.46 — — — — — —

VTMS22%e 7.63 0.74 3.04 0.20 0.12 1.07 0.36 0.8 —

(35) (32) (15) (70) (15)VTMS46%

e 9.40 0.37 1.38 0.38 0.24 1.75 1.12 1.5 0.29(68) (69) (16) (59) (25)

VTES29%e 1.63 0.57 2.83 0.29 0.54 1.31 0 1.2 —

(50) (37) (45) (55)VTES33%

e 13.44 0.64 2.53 0.25 0.48 1.63 0 1.3 —(44) (43) (37) (63)

ATMS21%e 1.38 0.71 2.84 0.21 0.34 1.43 0 1.1 —

(38) (37) (32) (68)ATMS43%

e 8.17 0.62 1.94 0.26 0.59 1.67 0.44 1.6 —(46) (57) (38) (53) (9)

ATES28%e 4.25 0.70 3.19 0.22 0.51 0.92 0 1.0 —

(38) (29) (53) (47)

The values in parentheses, tQx graft

b and tTx graft,c are respectively the ratios (in %) of the Q2 and Q3 grafted silanols and the proportions of the T1 , T2 , and T3 grafted

structures.a GSilane Å GSilane(T1) / GSilane(T2) / GSilane(T3) / GSilane(M) .

b tQx graftÅ

G0OH(Qx) 0 GOH(Qx)

G0OH(Qx)

r100.

c tTx graftÅ

Gsilane(Tx)

xrngraftr100, with ngraft the number of moles of alkenyl grafts per m2 of silica surface (in mmol/m2).

ngraft Å GSilane(T1) / 12 GSilane(T2) / 1

3 GSilane(T3).d Ungrafted silica. G0 Å 5.6 mmol/m2 of silica surface, with G0 Å G0

OH(Q2) / G0OH(Q3).

e Grafted silica. The percentage given in subscript is the grafting yield T determined by the Wijs titration.

8G66

022T/8G

66$$022T12-11-97

12:07:37polca

W:

Poly

Chem


the silica, and (2) condensation between neighboring be G *2Å 1.1 and G 92Å 0.2 (T1 and T2 concentrationsdetermined according to the normal calculationmethoxy groups from two different grafted struc-

tures [eq. (14)]. The increase of the initial coupling are respectively 0.6 and 1.3 mmol/m2), whichleads to a value of ngraft Å GT1 / (G *T2 /2) / G 9T2agent concentration leads to an enhancement of the

graft density at the silica surface and so, to a de- Å1.35 mmol/m2 instead of 1.2 according to the nor-crease of the distance between two neighboring mal calculation (1.6 mmol/m2 with the Wijsgrafted structures, which therefore increases the method).probability of condensation involving two neigh-boring alkoxy groups (giving the T3 structure).

Polyisoprene-Grafted Silica Gels

13C CP/MAS NMR

The silica gels obtained after grafting of trimeth-oxysilyl (or triethoxysilyl)-terminated polyiso-prene were characterized by means of 13C CP/MAS NMR. Spectra are similar to those of thecorresponding polymeric coupling agent per-formed in CDCl3 solution (Fig. 5). A good agree-ment between the chemical shifts of the carbonsof the starting ungrafted polyisoprenes and thoseof their homologues of the grafted structures isobserved.

(14)

©O©Si©O©Si©O©Si©O©

Si

O O

O

O

R9

26 MeOH

O

Si

R9

Si

OMeOMe

MeO

R9

Si

OMeOMe

MeO

R9

©O©Si©O©Si©O©

HO OH OH

The linear polyisoprene structures at the silicasurface were easily identified according to their

The graft concentrations (defined as ngraft ) de- chemical shifts (Table IV). It is noticed that thetermined by 29Si NMR were also compared to conversion of the alkoxy groups is never completethose obtained according to the Wijs procedure because additional signals, characteristic of these(Fig. 4). The yields are approximately similar, functional groups, are always observed on the solid-which shows the validity and the coherency of the state NMR spectra of the polyisoprene-grafted silicaquantitative 29Si NMR, even if the values ob- gels. The results are in accordance with those foundtained according to this last method are always for the previous alkenyltrialkoxysilanes.lower (É 30%) than those calculated by using theWijs method. This is probably due to the approxi-

29Si CP/MAS NMRmation introduced in the calculation of the graftedsilane percentage. In fact, the relation used to cal- 29Si CP/MAS NMR spectra of ungrafted silica (K0)culate ngraft (see Table III) does not take the side

and polyisoprene-grafted silica gels prepared byreaction between neighboring alkoxy groups fromreaction between Kieselgel S silica and linear 1,4-different grafted structures into account.polyisoprenes having a trimethoxy (PI/SiOMe)-or triethoxy (PI/SiOEt)-silyl extremity are shownin Figure 6. In comparison with the ungrafted K0

spectrum, the spectra of the polyisoprene-graftedsilica gels show additional small peaks in therange 050 to 080 ppm, which constitute the proofof the grafting. At the same time, a decrease of

©O©Si©O©Si©O©

Si

T9¤ T 0¤

OO

RO R9

©O©Si©O©Si©O©

O

OO

SiSiRO OR

R9 R9

the intensities of the signals assigned respectivelyto Q2 geminal and Q3 free silanol groups, as wellas an increase of the amount of siloxane groupsTo justify this explanation, a modification was

brought to the calculation method so that the side (Q4) , are noticed.As for the alkenyl-bonded silica gels, a quanti-condensation reaction described above could be

taken into account. Then, supposing that the side tative determination of the different structurespresent at the surface of the polyisoprene-graftedcondensation reaction represents about 15% of all

the reactions, it was demonstrated, in the case of silica gels was performed by using 29Si CP/MASNMR peak area simulations. The calculationsVTES29%, that T *2 and T 92 concentrations would


448 DEROUET ET AL.

Figure 4. Grafting of the alkenyltrialkoxysilanes onto Kieselgel S: comparison of thealkenyl graft concentrations calculated from 29Si CP/MAS NMR spectra with thosemeasured according to the Wijs titration (the percentages given in subscript is thegrafting yield determined by Wijs titration).

were made on four polyisoprene-grafted silica gels pling agent, a part of the silica mixture was sub-mitted to a HMDS end-capping treatment. After(Table V):

The first two [respectively so-called PI/Si- purification by extraction of the ungrafted reagentwith dichloromethane in a Soxhlet, the two poly-OMe6% and PI/SiOMe6% (after end-capping)]

were the result of a same grafting reaction with isoprene-grafted silica gels [eq. (15)] were ana-lyzed by 29Si CP/MAS NMR and their spectratrimethoxysilyl-terminated polyisoprene. After

grafting by reaction with the polyisoprene cou- were compared.

(15)2MeOH

(MeO)‹Si©CH¤

Si©OH

O

O

Si©OH

K‚ PI/SiOMe6%

O

Si©OH

O

Si©OH

O

O

PI/SiOMe6% (after end-capping)

5 polyisoprenyl graft

Si©O©Si

O MeO

OMe

CH¤

Si©OH

O2NH‹

HMDS

Si©OH

O

O

Si©O©Si

O MeO

OMe

CH¤

Si©O©Si(Me)‹

O

In comparison with the spectrum of the non- ppm, which characterizes the chemical bonding oftrimethylsilyl groups.end-capped product, the 29Si CP/MAS NMR spec-

trum of the end-capped polyisoprene-grafted sil- On the other hand, the data summarized inTable V show that the grafting yields T are veryica gel (Fig. 6) shows an additional peak at /14



Figure 5. 13C NMR spectra of triethoxysilyl (a,b) and trimethoxysilyl (c,d) termi-nated polyisoprenes, respectively, before (standard solution NMR in CDCl3) , and after(CP/MAS NMR) grafting onto silica.

low (around 6%). Moreover, the graft concentra- more reactive than Q3 free silanols (5%). Regard-ing their reactivities towards HMDS, it is thetions ngraft determined from the quantitative 29Si

NMR spectra are near to those obtained by the same thing.The last two [so-called PI/SiOEt3%) and PI/Si-Wijs titration or elementary analysis. Comparing

to the K0 silica spectrum, a significative decrease OEt4%] were the results of grafting of triethoxysi-lyl-terminated polyisoprene at various initial con-of the peak surface assigned to geminal and free

silanols is observed. This evolution, which de- centrations.The results summarized in Table V show thatpends on the nature of the silanols, clearly shows

that the grafting of the polymer onto the silica the grafting yields are lower than those obtainedwith the trimethoxysilyl derivative (less thantakes place. Q2 geminal silanols (18%) appear


450 DEROUET ET AL.

OH

dCa

Chemical Shifts

Table IV.≥¡£C-NMR Chemical Shifts (d in ppm) of the Polyisoprene-Grafted Silica Gels (CP/MAS NMR)

OCH¤CH‹

CH‹

C®CH

CH¤©

©CH¤

H‹C

a

d CH¤CH¤

c

e d

CH¤a

a

b

b c

C®CH©CH¤ CH¤©

H‹Ce

a d

b c

b

O©©Si©OH

OH OCH‹a

O©©Si©

40.0

30.0

32.0

58.5

50.0

dCb

135.0

≤30.0

135.0

≤18.0

—

dCc

125.0

≤22.0

125.0

—

—

dCd

27.0

14.0

26.0

—

—

dCe

16.0

—

26.0

—

—

1,4-trans isoprene unit

Butyl extremity

1,4-cis isoprene unit

Ethoxysilyl group

Methoxysilyl groupOH

5%). But, the main remark concerns the incoher- quently, the calculations are systematically al-ence of the various data obtained from analysis of tered when secondary reactions occur during thethe 29Si CP/MAS NMR spectra. The ngraft values grafting procedure. In the present case, it is prob-calculated from the 29Si CP/MAS NMR spectra able that the grafting reaction is disturbed byare very different of those obtained from the other some hydrolysis reactions, which can be easilymethods. Moreover, the ratio values of geminal justified according to the presence of basic com-and free silanols having reacted are quite aber- pounds (lithium hydroxide and lithium ethoxide)rant. Indeed, it is surprising to notice that for one in the medium. So, calculation of the proportionexperiment (case of PI/SiOEt3%) no Q2 silanol has of the final silica geminal silanols could be alteredreacted with the ethoxysilyl functions, even by the presence of new groups, as for instancethough for the other (case of PI/SiOEt4%) realized {Si(OH)2 characterized by a peak at090.2 ppm.with a higher coupling agent concentration, theproportion of Q2 silanols having reacted is supe-rior to the initial content of Q2 silanols in the un-grafted K0 silica. A possible explanation to justifythese altered results can be given by consideringthe possible existence of secondary processes. In-deed, the used calculation method do not take thepossible secondary processes into account. Conse-

CH¤

Si©O

O

O

5 polyisoprenyl graft

Si

EtO

OEt250 to 280 ppm

CH¤

Si©O

O

O

Si

HO

OH290.2 ppm



Figure 6. 29Si CP/MAS NMR spectra of the various polyisoprene-grafted silica gels;(K0) ungrafted silica, (a) PI/SiOMe, and (b) PI/SiOEt.


452 DEROUET ET AL.

Table V. 29Si CP/MAS NMR Characterization of the Polyisoprenyl-Grafted Silica Gel Surfaces:Comparison of the Grafting Yield Values with Those Determined from Another Methods(Wijs Titration and Elementary Analysis)

Concentrations (mmol/m2 of Silica)

29Si CP/MAS NMR

Wijs EAc

GOH(Q3)GOH(Q2)Gi Coupling AgentReference Product (mmol/m2 of Silica) (tQ2graft

) (tQ3graft) GOH(Silane) ngraft GSilane(M) ngraft ngraft

K0a 1.14 4.46 — — — — —

PI/SiOMe6%b 1.60 0.91 4.17 0.12 0.31 — 0.34 0.35

(18) (5)PI/SiOMe6%

b 1.60 0.81 3.85 0.17 0.37 0.39 0.34 0.35(after end-capping) (28) (12)PI/SiOEt3%

b 1.75 1.14 4.02 0 0.31 — 0.18 0.19(0) (8)

PI/SiOEt4%b 2.40 2.11 2.93 0 1.01 — 0.23 0.30

(120) (34)

The values in parentheses, tQxgraft, are the ratios (in %) of grafted silanols (see Table III).

a Ungrafted silica.b Grafted silica. The percentage given in subscript is the grafting yield T determined by the Wijs titration.c Elementary analysis.

CONCLUSIONS trations of the respective residual silanols gave agood idea on their reactivities towards the variousalkoxy functions of the coupling agent. Moreover,CP/MAS NMR spectroscopy proved to be a usefulthe relatively good agreement noticed betweentechnique to clearly detect and identify the poly-the grafting yield values calculated from the 29Simeric structures chemically grafted onto a silicaCP/MAS NMR spectra and those determined bysurface and so, to confirm a grafting onto silicausing other methods, such as Wijs titration or ele-particles.mentary analysis, proved the validity of the calcu-In 13C-NMR, a good agreement was noted be-lation method used to determine the graftingtween the carbon chemical shifts of the startingyields from the NMR spectra. However, the calcu-coupling agents, simple alkenyltrialkoxysilanes,lation method set up to determine the graftingor trialkoxysilyl-terminated polyisoprenes (ana-yields from 29Si CP/MAS NMR spectra of the al-lyzed in CDCl3 solution), and those of the respec-kenyl-bonded or polyisoprene-grafted silica gels,tive structures grafted onto silica (analyzed atinvolves that the grafting reaction (reaction in-solid state).volving only silanols at the silica surface and al-On the other hand, 29Si CP/MAS NMR was suc-koxysilyl functions of the coupling agent) is notcessfully used to get a well-defined characteriza-altered by secondary processes (for instance, thetion of the structures formed at the silica surfacereactions between neighboring alkoxy groupsand to obtain informations about the reactivitiesfrom different grafted structures and the hydroly-of the different silanol groups (geminal and free)sis reactions), which is an important limitation.towards the alkoxy functions of the various cou-This was established during the study realizedpling agents. By this means and using a quantita-with the triethoxysilyl-terminated polyisoprenes,tive analysis based on peak area simulations, itwhere it was demonstrated that grafting can bewas possible to conveniently determine the rightdisturbed by the presence of basic compounds.percentages of geminal and free residual silanol

groups, as well as the various grafted structuresformed onto the silica surface (T1 , T2 , and T3 FORMULAS AND ABBREVIATIONSstructures, which result, respectively, from the re-action of 1, 2, and 3 alkoxy groups of a same start- ATES—allyltriethoxysilane; ATMS—allyltrimeth-

oxysilane; HMDS—hexamethyldisilazane; PI/Si-ing coupling agent). The knowledge of the concen-



Schmid, K. Albert, and E. Bayer, Chromato-OEt—triethoxysilyl-terminated polyisoprene; PI/graphia, 35, 403 (1993).SiOMe—trimethoxysilyl-terminated polyisoprene;

12. V. Vangani and A. K. Rakshit, Angew. Makromol.VTES—vinyltriethoxysilane; VTMS—vinyltri-Chem., 220, 21 (1994).methoxysilane.

13. S. Forgeard, D. Derouet, and J. C. Brosse, Mac-romol. Chem. Phys., to appear.

14. J. P. Wolff, Manuel d’analyse des corps gras, vol.VIII, Azoulay, Paris, 1968, p. 114.REFERENCES AND NOTES

15. D. David, R. Caplain, and B. Agius, Methodesusuelles de caracterisation des surfaces. Mesure des

1. K. Albert, R. Brindle, J. Schmid, R. Buszewski, and surfaces specifiques, Eyrolles, Paris, 1988, p. 290.E. Bayer, Chromatographia, 38, 283 (1994). 16. M. Sato, Y. Kanbayashi, K. Kobayashi, and Y.

2. R. Audebert and C. Quivoron, Bull. Soc. Chim. Fr., Sone, J. Catal., 7, 342 (1976).6, 1107 (1985). 17. S. Bywater, Prog. Polym. Sci., 19, 287 (1994).

3. S. Heron and A. Tchapla, Analusis, 21, 327 (1993). 18. A. B. Scholten, H. G. Janssen, J. W. de Haan, and4. J. P. Mathew and M. Srinivasan, Polym. Int., 29, C. A. Cramers, J. High Resolut. Chromatogr., 17,

179 (1992). 77 (1994).5. E. Carlier, A. Revillon, A. Guyot, and P. Baum- 19. K. Albert and E. Bayer, J. Chromatogr., 544, 345

gartner, React. Polym., 21, 15 (1993). (1991).6. J. P. Mathew and M. Srinivasan, Polym. Int., 24, 20. D. W. Sindorf and G. E. Maciel, J. Am. Chem. Soc.,

249 (1991). 105, 3767 (1983).7. H. Hommel, A. Touhami, and A. P. Legrand, Mak- 21. T. J. Barton and P. Boudjouk, Silicon-Based Poly-

romol. Chem., 194, 879 (1993). mer Science: A Comprehensive Resource, ACS8. M. P. Wagner, Rubber Chem. Technol., 49, 703 Symp. Ser., vol. 224, J. M. Zeigler and F. W. G.

(1976). Fearon, Eds., American Chemical Society, Wash-9. A. Vidal and J. B. Donnet, Bull. Soc. Chim. Fr., 6, ington, DC, 1990, p. 3.

1088 (1985). 22. A. R. Bassindale and P. G. Taylor, The Chemistry10. N. Tsubokawa, T. Kimoto, and K. Koyama, Colloid of Organic Silicon Compounds, Vol 1, S. Patai and

Polym. Sci., 271, 240 (1995). Z. Rappoport, Eds., J. Wiley & Sons, Chichester,UK, 1989, p. 809.11. M. Hanson, B. Eray, K. Unger, A. V. Neimark, J.


Documents

Application of solid-state NMR (13C and 29Si CP/MAS NMR) spectroscopy to the characterization of alkenyltrialkoxysilane and trialkoxysilyl-terminated polyisoprene grafting onto silica